metadata

base_model: allenai/scibert_scivocab_uncased
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - precision
  - recall
  - f1
model-index:
  - name: SciBERT_100K_steps
    results: []

SciBERT_100K_steps

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: 0.0144
Accuracy: 0.9947
Precision: 0.7850
Recall: 0.6355
F1: 0.7024
Hamming: 0.0053

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: 32
eval_batch_size: 32
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: 100000

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Hamming
0.1681	0.08	5000	0.0487	0.9902	0.0	0.0	0.0	0.0098
0.032	0.16	10000	0.0223	0.9930	0.8068	0.3728	0.5100	0.0070
0.0201	0.24	15000	0.0186	0.9937	0.7815	0.4970	0.6076	0.0063
0.018	0.32	20000	0.0172	0.9941	0.7763	0.5550	0.6472	0.0059
0.017	0.4	25000	0.0166	0.9942	0.7864	0.5624	0.6558	0.0058
0.0166	0.47	30000	0.0163	0.9943	0.7707	0.5880	0.6671	0.0057
0.0163	0.55	35000	0.0160	0.9943	0.7802	0.5809	0.6659	0.0057
0.0159	0.63	40000	0.0158	0.9944	0.7719	0.6012	0.6759	0.0056
0.0157	0.71	45000	0.0155	0.9945	0.7750	0.6104	0.6829	0.0055
0.0154	0.79	50000	0.0153	0.9945	0.7734	0.6202	0.6884	0.0055
0.0153	0.87	55000	0.0151	0.9945	0.7823	0.6072	0.6837	0.0055
0.0152	0.95	60000	0.0151	0.9945	0.7813	0.6124	0.6866	0.0055
0.0148	1.03	65000	0.0149	0.9946	0.7843	0.6208	0.6930	0.0054
0.0143	1.11	70000	0.0148	0.9946	0.7802	0.6231	0.6929	0.0054
0.0142	1.19	75000	0.0148	0.9946	0.7714	0.6377	0.6982	0.0054
0.0141	1.27	80000	0.0146	0.9947	0.7837	0.6281	0.6973	0.0053
0.0141	1.34	85000	0.0146	0.9947	0.7836	0.6374	0.7030	0.0053
0.014	1.42	90000	0.0145	0.9947	0.7859	0.6326	0.7010	0.0053
0.0139	1.5	95000	0.0145	0.9947	0.7875	0.6317	0.7010	0.0053
0.0139	1.58	100000	0.0144	0.9947	0.7850	0.6355	0.7024	0.0053

Framework versions

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