cellate2.0_grouped_txt-tapt_base-LR_1e-05

This model is a fine-tuned version of microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.2581
• Accuracy: 0.7389

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 3407
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-06 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.3595	1.0	6	1.2848	0.7278
1.3505	2.0	12	1.2635	0.7375
1.3674	3.0	18	1.3147	0.7406
1.2867	4.0	24	1.2947	0.7277
1.2594	5.0	30	1.3362	0.7246
1.2579	6.0	36	1.3625	0.7243
1.306	7.0	42	1.2952	0.7280
1.246	8.0	48	1.2754	0.7266
1.2535	9.0	54	1.2704	0.7289
1.2113	10.0	60	1.3327	0.7147
1.2085	11.0	66	1.2989	0.7317
1.2484	12.0	72	1.2919	0.7283
1.2019	13.0	78	1.4255	0.7158
1.1917	14.0	84	1.2635	0.7224
1.2182	15.0	90	1.3473	0.7179
1.1995	16.0	96	1.3205	0.7196
1.3415	16.7273	100	1.2581	0.7389

Framework versions

• Transformers 4.48.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.2
• Tokenizers 0.21.0