fresh-2-layer-piqa10000-distill-of-fresh-2-layer-gpqa_EVAL_gpqa

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 12.9939
• Accuracy: 0.5101

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: 0.0005
• train_batch_size: 32
• eval_batch_size: 32
• seed: 321
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 5000

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.32	100	19.1271	0.2879
No log	0.64	200	13.9385	0.3687
No log	0.96	300	13.3742	0.4040
No log	1.28	400	13.6512	0.4293
2.752	1.6	500	13.0455	0.3838
2.752	1.92	600	15.1798	0.4293
2.752	2.24	700	13.8055	0.4545
2.752	2.56	800	13.9917	0.4394
2.752	2.88	900	13.5923	0.4899
0.9799	3.19	1000	13.3283	0.4545
0.9799	3.51	1100	12.5037	0.4798
0.9799	3.83	1200	13.9429	0.4545
0.9799	4.15	1300	12.9939	0.5101
0.9799	4.47	1400	13.7332	0.4899
0.5295	4.79	1500	13.9869	0.4596
0.5295	5.11	1600	12.5349	0.4747
0.5295	5.43	1700	12.5107	0.5
0.5295	5.75	1800	13.0188	0.5

Framework versions

• Transformers 4.34.0.dev0
• Pytorch 2.0.1+cu117
• Datasets 2.14.5
• Tokenizers 0.14.0