reward-model

This model is a fine-tuned version of Qwen/Qwen2-0.5B on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5217
• Accuracy: 0.727

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• total_train_batch_size: 64
• total_eval_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 1.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.636	0.0516	50	0.6010	0.688
0.5793	0.1032	100	0.5676	0.703
0.5807	0.1548	150	0.5732	0.705
0.5572	0.2064	200	0.5513	0.706
0.5695	0.2580	250	0.5472	0.718
0.5596	0.3096	300	0.5283	0.723
0.54	0.3612	350	0.5445	0.715
0.5291	0.4128	400	0.5387	0.722
0.539	0.4644	450	0.5461	0.726
0.5248	0.5160	500	0.5402	0.724
0.5263	0.5676	550	0.5271	0.726
0.5222	0.6192	600	0.5238	0.724
0.5259	0.6708	650	0.5200	0.728
0.5118	0.7224	700	0.5190	0.728
0.513	0.7740	750	0.5213	0.731
0.5141	0.8256	800	0.5253	0.729
0.5197	0.8772	850	0.5256	0.724
0.4968	0.9288	900	0.5231	0.726
0.4983	0.9804	950	0.5217	0.727

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1