ben0it8

sh -c "$(curl -fsSL https://raw.githubusercontent.com/robbyrussell/oh-my-zsh/master/tools/install.sh)"
echo "ZSH=$HOME/.oh-my-zsh" >> ~/.zshrc
echo "source $ZSH/oh-my-zsh.sh" >> ~/.zshrc
echo "ZSH_THEME='robbyrussel'" >> ~/.zshrc
echo "plugins = (git python osx web-search vi-mode dotenv" >> ~/.zshrc
echo "alias config='/usr/bin/git --git-dir=$HOME/.cfg/ --work-tree=$HOME'" >> ~/.zshrc
source ~/.zshrc
echo ".cfg" >> .gitignore
git clone --bare https://github.com/ben0it8/dotfiles.git .cfg/
config checkout

ben0it8

import os
import requests
import tarfile
from tqdm import tqdm

# path to data 
DATA_DIR = os.path.abspath('./data')

# path to IMDB
IMDB_DIR = os.path.join(DATA_DIR, "imdb5k")

ben0it8

import pandas as pd
import re

# text and label column names
TEXT_COL = "text"
LABEL_COL = "label"

def clean_html(text: str):
    "remove html tags and whitespaces"
    cleanr = re.compile('<.*?>')

ben0it8

import torch
from torch.utils.data import TensorDataset, random_split, DataLoader
import numpy as np
import warnings
from tqdm import tqdm_notebook as tqdm
from typing import Tuple

NUM_MAX_POSITIONS = 256
BATCH_SIZE = 32

ben0it8

from concurrent.futures import ProcessPoolExecutor
from multiprocessing import cpu_count
from itertools import repeat

num_cores = cpu_count()

def process_row(processor, row):
    return processor.process_example((row[1][LABEL_COL], row[1][TEXT_COL]))

def create_dataloader(df: pd.DataFrame,

ben0it8

from collections import namedtuple
import torch

LOG_DIR = "./logs/"
CACHE_DIR = "./cache/"

device = "cuda" if torch.cuda.is_available() else "cpu"

FineTuningConfig = namedtuple('FineTuningConfig',
      field_names="num_classes, dropout, init_range, batch_size, lr, max_norm,"

ben0it8

import torch.nn as nn
import torch

class Transformer(nn.Module):
    "Adopted from https://github.com/huggingface/naacl_transfer_learning_tutorial"
  
    def __init__(self, embed_dim, hidden_dim, num_embeddings, num_max_positions, 
                 num_heads, num_layers, dropout, causal):
        super().__init__()
        self.causal = causal

ben0it8

from pytorch_transformers import cached_path

# download pre-trained model and config
state_dict = torch.load(cached_path("https://s3.amazonaws.com/models.huggingface.co/"
                                    "naacl-2019-tutorial/model_checkpoint.pth"), map_location='cpu')

config = torch.load(cached_path("https://s3.amazonaws.com/models.huggingface.co/"
                                        "naacl-2019-tutorial/model_training_args.bin"))

# init model: Transformer base + classifier head

ben0it8

from ignite.engine import Engine, Events
from ignite.metrics import RunningAverage, Accuracy 
from ignite.handlers import ModelCheckpoint
from ignite.contrib.handlers import CosineAnnealingScheduler, PiecewiseLinear, create_lr_scheduler_with_warmup, ProgressBar
import torch.nn.functional as F
from pytorch_transformers.optimization import AdamW

# Bert optimizer
optimizer = AdamW(model.parameters(), lr=finetuning_config.lr, correct_bias=False) 

ben0it8

# fit the model on `train_dl`"
trainer.run(train_dl, max_epochs=finetuning_config.n_epochs)

# save model weights
torch.save(model.state_dict(), os.path.join(finetuning_config.log_dir, "model_weights.pth"))

# evaluate the model on `test_dl`"
evaluator.run(test_dl)
print(f"Test accuracy: {100*evaluator.state.metrics['accuracy']:.3f}")