vanilla_dpo_iter_4

This model is a fine-tuned version of YYYYYYibo/vanilla_dpo_iter_3 on the updated and the original datasets. It achieves the following results on the evaluation set:

• Loss: 0.5941
• Rewards/chosen: -0.4399
• Rewards/rejected: -0.7506
• Rewards/accuracies: 0.7020
• Rewards/margins: 0.3107
• Logps/rejected: -356.7067
• Logps/chosen: -337.8192
• Logits/rejected: -2.0692
• Logits/chosen: -2.2237

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
0.5962	0.64	100	0.5941	-0.4399	-0.7506	0.7020	0.3107	-356.7067	-337.8192	-2.0692	-2.2237

Framework versions

• PEFT 0.7.1
• Transformers 4.36.2
• Pytorch 2.3.0+cu121
• Datasets 2.14.6
• Tokenizers 0.15.2