llama_DPO_model

This model is a fine-tuned version of meta-llama/Llama-2-7b-hf on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2506
• Rewards/chosen: 0.2764
• Rewards/rejected: -1.0388
• Rewards/accuracies: 1.0
• Rewards/margins: 1.3152
• Logps/rejected: -194.5943
• Logps/chosen: -156.0318
• Logits/rejected: -1.0532
• Logits/chosen: -0.8577

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: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.3358	0.79	200	0.3244	0.2277	-0.7696	1.0	0.9973	-191.9022	-156.5185	-1.0547	-0.8590
0.2428	1.59	400	0.2506	0.2764	-1.0388	1.0	1.3152	-194.5943	-156.0318	-1.0532	-0.8577

Framework versions

• PEFT 0.8.2
• Transformers 4.38.1
• Pytorch 2.2.0+cu118
• Datasets 2.17.1
• Tokenizers 0.15.2