SciBERT_AsymmetricLoss_25K_bs64_P4_N1

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

• Loss: 30.2502
• Accuracy: 0.9871
• Precision: 0.4247
• Recall: 0.8998
• F1: 0.5770
• Hamming: 0.0129

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: 2e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 25000

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Hamming
36.6287	0.16	5000	34.9978	0.9852	0.3863	0.8728	0.5355	0.0148
33.8929	0.32	10000	32.4942	0.9857	0.3958	0.8901	0.5480	0.0143
32.5419	0.47	15000	31.3170	0.9867	0.4162	0.8941	0.5680	0.0133
31.565	0.63	20000	30.6092	0.9869	0.4201	0.8975	0.5723	0.0131
31.105	0.79	25000	30.2502	0.9871	0.4247	0.8998	0.5770	0.0129

Framework versions

• Transformers 4.35.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.7.1
• Tokenizers 0.14.1