lmind_nq_train1000_eval500_v1_recite_qa_gpt2-xl

This model is a fine-tuned version of gpt2-xl on the tyzhu/lmind_nq_train1000_eval500_v1_recite_qa dataset. It achieves the following results on the evaluation set:

• Loss: 0.3712
• Accuracy: 0.8446

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: 3e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant
• num_epochs: 20.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.1808	1.0	154	1.8207	0.6294
1.4349	2.0	308	1.2914	0.6957
0.8529	3.0	462	0.8566	0.7584
0.4829	4.0	616	0.5789	0.8049
0.2798	5.0	770	0.4421	0.8269
0.1904	6.0	924	0.3886	0.8373
0.1351	7.0	1078	0.3645	0.8411
0.1053	8.0	1232	0.3556	0.8431
0.0889	9.0	1386	0.3492	0.8432
0.0801	10.0	1540	0.3630	0.8441
0.0768	11.0	1694	0.3605	0.8445
0.0729	12.0	1848	0.3607	0.8444
0.0706	13.0	2002	0.3570	0.8450
0.07	14.0	2156	0.3634	0.8447
0.0697	15.0	2310	0.3627	0.8447
0.0681	16.0	2464	0.3645	0.8446
0.0685	17.0	2618	0.3652	0.8443
0.0683	18.0	2772	0.3646	0.8442
0.0665	19.0	2926	0.3732	0.8444
0.0659	20.0	3080	0.3712	0.8446

Framework versions

• Transformers 4.34.0
• Pytorch 2.1.0+cu121
• Datasets 2.14.5
• Tokenizers 0.14.1