vanilla_doff_dpo_iter_2

This model is a fine-tuned version of YYYYYYibo/vanilla_doff_dpo_iter_1 on the updated and the original datasets. It achieves the following results on the evaluation set:

• Loss: 0.6693
• Rewards/chosen: -0.1884
• Rewards/rejected: -0.2428
• Rewards/accuracies: 0.5520
• Rewards/margins: 0.0543
• Logps/rejected: -294.6735
• Logps/chosen: -310.7413
• Logits/rejected: -2.5815
• Logits/chosen: -2.6792

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-06
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 16
• total_train_batch_size: 128
• total_eval_batch_size: 8
• 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
0.6615	0.65	100	0.6693	-0.1884	-0.2428	0.5520	0.0543	-294.6735	-310.7413	-2.5815	-2.6792

Framework versions

• PEFT 0.7.1
• Transformers 4.36.2
• Pytorch 2.1.2+cu121
• Datasets 2.14.6
• Tokenizers 0.15.2