lmind_hotpot_train8000_eval7405_v1_recite_qa_gpt2-xl

This model is a fine-tuned version of gpt2-xl on the tyzhu/lmind_hotpot_train8000_eval7405_v1_recite_qa dataset. It achieves the following results on the evaluation set:

• Loss: 0.4650
• Accuracy: 0.7664

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	Accuracy	Validation Loss
1.9537	1.0	2179	0.6693	1.5623
1.4734	2.0	4358	0.6924	1.2099
1.0665	3.0	6537	0.7147	0.9178
0.7684	4.0	8716	0.7331	0.6988
0.548	5.0	10895	0.7466	0.5567
0.4039	6.0	13074	0.7551	0.4728
0.3044	7.0	15253	0.7600	0.4376
0.2446	8.0	17432	0.7628	0.4220
0.2039	9.0	19611	0.7642	0.4190
0.1787	10.0	21790	0.7649	0.4250
0.1652	11.0	23969	0.7654	0.4295
0.154	12.0	26148	0.7655	0.4366
0.1441	13.0	28327	0.7657	0.4429
0.143	14.0	30506	0.7657	0.4418
0.1366	15.0	32685	0.7661	0.4469
0.1335	16.0	34864	0.7661	0.4517
0.1299	17.0	37043	0.7662	0.4534
0.1271	18.0	39222	0.7664	0.4579
0.1268	19.0	41401	0.7664	0.4556
0.1238	20.0	43580	0.4650	0.7664

Framework versions

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