thak123's picture
Update config.py
c3e7d26
raw
history blame
948 Bytes
import transformers
import os
import torch
import requests
MAX_LEN = 150 #256
TRAIN_BATCH_SIZE = 8
VALID_BATCH_SIZE = 4
EPOCHS = 5
# Folder to contain all the datasets
from huggingface_hub import hf_hub_download
DATASET_LOCATION = "" #
print("hi")
MODEL_PATH = hf_hub_download(repo_id="thak123/bert-emoji-latvian-twitter-classifier", filename="model.bin")
# from huggingface_hub import snapshot_download
# snapshot_download(repo_id="thak123/bert-emoji-latvian-twitter-classifier", allow_patterns="*.bin")
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
# 7 EPOCH Version
BERT_PATH = "FFZG-cleopatra/bert-emoji-latvian-twitter"
# TODO check if lower casing is required
# BertTokenizer
TOKENIZER = transformers.BertTokenizer.from_pretrained(
BERT_PATH,
do_lower_case=True
)
####################################################################################################################################