CeLLaTe-AL-Testbase_adapted_tok

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:

Model description

More information needed

More information needed

More information needed

The following hyperparameters were used during training:

learning_rate: 2e-05
train_batch_size: 16
eval_batch_size: 16
seed: 3407
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 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 Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
1.0711	1.0	115	0.4380	0.2039	0.1800	0.1912	0.8599
0.2579	2.0	230	0.2830	0.4117	0.3991	0.4053	0.9112
0.1083	3.0	345	0.2841	0.5199	0.4894	0.5042	0.9222
0.0707	4.0	460	0.2899	0.5291	0.5481	0.5384	0.9240
0.0472	5.0	575	0.2954	0.5816	0.5354	0.5576	0.9297
0.0336	6.0	690	0.3135	0.5902	0.5354	0.5615	0.9296
0.0232	7.0	805	0.3596	0.6125	0.5419	0.5750	0.9299
0.0178	8.0	920	0.3412	0.5675	0.5903	0.5787	0.9285
0.0181	9.0	1035	0.3621	0.6100	0.5470	0.5768	0.9296
0.0132	10.0	1150	0.3638	0.5904	0.5733	0.5817	0.9302
0.0106	11.0	1265	0.4134	0.6078	0.5556	0.5805	0.9303
0.0088	12.0	1380	0.4197	0.6023	0.5579	0.5793	0.9297
0.0071	13.0	1495	0.4162	0.5947	0.5712	0.5828	0.9301
0.0067	14.0	1610	0.4196	0.6077	0.5518	0.5784	0.9305
0.005	15.0	1725	0.4391	0.5892	0.5661	0.5774	0.9295
0.0047	16.0	1840	0.4424	0.5872	0.5682	0.5775	0.9295
0.0042	17.0	1955	0.4632	0.5939	0.5539	0.5732	0.9293
0.0037	18.0	2070	0.4541	0.5971	0.5692	0.5828	0.9300
0.0041	19.0	2185	0.4611	0.6069	0.5631	0.5842	0.9307
0.0034	20.0	2300	0.4573	0.5991	0.5668	0.5825	0.9301

Safetensors

Model size

0.1B params

Tensor type

F32

Base model

Finetuned

(120)

this model