travishsu

# uvx --with "git+https://github.com/travishsu/mimm.git" --with numpy ipython
import mlx.core as mx
from mimm import get_model, list_models

model = get_model('resnet50', pretrained=True)
img = mx.random.normal((3, 256, 256, 3))
features = model.features(img)
for layer, feature in features.items():
    print(layer, feature.shape)

travishsu

# uvx --with torch --with torchvision ipython
from torchvision.models import get_model
model = get_model('resnet50', pretrained=True)

# https://pytorch.org/vision/main/feature_extraction.html
from torchvision.models.feature_extraction import create_feature_extractor
return_nodes = {f'layer{i}': f'layer{i}' for i in range(1, 5)}
extractor = create_feature_extractor(model, return_nodes)

import torch

travishsu

# Tic Tac Toe Python implementation

import random
from math import exp

from tqdm import tqdm

NN_INPUT_SIZE = 18
NN_HIDDEN_SIZE = 100
NN_OUTPUT_SIZE = 9

travishsu

import os
import glob
import sys
from PyQt5.QtWidgets import QApplication, QWidget, QVBoxLayout, QPushButton, QListWidget, QHBoxLayout, QFileDialog


def compare_folders(dir1, dir2):
    files_in_dir1 = [
        f.replace('._', '').replace(dir1 + '/', '')
        for f in glob.glob(dir1 + '/**/*', recursive=True)

travishsu

def pascal(n):
    L = n + n - 1
    rows = []
    row = [1]
    rows.append(row)
    while len(rows) < n:
        row_cat = [0] + row + [0]
        row = [row_cat[i] + row_cat[i+1] for i in range(len(row_cat)-1)]
        rows.append(row)
    for row_idx, row in enumerate(rows):

travishsu

travishsu

[[runners]]
  pre_build_script = "mkdir -p $HOME/.docker/ && echo \"{ \\\"proxies\\\": { \\\"default\\\": { \\\"httpProxy\\\": \\\"$HTTP_PROXY\\\", \\\"httpsProxy\\\": \\\"$HTTPS_PROXY\\\", \\\"noProxy\\\": \\\"$NO_PROXY\\\" } } }\" > $HOME/.docker/config.json"
  
  pre_clone_script = "git config --global http.proxy $HTTP_PROXY; git config --global https.proxy $HTTPS_PROXY"
environment = ["https_proxy=http://docker0_interface_ip:3128", "http_proxy=http://docker0_interface_ip:3128", "HTTPS_PROXY=docker0_interface_ip:3128", "HTTP_PROXY=docker0_interface_ip:3128"]

travishsu

import os
import json
import subprocess
import numpy as np
import pandas as pd
from skimage.measure import find_contours


class CocoDatasetHandler:
    def __init__(self, jsonpath, imgpath):

travishsu

class YourDataset(torch.utils.data.Dataset):
    def __init__(self, ...):
        ...

   def __len__(self):
       ...
       return num_all_examples

   def __getitem__(self, idx):
       ...

travishsu

x = tf.placeholder(.....)

# Scope: 運算區塊
with tf.name_scope("operation_scope"):
    weight = ...
    l1 = op(x, w)
    ...
    ...
    tf.summary.histogram("weight", weight) # 可以看 weight 值的分佈，只是我還不清楚分佈是指什麼