roberta_year_as_text_better_finetunning

This commit is contained in:
kubapok 2021-10-07 08:54:19 +02:00
parent fd0cae92e5
commit 1e7965e478
4 changed files with 28174 additions and 28174 deletions

File diff suppressed because it is too large Load Diff

View File

@ -21,7 +21,7 @@ model = AutoModelForSequenceClassification.from_pretrained('roberta-ireland').cu
from transformers import TrainingArguments from transformers import TrainingArguments
training_args = TrainingArguments("test_trainer", training_args = TrainingArguments("roberta-ireland",
per_device_train_batch_size=4, per_device_train_batch_size=4,
per_device_eval_batch_size=4, per_device_eval_batch_size=4,
evaluation_strategy='steps', evaluation_strategy='steps',

View File

@ -1 +1 @@
python run_glue.py --model_name_or_path roberta-base --train_file /media/kuba/ssdsam/gonito/ireland-news/train/huggingface_format_year.csv --validation_file /media/kuba/ssdsam/gonito/ireland-news/dev-0/huggingface_format_year.csv --do_train --do_eval --max_seq_length 64 --per_device_train_batch_size 32 --learning_rate 2e-5 --num_train_epochs 3 --save_steps=5000 --eval_steps=5000 --evaluation_strategy steps --output_dir ./roberta-ireland python run_glue.py --model_name_or_path roberta-base --train_file /media/kuba/ssdsam/gonito/ireland-news/train/huggingface_format_year_as_text.csv --validation_file /media/kuba/ssdsam/gonito/ireland-news/dev-0/huggingface_format_year_as_text.csv --do_train --do_eval --max_seq_length 64 --per_device_train_batch_size 32 --learning_rate 2e-5 --num_train_epochs 3 --save_steps=5000 --eval_steps=5000 --evaluation_strategy steps --output_dir ./roberta-ireland

File diff suppressed because it is too large Load Diff