SOMD-scibert-stage2-v1

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

• Loss: 0.0115

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: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	202	0.4607
No log	1.99	404	0.2408
0.5891	2.99	606	0.1086
0.5891	3.98	808	0.0879
0.1352	4.98	1010	0.0505
0.1352	5.97	1212	0.0286
0.1352	6.97	1414	0.0262
0.0541	7.96	1616	0.0231
0.0541	8.96	1818	0.0224
0.0395	9.95	2020	0.0217
0.0395	10.95	2222	0.0191
0.0395	11.94	2424	0.0156
0.0303	12.94	2626	0.0164
0.0303	13.93	2828	0.0146
0.0226	14.93	3030	0.0124
0.0226	15.92	3232	0.0127
0.0226	16.92	3434	0.0115
0.0176	17.91	3636	0.0123
0.0176	18.91	3838	0.0115
0.0146	19.9	4040	0.0115

Framework versions

• Transformers 4.38.1
• Pytorch 2.1.0+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2