Skip to content

Instantly share code, notes, and snippets.

@taroushirani

taroushirani/nnsvs_test_nit_song070_svs_world_cnn_mdn.ipynb

Last active November 8, 2020 07:06
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save taroushirani/e6f91ae272b90ca1dcd1e261044a14eb to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save taroushirani/e6f91ae272b90ca1dcd1e261044a14eb to your computer and use it in GitHub Desktop.
Download ZIP
nnsvs_test_nit_song070_svs_world_cnn_mdn
Display the source blob
Display the rendered blob
Raw
nnsvs_test_nit_song070_svs_world_cnn_mdn.ipynb
{
"nbformat": 4,
"nbformat_minor": 0,
"metadata": {
"colab": {
"name": "nnsvs_test_nit_song070_svs_world_cnn_mdn",
"provenance": [],
"collapsed_sections": [],
"mount_file_id": "12HbEBcuG8pRRDY0w9QECR16MVyqlitsN",
"authorship_tag": "ABX9TyOdPnsT/gy+Jv9Gb7bqS0Nq",
"include_colab_link": true
},
"kernelspec": {
"name": "python3",
"display_name": "Python 3"
},
"accelerator": "GPU"
},
"cells": [
{
"cell_type": "markdown",
"metadata": {
"id": "view-in-github",
"colab_type": "text"
},
"source": [
"<a href=\"https://colab.research.google.com/gist/taroushirani/e6f91ae272b90ca1dcd1e261044a14eb/nnsvs_test_nit_song070_svs_world_cnn_mdn.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
]
},
{
"cell_type": "code",
"metadata": {
"id": "FCY9PjEUXT5i",
"outputId": "42fa7d31-ca83-4d49-9106-6ba3799223bb",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 311
}
},
"source": [
"! pip install -U numpy cython"
],
"execution_count": 1,
"outputs": [
{
"output_type": "stream",
"text": [
"Collecting numpy\n",
"\u001b[?25l Downloading https://files.pythonhosted.org/packages/87/86/753182c9085ba4936c0076269a571613387cdb77ae2bf537448bfd63472c/numpy-1.19.4-cp36-cp36m-manylinux2010_x86_64.whl (14.5MB)\n",
"\u001b[K |████████████████████████████████| 14.5MB 231kB/s \n",
"\u001b[?25hRequirement already up-to-date: cython in /usr/local/lib/python3.6/dist-packages (0.29.21)\n",
"\u001b[31mERROR: tensorflow 2.3.0 has requirement numpy<1.19.0,>=1.16.0, but you'll have numpy 1.19.4 which is incompatible.\u001b[0m\n",
"\u001b[31mERROR: datascience 0.10.6 has requirement folium==0.2.1, but you'll have folium 0.8.3 which is incompatible.\u001b[0m\n",
"\u001b[31mERROR: albumentations 0.1.12 has requirement imgaug<0.2.7,>=0.2.5, but you'll have imgaug 0.2.9 which is incompatible.\u001b[0m\n",
"Installing collected packages: numpy\n",
" Found existing installation: numpy 1.18.5\n",
" Uninstalling numpy-1.18.5:\n",
" Successfully uninstalled numpy-1.18.5\n",
"Successfully installed numpy-1.19.4\n"
],
"name": "stdout"
},
{
"output_type": "display_data",
"data": {
"application/vnd.colab-display-data+json": {
"pip_warning": {
"packages": [
"numpy"
]
}
}
},
"metadata": {
"tags": []
}
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "iV4ghgxzXaNt",
"outputId": "3068816b-5181-43b3-c386-e92a78289966",
"colab": {
"base_uri": "https://localhost:8080/"
}
},
"source": [
"! git clone -q https://github.com/r9y9/hts_engine_API\n",
"! cd hts_engine_API/src && ./waf configure --prefix=/usr/ && ./waf build > hts_engine_API_build.log 2>&1 && ./waf install\n",
"! git clone -q https://github.com/r9y9/sinsy\n",
"! cd sinsy/src/ && mkdir -p build && cd build && cmake -DCMAKE_BUILD_TYPE=Release -DBUILD_SHARED_LIBS=ON -DCMAKE_INSTALL_PREFIX=/usr/ .. && make -j > sinsy_build.log 2>&1 && make install"
],
"execution_count": 2,
"outputs": [
{
"output_type": "stream",
"text": [
"\u001b[32m\u001b[0mSetting top to :\u001b[0m \u001b[0m\u001b[32m\u001b[32m/content/hts_engine_API/src\u001b[0m \u001b[0m\n",
"\u001b[32m\u001b[0mSetting out to :\u001b[0m \u001b[0m\u001b[32m\u001b[32m/content/hts_engine_API/src/build\u001b[0m \u001b[0m\n",
"\u001b[32m\u001b[0mChecking for waf version in 1.7.11-2.1.0 :\u001b[0m \u001b[0m\u001b[32m\u001b[32mok\u001b[0m \u001b[0m\n",
"\u001b[32m\u001b[0mChecking for 'gcc' (C compiler) :\u001b[0m \u001b[0m\u001b[32m\u001b[32m/usr/bin/gcc\u001b[0m \u001b[0m\n",
"\u001b[32m\u001b[0mChecking for header stdlib.h :\u001b[0m \u001b[0m\u001b[32m\u001b[32myes\u001b[0m \u001b[0m\n",
"\u001b[32m\u001b[0mChecking for header string.h :\u001b[0m \u001b[0m\u001b[32m\u001b[32myes\u001b[0m \u001b[0m\n",
"\n",
"hts_engine_API has been configured as follows:\n",
"\n",
"[Build information]\n",
"Package: hts_engine_API-1.0.9\n",
"build (compile on): x86_64-linux\n",
"host endian: little\n",
"Compiler: gcc\n",
"Compiler version: 7.5.0\n",
"CFLAGS: -O2 -Wall -fno-common -Wstrict-prototypes\n",
"\n",
"\u001b[32m'configure' finished successfully (0.858s)\u001b[0m\n",
"\u001b[32mWaf: Entering directory `/content/hts_engine_API/src/build'\u001b[0m\n",
"/usr/bin/gcc\n",
"/usr/bin/gcc\n",
"\u001b[32m\u001b[0m+ install \u001b[01;34m/usr/include/HTS_hidden.h\u001b[0m (from lib/HTS_hidden.h)\u001b[0m\n",
"\u001b[32m\u001b[0m+ install \u001b[01;34m/usr/include/HTS_engine.h\u001b[0m (from include/HTS_engine.h)\u001b[0m\n",
"\u001b[32m\u001b[0m+ symlink \u001b[01;34m/usr/lib/libhts_engine_API.so\u001b[0m (to libhts_engine_API.so.1.0.9)\u001b[0m\n",
"\u001b[32m\u001b[0m+ install \u001b[01;34m/usr/lib/libhts_engine_API.so.1.0.9\u001b[0m (from build/lib/libhts_engine_API.so)\u001b[0m\n",
"\u001b[32m\u001b[0m+ symlink \u001b[01;34m/usr/lib/libhts_engine_API.so.1\u001b[0m (to libhts_engine_API.so.1.0.9)\u001b[0m\n",
"\u001b[32m\u001b[0m+ install \u001b[01;34m/usr/bin/hts_engine\u001b[0m (from build/bin/hts_engine)\u001b[0m\n",
"\u001b[32m\u001b[0m+ install \u001b[01;34m/usr/lib/pkgconfig/hts_engine_API.pc\u001b[0m (from build/hts_engine_API.pc)\u001b[0m\n",
"\u001b[32mWaf: Leaving directory `/content/hts_engine_API/src/build'\u001b[0m\n",
"\u001b[32m'install' finished successfully (0.061s)\u001b[0m\n",
"-- The C compiler identification is GNU 7.5.0\n",
"-- The CXX compiler identification is GNU 7.5.0\n",
"-- Check for working C compiler: /usr/bin/cc\n",
"-- Check for working C compiler: /usr/bin/cc -- works\n",
"-- Detecting C compiler ABI info\n",
"-- Detecting C compiler ABI info - done\n",
"-- Detecting C compile features\n",
"-- Detecting C compile features - done\n",
"-- Check for working CXX compiler: /usr/bin/c++\n",
"-- Check for working CXX compiler: /usr/bin/c++ -- works\n",
"-- Detecting CXX compiler ABI info\n",
"-- Detecting CXX compiler ABI info - done\n",
"-- Detecting CXX compile features\n",
"-- Detecting CXX compile features - done\n",
"-- Configuring done\n",
"-- Generating done\n",
"-- Build files have been written to: /content/sinsy/src/build\n",
"[ 95%] Built target sinsy\n",
"[100%] Built target sinsy-bin\n",
"\u001b[36mInstall the project...\u001b[0m\n",
"-- Install configuration: \"Release\"\n",
"-- Installing: /usr/lib/libsinsy.so.0.9.2\n",
"-- Installing: /usr/lib/libsinsy.so.0.9\n",
"-- Installing: /usr/lib/libsinsy.so\n",
"-- Installing: /usr/bin/sinsy\n",
"-- Set runtime path of \"/usr/bin/sinsy\" to \"\"\n",
"-- Installing: /usr/include/sinsy\n",
"-- Installing: /usr/include/sinsy/sinsy.h\n",
"-- Installing: /usr/include/sinsy/ILabelOutput.h\n",
"-- Installing: /usr/include/sinsy/LabelStrings.h\n",
"-- Installing: /usr/lib/sinsy/dic\n",
"-- Installing: /usr/lib/sinsy/dic/japanese.shift_jis.table\n",
"-- Installing: /usr/lib/sinsy/dic/japanese.utf_8.table\n",
"-- Installing: /usr/lib/sinsy/dic/japanese.euc_jp.conf\n",
"-- Installing: /usr/lib/sinsy/dic/japanese.shift_jis.conf\n",
"-- Installing: /usr/lib/sinsy/dic/japanese.macron\n",
"-- Installing: /usr/lib/sinsy/dic/japanese.euc_jp.table\n",
"-- Installing: /usr/lib/sinsy/dic/japanese.utf_8.conf\n",
"-- Installing: /usr/lib/pkgconfig/sinsy.pc\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "UdRQ5pMuYtFj",
"outputId": "3b802a13-075d-4209-b2b1-1c1a42228a35",
"colab": {
"base_uri": "https://localhost:8080/"
}
},
"source": [
"! git clone -q https://github.com/r9y9/pysinsy\n",
"! cd pysinsy && export SINSY_INSTALL_PREFIX=/usr/ && pip install -q .\n",
"! git clone -q https://github.com/r9y9/nnmnkwii\n",
"! cd nnmnkwii && pip install -q .\n",
"! git clone -b cnn_mdn_test -q https://github.com/taroushirani/nnsvs\n",
"! cd nnsvs && pip install -q ."
],
"execution_count": 3,
"outputs": [
{
"output_type": "stream",
"text": [
" Building wheel for pysinsy (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
"\u001b[K |████████████████████████████████| 419kB 6.3MB/s \n",
"\u001b[K |████████████████████████████████| 368kB 13.6MB/s \n",
"\u001b[?25h Building wheel for nnmnkwii (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
" Building wheel for pysptk (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
" Building wheel for bandmat (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
"\u001b[K |████████████████████████████████| 7.6MB 4.6MB/s \n",
"\u001b[K |████████████████████████████████| 122kB 51.3MB/s \n",
"\u001b[K |████████████████████████████████| 184kB 55.1MB/s \n",
"\u001b[K |████████████████████████████████| 225kB 48.6MB/s \n",
"\u001b[K |████████████████████████████████| 112kB 13.8MB/s \n",
"\u001b[K |████████████████████████████████| 51kB 8.5MB/s \n",
"\u001b[K |████████████████████████████████| 276kB 42.2MB/s \n",
"\u001b[?25h Building wheel for nnsvs (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
" Building wheel for librosa (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
" Building wheel for pyworld (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
" Building wheel for antlr4-python3-runtime (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
" Building wheel for PyYAML (setup.py) ... \u001b[?25l\u001b[?25hdone\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "UAHxQFcMOBR2",
"outputId": "bebdc119-d168-47bf-d308-43d2071c09bf",
"colab": {
"base_uri": "https://localhost:8080/"
}
},
"source": [
"from google.colab import drive\n",
"drive.mount('/content/drive')"
],
"execution_count": 4,
"outputs": [
{
"output_type": "stream",
"text": [
"Mounted at /content/drive\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "7b6A4YPEOEjt"
},
"source": [
"!ln -s \"/content/drive/My Drive\" /content/gdrive"
],
"execution_count": 5,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "nJHNL9htIP4D"
},
"source": [
"RECIPE_ROOT=\"/content/nnsvs/egs/nit-song070/svs-world-cnn-mdn\""
],
"execution_count": 7,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"id": "QnYwVu94gjF4",
"outputId": "97743f4c-b900-4e9a-a249-b257465f0315",
"colab": {
"base_uri": "https://localhost:8080/"
}
},
"source": [
"! cd $RECIPE_ROOT && bash run.sh --stage -1 --stop-stage 1"
],
"execution_count": 9,
"outputs": [
{
"output_type": "stream",
"text": [
"stage 0: Data preparation\n",
"Prepare data for time-lag models\n",
"nitech_jp_song070_f001_003.lab: Global offset (in sec): -0.04\n",
"nitech_jp_song070_f001_003.lab: 3/85 time-lags are excluded.\n",
"nitech_jp_song070_f001_004.lab: Global offset (in sec): -0.06999999999999999\n",
"nitech_jp_song070_f001_004.lab: 10/73 time-lags are excluded.\n",
"nitech_jp_song070_f001_007.lab: Global offset (in sec): -0.049999999999999996\n",
"nitech_jp_song070_f001_007.lab: 3/67 time-lags are excluded.\n",
"nitech_jp_song070_f001_010.lab: Global offset (in sec): -0.01\n",
"nitech_jp_song070_f001_010.lab: 9/98 time-lags are excluded.\n",
"nitech_jp_song070_f001_012.lab: Global offset (in sec): -0.06\n",
"nitech_jp_song070_f001_012.lab: 7/211 time-lags are excluded.\n",
"nitech_jp_song070_f001_014.lab: Global offset (in sec): -0.04\n",
"nitech_jp_song070_f001_014.lab: 3/135 time-lags are excluded.\n",
"nitech_jp_song070_f001_015.lab: Global offset (in sec): -0.09\n",
"nitech_jp_song070_f001_015.lab: 5/63 time-lags are excluded.\n",
"nitech_jp_song070_f001_016.lab: Global offset (in sec): -0.02\n",
"nitech_jp_song070_f001_016.lab: 10/100 time-lags are excluded.\n",
"nitech_jp_song070_f001_019.lab: Global offset (in sec): -0.06\n",
"nitech_jp_song070_f001_019.lab: 4/115 time-lags are excluded.\n",
"nitech_jp_song070_f001_020.lab: Global offset (in sec): -0.13\n",
"nitech_jp_song070_f001_020.lab: 8/77 time-lags are excluded.\n",
"nitech_jp_song070_f001_021.lab: Global offset (in sec): -0.024999999999999998\n",
"nitech_jp_song070_f001_021.lab: 2/59 time-lags are excluded.\n",
"nitech_jp_song070_f001_022.lab: Global offset (in sec): -0.034999999999999996\n",
"nitech_jp_song070_f001_022.lab: 17/119 time-lags are excluded.\n",
"nitech_jp_song070_f001_023.lab: Global offset (in sec): -0.034999999999999996\n",
"nitech_jp_song070_f001_023.lab: 6/129 time-lags are excluded.\n",
"nitech_jp_song070_f001_025.lab: 20/137 time-lags are excluded.\n",
"nitech_jp_song070_f001_028.lab: Global offset (in sec): -0.055\n",
"nitech_jp_song070_f001_028.lab: 9/86 time-lags are excluded.\n",
"nitech_jp_song070_f001_029.lab: Global offset (in sec): -0.04\n",
"nitech_jp_song070_f001_029.lab: 3/55 time-lags are excluded.\n",
"nitech_jp_song070_f001_030.lab: Global offset (in sec): -0.01\n",
"nitech_jp_song070_f001_030.lab: 21/131 time-lags are excluded.\n",
"nitech_jp_song070_f001_037.lab: Global offset (in sec): -0.045\n",
"nitech_jp_song070_f001_037.lab: 21/187 time-lags are excluded.\n",
"nitech_jp_song070_f001_039.lab: Global offset (in sec): -0.075\n",
"nitech_jp_song070_f001_039.lab: 15/197 time-lags are excluded.\n",
"nitech_jp_song070_f001_040.lab: Global offset (in sec): -0.015\n",
"nitech_jp_song070_f001_040.lab: 11/148 time-lags are excluded.\n",
"nitech_jp_song070_f001_041.lab: Global offset (in sec): -0.08499999999999999\n",
"nitech_jp_song070_f001_041.lab: 9/251 time-lags are excluded.\n",
"nitech_jp_song070_f001_045.lab: Global offset (in sec): -0.055\n",
"nitech_jp_song070_f001_045.lab: 14/329 time-lags are excluded.\n",
"nitech_jp_song070_f001_048.lab: Global offset (in sec): -0.049999999999999996\n",
"nitech_jp_song070_f001_048.lab: 9/241 time-lags are excluded.\n",
"nitech_jp_song070_f001_050.lab: Global offset (in sec): -0.02\n",
"nitech_jp_song070_f001_050.lab: 2/90 time-lags are excluded.\n",
"nitech_jp_song070_f001_051.lab: Global offset (in sec): -0.065\n",
"nitech_jp_song070_f001_051.lab: 5/117 time-lags are excluded.\n",
"nitech_jp_song070_f001_054.lab: Global offset (in sec): -0.03\n",
"nitech_jp_song070_f001_054.lab: 6/55 time-lags are excluded.\n",
"nitech_jp_song070_f001_055.lab: Global offset (in sec): -0.06\n",
"nitech_jp_song070_f001_055.lab: 3/108 time-lags are excluded.\n",
"nitech_jp_song070_f001_059.lab: Global offset (in sec): 0.024999999999999998\n",
"nitech_jp_song070_f001_059.lab: 12/60 time-lags are excluded.\n",
"nitech_jp_song070_f001_060.lab: Global offset (in sec): -0.03\n",
"nitech_jp_song070_f001_060.lab: 5/118 time-lags are excluded.\n",
"nitech_jp_song070_f001_063.lab: Global offset (in sec): -0.08499999999999999\n",
"nitech_jp_song070_f001_063.lab: 4/38 time-lags are excluded.\n",
"nitech_jp_song070_f001_070.lab: Global offset (in sec): -0.02\n",
"nitech_jp_song070_f001_070.lab: 3/31 time-lags are excluded.\n",
"Prepare data for duration models\n",
"Prepare data for acoustic models\n",
"train/dev/eval split\n",
"stage 1: Feature generation\n",
"++ nnsvs-prepare-features utt_list=data/list/train_no_dev.list out_dir=dump/yoko/org/train_no_dev/ question_path=../../_common/hed/jp_qst001_nnsvs.hed timelag=defaults duration=defaults acoustic=static_deltadelta\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:26:46,973\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"utt_list: data/list/train_no_dev.list\n",
"out_dir: dump/yoko/org/train_no_dev/\n",
"question_path: ../../_common/hed/jp_qst001_nnsvs.hed\n",
"log_f0_conditioning: true\n",
"timelag:\n",
" enabled: true\n",
" question_path: null\n",
" label_phone_score_dir: data/timelag/label_phone_score\n",
" label_phone_align_dir: data/timelag/label_phone_align\n",
"duration:\n",
" enabled: true\n",
" question_path: null\n",
" label_dir: data/duration/label_phone_align\n",
"acoustic:\n",
" enabled: true\n",
" question_path: null\n",
" wav_dir: data/acoustic/wav\n",
" label_dir: data/acoustic/label_phone_align\n",
" subphone_features: coarse_coding\n",
" f0_floor: 150\n",
" f0_ceil: 700\n",
" use_harvest: true\n",
" frame_period: 5\n",
" mgc_order: 59\n",
" num_windows: 3\n",
" relative_f0: true\n",
" interp_unvoiced_aperiodicity: true\n",
"\u001b[0m\n",
"[\u001b[36m2020-11-08 06:26:47,078\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/train_no_dev/in_timelag\u001b[0m\n",
"[\u001b[36m2020-11-08 06:26:47,079\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/train_no_dev/out_timelag\u001b[0m\n",
"[\u001b[36m2020-11-08 06:26:47,079\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/train_no_dev/in_duration\u001b[0m\n",
"[\u001b[36m2020-11-08 06:26:47,079\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/train_no_dev/out_duration\u001b[0m\n",
"[\u001b[36m2020-11-08 06:26:47,079\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/train_no_dev/in_acoustic\u001b[0m\n",
"[\u001b[36m2020-11-08 06:26:47,080\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/train_no_dev/out_acoustic\u001b[0m\n",
"[\u001b[36m2020-11-08 06:26:47,114\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Timelag linguistic feature dim: 420\u001b[0m\n",
"[\u001b[36m2020-11-08 06:26:47,115\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Timelag feature dim: 1\u001b[0m\n",
"100% 29/29 [00:01<00:00, 15.65it/s]\n",
"[\u001b[36m2020-11-08 06:26:49,039\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Duration linguistic feature dim: 420\u001b[0m\n",
"[\u001b[36m2020-11-08 06:26:49,040\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Duration feature dim: 1\u001b[0m\n",
"100% 29/29 [00:03<00:00, 8.28it/s]\n",
"[\u001b[36m2020-11-08 06:26:52,674\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Acoustic linguistic feature dim: 424\u001b[0m\n",
"[\u001b[36m2020-11-08 06:27:08,979\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Acoustic feature dim: 199\u001b[0m\n",
"100% 29/29 [11:38<00:00, 24.10s/it]\n",
"++ set +x\n",
"++ nnsvs-prepare-features utt_list=data/list/dev.list out_dir=dump/yoko/org/dev/ question_path=../../_common/hed/jp_qst001_nnsvs.hed timelag=defaults duration=defaults acoustic=static_deltadelta\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:38:50,292\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"utt_list: data/list/dev.list\n",
"out_dir: dump/yoko/org/dev/\n",
"question_path: ../../_common/hed/jp_qst001_nnsvs.hed\n",
"log_f0_conditioning: true\n",
"timelag:\n",
" enabled: true\n",
" question_path: null\n",
" label_phone_score_dir: data/timelag/label_phone_score\n",
" label_phone_align_dir: data/timelag/label_phone_align\n",
"duration:\n",
" enabled: true\n",
" question_path: null\n",
" label_dir: data/duration/label_phone_align\n",
"acoustic:\n",
" enabled: true\n",
" question_path: null\n",
" wav_dir: data/acoustic/wav\n",
" label_dir: data/acoustic/label_phone_align\n",
" subphone_features: coarse_coding\n",
" f0_floor: 150\n",
" f0_ceil: 700\n",
" use_harvest: true\n",
" frame_period: 5\n",
" mgc_order: 59\n",
" num_windows: 3\n",
" relative_f0: true\n",
" interp_unvoiced_aperiodicity: true\n",
"\u001b[0m\n",
"[\u001b[36m2020-11-08 06:38:50,379\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/dev/in_timelag\u001b[0m\n",
"[\u001b[36m2020-11-08 06:38:50,379\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/dev/out_timelag\u001b[0m\n",
"[\u001b[36m2020-11-08 06:38:50,380\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/dev/in_duration\u001b[0m\n",
"[\u001b[36m2020-11-08 06:38:50,380\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/dev/out_duration\u001b[0m\n",
"[\u001b[36m2020-11-08 06:38:50,380\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/dev/in_acoustic\u001b[0m\n",
"[\u001b[36m2020-11-08 06:38:50,380\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/dev/out_acoustic\u001b[0m\n",
"[\u001b[36m2020-11-08 06:38:50,418\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Timelag linguistic feature dim: 420\u001b[0m\n",
"[\u001b[36m2020-11-08 06:38:50,418\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Timelag feature dim: 1\u001b[0m\n",
"100% 1/1 [00:00<00:00, 27.36it/s]\n",
"[\u001b[36m2020-11-08 06:38:50,525\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Duration linguistic feature dim: 420\u001b[0m\n",
"[\u001b[36m2020-11-08 06:38:50,525\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Duration feature dim: 1\u001b[0m\n",
"100% 1/1 [00:00<00:00, 13.14it/s]\n",
"[\u001b[36m2020-11-08 06:38:50,709\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Acoustic linguistic feature dim: 424\u001b[0m\n",
"[\u001b[36m2020-11-08 06:38:59,430\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Acoustic feature dim: 199\u001b[0m\n",
"100% 1/1 [00:08<00:00, 8.84s/it]\n",
"++ set +x\n",
"++ nnsvs-prepare-features utt_list=data/list/eval.list out_dir=dump/yoko/org/eval/ question_path=../../_common/hed/jp_qst001_nnsvs.hed timelag=defaults duration=defaults acoustic=static_deltadelta\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:10,287\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"utt_list: data/list/eval.list\n",
"out_dir: dump/yoko/org/eval/\n",
"question_path: ../../_common/hed/jp_qst001_nnsvs.hed\n",
"log_f0_conditioning: true\n",
"timelag:\n",
" enabled: true\n",
" question_path: null\n",
" label_phone_score_dir: data/timelag/label_phone_score\n",
" label_phone_align_dir: data/timelag/label_phone_align\n",
"duration:\n",
" enabled: true\n",
" question_path: null\n",
" label_dir: data/duration/label_phone_align\n",
"acoustic:\n",
" enabled: true\n",
" question_path: null\n",
" wav_dir: data/acoustic/wav\n",
" label_dir: data/acoustic/label_phone_align\n",
" subphone_features: coarse_coding\n",
" f0_floor: 150\n",
" f0_ceil: 700\n",
" use_harvest: true\n",
" frame_period: 5\n",
" mgc_order: 59\n",
" num_windows: 3\n",
" relative_f0: true\n",
" interp_unvoiced_aperiodicity: true\n",
"\u001b[0m\n",
"[\u001b[36m2020-11-08 06:39:10,381\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/eval/in_timelag\u001b[0m\n",
"[\u001b[36m2020-11-08 06:39:10,381\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/eval/out_timelag\u001b[0m\n",
"[\u001b[36m2020-11-08 06:39:10,381\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/eval/in_duration\u001b[0m\n",
"[\u001b[36m2020-11-08 06:39:10,381\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/eval/out_duration\u001b[0m\n",
"[\u001b[36m2020-11-08 06:39:10,381\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/eval/in_acoustic\u001b[0m\n",
"[\u001b[36m2020-11-08 06:39:10,382\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - mkdirs: /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/dump/yoko/org/eval/out_acoustic\u001b[0m\n",
"[\u001b[36m2020-11-08 06:39:10,426\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Timelag linguistic feature dim: 420\u001b[0m\n",
"[\u001b[36m2020-11-08 06:39:10,427\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Timelag feature dim: 1\u001b[0m\n",
"100% 1/1 [00:00<00:00, 20.03it/s]\n",
"[\u001b[36m2020-11-08 06:39:10,557\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Duration linguistic feature dim: 420\u001b[0m\n",
"[\u001b[36m2020-11-08 06:39:10,557\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Duration feature dim: 1\u001b[0m\n",
"100% 1/1 [00:00<00:00, 12.65it/s]\n",
"[\u001b[36m2020-11-08 06:39:10,756\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Acoustic linguistic feature dim: 424\u001b[0m\n",
"[\u001b[36m2020-11-08 06:39:21,460\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Acoustic feature dim: 199\u001b[0m\n",
"100% 1/1 [00:10<00:00, 10.72s/it]\n",
"++ set +x\n",
"++ nnsvs-fit-scaler list_path=train_list.txt scaler.class=sklearn.preprocessing.MinMaxScaler out_path=dump/yoko/org/in_timelag_scaler.joblib\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[2020-11-08 06:39:33,168][nnsvs][INFO] - verbose: 100\n",
"scaler:\n",
" class: sklearn.preprocessing.MinMaxScaler\n",
" params: {}\n",
"list_path: train_list.txt\n",
"out_path: dump/yoko/org/in_timelag_scaler.joblib\n",
"\n",
"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py:632: UserWarning: \n",
"Config key 'class' is deprecated since Hydra 1.0 and will be removed in Hydra 1.1.\n",
"Use '_target_' instead of 'class'.\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/object_instantiation_changes\n",
" warnings.warn(message=msg, category=UserWarning)\n",
"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py:577: UserWarning: \n",
"Field 'params' is deprecated since Hydra 1.0 and will be removed in Hydra 1.1.\n",
"Inline the content of params directly at the containing node.\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/object_instantiation_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"[2020-11-08 06:39:33,208][nnsvs][INFO] - data min:\n",
"[ 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 5.3936276 5.3936276 5.3936276\n",
" 1. 1. -1. -1. -1. -1.\n",
" -1. 1. 1. 1. -1. -1.\n",
" -1. -1. -1. -1. -1. -1.\n",
" -1. 0. 1. 15. 6. 0.\n",
" 0. -1. -1. -1. -1. -1.\n",
" -1. -1. -1. -1. -1. -1.\n",
" -1. -1. -1. -1. -1. -1.\n",
" -1. -1. -1. -1. -1. -1.\n",
" -1. -1. -1. -1. -1. -12.\n",
" -12. -1. -1. -1. -1. -1. ]\n",
"[2020-11-08 06:39:33,211][nnsvs][INFO] - data max:\n",
"[ 1. 1. 0. 1. 1. 1.\n",
" 0. 1. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 1. 1.\n",
" 1. 1. 0. 0. 1. 0.\n",
" 0. 0. 0. 0. 1. 0.\n",
" 0. 0. 0. 0. 1. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 1.\n",
" 0. 0. 1. 0. 1. 0.\n",
" 0. 1. 0. 0. 1. 0.\n",
" 0. 0. 0. 0. 1. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 1. 0. 1. 0. 0. 0.\n",
" 0. 0. 1. 0. 0. 0.\n",
" 1. 0. 0. 0. 0. 1.\n",
" 1. 1. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 0. 0. 0. 1.\n",
" 0. 0. 1. 1. 0. 0.\n",
" 1. 0. 1. 0. 1. 1.\n",
" 1. 0. 1. 0. 1. 1.\n",
" 1. 1. 1. 0. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 0. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 0. 0. 0. 1.\n",
" 0. 0. 1. 1. 1. 0.\n",
" 1. 1. 1. 0. 1. 1.\n",
" 1. 0. 1. 0. 1. 1.\n",
" 1. 1. 1. 0. 1. 0.\n",
" 1. 0. 1. 1. 1. 0.\n",
" 1. 1. 0. 1. 1. 1.\n",
" 1. 0. 1. 1. 1. 0.\n",
" 0. 1. 0. 0. 0. 0.\n",
" 0. 0. 1. 0. 0. 0.\n",
" 0. 0. 0. 1. 0. 0.\n",
" 0. 0. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 0.\n",
" 0. 1. 1. 1. 0. 1.\n",
" 1. 0. 0. 1. 1. 1.\n",
" 0. 1. 1. 0. 0. 0.\n",
" 0. 1. 1. 1. 1. 1.\n",
" 1. 0. 0. 1. 1. 1.\n",
" 0. 1. 1. 6.491111 6.491111 6.491111\n",
" 1. 3. -1. 1. 3. 2.\n",
" 1. 3. 1. 2. 3. 2.\n",
" 2. 11. 11. 2. 240. 96.\n",
" 11. 11. 2. 390. 156. 0.\n",
" 0. 9. 1. 99. 6. 99.\n",
" 24. 9. 9. 99. 99. 99.\n",
" 99. 12. 12. 50. 53. 204.\n",
" 216. 95. 100. 9. 9. 42.\n",
" 54. 168. 216. 87. 100. 12.\n",
" 12. 11. 11. 2. 180. 72. ]\n",
"++ set +x\n",
"'dump/yoko/org/in_timelag_scaler.joblib' -> 'dump/yoko/norm/in_timelag_scaler.joblib'\n",
"++ nnsvs-fit-scaler list_path=train_list.txt scaler.class=sklearn.preprocessing.MinMaxScaler out_path=dump/yoko/org/in_duration_scaler.joblib\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[2020-11-08 06:39:34,073][nnsvs][INFO] - verbose: 100\n",
"scaler:\n",
" class: sklearn.preprocessing.MinMaxScaler\n",
" params: {}\n",
"list_path: train_list.txt\n",
"out_path: dump/yoko/org/in_duration_scaler.joblib\n",
"\n",
"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py:632: UserWarning: \n",
"Config key 'class' is deprecated since Hydra 1.0 and will be removed in Hydra 1.1.\n",
"Use '_target_' instead of 'class'.\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/object_instantiation_changes\n",
" warnings.warn(message=msg, category=UserWarning)\n",
"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py:577: UserWarning: \n",
"Field 'params' is deprecated since Hydra 1.0 and will be removed in Hydra 1.1.\n",
"Inline the content of params directly at the containing node.\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/object_instantiation_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"[2020-11-08 06:39:34,121][nnsvs][INFO] - data min:\n",
"[ 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 5.3936276 5.3936276 5.3936276\n",
" 1. 1. -1. -1. -1. -1.\n",
" -1. 1. 1. 1. -1. -1.\n",
" -1. -1. -1. -1. -1. -1.\n",
" -1. 0. 1. 15. 6. 0.\n",
" 0. -1. -1. -1. -1. -1.\n",
" -1. -1. -1. -1. -1. -1.\n",
" -1. -1. -1. -1. -1. -1.\n",
" -1. -1. -1. -1. -1. -1.\n",
" -1. -1. -1. -1. -1. -12.\n",
" -12. -1. -1. -1. -1. -1. ]\n",
"[2020-11-08 06:39:34,124][nnsvs][INFO] - data max:\n",
"[ 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 0. 0. 0. 1.\n",
" 0. 0. 1. 1. 1. 0.\n",
" 1. 1. 1. 0. 1. 1.\n",
" 1. 0. 1. 0. 1. 1.\n",
" 1. 1. 1. 0. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 0. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 0. 0. 0. 1.\n",
" 0. 0. 1. 1. 1. 0.\n",
" 1. 1. 1. 0. 1. 1.\n",
" 1. 0. 1. 0. 1. 1.\n",
" 1. 1. 1. 0. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 0. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 0. 0. 0. 1.\n",
" 0. 0. 1. 1. 1. 0.\n",
" 1. 1. 1. 0. 1. 1.\n",
" 1. 0. 1. 0. 1. 1.\n",
" 1. 1. 1. 0. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 1.\n",
" 1. 0. 1. 1. 1. 0.\n",
" 0. 1. 0. 0. 0. 0.\n",
" 0. 0. 1. 0. 0. 0.\n",
" 0. 0. 0. 1. 0. 0.\n",
" 0. 0. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1. 0.\n",
" 0. 1. 1. 1. 0. 1.\n",
" 1. 0. 0. 1. 1. 1.\n",
" 0. 1. 1. 0. 0. 0.\n",
" 0. 1. 1. 1. 1. 1.\n",
" 1. 0. 0. 1. 1. 1.\n",
" 0. 1. 1. 6.491111 6.491111 6.491111\n",
" 3. 3. -1. 1. 3. 2.\n",
" 2. 3. 2. 2. 3. 2.\n",
" 2. 11. 11. 2. 240. 96.\n",
" 11. 11. 2. 390. 156. 0.\n",
" 0. 9. 1. 99. 6. 99.\n",
" 24. 9. 9. 99. 99. 99.\n",
" 99. 12. 12. 50. 53. 204.\n",
" 216. 95. 100. 9. 9. 42.\n",
" 54. 168. 216. 87. 100. 12.\n",
" 12. 11. 11. 2. 240. 96. ]\n",
"++ set +x\n",
"'dump/yoko/org/in_duration_scaler.joblib' -> 'dump/yoko/norm/in_duration_scaler.joblib'\n",
"++ nnsvs-fit-scaler list_path=train_list.txt scaler.class=sklearn.preprocessing.MinMaxScaler out_path=dump/yoko/org/in_acoustic_scaler.joblib\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[2020-11-08 06:39:35,011][nnsvs][INFO] - verbose: 100\n",
"scaler:\n",
" class: sklearn.preprocessing.MinMaxScaler\n",
" params: {}\n",
"list_path: train_list.txt\n",
"out_path: dump/yoko/org/in_acoustic_scaler.joblib\n",
"\n",
"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py:632: UserWarning: \n",
"Config key 'class' is deprecated since Hydra 1.0 and will be removed in Hydra 1.1.\n",
"Use '_target_' instead of 'class'.\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/object_instantiation_changes\n",
" warnings.warn(message=msg, category=UserWarning)\n",
"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py:577: UserWarning: \n",
"Field 'params' is deprecated since Hydra 1.0 and will be removed in Hydra 1.1.\n",
"Inline the content of params directly at the containing node.\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/object_instantiation_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"[2020-11-08 06:39:35,635][nnsvs][INFO] - data min:\n",
"[ 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 0. 0. 0.\n",
" 0. 0. 5.3936276 5.3936276 5.3936276\n",
" 1. 1. -1. -1. -1.\n",
" -1. -1. 1. 1. 1.\n",
" -1. -1. -1. -1. -1.\n",
" -1. -1. -1. -1. 0.\n",
" 1. 15. 6. 0. 0.\n",
" -1. -1. -1. -1. -1.\n",
" -1. -1. -1. -1. -1.\n",
" -1. -1. -1. -1. -1.\n",
" -1. -1. -1. -1. -1.\n",
" -1. -1. -1. -1. -1.\n",
" -1. -1. -1. -12. -12.\n",
" -1. -1. -1. -1. -1.\n",
" 0.04404987 0.45900714 0.04404987 5. ]\n",
"[2020-11-08 06:39:35,638][nnsvs][INFO] - data max:\n",
"[ 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 0.\n",
" 1. 1. 0. 1. 1.\n",
" 0. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 0. 0. 0. 1.\n",
" 0. 0. 1. 1. 1.\n",
" 0. 1. 1. 1. 0.\n",
" 1. 1. 1. 0. 1.\n",
" 0. 1. 1. 1. 1.\n",
" 1. 0. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 0. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 0. 1. 1. 0.\n",
" 1. 1. 0. 1. 1.\n",
" 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 0. 0. 0.\n",
" 1. 0. 0. 1. 1.\n",
" 1. 0. 1. 1. 1.\n",
" 0. 1. 1. 1. 0.\n",
" 1. 0. 1. 1. 1.\n",
" 1. 1. 0. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 0. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 0. 1. 1.\n",
" 0. 1. 1. 0. 1.\n",
" 1. 0. 1. 1. 0.\n",
" 1. 1. 0. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 0. 0.\n",
" 0. 1. 0. 0. 1.\n",
" 1. 1. 0. 1. 1.\n",
" 1. 0. 1. 1. 1.\n",
" 0. 1. 0. 1. 1.\n",
" 1. 1. 1. 0. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 1. 0.\n",
" 1. 1. 1. 0. 0.\n",
" 1. 0. 0. 0. 0.\n",
" 0. 0. 1. 0. 0.\n",
" 0. 0. 0. 0. 1.\n",
" 0. 0. 0. 0. 1.\n",
" 1. 1. 1. 1. 1.\n",
" 1. 1. 1. 0. 0.\n",
" 1. 1. 1. 0. 1.\n",
" 1. 0. 0. 1. 1.\n",
" 1. 0. 1. 1. 0.\n",
" 0. 0. 0. 1. 1.\n",
" 1. 1. 1. 1. 0.\n",
" 0. 1. 1. 1. 0.\n",
" 1. 1. 6.491111 6.491111 6.491111\n",
" 3. 3. -1. 1. 3.\n",
" 2. 2. 3. 2. 2.\n",
" 3. 2. 2. 11. 11.\n",
" 2. 240. 96. 11. 11.\n",
" 2. 390. 156. 0. 0.\n",
" 9. 1. 99. 6. 99.\n",
" 24. 9. 9. 99. 99.\n",
" 99. 99. 12. 12. 50.\n",
" 53. 204. 216. 95. 100.\n",
" 9. 9. 42. 54. 168.\n",
" 216. 87. 100. 12. 12.\n",
" 11. 11. 2. 240. 96.\n",
" 0.99733615 0.99733615 0.99733615 786. ]\n",
"++ set +x\n",
"'dump/yoko/org/in_acoustic_scaler.joblib' -> 'dump/yoko/norm/in_acoustic_scaler.joblib'\n",
"++ nnsvs-fit-scaler list_path=train_list.txt scaler.class=sklearn.preprocessing.StandardScaler out_path=dump/yoko/org/out_timelag_scaler.joblib\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[2020-11-08 06:39:36,495][nnsvs][INFO] - verbose: 100\n",
"scaler:\n",
" class: sklearn.preprocessing.StandardScaler\n",
" params: {}\n",
"list_path: train_list.txt\n",
"out_path: dump/yoko/org/out_timelag_scaler.joblib\n",
"\n",
"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py:632: UserWarning: \n",
"Config key 'class' is deprecated since Hydra 1.0 and will be removed in Hydra 1.1.\n",
"Use '_target_' instead of 'class'.\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/object_instantiation_changes\n",
" warnings.warn(message=msg, category=UserWarning)\n",
"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py:577: UserWarning: \n",
"Field 'params' is deprecated since Hydra 1.0 and will be removed in Hydra 1.1.\n",
"Inline the content of params directly at the containing node.\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/object_instantiation_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"[2020-11-08 06:39:36,530][nnsvs][INFO] - mean:\n",
"[-0.45280495]\n",
"[2020-11-08 06:39:36,530][nnsvs][INFO] - std:\n",
"[9.00099378]\n",
"++ set +x\n",
"'dump/yoko/org/out_timelag_scaler.joblib' -> 'dump/yoko/norm/out_timelag_scaler.joblib'\n",
"++ nnsvs-fit-scaler list_path=train_list.txt scaler.class=sklearn.preprocessing.StandardScaler out_path=dump/yoko/org/out_duration_scaler.joblib\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[2020-11-08 06:39:37,383][nnsvs][INFO] - verbose: 100\n",
"scaler:\n",
" class: sklearn.preprocessing.StandardScaler\n",
" params: {}\n",
"list_path: train_list.txt\n",
"out_path: dump/yoko/org/out_duration_scaler.joblib\n",
"\n",
"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py:632: UserWarning: \n",
"Config key 'class' is deprecated since Hydra 1.0 and will be removed in Hydra 1.1.\n",
"Use '_target_' instead of 'class'.\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/object_instantiation_changes\n",
" warnings.warn(message=msg, category=UserWarning)\n",
"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py:577: UserWarning: \n",
"Field 'params' is deprecated since Hydra 1.0 and will be removed in Hydra 1.1.\n",
"Inline the content of params directly at the containing node.\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/object_instantiation_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"[2020-11-08 06:39:37,418][nnsvs][INFO] - mean:\n",
"[56.6853255]\n",
"[2020-11-08 06:39:37,419][nnsvs][INFO] - std:\n",
"[63.38523054]\n",
"++ set +x\n",
"'dump/yoko/org/out_duration_scaler.joblib' -> 'dump/yoko/norm/out_duration_scaler.joblib'\n",
"++ nnsvs-fit-scaler list_path=train_list.txt scaler.class=sklearn.preprocessing.StandardScaler out_path=dump/yoko/org/out_acoustic_scaler.joblib\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[2020-11-08 06:39:38,269][nnsvs][INFO] - verbose: 100\n",
"scaler:\n",
" class: sklearn.preprocessing.StandardScaler\n",
" params: {}\n",
"list_path: train_list.txt\n",
"out_path: dump/yoko/org/out_acoustic_scaler.joblib\n",
"\n",
"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py:632: UserWarning: \n",
"Config key 'class' is deprecated since Hydra 1.0 and will be removed in Hydra 1.1.\n",
"Use '_target_' instead of 'class'.\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/object_instantiation_changes\n",
" warnings.warn(message=msg, category=UserWarning)\n",
"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py:577: UserWarning: \n",
"Field 'params' is deprecated since Hydra 1.0 and will be removed in Hydra 1.1.\n",
"Inline the content of params directly at the containing node.\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/object_instantiation_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"[2020-11-08 06:39:39,145][nnsvs][INFO] - mean:\n",
"[ 4.78342476e+00 2.74917070e+00 5.12944481e-03 3.87839949e-01\n",
" 1.79770210e-01 5.02097549e-01 1.09037803e-01 -3.69180404e-01\n",
" -1.12757253e-01 4.07469283e-01 -5.19477447e-02 -5.69876404e-02\n",
" -6.74684293e-02 4.76904200e-02 1.59106203e-02 -1.03088850e-01\n",
" -5.60067584e-03 2.36715079e-02 -1.77116500e-02 -6.84369415e-02\n",
" 1.59850589e-02 5.95883622e-02 -1.06173976e-01 -8.29753426e-03\n",
" 3.75368433e-02 2.85884018e-02 -2.87232162e-02 -2.59581970e-03\n",
" 1.43792494e-02 2.34217371e-02 -3.46846127e-02 6.57554681e-03\n",
" 2.45415648e-02 -1.85503584e-02 -1.00396884e-02 2.93056715e-02\n",
" -2.86606153e-02 2.84990011e-02 -2.64671161e-02 1.62423030e-02\n",
" -5.62492999e-03 5.96119650e-03 -1.16559812e-02 1.41650105e-02\n",
" -7.80952468e-03 4.07801437e-04 3.37590994e-03 -1.45775688e-03\n",
" -4.72142315e-03 1.10249554e-02 -1.55241146e-02 1.51586779e-02\n",
" -1.04680051e-02 4.89908085e-03 -4.40329662e-03 6.77060364e-03\n",
" -9.32549976e-03 9.72182370e-03 -8.79440886e-03 5.91195448e-03\n",
" 1.84514589e-05 3.95044097e-06 -2.28347251e-06 -3.41902406e-06\n",
" -3.37611532e-06 -1.80415422e-06 -3.43702772e-07 2.02945576e-06\n",
" 4.26354687e-07 8.24248608e-07 2.55353231e-06 9.18584668e-07\n",
" 1.21511583e-07 1.16102102e-06 -1.39680575e-06 -2.58783959e-06\n",
" 3.54993308e-07 1.06804218e-07 5.52188839e-07 -2.85887659e-07\n",
" -5.87276698e-07 1.13786274e-06 -4.20122333e-07 -6.86834775e-07\n",
" 6.52838214e-07 4.31229333e-07 -9.10151349e-07 -4.44298668e-08\n",
" 9.83112027e-07 -8.79199653e-07 4.29362409e-07 -1.80560223e-07\n",
" 1.72436753e-08 2.18989844e-07 -4.47299293e-07 4.16705731e-07\n",
" -6.87272169e-08 2.32123230e-08 -6.09104782e-07 1.05671025e-06\n",
" -7.03253721e-07 -6.75806354e-08 6.31020456e-07 -4.43689651e-07\n",
" 3.91117472e-08 2.43488991e-07 -3.40813530e-07 1.38948634e-07\n",
" 9.51343442e-08 -4.01862438e-07 6.19363023e-07 -7.15181136e-07\n",
" 7.25077747e-07 -5.92106741e-07 3.17084086e-07 7.25970745e-09\n",
" -2.37505339e-07 3.96175724e-07 -3.90214804e-07 1.77717670e-07\n",
" -1.37776567e-04 -1.29514985e-04 -7.51909570e-05 -5.63619473e-05\n",
" -4.31658714e-05 -3.96452713e-05 -3.29980932e-05 -2.52508901e-05\n",
" -2.11055235e-05 -1.59377686e-05 -8.02098532e-06 -3.93103627e-06\n",
" 2.14098375e-06 2.37650750e-06 1.03344016e-05 5.21869308e-06\n",
" 3.98872382e-06 5.69920779e-06 3.97810054e-06 4.45114275e-06\n",
" 4.51568467e-06 8.90029898e-07 1.10136751e-06 1.07411846e-06\n",
" 4.64741150e-08 -2.59378207e-06 4.64360583e-06 -1.53321347e-06\n",
" -1.92082370e-06 1.74312744e-06 6.69793052e-07 -3.07256990e-06\n",
" 4.29837843e-06 -2.98688853e-06 1.40043280e-06 -6.54190444e-08\n",
" -5.04659863e-07 -3.68352828e-07 7.12388865e-07 -1.95612580e-06\n",
" 1.20694091e-06 -2.06889190e-06 1.38537713e-06 -2.13890317e-06\n",
" 1.65125445e-06 -1.78639262e-06 1.26129285e-06 -1.12448021e-06\n",
" 1.20838176e-06 -1.15571979e-06 9.58750724e-07 -3.69277872e-07\n",
" -2.70407908e-08 2.86179946e-07 2.01399983e-07 -6.27553551e-07\n",
" 7.42895229e-07 -7.81915737e-07 7.85378220e-07 -8.83287144e-07\n",
" -3.14993121e-02 -1.29213667e-08 1.00635497e-05 7.93921721e-01\n",
" -1.44323971e+01 -1.00994752e+01 -5.43791130e+00 -3.26144403e+00\n",
" -5.64111166e+00 8.00034804e-06 4.03115374e-07 1.08807099e-06\n",
" 1.42748894e-05 2.56005217e-05 3.40196748e-04 4.11871934e-04\n",
" 3.50853310e-04 4.19679208e-04 3.63122763e-04]\n",
"[2020-11-08 06:39:39,147][nnsvs][INFO] - std:\n",
"[1.73399012 1.01464088 0.47228913 0.34532917 0.34103896 0.33398652\n",
" 0.35581104 0.42783709 0.27665228 0.27945993 0.21285026 0.16957813\n",
" 0.17424989 0.2066463 0.17718029 0.15568826 0.1408696 0.13214229\n",
" 0.13318904 0.12406065 0.13132254 0.12752999 0.11620726 0.11531741\n",
" 0.10502666 0.09521375 0.09558436 0.09274859 0.08921838 0.08453486\n",
" 0.08322494 0.08569423 0.08540422 0.08820595 0.08498066 0.08221735\n",
" 0.08252367 0.08023513 0.08018691 0.07797202 0.07668795 0.07128373\n",
" 0.06724783 0.06460587 0.0613163 0.05933368 0.05858136 0.0576386\n",
" 0.05761079 0.05688922 0.05528237 0.05334595 0.05184026 0.05065138\n",
" 0.0499082 0.04945128 0.04908737 0.0482157 0.04696458 0.04579251\n",
" 0.21010084 0.13759226 0.10509428 0.08157427 0.07531109 0.070916\n",
" 0.07051921 0.07388144 0.06462866 0.06394295 0.06065996 0.05723132\n",
" 0.05802181 0.05565779 0.05440932 0.05284463 0.05131 0.0496664\n",
" 0.04821025 0.04669917 0.04596433 0.04415716 0.0433885 0.04195605\n",
" 0.04093269 0.03949937 0.03880426 0.0378889 0.03691353 0.03580254\n",
" 0.03513868 0.03423714 0.03360242 0.03316704 0.03261716 0.03191549\n",
" 0.03144512 0.03097298 0.03047856 0.02999397 0.02950807 0.02889084\n",
" 0.02833144 0.02784387 0.02743354 0.0270057 0.02660804 0.02616997\n",
" 0.02578787 0.02544988 0.02513642 0.02477436 0.02440032 0.02410925\n",
" 0.02385252 0.02355782 0.023292 0.02299542 0.02265909 0.02231298\n",
" 0.37183484 0.27290992 0.25325313 0.22046325 0.2090978 0.20058698\n",
" 0.19726328 0.19862739 0.18772269 0.18299982 0.17638277 0.16827651\n",
" 0.16929123 0.16499408 0.16099641 0.15591313 0.15103185 0.1460962\n",
" 0.14211302 0.13723963 0.13474449 0.12966722 0.12682478 0.12350306\n",
" 0.12059602 0.11723613 0.11473051 0.11191842 0.10967378 0.10689813\n",
" 0.10463616 0.1023561 0.10069151 0.09898355 0.09711368 0.09507831\n",
" 0.09373526 0.09225308 0.09099764 0.08944663 0.08827378 0.08675291\n",
" 0.08527685 0.08392321 0.08300574 0.08203382 0.0808053 0.07960657\n",
" 0.0786723 0.07762875 0.07658933 0.07569998 0.07479483 0.0741078\n",
" 0.07321134 0.07220156 0.07134968 0.07064027 0.06978393 0.0688563\n",
" 0.09970091 0.01687954 0.02238207 0.40448736 8.19504634 6.8844785\n",
" 4.13302028 2.3768468 4.73236516 1.6059199 1.85620703 1.84591274\n",
" 1.47699348 1.92961435 3.85211018 5.33611497 5.86728126 4.99254793\n",
" 6.0215684 ]\n",
"++ set +x\n",
"'dump/yoko/org/out_acoustic_scaler.joblib' -> 'dump/yoko/norm/out_acoustic_scaler.joblib'\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/train_no_dev/in_timelag/ scaler_path=dump/yoko/org/in_timelag_scaler.joblib out_dir=dump/yoko/norm/train_no_dev/in_timelag/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:40,046\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/train_no_dev/in_timelag/\n",
"out_dir: dump/yoko/norm/train_no_dev/in_timelag/\n",
"scaler_path: dump/yoko/org/in_timelag_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 29/29 [00:00<00:00, 441.51it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/train_no_dev/in_duration/ scaler_path=dump/yoko/org/in_duration_scaler.joblib out_dir=dump/yoko/norm/train_no_dev/in_duration/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:41,027\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/train_no_dev/in_duration/\n",
"out_dir: dump/yoko/norm/train_no_dev/in_duration/\n",
"scaler_path: dump/yoko/org/in_duration_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 29/29 [00:00<00:00, 416.27it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/train_no_dev/in_acoustic/ scaler_path=dump/yoko/org/in_acoustic_scaler.joblib out_dir=dump/yoko/norm/train_no_dev/in_acoustic/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:42,047\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/train_no_dev/in_acoustic/\n",
"out_dir: dump/yoko/norm/train_no_dev/in_acoustic/\n",
"scaler_path: dump/yoko/org/in_acoustic_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 29/29 [00:01<00:00, 14.55it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/train_no_dev/out_timelag/ scaler_path=dump/yoko/org/out_timelag_scaler.joblib out_dir=dump/yoko/norm/train_no_dev/out_timelag/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:45,117\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/train_no_dev/out_timelag/\n",
"out_dir: dump/yoko/norm/train_no_dev/out_timelag/\n",
"scaler_path: dump/yoko/org/out_timelag_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 29/29 [00:00<00:00, 661.72it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/train_no_dev/out_duration/ scaler_path=dump/yoko/org/out_duration_scaler.joblib out_dir=dump/yoko/norm/train_no_dev/out_duration/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:46,071\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/train_no_dev/out_duration/\n",
"out_dir: dump/yoko/norm/train_no_dev/out_duration/\n",
"scaler_path: dump/yoko/org/out_duration_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 29/29 [00:00<00:00, 672.32it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/train_no_dev/out_acoustic/ scaler_path=dump/yoko/org/out_acoustic_scaler.joblib out_dir=dump/yoko/norm/train_no_dev/out_acoustic/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:47,047\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/train_no_dev/out_acoustic/\n",
"out_dir: dump/yoko/norm/train_no_dev/out_acoustic/\n",
"scaler_path: dump/yoko/org/out_acoustic_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 29/29 [00:01<00:00, 14.62it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/dev/in_timelag/ scaler_path=dump/yoko/org/in_timelag_scaler.joblib out_dir=dump/yoko/norm/dev/in_timelag/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:50,138\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/dev/in_timelag/\n",
"out_dir: dump/yoko/norm/dev/in_timelag/\n",
"scaler_path: dump/yoko/org/in_timelag_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 104.69it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/dev/in_duration/ scaler_path=dump/yoko/org/in_duration_scaler.joblib out_dir=dump/yoko/norm/dev/in_duration/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:51,095\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/dev/in_duration/\n",
"out_dir: dump/yoko/norm/dev/in_duration/\n",
"scaler_path: dump/yoko/org/in_duration_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 126.64it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/dev/in_acoustic/ scaler_path=dump/yoko/org/in_acoustic_scaler.joblib out_dir=dump/yoko/norm/dev/in_acoustic/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:52,032\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/dev/in_acoustic/\n",
"out_dir: dump/yoko/norm/dev/in_acoustic/\n",
"scaler_path: dump/yoko/org/in_acoustic_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 37.69it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/dev/out_timelag/ scaler_path=dump/yoko/org/out_timelag_scaler.joblib out_dir=dump/yoko/norm/dev/out_timelag/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:53,018\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/dev/out_timelag/\n",
"out_dir: dump/yoko/norm/dev/out_timelag/\n",
"scaler_path: dump/yoko/org/out_timelag_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 118.30it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/dev/out_duration/ scaler_path=dump/yoko/org/out_duration_scaler.joblib out_dir=dump/yoko/norm/dev/out_duration/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:53,952\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/dev/out_duration/\n",
"out_dir: dump/yoko/norm/dev/out_duration/\n",
"scaler_path: dump/yoko/org/out_duration_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 127.79it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/dev/out_acoustic/ scaler_path=dump/yoko/org/out_acoustic_scaler.joblib out_dir=dump/yoko/norm/dev/out_acoustic/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:54,885\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/dev/out_acoustic/\n",
"out_dir: dump/yoko/norm/dev/out_acoustic/\n",
"scaler_path: dump/yoko/org/out_acoustic_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 36.69it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/eval/in_timelag/ scaler_path=dump/yoko/org/in_timelag_scaler.joblib out_dir=dump/yoko/norm/eval/in_timelag/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:55,835\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/eval/in_timelag/\n",
"out_dir: dump/yoko/norm/eval/in_timelag/\n",
"scaler_path: dump/yoko/org/in_timelag_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 122.59it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/eval/in_duration/ scaler_path=dump/yoko/org/in_duration_scaler.joblib out_dir=dump/yoko/norm/eval/in_duration/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:56,805\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/eval/in_duration/\n",
"out_dir: dump/yoko/norm/eval/in_duration/\n",
"scaler_path: dump/yoko/org/in_duration_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 162.01it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/eval/in_acoustic/ scaler_path=dump/yoko/org/in_acoustic_scaler.joblib out_dir=dump/yoko/norm/eval/in_acoustic/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:57,750\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/eval/in_acoustic/\n",
"out_dir: dump/yoko/norm/eval/in_acoustic/\n",
"scaler_path: dump/yoko/org/in_acoustic_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 41.63it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/eval/out_timelag/ scaler_path=dump/yoko/org/out_timelag_scaler.joblib out_dir=dump/yoko/norm/eval/out_timelag/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:58,695\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/eval/out_timelag/\n",
"out_dir: dump/yoko/norm/eval/out_timelag/\n",
"scaler_path: dump/yoko/org/out_timelag_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 137.34it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/eval/out_duration/ scaler_path=dump/yoko/org/out_duration_scaler.joblib out_dir=dump/yoko/norm/eval/out_duration/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:39:59,631\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/eval/out_duration/\n",
"out_dir: dump/yoko/norm/eval/out_duration/\n",
"scaler_path: dump/yoko/org/out_duration_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 147.08it/s]\n",
"++ set +x\n",
"++ nnsvs-preprocess-normalize in_dir=dump/yoko/org/eval/out_acoustic/ scaler_path=dump/yoko/org/out_acoustic_scaler.joblib out_dir=dump/yoko/norm/eval/out_acoustic/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:40:00,572\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/org/eval/out_acoustic/\n",
"out_dir: dump/yoko/norm/eval/out_acoustic/\n",
"scaler_path: dump/yoko/org/out_acoustic_scaler.joblib\n",
"inverse: false\n",
"num_workers: 4\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 42.72it/s]\n",
"++ set +x\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "andN-LWzC9-l",
"outputId": "6449b530-bba2-40d8-eddd-4c2db791563e",
"colab": {
"base_uri": "https://localhost:8080/"
}
},
"source": [
"! cd $RECIPE_ROOT && bash run.sh --stage 2 --stop-stage 2"
],
"execution_count": 10,
"outputs": [
{
"output_type": "stream",
"text": [
"stage 2: Training time-lag model\n",
"++ nnsvs-train --config-dir conf/train data.train_no_dev.in_dir=dump/yoko/norm/train_no_dev/in_timelag/ data.train_no_dev.out_dir=dump/yoko/norm/train_no_dev/out_timelag/ data.dev.in_dir=dump/yoko/norm/dev/in_timelag/ data.dev.out_dir=dump/yoko/norm/dev/out_timelag/ model=timelag_mdn train.out_dir=exp/yoko/timelag data.batch_size=2 resume.checkpoint=\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:40:02,401\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"data:\n",
" train_no_dev:\n",
" in_dir: dump/yoko/norm/train_no_dev/in_timelag/\n",
" out_dir: dump/yoko/norm/train_no_dev/out_timelag/\n",
" dev:\n",
" in_dir: dump/yoko/norm/dev/in_timelag/\n",
" out_dir: dump/yoko/norm/dev/out_timelag/\n",
" num_workers: 2\n",
" batch_size: 2\n",
" pin_memory: true\n",
"optim:\n",
" optimizer:\n",
" name: Adam\n",
" params:\n",
" lr: 0.001\n",
" betas:\n",
" - 0.5\n",
" - 0.999\n",
" weight_decay: 0.0\n",
" lr_scheduler:\n",
" name: StepLR\n",
" params:\n",
" step_size: 20\n",
" gamma: 0.5\n",
"train:\n",
" out_dir: exp/yoko/timelag\n",
" nepochs: 50\n",
" checkpoint_epoch_interval: 20\n",
" stream_wise_loss: false\n",
" use_detect_anomaly: true\n",
"resume:\n",
" checkpoint: ''\n",
" load_optimizer: false\n",
"cudnn:\n",
" benchmark: false\n",
" deterministic: false\n",
"model:\n",
" stream_sizes:\n",
" - 1\n",
" has_dynamic_features:\n",
" - false\n",
" stream_weights:\n",
" - 1\n",
" netG:\n",
" _target_: nnsvs.model.MDN\n",
" in_dim: 420\n",
" out_dim: 1\n",
" hidden_dim: 1024\n",
" num_layers: 4\n",
" dropout: 0.5\n",
" num_gaussians: 4\n",
"\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:02,401\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - cudnn.deterministic: False\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:02,401\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - cudnn.benchmark: False\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:02,402\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Set to use torch.autograd.detect_anomaly\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,043\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 113, 420]), torch.Size([2, 113, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,044\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 204, 420]), torch.Size([2, 204, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,051\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 117, 420]), torch.Size([2, 117, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,052\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 166, 420]), torch.Size([2, 166, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,058\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 123, 420]), torch.Size([2, 123, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,062\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 242, 420]), torch.Size([2, 242, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,064\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 102, 420]), torch.Size([2, 102, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,068\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 89, 420]), torch.Size([2, 89, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,070\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 232, 420]), torch.Size([2, 232, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,073\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 315, 420]), torch.Size([2, 315, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,082\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 182, 420]), torch.Size([2, 182, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,082\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 111, 420]), torch.Size([2, 111, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,087\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 105, 420]), torch.Size([2, 105, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,087\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 77, 420]), torch.Size([2, 77, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,089\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([1, 110, 420]), torch.Size([1, 110, 1]), torch.Size([1])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,259\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([1, 63, 420]), torch.Size([1, 63, 1]), torch.Size([1])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:13,316\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Start utterance-wise training...\u001b[0m\n",
" 0% 0/50 [00:00<?, ?it/s][\u001b[36m2020-11-08 06:40:14,097\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 1]: loss 1.3745381355285644\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:14,227\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 1]: loss 1.3296167850494385\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:14,318\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 1.3296167850494385: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/timelag/best_loss.pth\u001b[0m\n",
" 2% 1/50 [00:01<00:49, 1.00s/it][\u001b[36m2020-11-08 06:40:14,843\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 2]: loss 1.330424698193868\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:14,970\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 2]: loss 1.4737210273742676\u001b[0m\n",
" 4% 2/50 [00:01<00:43, 1.12it/s][\u001b[36m2020-11-08 06:40:15,460\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 3]: loss 1.2115388552347819\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:15,591\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 3]: loss 1.5297843217849731\u001b[0m\n",
" 6% 3/50 [00:02<00:38, 1.23it/s][\u001b[36m2020-11-08 06:40:16,093\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 4]: loss 1.2365776618321738\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:16,223\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 4]: loss 1.3373641967773438\u001b[0m\n",
" 8% 4/50 [00:02<00:34, 1.32it/s][\u001b[36m2020-11-08 06:40:16,738\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 5]: loss 1.1931703686714172\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:16,866\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 5]: loss 1.226229190826416\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:16,960\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 1.226229190826416: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/timelag/best_loss.pth\u001b[0m\n",
" 10% 5/50 [00:03<00:33, 1.33it/s][\u001b[36m2020-11-08 06:40:17,480\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 6]: loss 1.1820381879806519\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:17,623\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 6]: loss 1.1745021343231201\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:18,218\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 1.1745021343231201: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/timelag/best_loss.pth\u001b[0m\n",
" 12% 6/50 [00:04<00:39, 1.11it/s][\u001b[36m2020-11-08 06:40:18,751\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 7]: loss 1.1750088135401409\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:18,884\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 7]: loss 1.3687212467193604\u001b[0m\n",
" 14% 7/50 [00:05<00:35, 1.20it/s][\u001b[36m2020-11-08 06:40:19,400\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 8]: loss 1.1262434800465901\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:19,530\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 8]: loss 1.6106655597686768\u001b[0m\n",
" 16% 8/50 [00:06<00:32, 1.29it/s][\u001b[36m2020-11-08 06:40:20,019\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 9]: loss 1.1006057937939961\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:20,144\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 9]: loss 1.2504684925079346\u001b[0m\n",
" 18% 9/50 [00:06<00:29, 1.37it/s][\u001b[36m2020-11-08 06:40:20,640\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 10]: loss 1.2001180926958719\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:20,768\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 10]: loss 1.1764816045761108\u001b[0m\n",
" 20% 10/50 [00:07<00:27, 1.44it/s][\u001b[36m2020-11-08 06:40:21,256\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 11]: loss 1.1581034342447916\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:21,378\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 11]: loss 1.6784672737121582\u001b[0m\n",
" 22% 11/50 [00:08<00:26, 1.49it/s][\u001b[36m2020-11-08 06:40:21,872\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 12]: loss 1.1027355909347534\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:21,997\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 12]: loss 1.2114170789718628\u001b[0m\n",
" 24% 12/50 [00:08<00:24, 1.53it/s][\u001b[36m2020-11-08 06:40:22,483\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 13]: loss 1.0876460353533426\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:22,626\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 13]: loss 1.751492977142334\u001b[0m\n",
" 26% 13/50 [00:09<00:23, 1.55it/s][\u001b[36m2020-11-08 06:40:23,146\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 14]: loss 1.1654582420984905\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:23,275\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 14]: loss 1.5063035488128662\u001b[0m\n",
" 28% 14/50 [00:09<00:23, 1.54it/s][\u001b[36m2020-11-08 06:40:23,775\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 15]: loss 1.1391226689020792\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:23,903\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 15]: loss 1.5750874280929565\u001b[0m\n",
" 30% 15/50 [00:10<00:22, 1.56it/s][\u001b[36m2020-11-08 06:40:24,403\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 16]: loss 1.0920168121655782\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:24,530\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 16]: loss 1.3197920322418213\u001b[0m\n",
" 32% 16/50 [00:11<00:21, 1.57it/s][\u001b[36m2020-11-08 06:40:25,041\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 17]: loss 1.0508599996566772\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:25,177\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 17]: loss 1.5382708311080933\u001b[0m\n",
" 34% 17/50 [00:11<00:21, 1.56it/s][\u001b[36m2020-11-08 06:40:25,685\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 18]: loss 1.0013518969217936\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:25,817\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 18]: loss 1.5805810689926147\u001b[0m\n",
" 36% 18/50 [00:12<00:20, 1.56it/s][\u001b[36m2020-11-08 06:40:26,303\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 19]: loss 0.9694705724716186\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:26,428\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 19]: loss 1.8374016284942627\u001b[0m\n",
" 38% 19/50 [00:13<00:19, 1.58it/s][\u001b[36m2020-11-08 06:40:26,930\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 20]: loss 0.9675236384073893\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:27,059\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 20]: loss 1.8533836603164673\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:27,152\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/timelag/checkpoint_epoch0020.pth\u001b[0m\n",
" 40% 20/50 [00:13<00:20, 1.49it/s][\u001b[36m2020-11-08 06:40:27,700\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 21]: loss 0.8335338572661082\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:27,879\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 21]: loss 1.3978370428085327\u001b[0m\n",
" 42% 21/50 [00:14<00:19, 1.48it/s][\u001b[36m2020-11-08 06:40:28,416\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 22]: loss 0.7983591516812643\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:28,543\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 22]: loss 2.1630611419677734\u001b[0m\n",
" 44% 22/50 [00:15<00:18, 1.49it/s][\u001b[36m2020-11-08 06:40:29,047\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 23]: loss 0.7893897453943889\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:29,180\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 23]: loss 1.5683197975158691\u001b[0m\n",
" 46% 23/50 [00:15<00:17, 1.51it/s][\u001b[36m2020-11-08 06:40:29,682\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 24]: loss 0.6838932116826375\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:29,816\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 24]: loss 1.5895270109176636\u001b[0m\n",
" 48% 24/50 [00:16<00:17, 1.53it/s][\u001b[36m2020-11-08 06:40:30,311\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 25]: loss 0.6700842161973317\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:30,440\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 25]: loss 1.9990490674972534\u001b[0m\n",
" 50% 25/50 [00:17<00:16, 1.55it/s][\u001b[36m2020-11-08 06:40:30,945\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 26]: loss 0.5961545914411545\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:31,074\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 26]: loss 2.3024139404296875\u001b[0m\n",
" 52% 26/50 [00:17<00:15, 1.56it/s][\u001b[36m2020-11-08 06:40:31,570\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 27]: loss 0.605980298233529\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:31,696\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 27]: loss 1.935489535331726\u001b[0m\n",
" 54% 27/50 [00:18<00:14, 1.57it/s][\u001b[36m2020-11-08 06:40:32,192\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 28]: loss 0.5435595353444417\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:32,321\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 28]: loss 2.0033721923828125\u001b[0m\n",
" 56% 28/50 [00:19<00:13, 1.58it/s][\u001b[36m2020-11-08 06:40:32,818\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 29]: loss 0.5449054419994355\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:32,951\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 29]: loss 2.09224796295166\u001b[0m\n",
" 58% 29/50 [00:19<00:13, 1.58it/s][\u001b[36m2020-11-08 06:40:33,461\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 30]: loss 0.49069395164648694\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:33,594\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 30]: loss 1.6056712865829468\u001b[0m\n",
" 60% 30/50 [00:20<00:12, 1.57it/s][\u001b[36m2020-11-08 06:40:34,103\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 31]: loss 0.5000701397657394\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:34,231\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 31]: loss 1.81768000125885\u001b[0m\n",
" 62% 31/50 [00:20<00:12, 1.57it/s][\u001b[36m2020-11-08 06:40:34,730\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 32]: loss 0.42523051301638287\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:34,860\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 32]: loss 1.7658101320266724\u001b[0m\n",
" 64% 32/50 [00:21<00:11, 1.58it/s][\u001b[36m2020-11-08 06:40:35,350\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 33]: loss 0.4266463587681452\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:35,477\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 33]: loss 2.5119454860687256\u001b[0m\n",
" 66% 33/50 [00:22<00:10, 1.59it/s][\u001b[36m2020-11-08 06:40:35,976\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 34]: loss 0.3491426470379035\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:36,103\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 34]: loss 1.8600057363510132\u001b[0m\n",
" 68% 34/50 [00:22<00:10, 1.59it/s][\u001b[36m2020-11-08 06:40:36,591\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 35]: loss 0.5926243484020233\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:36,721\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 35]: loss 2.332580804824829\u001b[0m\n",
" 70% 35/50 [00:23<00:09, 1.60it/s][\u001b[36m2020-11-08 06:40:37,223\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 36]: loss 0.41532623767852783\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:37,358\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 36]: loss 2.3183701038360596\u001b[0m\n",
" 72% 36/50 [00:24<00:08, 1.59it/s][\u001b[36m2020-11-08 06:40:37,878\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 37]: loss 0.3330076664686203\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:38,007\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 37]: loss 2.6100611686706543\u001b[0m\n",
" 74% 37/50 [00:24<00:08, 1.58it/s][\u001b[36m2020-11-08 06:40:38,503\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 38]: loss 0.33520599007606505\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:38,630\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 38]: loss 2.2604033946990967\u001b[0m\n",
" 76% 38/50 [00:25<00:07, 1.58it/s][\u001b[36m2020-11-08 06:40:39,121\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 39]: loss 0.2598566909631093\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:39,251\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 39]: loss 2.2122538089752197\u001b[0m\n",
" 78% 39/50 [00:25<00:06, 1.59it/s][\u001b[36m2020-11-08 06:40:39,762\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 40]: loss 0.2391318589448929\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:39,888\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 40]: loss 2.4751102924346924\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:39,971\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/timelag/checkpoint_epoch0040.pth\u001b[0m\n",
" 80% 40/50 [00:26<00:06, 1.49it/s][\u001b[36m2020-11-08 06:40:40,531\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 41]: loss 0.2560277412335078\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:40,661\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 41]: loss 1.9318873882293701\u001b[0m\n",
" 82% 41/50 [00:27<00:05, 1.51it/s][\u001b[36m2020-11-08 06:40:41,167\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 42]: loss 0.1704696031908194\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:41,292\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 42]: loss 2.8476011753082275\u001b[0m\n",
" 84% 42/50 [00:27<00:05, 1.53it/s][\u001b[36m2020-11-08 06:40:41,782\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 43]: loss -0.08641724648574987\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:41,910\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 43]: loss 2.9160380363464355\u001b[0m\n",
" 86% 43/50 [00:28<00:04, 1.56it/s][\u001b[36m2020-11-08 06:40:42,426\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 44]: loss -0.112529675103724\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:42,556\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 44]: loss 3.726233720779419\u001b[0m\n",
" 88% 44/50 [00:29<00:03, 1.56it/s][\u001b[36m2020-11-08 06:40:43,051\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 45]: loss -0.11099585853517055\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:43,179\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 45]: loss 2.8829026222229004\u001b[0m\n",
" 90% 45/50 [00:29<00:03, 1.57it/s][\u001b[36m2020-11-08 06:40:43,686\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 46]: loss -0.16384430152053633\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:43,823\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 46]: loss 3.9400243759155273\u001b[0m\n",
" 92% 46/50 [00:30<00:02, 1.56it/s][\u001b[36m2020-11-08 06:40:44,329\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 47]: loss -0.0900291426728169\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:44,461\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 47]: loss 3.240088701248169\u001b[0m\n",
" 94% 47/50 [00:31<00:01, 1.57it/s][\u001b[36m2020-11-08 06:40:44,967\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 48]: loss -0.10202263686805964\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:45,099\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 48]: loss 3.0919504165649414\u001b[0m\n",
" 96% 48/50 [00:31<00:01, 1.57it/s][\u001b[36m2020-11-08 06:40:45,582\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 49]: loss -0.10511299011607965\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:45,709\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 49]: loss 3.5698001384735107\u001b[0m\n",
" 98% 49/50 [00:32<00:00, 1.59it/s][\u001b[36m2020-11-08 06:40:46,198\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 50]: loss -0.06097584267457326\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:46,331\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 50]: loss 3.005959987640381\u001b[0m\n",
"100% 50/50 [00:33<00:00, 1.51it/s]\n",
"[\u001b[36m2020-11-08 06:40:46,414\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/timelag/checkpoint_epoch0050.pth\u001b[0m\n",
"[\u001b[36m2020-11-08 06:40:46,496\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - The best loss was 1.1745021343231201\u001b[0m\n",
"++ set +x\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "js0FLYInDEGL",
"outputId": "d171ce02-1588-4825-9e47-ba6be4e02ea5",
"colab": {
"base_uri": "https://localhost:8080/"
}
},
"source": [
"! cd $RECIPE_ROOT && bash run.sh --stage 3 --stop-stage 3"
],
"execution_count": 11,
"outputs": [
{
"output_type": "stream",
"text": [
"stage 3: Training duration model\n",
"++ nnsvs-train --config-dir conf/train data.train_no_dev.in_dir=dump/yoko/norm/train_no_dev/in_duration/ data.train_no_dev.out_dir=dump/yoko/norm/train_no_dev/out_duration/ data.dev.in_dir=dump/yoko/norm/dev/in_duration/ data.dev.out_dir=dump/yoko/norm/dev/out_duration/ model=duration_mdn train.out_dir=exp/yoko/duration data.batch_size=2 resume.checkpoint=\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:42:38,831\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"data:\n",
" train_no_dev:\n",
" in_dir: dump/yoko/norm/train_no_dev/in_duration/\n",
" out_dir: dump/yoko/norm/train_no_dev/out_duration/\n",
" dev:\n",
" in_dir: dump/yoko/norm/dev/in_duration/\n",
" out_dir: dump/yoko/norm/dev/out_duration/\n",
" num_workers: 2\n",
" batch_size: 2\n",
" pin_memory: true\n",
"optim:\n",
" optimizer:\n",
" name: Adam\n",
" params:\n",
" lr: 0.001\n",
" betas:\n",
" - 0.5\n",
" - 0.999\n",
" weight_decay: 0.0\n",
" lr_scheduler:\n",
" name: StepLR\n",
" params:\n",
" step_size: 20\n",
" gamma: 0.5\n",
"train:\n",
" out_dir: exp/yoko/duration\n",
" nepochs: 50\n",
" checkpoint_epoch_interval: 20\n",
" stream_wise_loss: false\n",
" use_detect_anomaly: true\n",
"resume:\n",
" checkpoint: ''\n",
" load_optimizer: false\n",
"cudnn:\n",
" benchmark: false\n",
" deterministic: false\n",
"model:\n",
" stream_sizes:\n",
" - 1\n",
" has_dynamic_features:\n",
" - false\n",
" stream_weights:\n",
" - 1\n",
" netG:\n",
" _target_: nnsvs.model.MDN\n",
" in_dim: 420\n",
" out_dim: 1\n",
" hidden_dim: 1024\n",
" num_layers: 4\n",
" dropout: 0.5\n",
" num_gaussians: 4\n",
"\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:38,831\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - cudnn.deterministic: False\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:38,831\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - cudnn.benchmark: False\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:38,832\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Set to use torch.autograd.detect_anomaly\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,676\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 468, 420]), torch.Size([2, 468, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,685\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 131, 420]), torch.Size([2, 131, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,686\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 205, 420]), torch.Size([2, 205, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,693\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 206, 420]), torch.Size([2, 206, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,693\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 238, 420]), torch.Size([2, 238, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,702\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 157, 420]), torch.Size([2, 157, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,705\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 335, 420]), torch.Size([2, 335, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,715\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 261, 420]), torch.Size([2, 261, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,715\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 455, 420]), torch.Size([2, 455, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,724\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 270, 420]), torch.Size([2, 270, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,727\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 227, 420]), torch.Size([2, 227, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,735\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 393, 420]), torch.Size([2, 393, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,735\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 230, 420]), torch.Size([2, 230, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,737\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 364, 420]), torch.Size([2, 364, 1]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,739\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([1, 577, 420]), torch.Size([1, 577, 1]), torch.Size([1])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,917\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([1, 135, 420]), torch.Size([1, 135, 1]), torch.Size([1])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:43,972\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Start utterance-wise training...\u001b[0m\n",
" 0% 0/50 [00:00<?, ?it/s][\u001b[36m2020-11-08 06:42:44,668\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 1]: loss 1.2194159587224325\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:44,804\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 1]: loss 2.2860889434814453\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:44,896\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 2.2860889434814453: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/duration/best_loss.pth\u001b[0m\n",
" 2% 1/50 [00:00<00:45, 1.09it/s][\u001b[36m2020-11-08 06:42:45,463\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 2]: loss 1.166808048884074\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:45,599\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 2]: loss 0.5686985850334167\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:45,699\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 0.5686985850334167: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/duration/best_loss.pth\u001b[0m\n",
" 4% 2/50 [00:01<00:42, 1.13it/s][\u001b[36m2020-11-08 06:42:46,254\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 3]: loss 0.7782502373059591\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:46,388\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 3]: loss 0.25162461400032043\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:46,957\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 0.25162461400032043: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/duration/best_loss.pth\u001b[0m\n",
" 6% 3/50 [00:02<00:46, 1.00it/s][\u001b[36m2020-11-08 06:42:47,552\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 4]: loss 0.7711286703745525\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:47,703\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 4]: loss 0.8084796071052551\u001b[0m\n",
" 8% 4/50 [00:03<00:42, 1.08it/s][\u001b[36m2020-11-08 06:42:48,275\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 5]: loss 0.6937808354695638\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:48,412\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 5]: loss 0.5929139256477356\u001b[0m\n",
" 10% 5/50 [00:04<00:38, 1.17it/s][\u001b[36m2020-11-08 06:42:48,983\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 6]: loss 0.8485954225063324\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:49,120\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 6]: loss 0.5819787979125977\u001b[0m\n",
" 12% 6/50 [00:05<00:35, 1.23it/s][\u001b[36m2020-11-08 06:42:49,686\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 7]: loss 0.7481428682804108\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:49,824\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 7]: loss 0.18472526967525482\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:49,924\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 0.18472526967525482: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/duration/best_loss.pth\u001b[0m\n",
" 14% 7/50 [00:05<00:34, 1.23it/s][\u001b[36m2020-11-08 06:42:50,499\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 8]: loss 0.6689550951123238\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:50,641\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 8]: loss 0.25903958082199097\u001b[0m\n",
" 16% 8/50 [00:06<00:32, 1.28it/s][\u001b[36m2020-11-08 06:42:51,215\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 9]: loss 0.5833607092499733\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:51,360\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 9]: loss 0.16170839965343475\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:51,463\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 0.16170839965343475: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/duration/best_loss.pth\u001b[0m\n",
" 18% 9/50 [00:07<00:32, 1.26it/s][\u001b[36m2020-11-08 06:42:52,032\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 10]: loss 0.5233926196893056\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:52,170\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 10]: loss 0.14153939485549927\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:52,751\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 0.14153939485549927: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/duration/best_loss.pth\u001b[0m\n",
" 20% 10/50 [00:08<00:37, 1.06it/s][\u001b[36m2020-11-08 06:42:53,341\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 11]: loss 0.2704586165646712\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:53,495\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 11]: loss 0.18012113869190216\u001b[0m\n",
" 22% 11/50 [00:09<00:34, 1.13it/s][\u001b[36m2020-11-08 06:42:54,072\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 12]: loss 0.10373166265587012\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:54,211\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 12]: loss -0.19085319340229034\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:54,316\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss -0.19085319340229034: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/duration/best_loss.pth\u001b[0m\n",
" 24% 12/50 [00:10<00:32, 1.16it/s][\u001b[36m2020-11-08 06:42:54,904\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 13]: loss 0.14537517751256626\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:55,040\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 13]: loss 0.11204147338867188\u001b[0m\n",
" 26% 13/50 [00:11<00:30, 1.22it/s][\u001b[36m2020-11-08 06:42:55,604\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 14]: loss 0.3105158448840181\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:55,738\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 14]: loss 0.2079266756772995\u001b[0m\n",
" 28% 14/50 [00:11<00:28, 1.27it/s][\u001b[36m2020-11-08 06:42:56,285\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 15]: loss 0.2450202410419782\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:56,424\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 15]: loss 0.4843209683895111\u001b[0m\n",
" 30% 15/50 [00:12<00:26, 1.32it/s][\u001b[36m2020-11-08 06:42:56,984\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 16]: loss 0.18127077960719665\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:57,116\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 16]: loss 0.030638616532087326\u001b[0m\n",
" 32% 16/50 [00:13<00:25, 1.36it/s][\u001b[36m2020-11-08 06:42:57,682\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 17]: loss -0.01950769325097402\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:57,825\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 17]: loss -0.26185154914855957\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:57,930\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss -0.26185154914855957: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/duration/best_loss.pth\u001b[0m\n",
" 34% 17/50 [00:13<00:25, 1.32it/s][\u001b[36m2020-11-08 06:42:58,509\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 18]: loss -0.015172491843501727\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:58,650\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 18]: loss -0.22730712592601776\u001b[0m\n",
" 36% 18/50 [00:14<00:23, 1.34it/s][\u001b[36m2020-11-08 06:42:59,231\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 19]: loss 0.3122451807061831\u001b[0m\n",
"[\u001b[36m2020-11-08 06:42:59,373\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 19]: loss -0.0829862579703331\u001b[0m\n",
" 38% 19/50 [00:15<00:22, 1.35it/s][\u001b[36m2020-11-08 06:42:59,944\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 20]: loss 0.13357394312818846\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:00,080\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 20]: loss -0.3061339855194092\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:00,180\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss -0.3061339855194092: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/duration/best_loss.pth\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:00,270\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/duration/checkpoint_epoch0020.pth\u001b[0m\n",
" 40% 20/50 [00:16<00:24, 1.25it/s][\u001b[36m2020-11-08 06:43:00,884\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 21]: loss -0.19545197983582815\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:01,031\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 21]: loss -0.18823018670082092\u001b[0m\n",
" 42% 21/50 [00:17<00:22, 1.29it/s][\u001b[36m2020-11-08 06:43:01,594\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 22]: loss -0.31819156805674237\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:01,731\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 22]: loss -0.33386415243148804\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:01,896\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss -0.33386415243148804: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/duration/best_loss.pth\u001b[0m\n",
" 44% 22/50 [00:17<00:22, 1.25it/s][\u001b[36m2020-11-08 06:43:02,507\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 23]: loss -0.35310772707064947\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:02,648\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 23]: loss -0.26860764622688293\u001b[0m\n",
" 46% 23/50 [00:18<00:21, 1.27it/s][\u001b[36m2020-11-08 06:43:03,211\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 24]: loss -0.34564225127299625\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:03,350\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 24]: loss -0.2489587515592575\u001b[0m\n",
" 48% 24/50 [00:19<00:19, 1.31it/s][\u001b[36m2020-11-08 06:43:03,918\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 25]: loss -0.4155164842804273\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:04,059\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 25]: loss -0.23731155693531036\u001b[0m\n",
" 50% 25/50 [00:20<00:18, 1.34it/s][\u001b[36m2020-11-08 06:43:04,619\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 26]: loss -0.5574668377637864\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:04,757\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 26]: loss -0.23123586177825928\u001b[0m\n",
" 52% 26/50 [00:20<00:17, 1.37it/s][\u001b[36m2020-11-08 06:43:05,301\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 27]: loss -0.5608094068864982\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:05,438\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 27]: loss -0.17303864657878876\u001b[0m\n",
" 54% 27/50 [00:21<00:16, 1.40it/s][\u001b[36m2020-11-08 06:43:05,998\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 28]: loss -0.5330424070358276\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:06,136\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 28]: loss 0.19807100296020508\u001b[0m\n",
" 56% 28/50 [00:22<00:15, 1.41it/s][\u001b[36m2020-11-08 06:43:06,707\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 29]: loss -0.5545578281084697\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:06,849\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 29]: loss 0.13198751211166382\u001b[0m\n",
" 58% 29/50 [00:22<00:14, 1.41it/s][\u001b[36m2020-11-08 06:43:07,410\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 30]: loss -0.4690553416808446\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:07,558\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 30]: loss -0.23092468082904816\u001b[0m\n",
" 60% 30/50 [00:23<00:14, 1.41it/s][\u001b[36m2020-11-08 06:43:08,119\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 31]: loss -0.5969082832336425\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:08,260\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 31]: loss -0.1347690373659134\u001b[0m\n",
" 62% 31/50 [00:24<00:13, 1.41it/s][\u001b[36m2020-11-08 06:43:08,827\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 32]: loss -0.663450175523758\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:08,971\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 32]: loss 0.13954021036624908\u001b[0m\n",
" 64% 32/50 [00:24<00:12, 1.41it/s][\u001b[36m2020-11-08 06:43:09,540\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 33]: loss -0.5637840261061986\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:09,689\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 33]: loss -0.023513955995440483\u001b[0m\n",
" 66% 33/50 [00:25<00:12, 1.40it/s][\u001b[36m2020-11-08 06:43:10,273\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 34]: loss -0.7025902489821116\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:10,413\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 34]: loss 0.2713871896266937\u001b[0m\n",
" 68% 34/50 [00:26<00:11, 1.40it/s][\u001b[36m2020-11-08 06:43:10,990\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 35]: loss -0.6814305618405342\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:11,126\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 35]: loss 0.2566107213497162\u001b[0m\n",
" 70% 35/50 [00:27<00:10, 1.40it/s][\u001b[36m2020-11-08 06:43:11,690\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 36]: loss -0.5727538297573725\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:11,828\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 36]: loss 0.004016953986138105\u001b[0m\n",
" 72% 36/50 [00:27<00:09, 1.41it/s][\u001b[36m2020-11-08 06:43:12,400\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 37]: loss -0.7272753735383352\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:12,537\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 37]: loss 0.25739920139312744\u001b[0m\n",
" 74% 37/50 [00:28<00:09, 1.41it/s][\u001b[36m2020-11-08 06:43:13,114\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 38]: loss -0.6849347005287806\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:13,258\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 38]: loss -0.16628797352313995\u001b[0m\n",
" 76% 38/50 [00:29<00:08, 1.40it/s][\u001b[36m2020-11-08 06:43:13,857\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 39]: loss -0.6137308021386464\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:13,992\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 39]: loss 0.5424676537513733\u001b[0m\n",
" 78% 39/50 [00:30<00:07, 1.39it/s][\u001b[36m2020-11-08 06:43:14,552\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 40]: loss -0.8041770656903585\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:14,693\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 40]: loss 0.47989118099212646\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:14,780\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/duration/checkpoint_epoch0040.pth\u001b[0m\n",
" 80% 40/50 [00:30<00:07, 1.33it/s][\u001b[36m2020-11-08 06:43:15,394\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 41]: loss -0.9146930138270061\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:15,535\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 41]: loss 0.08616559952497482\u001b[0m\n",
" 82% 41/50 [00:31<00:06, 1.35it/s][\u001b[36m2020-11-08 06:43:16,161\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 42]: loss -1.0497997601826985\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:16,298\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 42]: loss 0.036433715373277664\u001b[0m\n",
" 84% 42/50 [00:32<00:05, 1.34it/s][\u001b[36m2020-11-08 06:43:16,866\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 43]: loss -1.1124594926834106\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:17,007\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 43]: loss 0.5006793737411499\u001b[0m\n",
" 86% 43/50 [00:33<00:05, 1.36it/s][\u001b[36m2020-11-08 06:43:17,569\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 44]: loss -1.0271951039632161\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:17,711\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 44]: loss 0.4405345022678375\u001b[0m\n",
" 88% 44/50 [00:33<00:04, 1.38it/s][\u001b[36m2020-11-08 06:43:18,265\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 45]: loss -1.1251414696375528\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:18,401\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 45]: loss 0.27562734484672546\u001b[0m\n",
" 90% 45/50 [00:34<00:03, 1.40it/s][\u001b[36m2020-11-08 06:43:18,976\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 46]: loss -1.1168710986773174\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:19,109\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 46]: loss 0.48207148909568787\u001b[0m\n",
" 92% 46/50 [00:35<00:02, 1.40it/s][\u001b[36m2020-11-08 06:43:19,651\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 47]: loss -1.116710368792216\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:19,792\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 47]: loss 0.2646186947822571\u001b[0m\n",
" 94% 47/50 [00:35<00:02, 1.42it/s][\u001b[36m2020-11-08 06:43:20,351\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 48]: loss -1.092997521162033\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:20,488\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 48]: loss 0.2359001785516739\u001b[0m\n",
" 96% 48/50 [00:36<00:01, 1.43it/s][\u001b[36m2020-11-08 06:43:21,063\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 49]: loss -1.0798675537109375\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:21,199\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 49]: loss 0.2798577845096588\u001b[0m\n",
" 98% 49/50 [00:37<00:00, 1.42it/s][\u001b[36m2020-11-08 06:43:21,769\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 50]: loss -1.1697866996129354\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:21,909\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 50]: loss 0.3523787558078766\u001b[0m\n",
"100% 50/50 [00:37<00:00, 1.32it/s]\n",
"[\u001b[36m2020-11-08 06:43:22,000\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/duration/checkpoint_epoch0050.pth\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:22,049\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - The best loss was -0.33386415243148804\u001b[0m\n",
"++ set +x\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "KQ1u_vYdDFwm",
"outputId": "779243f0-6d12-4ce5-f11f-876aff2e1b2d",
"colab": {
"base_uri": "https://localhost:8080/"
}
},
"source": [
"! cd $RECIPE_ROOT && bash run.sh --stage 4 --stop-stage 4"
],
"execution_count": 12,
"outputs": [
{
"output_type": "stream",
"text": [
"stage 4: Training acoustic model\n",
"++ nnsvs-train --config-dir conf/train data.train_no_dev.in_dir=dump/yoko/norm/train_no_dev/in_acoustic/ data.train_no_dev.out_dir=dump/yoko/norm/train_no_dev/out_acoustic/ data.dev.in_dir=dump/yoko/norm/dev/in_acoustic/ data.dev.out_dir=dump/yoko/norm/dev/out_acoustic/ model=acoustic_cnn_mdn train.out_dir=exp/yoko/acoustic data.batch_size=2 resume.checkpoint=\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:43:24,293\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"data:\n",
" train_no_dev:\n",
" in_dir: dump/yoko/norm/train_no_dev/in_acoustic/\n",
" out_dir: dump/yoko/norm/train_no_dev/out_acoustic/\n",
" dev:\n",
" in_dir: dump/yoko/norm/dev/in_acoustic/\n",
" out_dir: dump/yoko/norm/dev/out_acoustic/\n",
" num_workers: 2\n",
" batch_size: 2\n",
" pin_memory: true\n",
"optim:\n",
" optimizer:\n",
" name: Adam\n",
" params:\n",
" lr: 0.001\n",
" betas:\n",
" - 0.5\n",
" - 0.999\n",
" weight_decay: 0.0\n",
" lr_scheduler:\n",
" name: StepLR\n",
" params:\n",
" step_size: 20\n",
" gamma: 0.5\n",
"train:\n",
" out_dir: exp/yoko/acoustic\n",
" nepochs: 50\n",
" checkpoint_epoch_interval: 20\n",
" stream_wise_loss: false\n",
" use_detect_anomaly: true\n",
"resume:\n",
" checkpoint: ''\n",
" load_optimizer: false\n",
"cudnn:\n",
" benchmark: false\n",
" deterministic: false\n",
"model:\n",
" stream_sizes:\n",
" - 180\n",
" - 3\n",
" - 1\n",
" - 15\n",
" has_dynamic_features:\n",
" - true\n",
" - true\n",
" - false\n",
" - true\n",
" num_windows: 3\n",
" stream_weights: null\n",
" netG:\n",
" _target_: nnsvs.model.Conv1dResnetMDN\n",
" in_dim: 424\n",
" out_dim: 199\n",
" hidden_dim: 256\n",
" num_layers: 6\n",
" dropout: 0.1\n",
" num_gaussians: 4\n",
"\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:24,293\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - cudnn.deterministic: False\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:24,294\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - cudnn.benchmark: False\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:24,294\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Set to use torch.autograd.detect_anomaly\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:29,789\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 16800, 424]), torch.Size([2, 16800, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:29,791\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 10080, 424]), torch.Size([2, 10080, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:30,087\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 18360, 424]), torch.Size([2, 18360, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:30,198\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 29760, 424]), torch.Size([2, 29760, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:31,147\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 6720, 424]), torch.Size([2, 6720, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:32,478\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 18000, 424]), torch.Size([2, 18000, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:32,672\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 16320, 424]), torch.Size([2, 16320, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:32,672\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 3360, 424]), torch.Size([2, 3360, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:35,243\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 16320, 424]), torch.Size([2, 16320, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:35,243\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 8160, 424]), torch.Size([2, 8160, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:36,906\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 12480, 424]), torch.Size([2, 12480, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:36,906\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 16318, 424]), torch.Size([2, 16318, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:38,303\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 12480, 424]), torch.Size([2, 12480, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:38,304\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 19920, 424]), torch.Size([2, 19920, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:39,168\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([1, 16320, 424]), torch.Size([1, 16320, 199]), torch.Size([1])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:39,342\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([1, 6240, 424]), torch.Size([1, 6240, 199]), torch.Size([1])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:39,396\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Start utterance-wise training...\u001b[0m\n",
" 0% 0/50 [00:00<?, ?it/s][\u001b[36m2020-11-08 06:43:51,922\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 1]: loss 279.9014078776042\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:52,571\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 1]: loss 278.20220947265625\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:52,678\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 278.20220947265625: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 2% 1/50 [00:13<10:50, 13.28s/it][\u001b[36m2020-11-08 06:43:58,765\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 2]: loss 265.63893229166666\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:59,039\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 2]: loss 265.9647521972656\u001b[0m\n",
"[\u001b[36m2020-11-08 06:43:59,153\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 265.9647521972656: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 4% 2/50 [00:19<08:59, 11.24s/it][\u001b[36m2020-11-08 06:44:04,904\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 3]: loss 255.97406412760418\u001b[0m\n",
"[\u001b[36m2020-11-08 06:44:05,188\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 3]: loss 262.90325927734375\u001b[0m\n",
"[\u001b[36m2020-11-08 06:44:05,303\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 262.90325927734375: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 6% 3/50 [00:25<07:36, 9.71s/it][\u001b[36m2020-11-08 06:44:10,982\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 4]: loss 255.89339192708334\u001b[0m\n",
"[\u001b[36m2020-11-08 06:44:11,239\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 4]: loss 261.5957946777344\u001b[0m\n",
"[\u001b[36m2020-11-08 06:44:11,358\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 261.5957946777344: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 8% 4/50 [00:31<06:36, 8.61s/it][\u001b[36m2020-11-08 06:44:18,062\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 5]: loss 255.82510070800782\u001b[0m\n",
"[\u001b[36m2020-11-08 06:44:18,326\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 5]: loss 262.5871276855469\u001b[0m\n",
" 10% 5/50 [00:38<06:05, 8.12s/it][\u001b[36m2020-11-08 06:44:24,505\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 6]: loss 253.6558858235677\u001b[0m\n",
"[\u001b[36m2020-11-08 06:44:24,777\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 6]: loss 263.4515380859375\u001b[0m\n",
" 12% 6/50 [00:45<05:35, 7.62s/it][\u001b[36m2020-11-08 06:44:30,452\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 7]: loss 253.0725860595703\u001b[0m\n",
"[\u001b[36m2020-11-08 06:44:30,711\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 7]: loss 262.64569091796875\u001b[0m\n",
" 14% 7/50 [00:51<05:05, 7.11s/it][\u001b[36m2020-11-08 06:44:36,620\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 8]: loss 255.2508514404297\u001b[0m\n",
"[\u001b[36m2020-11-08 06:44:36,885\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 8]: loss 265.374755859375\u001b[0m\n",
" 16% 8/50 [00:57<04:46, 6.83s/it][\u001b[36m2020-11-08 06:44:42,503\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 9]: loss 256.0583872477213\u001b[0m\n",
"[\u001b[36m2020-11-08 06:44:42,764\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 9]: loss 264.1356506347656\u001b[0m\n",
" 18% 9/50 [01:03<04:28, 6.55s/it][\u001b[36m2020-11-08 06:44:48,669\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 10]: loss 256.6583231608073\u001b[0m\n",
"[\u001b[36m2020-11-08 06:44:48,930\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 10]: loss 261.0714416503906\u001b[0m\n",
"[\u001b[36m2020-11-08 06:44:49,045\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 261.0714416503906: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 20% 10/50 [01:09<04:18, 6.47s/it][\u001b[36m2020-11-08 06:44:54,849\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 11]: loss 252.54214579264323\u001b[0m\n",
"[\u001b[36m2020-11-08 06:44:55,105\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 11]: loss 261.39434814453125\u001b[0m\n",
" 22% 11/50 [01:15<04:07, 6.34s/it][\u001b[36m2020-11-08 06:45:00,984\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 12]: loss 252.43341267903645\u001b[0m\n",
"[\u001b[36m2020-11-08 06:45:01,242\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 12]: loss 261.2166748046875\u001b[0m\n",
" 24% 12/50 [01:21<03:58, 6.28s/it][\u001b[36m2020-11-08 06:45:07,226\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 13]: loss 252.6378153483073\u001b[0m\n",
"[\u001b[36m2020-11-08 06:45:07,500\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 13]: loss 259.8487854003906\u001b[0m\n",
"[\u001b[36m2020-11-08 06:45:07,612\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 259.8487854003906: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 26% 13/50 [01:28<03:53, 6.31s/it][\u001b[36m2020-11-08 06:45:13,678\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 14]: loss 253.1966786702474\u001b[0m\n",
"[\u001b[36m2020-11-08 06:45:13,943\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 14]: loss 263.7522888183594\u001b[0m\n",
" 28% 14/50 [01:34<03:47, 6.32s/it][\u001b[36m2020-11-08 06:45:19,630\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 15]: loss 288.96129455566404\u001b[0m\n",
"[\u001b[36m2020-11-08 06:45:19,897\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 15]: loss 738.9043579101562\u001b[0m\n",
" 30% 15/50 [01:40<03:37, 6.21s/it][W python_anomaly_mode.cpp:104] Warning: Error detected in PowBackward0. Traceback of forward call that caused the error:\n",
" File \"/usr/local/bin/nnsvs-train\", line 8, in <module>\n",
" sys.exit(entry())\n",
" File \"/usr/local/lib/python3.6/dist-packages/nnsvs/bin/train.py\", line 275, in entry\n",
" my_app()\n",
" File \"/usr/local/lib/python3.6/dist-packages/hydra/main.py\", line 37, in decorated_main\n",
" strict=strict,\n",
" File \"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py\", line 347, in _run_hydra\n",
" lambda: hydra.run(\n",
" File \"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py\", line 198, in run_and_report\n",
" return func()\n",
" File \"/usr/local/lib/python3.6/dist-packages/hydra/_internal/utils.py\", line 350, in <lambda>\n",
" overrides=args.overrides,\n",
" File \"/usr/local/lib/python3.6/dist-packages/hydra/_internal/hydra.py\", line 112, in run\n",
" configure_logging=with_log_configuration,\n",
" File \"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py\", line 125, in run_job\n",
" ret.return_value = task_function(task_cfg)\n",
" File \"/usr/local/lib/python3.6/dist-packages/nnsvs/bin/train.py\", line 271, in my_app\n",
" train_loop(config, device, model, optimizer, lr_scheduler, data_loaders)\n",
" File \"/usr/local/lib/python3.6/dist-packages/nnsvs/bin/train.py\", line 175, in train_loop\n",
" loss = mdn_loss(pi, sigma, mu, y, reduce=False).masked_select(mask).mean()\n",
" File \"/usr/local/lib/python3.6/dist-packages/nnsvs/mdn.py\", line 94, in mdn_loss\n",
" log_prob = dist.log_prob(target)\n",
" File \"/usr/local/lib/python3.6/dist-packages/torch/distributions/normal.py\", line 74, in log_prob\n",
" var = (self.scale ** 2)\n",
" (function _print_stack)\n",
"Traceback (most recent call last):\n",
" File \"/usr/local/lib/python3.6/dist-packages/nnsvs/bin/train.py\", line 271, in my_app\n",
" train_loop(config, device, model, optimizer, lr_scheduler, data_loaders)\n",
" File \"/usr/local/lib/python3.6/dist-packages/nnsvs/bin/train.py\", line 199, in train_loop\n",
" loss.backward()\n",
" File \"/usr/local/lib/python3.6/dist-packages/torch/tensor.py\", line 221, in backward\n",
" torch.autograd.backward(self, gradient, retain_graph, create_graph)\n",
" File \"/usr/local/lib/python3.6/dist-packages/torch/autograd/__init__.py\", line 132, in backward\n",
" allow_unreachable=True) # allow_unreachable flag\n",
"RuntimeError: Function 'PowBackward0' returned nan values in its 0th output.\n",
"\n",
"Set the environment variable HYDRA_FULL_ERROR=1 for a complete stack trace.\n",
" 30% 15/50 [01:42<03:58, 6.82s/it]\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "yPE9iDlPg9AS",
"outputId": "c764a08e-f4bc-4a8d-f454-eb57bb23383d",
"colab": {
"base_uri": "https://localhost:8080/"
}
},
"source": [
"! cd nnsvs && git fetch origin\n",
"! cd nnsvs && git checkout cnn_mdn_test_fixed\n",
"! cd nnsvs && pip install . -q"
],
"execution_count": 13,
"outputs": [
{
"output_type": "stream",
"text": [
"remote: Enumerating objects: 21, done.\u001b[K\n",
"remote: Counting objects: 4% (1/21)\u001b[K\rremote: Counting objects: 9% (2/21)\u001b[K\rremote: Counting objects: 14% (3/21)\u001b[K\rremote: Counting objects: 19% (4/21)\u001b[K\rremote: Counting objects: 23% (5/21)\u001b[K\rremote: Counting objects: 28% (6/21)\u001b[K\rremote: Counting objects: 33% (7/21)\u001b[K\rremote: Counting objects: 38% (8/21)\u001b[K\rremote: Counting objects: 42% (9/21)\u001b[K\rremote: Counting objects: 47% (10/21)\u001b[K\rremote: Counting objects: 52% (11/21)\u001b[K\rremote: Counting objects: 57% (12/21)\u001b[K\rremote: Counting objects: 61% (13/21)\u001b[K\rremote: Counting objects: 66% (14/21)\u001b[K\rremote: Counting objects: 71% (15/21)\u001b[K\rremote: Counting objects: 76% (16/21)\u001b[K\rremote: Counting objects: 80% (17/21)\u001b[K\rremote: Counting objects: 85% (18/21)\u001b[K\rremote: Counting objects: 90% (19/21)\u001b[K\rremote: Counting objects: 95% (20/21)\u001b[K\rremote: Counting objects: 100% (21/21)\u001b[K\rremote: Counting objects: 100% (21/21), done.\u001b[K\n",
"remote: Compressing objects: 14% (1/7)\u001b[K\rremote: Compressing objects: 28% (2/7)\u001b[K\rremote: Compressing objects: 42% (3/7)\u001b[K\rremote: Compressing objects: 57% (4/7)\u001b[K\rremote: Compressing objects: 71% (5/7)\u001b[K\rremote: Compressing objects: 85% (6/7)\u001b[K\rremote: Compressing objects: 100% (7/7)\u001b[K\rremote: Compressing objects: 100% (7/7), done.\u001b[K\n",
"remote: Total 21 (delta 14), reused 20 (delta 13), pack-reused 0\u001b[K\n",
"Unpacking objects: 4% (1/21) \rUnpacking objects: 9% (2/21) \rUnpacking objects: 14% (3/21) \rUnpacking objects: 19% (4/21) \rUnpacking objects: 23% (5/21) \rUnpacking objects: 28% (6/21) \rUnpacking objects: 33% (7/21) \rUnpacking objects: 38% (8/21) \rUnpacking objects: 42% (9/21) \rUnpacking objects: 47% (10/21) \rUnpacking objects: 52% (11/21) \rUnpacking objects: 57% (12/21) \rUnpacking objects: 61% (13/21) \rUnpacking objects: 66% (14/21) \rUnpacking objects: 71% (15/21) \rUnpacking objects: 76% (16/21) \rUnpacking objects: 80% (17/21) \rUnpacking objects: 85% (18/21) \rUnpacking objects: 90% (19/21) \rUnpacking objects: 95% (20/21) \rUnpacking objects: 100% (21/21) \rUnpacking objects: 100% (21/21), done.\n",
"From https://github.com/taroushirani/nnsvs\n",
" 9e3c19f..7d94ecf cnn_mdn_test -> origin/cnn_mdn_test\n",
" * [new branch] cnn_mdn_dev_fixed -> origin/cnn_mdn_dev_fixed\n",
" * [new branch] cnn_mdn_test_fixed -> origin/cnn_mdn_test_fixed\n",
" * [new branch] fix_mdn_loss -> origin/fix_mdn_loss\n",
"D\tegs/nit-song070/svs-world-conv/local/data_prep.py\n",
"Branch 'cnn_mdn_test_fixed' set up to track remote branch 'cnn_mdn_test_fixed' from 'origin'.\n",
"Switched to a new branch 'cnn_mdn_test_fixed'\n",
" Building wheel for nnsvs (setup.py) ... \u001b[?25l\u001b[?25hdone\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "oHcn9Pmpm6IJ",
"outputId": "adfb4d67-0296-4e48-d69c-247d80924978",
"colab": {
"base_uri": "https://localhost:8080/"
}
},
"source": [
"! cd $RECIPE_ROOT && bash run.sh --stage 4 --stop-stage 4"
],
"execution_count": 14,
"outputs": [
{
"output_type": "stream",
"text": [
"stage 4: Training acoustic model\n",
"++ nnsvs-train --config-dir conf/train data.train_no_dev.in_dir=dump/yoko/norm/train_no_dev/in_acoustic/ data.train_no_dev.out_dir=dump/yoko/norm/train_no_dev/out_acoustic/ data.dev.in_dir=dump/yoko/norm/dev/in_acoustic/ data.dev.out_dir=dump/yoko/norm/dev/out_acoustic/ model=acoustic_cnn_mdn train.out_dir=exp/yoko/acoustic data.batch_size=2 resume.checkpoint=\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:47:11,250\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"data:\n",
" train_no_dev:\n",
" in_dir: dump/yoko/norm/train_no_dev/in_acoustic/\n",
" out_dir: dump/yoko/norm/train_no_dev/out_acoustic/\n",
" dev:\n",
" in_dir: dump/yoko/norm/dev/in_acoustic/\n",
" out_dir: dump/yoko/norm/dev/out_acoustic/\n",
" num_workers: 2\n",
" batch_size: 2\n",
" pin_memory: true\n",
"optim:\n",
" optimizer:\n",
" name: Adam\n",
" params:\n",
" lr: 0.001\n",
" betas:\n",
" - 0.5\n",
" - 0.999\n",
" weight_decay: 0.0\n",
" lr_scheduler:\n",
" name: StepLR\n",
" params:\n",
" step_size: 20\n",
" gamma: 0.5\n",
"train:\n",
" out_dir: exp/yoko/acoustic\n",
" nepochs: 50\n",
" checkpoint_epoch_interval: 20\n",
" stream_wise_loss: false\n",
" use_detect_anomaly: true\n",
"resume:\n",
" checkpoint: ''\n",
" load_optimizer: false\n",
"cudnn:\n",
" benchmark: false\n",
" deterministic: false\n",
"model:\n",
" stream_sizes:\n",
" - 180\n",
" - 3\n",
" - 1\n",
" - 15\n",
" has_dynamic_features:\n",
" - true\n",
" - true\n",
" - false\n",
" - true\n",
" num_windows: 3\n",
" stream_weights: null\n",
" netG:\n",
" _target_: nnsvs.model.Conv1dResnetMDN\n",
" in_dim: 424\n",
" out_dim: 199\n",
" hidden_dim: 256\n",
" num_layers: 6\n",
" dropout: 0.1\n",
" num_gaussians: 4\n",
"\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:11,251\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - cudnn.deterministic: False\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:11,251\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - cudnn.benchmark: False\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:11,251\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Set to use torch.autograd.detect_anomaly\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:16,676\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 9120, 424]), torch.Size([2, 9120, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:17,031\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 16320, 424]), torch.Size([2, 16320, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:17,132\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 16318, 424]), torch.Size([2, 16318, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:17,535\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 17760, 424]), torch.Size([2, 17760, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:17,728\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 16800, 424]), torch.Size([2, 16800, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:18,236\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 18360, 424]), torch.Size([2, 18360, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:18,751\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 16320, 424]), torch.Size([2, 16320, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:19,527\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 29760, 424]), torch.Size([2, 29760, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:19,527\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 15840, 424]), torch.Size([2, 15840, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:19,898\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 19680, 424]), torch.Size([2, 19680, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:20,029\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 16306, 424]), torch.Size([2, 16306, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:20,450\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 19920, 424]), torch.Size([2, 19920, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:20,450\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 8640, 424]), torch.Size([2, 8640, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:20,584\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([2, 18000, 424]), torch.Size([2, 18000, 199]), torch.Size([2])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:20,584\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([1, 16320, 424]), torch.Size([1, 16320, 199]), torch.Size([1])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:20,802\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - torch.Size([1, 6240, 424]), torch.Size([1, 6240, 199]), torch.Size([1])\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:20,866\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Start utterance-wise training...\u001b[0m\n",
" 0% 0/50 [00:00<?, ?it/s][\u001b[36m2020-11-08 06:47:29,478\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 1]: loss 290.66332194010414\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:29,770\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 1]: loss 286.694091796875\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:29,884\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 286.694091796875: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 2% 1/50 [00:09<07:21, 9.01s/it][\u001b[36m2020-11-08 06:47:36,070\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 2]: loss 268.7448028564453\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:36,372\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 2]: loss 278.47576904296875\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:36,485\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 278.47576904296875: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 4% 2/50 [00:15<06:37, 8.29s/it][\u001b[36m2020-11-08 06:47:42,691\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 3]: loss 261.20750732421874\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:42,990\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 3]: loss 268.23101806640625\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:43,102\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 268.23101806640625: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 6% 3/50 [00:22<06:05, 7.79s/it][\u001b[36m2020-11-08 06:47:49,253\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 4]: loss 257.40777893066405\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:49,560\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 4]: loss 264.79400634765625\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:49,677\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 264.79400634765625: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 8% 4/50 [00:28<05:41, 7.42s/it][\u001b[36m2020-11-08 06:47:55,869\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 5]: loss 255.5629689534505\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:56,163\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 5]: loss 262.9654235839844\u001b[0m\n",
"[\u001b[36m2020-11-08 06:47:56,280\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 262.9654235839844: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 10% 5/50 [00:35<05:22, 7.18s/it][\u001b[36m2020-11-08 06:48:02,802\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 6]: loss 254.5532430013021\u001b[0m\n",
"[\u001b[36m2020-11-08 06:48:03,094\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 6]: loss 264.51214599609375\u001b[0m\n",
" 12% 6/50 [00:42<05:10, 7.07s/it][\u001b[36m2020-11-08 06:48:09,123\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 7]: loss 255.44750569661457\u001b[0m\n",
"[\u001b[36m2020-11-08 06:48:09,422\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 7]: loss 261.703125\u001b[0m\n",
"[\u001b[36m2020-11-08 06:48:09,538\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 261.703125: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 14% 7/50 [00:48<04:55, 6.88s/it][\u001b[36m2020-11-08 06:48:15,796\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 8]: loss 254.5531748453776\u001b[0m\n",
"[\u001b[36m2020-11-08 06:48:16,089\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 8]: loss 264.3781433105469\u001b[0m\n",
" 16% 8/50 [00:55<04:44, 6.78s/it][\u001b[36m2020-11-08 06:48:22,506\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 9]: loss 256.7108154296875\u001b[0m\n",
"[\u001b[36m2020-11-08 06:48:22,800\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 9]: loss 262.3298645019531\u001b[0m\n",
" 18% 9/50 [01:01<04:37, 6.76s/it][\u001b[36m2020-11-08 06:48:29,287\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 10]: loss 253.41836547851562\u001b[0m\n",
"[\u001b[36m2020-11-08 06:48:29,580\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 10]: loss 261.201904296875\u001b[0m\n",
"[\u001b[36m2020-11-08 06:48:29,692\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 261.201904296875: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 20% 10/50 [01:08<04:31, 6.80s/it][\u001b[36m2020-11-08 06:48:35,956\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 11]: loss 253.29989217122395\u001b[0m\n",
"[\u001b[36m2020-11-08 06:48:36,252\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 11]: loss 266.64764404296875\u001b[0m\n",
" 22% 11/50 [01:15<04:22, 6.73s/it][\u001b[36m2020-11-08 06:48:42,626\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 12]: loss 256.0811696370443\u001b[0m\n",
"[\u001b[36m2020-11-08 06:48:42,915\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 12]: loss 260.91650390625\u001b[0m\n",
"[\u001b[36m2020-11-08 06:48:43,027\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 260.91650390625: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 24% 12/50 [01:22<04:16, 6.74s/it][\u001b[36m2020-11-08 06:48:49,282\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 13]: loss 253.55648701985677\u001b[0m\n",
"[\u001b[36m2020-11-08 06:48:49,571\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 13]: loss 260.2218933105469\u001b[0m\n",
"[\u001b[36m2020-11-08 06:48:49,687\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 260.2218933105469: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 26% 13/50 [01:28<04:08, 6.72s/it][\u001b[36m2020-11-08 06:48:55,970\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 14]: loss 254.81593424479166\u001b[0m\n",
"[\u001b[36m2020-11-08 06:48:56,268\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 14]: loss 265.8514404296875\u001b[0m\n",
" 28% 14/50 [01:35<04:00, 6.68s/it][\u001b[36m2020-11-08 06:49:02,650\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 15]: loss 256.91652628580727\u001b[0m\n",
"[\u001b[36m2020-11-08 06:49:02,944\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 15]: loss 263.7359313964844\u001b[0m\n",
" 30% 15/50 [01:42<03:53, 6.68s/it][\u001b[36m2020-11-08 06:49:09,230\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 16]: loss 255.08655598958333\u001b[0m\n",
"[\u001b[36m2020-11-08 06:49:09,519\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 16]: loss 262.5855407714844\u001b[0m\n",
" 32% 16/50 [01:48<03:45, 6.65s/it][\u001b[36m2020-11-08 06:49:15,813\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 17]: loss 254.9052530924479\u001b[0m\n",
"[\u001b[36m2020-11-08 06:49:16,110\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 17]: loss 261.9363098144531\u001b[0m\n",
" 34% 17/50 [01:55<03:38, 6.63s/it][\u001b[36m2020-11-08 06:49:22,627\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 18]: loss 253.19541625976564\u001b[0m\n",
"[\u001b[36m2020-11-08 06:49:22,917\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 18]: loss 260.76806640625\u001b[0m\n",
" 36% 18/50 [02:02<03:33, 6.68s/it][\u001b[36m2020-11-08 06:49:29,510\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 19]: loss 257.0956797281901\u001b[0m\n",
"[\u001b[36m2020-11-08 06:49:29,808\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 19]: loss 263.9874572753906\u001b[0m\n",
" 38% 19/50 [02:08<03:29, 6.75s/it][\u001b[36m2020-11-08 06:49:36,050\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 20]: loss 255.02989095052084\u001b[0m\n",
"[\u001b[36m2020-11-08 06:49:36,345\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 20]: loss 266.4212341308594\u001b[0m\n",
"[\u001b[36m2020-11-08 06:49:36,446\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/checkpoint_epoch0020.pth\u001b[0m\n",
" 40% 20/50 [02:15<03:21, 6.73s/it][\u001b[36m2020-11-08 06:49:42,935\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 21]: loss 253.53114013671876\u001b[0m\n",
"[\u001b[36m2020-11-08 06:49:43,222\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 21]: loss 261.0312805175781\u001b[0m\n",
" 42% 21/50 [02:22<03:15, 6.73s/it][\u001b[36m2020-11-08 06:49:49,343\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 22]: loss 252.14724731445312\u001b[0m\n",
"[\u001b[36m2020-11-08 06:49:49,638\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 22]: loss 263.2264709472656\u001b[0m\n",
" 44% 22/50 [02:28<03:05, 6.63s/it][\u001b[36m2020-11-08 06:49:56,148\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 23]: loss 250.82276814778646\u001b[0m\n",
"[\u001b[36m2020-11-08 06:49:56,436\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 23]: loss 259.16339111328125\u001b[0m\n",
"[\u001b[36m2020-11-08 06:49:56,550\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 259.16339111328125: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 46% 23/50 [02:35<03:01, 6.72s/it][\u001b[36m2020-11-08 06:50:02,955\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 24]: loss 250.22264811197917\u001b[0m\n",
"[\u001b[36m2020-11-08 06:50:03,241\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 24]: loss 258.8886413574219\u001b[0m\n",
"[\u001b[36m2020-11-08 06:50:03,354\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 258.8886413574219: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 48% 24/50 [02:42<02:55, 6.74s/it][\u001b[36m2020-11-08 06:50:09,656\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 25]: loss 251.36710510253906\u001b[0m\n",
"[\u001b[36m2020-11-08 06:50:09,952\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 25]: loss 258.0323791503906\u001b[0m\n",
"[\u001b[36m2020-11-08 06:50:10,065\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 258.0323791503906: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 50% 25/50 [02:49<02:48, 6.73s/it][\u001b[36m2020-11-08 06:50:16,489\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 26]: loss 249.34935506184897\u001b[0m\n",
"[\u001b[36m2020-11-08 06:50:16,782\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 26]: loss 258.8794860839844\u001b[0m\n",
" 52% 26/50 [02:55<02:41, 6.73s/it][\u001b[36m2020-11-08 06:50:23,235\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 27]: loss 248.94583638509116\u001b[0m\n",
"[\u001b[36m2020-11-08 06:50:23,522\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 27]: loss 259.0494079589844\u001b[0m\n",
" 54% 27/50 [03:02<02:34, 6.73s/it][\u001b[36m2020-11-08 06:50:30,073\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 28]: loss 249.069438680013\u001b[0m\n",
"[\u001b[36m2020-11-08 06:50:30,369\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 28]: loss 257.54437255859375\u001b[0m\n",
"[\u001b[36m2020-11-08 06:50:30,480\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 257.54437255859375: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 56% 28/50 [03:09<02:29, 6.80s/it][\u001b[36m2020-11-08 06:50:36,662\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 29]: loss 248.96646219889323\u001b[0m\n",
"[\u001b[36m2020-11-08 06:50:36,966\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 29]: loss 258.36297607421875\u001b[0m\n",
" 58% 29/50 [03:16<02:20, 6.71s/it][\u001b[36m2020-11-08 06:50:43,089\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 30]: loss 249.55107727050782\u001b[0m\n",
"[\u001b[36m2020-11-08 06:50:43,392\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 30]: loss 258.1591796875\u001b[0m\n",
" 60% 30/50 [03:22<02:12, 6.62s/it][\u001b[36m2020-11-08 06:50:50,662\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 31]: loss 248.5025400797526\u001b[0m\n",
"[\u001b[36m2020-11-08 06:50:50,951\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 31]: loss 259.5970153808594\u001b[0m\n",
" 62% 31/50 [03:30<02:11, 6.90s/it][\u001b[36m2020-11-08 06:50:57,287\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 32]: loss 249.64814351399738\u001b[0m\n",
"[\u001b[36m2020-11-08 06:50:57,582\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 32]: loss 256.8604736328125\u001b[0m\n",
"[\u001b[36m2020-11-08 06:50:57,695\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 256.8604736328125: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 64% 32/50 [03:36<02:03, 6.86s/it][\u001b[36m2020-11-08 06:51:04,113\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 33]: loss 249.5083241780599\u001b[0m\n",
"[\u001b[36m2020-11-08 06:51:04,400\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 33]: loss 257.20147705078125\u001b[0m\n",
" 66% 33/50 [03:43<01:55, 6.81s/it][\u001b[36m2020-11-08 06:51:11,289\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 34]: loss 248.32440795898438\u001b[0m\n",
"[\u001b[36m2020-11-08 06:51:11,578\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 34]: loss 256.60321044921875\u001b[0m\n",
"[\u001b[36m2020-11-08 06:51:11,691\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 256.60321044921875: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 68% 34/50 [03:50<01:51, 6.95s/it][\u001b[36m2020-11-08 06:51:18,166\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 35]: loss 247.54581807454426\u001b[0m\n",
"[\u001b[36m2020-11-08 06:51:18,459\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 35]: loss 255.73919677734375\u001b[0m\n",
"[\u001b[36m2020-11-08 06:51:18,580\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 255.73919677734375: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 70% 35/50 [03:57<01:44, 6.93s/it][\u001b[36m2020-11-08 06:51:24,930\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 36]: loss 247.2222676595052\u001b[0m\n",
"[\u001b[36m2020-11-08 06:51:25,221\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 36]: loss 258.358154296875\u001b[0m\n",
" 72% 36/50 [04:04<01:35, 6.85s/it][\u001b[36m2020-11-08 06:51:31,690\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 37]: loss 247.72347920735677\u001b[0m\n",
"[\u001b[36m2020-11-08 06:51:31,981\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 37]: loss 258.5198669433594\u001b[0m\n",
" 74% 37/50 [04:11<01:28, 6.82s/it][\u001b[36m2020-11-08 06:51:38,309\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 38]: loss 247.21583353678386\u001b[0m\n",
"[\u001b[36m2020-11-08 06:51:38,595\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 38]: loss 256.7582092285156\u001b[0m\n",
" 76% 38/50 [04:17<01:21, 6.76s/it][\u001b[36m2020-11-08 06:51:44,863\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 39]: loss 246.56411641438802\u001b[0m\n",
"[\u001b[36m2020-11-08 06:51:45,152\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 39]: loss 257.2162780761719\u001b[0m\n",
" 78% 39/50 [04:24<01:13, 6.70s/it][\u001b[36m2020-11-08 06:51:51,503\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 40]: loss 246.82678629557293\u001b[0m\n",
"[\u001b[36m2020-11-08 06:51:51,794\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 40]: loss 260.40899658203125\u001b[0m\n",
"[\u001b[36m2020-11-08 06:51:51,896\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/checkpoint_epoch0040.pth\u001b[0m\n",
" 80% 40/50 [04:31<01:07, 6.73s/it][\u001b[36m2020-11-08 06:51:58,912\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 41]: loss 245.7894744873047\u001b[0m\n",
"[\u001b[36m2020-11-08 06:51:59,202\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 41]: loss 254.4222412109375\u001b[0m\n",
"[\u001b[36m2020-11-08 06:51:59,316\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 254.4222412109375: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 82% 41/50 [04:38<01:02, 6.92s/it][\u001b[36m2020-11-08 06:52:05,527\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 42]: loss 245.47579243977864\u001b[0m\n",
"[\u001b[36m2020-11-08 06:52:05,829\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 42]: loss 254.70584106445312\u001b[0m\n",
" 84% 42/50 [04:44<00:54, 6.80s/it][\u001b[36m2020-11-08 06:52:12,117\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 43]: loss 245.03537089029948\u001b[0m\n",
"[\u001b[36m2020-11-08 06:52:12,412\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 43]: loss 253.3751220703125\u001b[0m\n",
"[\u001b[36m2020-11-08 06:52:12,523\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 253.3751220703125: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 86% 43/50 [04:51<00:47, 6.77s/it][\u001b[36m2020-11-08 06:52:18,686\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 44]: loss 246.16766052246095\u001b[0m\n",
"[\u001b[36m2020-11-08 06:52:18,983\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 44]: loss 253.8214874267578\u001b[0m\n",
" 88% 44/50 [04:58<00:40, 6.67s/it][\u001b[36m2020-11-08 06:52:25,286\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 45]: loss 244.8346201578776\u001b[0m\n",
"[\u001b[36m2020-11-08 06:52:25,575\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 45]: loss 255.14019775390625\u001b[0m\n",
" 90% 45/50 [05:04<00:33, 6.65s/it][\u001b[36m2020-11-08 06:52:32,038\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 46]: loss 244.4345906575521\u001b[0m\n",
"[\u001b[36m2020-11-08 06:52:32,332\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 46]: loss 254.6049346923828\u001b[0m\n",
" 92% 46/50 [05:11<00:26, 6.68s/it][\u001b[36m2020-11-08 06:52:38,698\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 47]: loss 244.48934020996094\u001b[0m\n",
"[\u001b[36m2020-11-08 06:52:38,994\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 47]: loss 253.2248077392578\u001b[0m\n",
"[\u001b[36m2020-11-08 06:52:39,107\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 253.2248077392578: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 94% 47/50 [05:18<00:20, 6.71s/it][\u001b[36m2020-11-08 06:52:45,729\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 48]: loss 244.05358479817707\u001b[0m\n",
"[\u001b[36m2020-11-08 06:52:46,023\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 48]: loss 252.91026306152344\u001b[0m\n",
"[\u001b[36m2020-11-08 06:52:46,134\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 252.91026306152344: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
" 96% 48/50 [05:25<00:13, 6.80s/it][\u001b[36m2020-11-08 06:52:52,488\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 49]: loss 245.50194193522137\u001b[0m\n",
"[\u001b[36m2020-11-08 06:52:52,778\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 49]: loss 253.00778198242188\u001b[0m\n",
" 98% 49/50 [05:31<00:06, 6.76s/it][\u001b[36m2020-11-08 06:52:58,951\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [train_no_dev] [Epoch 50]: loss 243.93228759765626\u001b[0m\n",
"[\u001b[36m2020-11-08 06:52:59,243\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [dev] [Epoch 50]: loss 252.7050323486328\u001b[0m\n",
"[\u001b[36m2020-11-08 06:52:59,355\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - [Best loss 252.7050323486328: checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/best_loss.pth\u001b[0m\n",
"100% 50/50 [05:38<00:00, 6.77s/it]\n",
"[\u001b[36m2020-11-08 06:52:59,451\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Checkpoint is saved at /content/nnsvs/egs/nit-song070/svs-world-cnn-mdn/exp/yoko/acoustic/checkpoint_epoch0050.pth\u001b[0m\n",
"[\u001b[36m2020-11-08 06:52:59,509\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - The best loss was 252.7050323486328\u001b[0m\n",
"++ set +x\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "sVU6ra7TDHEZ",
"outputId": "6113ca09-3387-4db7-cd66-2c5c0287ea72",
"colab": {
"base_uri": "https://localhost:8080/"
}
},
"source": [
"! cd $RECIPE_ROOT && bash run.sh --stage 5 --stop-stage 5"
],
"execution_count": 15,
"outputs": [
{
"output_type": "stream",
"text": [
"stage 5: Generate features from timelag/duration/acoustic models\n",
"++ nnsvs-generate model.checkpoint=exp/yoko/timelag/latest.pth model.model_yaml=exp/yoko/timelag/model.yaml out_scaler_path=dump/yoko/norm/out_timelag_scaler.joblib in_dir=dump/yoko/norm/dev/in_timelag/ out_dir=exp/yoko/timelag/predicted/dev/latest/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:53:02,288\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/norm/dev/in_timelag/\n",
"out_dir: exp/yoko/timelag/predicted/dev/latest/\n",
"out_scaler_path: dump/yoko/norm/out_timelag_scaler.joblib\n",
"model:\n",
" checkpoint: exp/yoko/timelag/latest.pth\n",
" model_yaml: exp/yoko/timelag/model.yaml\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 55.70it/s]\n",
"++ set +x\n",
"++ nnsvs-generate model.checkpoint=exp/yoko/duration/latest.pth model.model_yaml=exp/yoko/duration/model.yaml out_scaler_path=dump/yoko/norm/out_duration_scaler.joblib in_dir=dump/yoko/norm/dev/in_duration/ out_dir=exp/yoko/duration/predicted/dev/latest/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:53:09,555\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/norm/dev/in_duration/\n",
"out_dir: exp/yoko/duration/predicted/dev/latest/\n",
"out_scaler_path: dump/yoko/norm/out_duration_scaler.joblib\n",
"model:\n",
" checkpoint: exp/yoko/duration/latest.pth\n",
" model_yaml: exp/yoko/duration/model.yaml\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 146.02it/s]\n",
"++ set +x\n",
"++ nnsvs-generate model.checkpoint=exp/yoko/acoustic/latest.pth model.model_yaml=exp/yoko/acoustic/model.yaml out_scaler_path=dump/yoko/norm/out_acoustic_scaler.joblib in_dir=dump/yoko/norm/dev/in_acoustic/ out_dir=exp/yoko/acoustic/predicted/dev/latest/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:53:17,638\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/norm/dev/in_acoustic/\n",
"out_dir: exp/yoko/acoustic/predicted/dev/latest/\n",
"out_scaler_path: dump/yoko/norm/out_acoustic_scaler.joblib\n",
"model:\n",
" checkpoint: exp/yoko/acoustic/latest.pth\n",
" model_yaml: exp/yoko/acoustic/model.yaml\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 8.32it/s]\n",
"++ set +x\n",
"++ nnsvs-generate model.checkpoint=exp/yoko/timelag/latest.pth model.model_yaml=exp/yoko/timelag/model.yaml out_scaler_path=dump/yoko/norm/out_timelag_scaler.joblib in_dir=dump/yoko/norm/eval/in_timelag/ out_dir=exp/yoko/timelag/predicted/eval/latest/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:53:24,866\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/norm/eval/in_timelag/\n",
"out_dir: exp/yoko/timelag/predicted/eval/latest/\n",
"out_scaler_path: dump/yoko/norm/out_timelag_scaler.joblib\n",
"model:\n",
" checkpoint: exp/yoko/timelag/latest.pth\n",
" model_yaml: exp/yoko/timelag/model.yaml\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 142.83it/s]\n",
"++ set +x\n",
"++ nnsvs-generate model.checkpoint=exp/yoko/duration/latest.pth model.model_yaml=exp/yoko/duration/model.yaml out_scaler_path=dump/yoko/norm/out_duration_scaler.joblib in_dir=dump/yoko/norm/eval/in_duration/ out_dir=exp/yoko/duration/predicted/eval/latest/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:53:31,822\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/norm/eval/in_duration/\n",
"out_dir: exp/yoko/duration/predicted/eval/latest/\n",
"out_scaler_path: dump/yoko/norm/out_duration_scaler.joblib\n",
"model:\n",
" checkpoint: exp/yoko/duration/latest.pth\n",
" model_yaml: exp/yoko/duration/model.yaml\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 135.85it/s]\n",
"++ set +x\n",
"++ nnsvs-generate model.checkpoint=exp/yoko/acoustic/latest.pth model.model_yaml=exp/yoko/acoustic/model.yaml out_scaler_path=dump/yoko/norm/out_acoustic_scaler.joblib in_dir=dump/yoko/norm/eval/in_acoustic/ out_dir=exp/yoko/acoustic/predicted/eval/latest/\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:53:38,819\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"in_dir: dump/yoko/norm/eval/in_acoustic/\n",
"out_dir: exp/yoko/acoustic/predicted/eval/latest/\n",
"out_scaler_path: dump/yoko/norm/out_acoustic_scaler.joblib\n",
"model:\n",
" checkpoint: exp/yoko/acoustic/latest.pth\n",
" model_yaml: exp/yoko/acoustic/model.yaml\n",
"\u001b[0m\n",
"100% 1/1 [00:00<00:00, 8.46it/s]\n",
"++ set +x\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "_GuQN0zEGXjA",
"outputId": "6400672c-1155-4bd9-f0d9-c8870f1084bd",
"colab": {
"base_uri": "https://localhost:8080/"
}
},
"source": [
"! cd $RECIPE_ROOT && bash run.sh --stage 6 --stop-stage 6"
],
"execution_count": 16,
"outputs": [
{
"output_type": "stream",
"text": [
"stage 6: Synthesis waveforms\n",
"++ nnsvs-synthesis question_path=../../_common/hed/jp_qst001_nnsvs.hed timelag=defaults duration=defaults acoustic=defaults timelag.checkpoint=exp/yoko/timelag/latest.pth timelag.in_scaler_path=dump/yoko/norm/in_timelag_scaler.joblib timelag.out_scaler_path=dump/yoko/norm/out_timelag_scaler.joblib timelag.model_yaml=exp/yoko/timelag/model.yaml duration.checkpoint=exp/yoko/duration/latest.pth duration.in_scaler_path=dump/yoko/norm/in_duration_scaler.joblib duration.out_scaler_path=dump/yoko/norm/out_duration_scaler.joblib duration.model_yaml=exp/yoko/duration/model.yaml acoustic.checkpoint=exp/yoko/acoustic/latest.pth acoustic.in_scaler_path=dump/yoko/norm/in_acoustic_scaler.joblib acoustic.out_scaler_path=dump/yoko/norm/out_acoustic_scaler.joblib acoustic.model_yaml=exp/yoko/acoustic/model.yaml utt_list=./data/list/dev.list in_dir=data/acoustic/label_phone_score/ out_dir=exp/yoko/synthesis/dev/latest/label_phone_score ground_truth_duration=false\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:53:46,239\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"device: cuda\n",
"utt_list: ./data/list/dev.list\n",
"in_dir: data/acoustic/label_phone_score/\n",
"out_dir: exp/yoko/synthesis/dev/latest/label_phone_score\n",
"label_path: null\n",
"out_wav_path: null\n",
"sample_rate: 48000\n",
"frame_period: 5\n",
"question_path: ../../_common/hed/jp_qst001_nnsvs.hed\n",
"log_f0_conditioning: true\n",
"ground_truth_duration: false\n",
"gain_normalize: false\n",
"stats_dir: null\n",
"model_dir: null\n",
"model_checkpoint: latest.pth\n",
"timelag:\n",
" question_path: null\n",
" checkpoint: exp/yoko/timelag/latest.pth\n",
" in_scaler_path: dump/yoko/norm/in_timelag_scaler.joblib\n",
" out_scaler_path: dump/yoko/norm/out_timelag_scaler.joblib\n",
" model_yaml: exp/yoko/timelag/model.yaml\n",
" allowed_range:\n",
" - -20\n",
" - 20\n",
" allowed_range_rest:\n",
" - -40\n",
" - 40\n",
"duration:\n",
" question_path: null\n",
" checkpoint: exp/yoko/duration/latest.pth\n",
" in_scaler_path: dump/yoko/norm/in_duration_scaler.joblib\n",
" out_scaler_path: dump/yoko/norm/out_duration_scaler.joblib\n",
" model_yaml: exp/yoko/duration/model.yaml\n",
"acoustic:\n",
" question_path: null\n",
" checkpoint: exp/yoko/acoustic/latest.pth\n",
" in_scaler_path: dump/yoko/norm/in_acoustic_scaler.joblib\n",
" out_scaler_path: dump/yoko/norm/out_acoustic_scaler.joblib\n",
" model_yaml: exp/yoko/acoustic/model.yaml\n",
" subphone_features: coarse_coding\n",
" relative_f0: true\n",
" post_filter: true\n",
"\u001b[0m\n",
"[\u001b[36m2020-11-08 06:53:51,115\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Processes 1 utterances...\u001b[0m\n",
"100% 1/1 [00:14<00:00, 14.57s/it]\n",
"++ set +x\n",
"++ nnsvs-synthesis question_path=../../_common/hed/jp_qst001_nnsvs.hed timelag=defaults duration=defaults acoustic=defaults timelag.checkpoint=exp/yoko/timelag/latest.pth timelag.in_scaler_path=dump/yoko/norm/in_timelag_scaler.joblib timelag.out_scaler_path=dump/yoko/norm/out_timelag_scaler.joblib timelag.model_yaml=exp/yoko/timelag/model.yaml duration.checkpoint=exp/yoko/duration/latest.pth duration.in_scaler_path=dump/yoko/norm/in_duration_scaler.joblib duration.out_scaler_path=dump/yoko/norm/out_duration_scaler.joblib duration.model_yaml=exp/yoko/duration/model.yaml acoustic.checkpoint=exp/yoko/acoustic/latest.pth acoustic.in_scaler_path=dump/yoko/norm/in_acoustic_scaler.joblib acoustic.out_scaler_path=dump/yoko/norm/out_acoustic_scaler.joblib acoustic.model_yaml=exp/yoko/acoustic/model.yaml utt_list=./data/list/dev.list in_dir=data/acoustic/label_phone_align/ out_dir=exp/yoko/synthesis/dev/latest/label_phone_align ground_truth_duration=true\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:54:08,208\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"device: cuda\n",
"utt_list: ./data/list/dev.list\n",
"in_dir: data/acoustic/label_phone_align/\n",
"out_dir: exp/yoko/synthesis/dev/latest/label_phone_align\n",
"label_path: null\n",
"out_wav_path: null\n",
"sample_rate: 48000\n",
"frame_period: 5\n",
"question_path: ../../_common/hed/jp_qst001_nnsvs.hed\n",
"log_f0_conditioning: true\n",
"ground_truth_duration: true\n",
"gain_normalize: false\n",
"stats_dir: null\n",
"model_dir: null\n",
"model_checkpoint: latest.pth\n",
"timelag:\n",
" question_path: null\n",
" checkpoint: exp/yoko/timelag/latest.pth\n",
" in_scaler_path: dump/yoko/norm/in_timelag_scaler.joblib\n",
" out_scaler_path: dump/yoko/norm/out_timelag_scaler.joblib\n",
" model_yaml: exp/yoko/timelag/model.yaml\n",
" allowed_range:\n",
" - -20\n",
" - 20\n",
" allowed_range_rest:\n",
" - -40\n",
" - 40\n",
"duration:\n",
" question_path: null\n",
" checkpoint: exp/yoko/duration/latest.pth\n",
" in_scaler_path: dump/yoko/norm/in_duration_scaler.joblib\n",
" out_scaler_path: dump/yoko/norm/out_duration_scaler.joblib\n",
" model_yaml: exp/yoko/duration/model.yaml\n",
"acoustic:\n",
" question_path: null\n",
" checkpoint: exp/yoko/acoustic/latest.pth\n",
" in_scaler_path: dump/yoko/norm/in_acoustic_scaler.joblib\n",
" out_scaler_path: dump/yoko/norm/out_acoustic_scaler.joblib\n",
" model_yaml: exp/yoko/acoustic/model.yaml\n",
" subphone_features: coarse_coding\n",
" relative_f0: true\n",
" post_filter: true\n",
"\u001b[0m\n",
"[\u001b[36m2020-11-08 06:54:13,142\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Processes 1 utterances...\u001b[0m\n",
"100% 1/1 [00:14<00:00, 14.20s/it]\n",
"++ set +x\n",
"++ nnsvs-synthesis question_path=../../_common/hed/jp_qst001_nnsvs.hed timelag=defaults duration=defaults acoustic=defaults timelag.checkpoint=exp/yoko/timelag/latest.pth timelag.in_scaler_path=dump/yoko/norm/in_timelag_scaler.joblib timelag.out_scaler_path=dump/yoko/norm/out_timelag_scaler.joblib timelag.model_yaml=exp/yoko/timelag/model.yaml duration.checkpoint=exp/yoko/duration/latest.pth duration.in_scaler_path=dump/yoko/norm/in_duration_scaler.joblib duration.out_scaler_path=dump/yoko/norm/out_duration_scaler.joblib duration.model_yaml=exp/yoko/duration/model.yaml acoustic.checkpoint=exp/yoko/acoustic/latest.pth acoustic.in_scaler_path=dump/yoko/norm/in_acoustic_scaler.joblib acoustic.out_scaler_path=dump/yoko/norm/out_acoustic_scaler.joblib acoustic.model_yaml=exp/yoko/acoustic/model.yaml utt_list=./data/list/eval.list in_dir=data/acoustic/label_phone_score/ out_dir=exp/yoko/synthesis/eval/latest/label_phone_score ground_truth_duration=false\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:54:29,873\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"device: cuda\n",
"utt_list: ./data/list/eval.list\n",
"in_dir: data/acoustic/label_phone_score/\n",
"out_dir: exp/yoko/synthesis/eval/latest/label_phone_score\n",
"label_path: null\n",
"out_wav_path: null\n",
"sample_rate: 48000\n",
"frame_period: 5\n",
"question_path: ../../_common/hed/jp_qst001_nnsvs.hed\n",
"log_f0_conditioning: true\n",
"ground_truth_duration: false\n",
"gain_normalize: false\n",
"stats_dir: null\n",
"model_dir: null\n",
"model_checkpoint: latest.pth\n",
"timelag:\n",
" question_path: null\n",
" checkpoint: exp/yoko/timelag/latest.pth\n",
" in_scaler_path: dump/yoko/norm/in_timelag_scaler.joblib\n",
" out_scaler_path: dump/yoko/norm/out_timelag_scaler.joblib\n",
" model_yaml: exp/yoko/timelag/model.yaml\n",
" allowed_range:\n",
" - -20\n",
" - 20\n",
" allowed_range_rest:\n",
" - -40\n",
" - 40\n",
"duration:\n",
" question_path: null\n",
" checkpoint: exp/yoko/duration/latest.pth\n",
" in_scaler_path: dump/yoko/norm/in_duration_scaler.joblib\n",
" out_scaler_path: dump/yoko/norm/out_duration_scaler.joblib\n",
" model_yaml: exp/yoko/duration/model.yaml\n",
"acoustic:\n",
" question_path: null\n",
" checkpoint: exp/yoko/acoustic/latest.pth\n",
" in_scaler_path: dump/yoko/norm/in_acoustic_scaler.joblib\n",
" out_scaler_path: dump/yoko/norm/out_acoustic_scaler.joblib\n",
" model_yaml: exp/yoko/acoustic/model.yaml\n",
" subphone_features: coarse_coding\n",
" relative_f0: true\n",
" post_filter: true\n",
"\u001b[0m\n",
"[\u001b[36m2020-11-08 06:54:34,751\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Processes 1 utterances...\u001b[0m\n",
"100% 1/1 [00:14<00:00, 14.18s/it]\n",
"++ set +x\n",
"++ nnsvs-synthesis question_path=../../_common/hed/jp_qst001_nnsvs.hed timelag=defaults duration=defaults acoustic=defaults timelag.checkpoint=exp/yoko/timelag/latest.pth timelag.in_scaler_path=dump/yoko/norm/in_timelag_scaler.joblib timelag.out_scaler_path=dump/yoko/norm/out_timelag_scaler.joblib timelag.model_yaml=exp/yoko/timelag/model.yaml duration.checkpoint=exp/yoko/duration/latest.pth duration.in_scaler_path=dump/yoko/norm/in_duration_scaler.joblib duration.out_scaler_path=dump/yoko/norm/out_duration_scaler.joblib duration.model_yaml=exp/yoko/duration/model.yaml acoustic.checkpoint=exp/yoko/acoustic/latest.pth acoustic.in_scaler_path=dump/yoko/norm/in_acoustic_scaler.joblib acoustic.out_scaler_path=dump/yoko/norm/out_acoustic_scaler.joblib acoustic.model_yaml=exp/yoko/acoustic/model.yaml utt_list=./data/list/eval.list in_dir=data/acoustic/label_phone_align/ out_dir=exp/yoko/synthesis/eval/latest/label_phone_align ground_truth_duration=true\n",
"/usr/local/lib/python3.6/dist-packages/hydra/core/utils.py:204: UserWarning: \n",
"Using config_path to specify the config name is deprecated, specify the config name via config_name\n",
"See https://hydra.cc/docs/next/upgrades/0.11_to_1.0/config_path_changes\n",
" warnings.warn(category=UserWarning, message=msg)\n",
"/usr/local/lib/python3.6/dist-packages/omegaconf/basecontainer.py:232: UserWarning: cfg.pretty() is deprecated and will be removed in a future version.\n",
"Use OmegaConf.to_yaml(cfg)\n",
"\n",
" category=UserWarning,\n",
"[\u001b[36m2020-11-08 06:54:51,508\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - verbose: 100\n",
"device: cuda\n",
"utt_list: ./data/list/eval.list\n",
"in_dir: data/acoustic/label_phone_align/\n",
"out_dir: exp/yoko/synthesis/eval/latest/label_phone_align\n",
"label_path: null\n",
"out_wav_path: null\n",
"sample_rate: 48000\n",
"frame_period: 5\n",
"question_path: ../../_common/hed/jp_qst001_nnsvs.hed\n",
"log_f0_conditioning: true\n",
"ground_truth_duration: true\n",
"gain_normalize: false\n",
"stats_dir: null\n",
"model_dir: null\n",
"model_checkpoint: latest.pth\n",
"timelag:\n",
" question_path: null\n",
" checkpoint: exp/yoko/timelag/latest.pth\n",
" in_scaler_path: dump/yoko/norm/in_timelag_scaler.joblib\n",
" out_scaler_path: dump/yoko/norm/out_timelag_scaler.joblib\n",
" model_yaml: exp/yoko/timelag/model.yaml\n",
" allowed_range:\n",
" - -20\n",
" - 20\n",
" allowed_range_rest:\n",
" - -40\n",
" - 40\n",
"duration:\n",
" question_path: null\n",
" checkpoint: exp/yoko/duration/latest.pth\n",
" in_scaler_path: dump/yoko/norm/in_duration_scaler.joblib\n",
" out_scaler_path: dump/yoko/norm/out_duration_scaler.joblib\n",
" model_yaml: exp/yoko/duration/model.yaml\n",
"acoustic:\n",
" question_path: null\n",
" checkpoint: exp/yoko/acoustic/latest.pth\n",
" in_scaler_path: dump/yoko/norm/in_acoustic_scaler.joblib\n",
" out_scaler_path: dump/yoko/norm/out_acoustic_scaler.joblib\n",
" model_yaml: exp/yoko/acoustic/model.yaml\n",
" subphone_features: coarse_coding\n",
" relative_f0: true\n",
" post_filter: true\n",
"\u001b[0m\n",
"[\u001b[36m2020-11-08 06:54:56,494\u001b[0m][\u001b[34mnnsvs\u001b[0m][\u001b[32mINFO\u001b[0m] - Processes 1 utterances...\u001b[0m\n",
"100% 1/1 [00:13<00:00, 13.98s/it]\n",
"++ set +x\n"
],
"name": "stdout"
}
]
}
]
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment