selective-pairrm-32754820

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.7284
• Rewards/chosen: -2.3186
• Rewards/rejected: -2.4353
• Rewards/accuracies: 0.5703
• Rewards/margins: 0.1167
• Logps/rejected: -555.1447
• Logps/chosen: -524.2090
• Logits/rejected: -2.5947
• Logits/chosen: -2.5961

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: 5e-07
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• total_eval_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.5696	0.32	100	0.7351	-2.0420	-2.1261	0.5781	0.0841	-524.2233	-496.5468	-2.5568	-2.5577
0.5384	0.64	200	0.7274	-2.2166	-2.3279	0.5781	0.1113	-544.4059	-514.0054	-2.6227	-2.6242
0.5331	0.96	300	0.7288	-2.3199	-2.4365	0.5703	0.1166	-555.2662	-524.3375	-2.5949	-2.5963

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2
• Datasets 2.14.6
• Tokenizers 0.15.0