lmind_nq_train6000_eval6489_v1_recite_qa_gpt2-xl

This model is a fine-tuned version of gpt2-xl on the tyzhu/lmind_nq_train6000_eval6489_v1_recite_qa dataset. It achieves the following results on the evaluation set:

• Loss: 0.3634
• Accuracy: 0.8783

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
2.1506	1.0	1058	1.7260	0.7131
1.5141	2.0	2116	1.2579	0.7600
0.9961	3.0	3174	0.8674	0.8056
0.6354	4.0	4232	0.6007	0.8397
0.4213	5.0	5290	0.4423	0.8612
0.283	6.0	6348	0.3741	0.8703
0.2072	7.0	7406	0.3511	0.8742
0.1641	8.0	8464	0.3441	0.8764
0.1365	9.0	9522	0.3439	0.8769
0.1225	10.0	10580	0.3467	0.8774
0.1129	11.0	11638	0.3479	0.8776
0.1074	12.0	12696	0.3505	0.8778
0.1026	13.0	13754	0.3498	0.8774
0.1	14.0	14812	0.3514	0.8780
0.0953	15.0	15870	0.3595	0.8782
0.0944	16.0	16928	0.3604	0.8781
0.0911	17.0	17986	0.3604	0.8781
0.0905	18.0	19044	0.3617	0.8781
0.0879	19.0	20102	0.3662	0.8784
0.0866	20.0	21160	0.3634	0.8783

Framework versions

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