You can deploy this model and start earning money today!

Mardiyyah/
CeLLaTe-tapt-pubmedbert-tokenizer-adapted_v2_lr-1e5

Name: CeLLaTe-tapt-pubmedbert-tokenizer-adapted_v2_lr-1e5
Availability: InStock

Views

No views yet

CeLLaTe-tapt-pubmedbert-tokenizer-adapted_v2_lr-1e5

This model is a fine-tuned version of microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext on the Mardiyyah/TAPT_CeLLaTe2.0 dataset. It achieves the following results on the evaluation set:

More information needed

More information needed

More information needed

The following hyperparameters were used during training:

learning_rate: 1e-05
train_batch_size: 32
eval_batch_size: 32
seed: 3407
gradient_accumulation_steps: 2
total_train_batch_size: 64
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: 0.06
num_epochs: 100
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Accuracy	Perplexity
1.3604	1.0	14	1.2099	0.7589	3.3532
1.379	2.0	28	1.2093	0.7611	3.3510
1.3573	3.0	42	1.2053	0.7575	3.3377
1.3451	4.0	56	1.2265	0.7551	3.4092
1.3249	5.0	70	1.2203	0.7572	3.3881
1.3186	6.0	84	1.2113	0.7549	3.3579
1.3065	7.0	98	1.1578	0.7646	3.1829
1.3073	8.0	112	1.2125	0.7521	3.3620
1.2854	9.0	126	1.1526	0.7605	3.1665
1.2812	10.0	140	1.1805	0.7573	3.2559
1.2996	11.0	154	1.1601	0.7548	3.1903
1.2608	12.0	168	1.1775	0.7599	3.2463
1.2642	13.0	182	1.2410	0.7528	3.4592
1.2615	14.0	196	1.1971	0.7567	3.3107