2c6-reward

This model is a fine-tuned version of gpt2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3850
• Accuracy: 0.8255

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: 1.41e-05
• train_batch_size: 256
• eval_batch_size: 256
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant_with_warmup
• lr_scheduler_warmup_steps: 100
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6638	0.1802	100	0.6210	0.6658
0.4826	0.3604	200	0.4749	0.7711
0.4566	0.5405	300	0.4489	0.7847
0.4471	0.7207	400	0.4414	0.7875
0.4399	0.9009	500	0.4287	0.7950
0.4295	1.0811	600	0.4238	0.7989
0.4099	1.2613	700	0.4168	0.8010
0.4154	1.4414	800	0.4109	0.8069
0.4233	1.6216	900	0.4071	0.8106
0.4013	1.8018	1000	0.4013	0.8144
0.4003	1.9820	1100	0.3996	0.8115
0.3778	2.1622	1200	0.4000	0.8140
0.3764	2.3423	1300	0.3942	0.8171
0.3767	2.5225	1400	0.3943	0.8195
0.3649	2.7027	1500	0.3902	0.8190
0.3612	2.8829	1600	0.3850	0.8255

Framework versions

• Transformers 4.41.2
• Pytorch 2.1.2
• Datasets 2.19.2
• Tokenizers 0.19.1