BioMedRoBERTa-finetuned-ner-pablo-just-classifier

This model is a fine-tuned version of emilyalsentzer/Bio_ClinicalBERT on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1228
• Precision: 0.6701
• Recall: 0.6809
• F1: 0.6754
• Accuracy: 0.9657

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: 0.01
• train_batch_size: 512
• eval_batch_size: 512
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 2048
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine_with_restarts
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	0.9697	16	0.2938	0.4425	0.5130	0.4751	0.9361
No log	2.0	33	0.1815	0.5546	0.5873	0.5705	0.9535
No log	2.9697	49	0.1617	0.5838	0.6189	0.6008	0.9575
No log	4.0	66	0.1482	0.6070	0.6396	0.6229	0.9602
No log	4.9697	82	0.1340	0.6465	0.6563	0.6513	0.9633
No log	6.0	99	0.1306	0.6561	0.6638	0.6599	0.9641
No log	6.9697	115	0.1290	0.6569	0.6705	0.6636	0.9645
No log	8.0	132	0.1246	0.6664	0.6794	0.6728	0.9654
No log	8.9697	148	0.1230	0.6699	0.6793	0.6745	0.9656
No log	9.6970	160	0.1228	0.6701	0.6809	0.6754	0.9657

Framework versions

• Transformers 4.44.1
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1