phi-2-gpo-test-longest-iter-random1-1

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

phi-2-gpo-test-longest-iter-random1-1

This model is a fine-tuned version of DUAL-GPO/phi-2-gpo-test-longest-iter-random1-0 on the HuggingFaceH4/ultrafeedback_binarized dataset. It achieves the following results on the evaluation set:

• Loss: 0.0012
• Rewards/chosen: -0.0006
• Rewards/rejected: -0.0003
• Rewards/accuracies: 0.4780
• Rewards/margins: -0.0003
• Logps/rejected: -278.6935
• Logps/chosen: -306.4963
• Logits/rejected: 0.0865
• Logits/chosen: -0.0098

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: 4
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 4

Training results

Framework versions

• PEFT 0.7.1
• Transformers 4.36.2
• Pytorch 2.2.1+cu121
• Datasets 2.14.6
• Tokenizers 0.15.2