File size: 676 Bytes
9880769 3a0714f 9880769 50d6cdc 9880769 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 |
import transformers
import os
import torch
MAX_LEN = 150 #256
TRAIN_BATCH_SIZE = 8
VALID_BATCH_SIZE = 4
EPOCHS = 5
# Folder to contain all the datasets
DATASET_LOCATION = "" #
MODEL_PATH = "thak123/bert-emoji-latvian-twitter-classifier"
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
)
####################################################################################################################################
|