• Model creator: https://huggingface.co/tanliboy/
• Original model: https://huggingface.co/tanliboy/lambda-llama-3-8b-ipo-test/

Original model description:

• alignment-handbook
• trl
• dpo
• generated_from_trainer
• trl
• dpo
• generated_from_trainer datasets:
• HuggingFaceH4/ultrafeedback_binarized
• tanliboy/orca_dpo_pairs model-index:
• name: lambda-llama-3-8b-ipo-test results: []

lambda-llama-3-8b-ipo-test

• Loss: 0.8931
• Rewards/chosen: -0.3610
• Rewards/rejected: -0.5883
• Rewards/accuracies: 0.7922
• Rewards/margins: 0.2272
• Logps/rejected: -3.1373
• Logps/chosen: -2.5334
• Logits/rejected: -2.9939
• Logits/chosen: -2.9244

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

• learning_rate: 5e-07
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• 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

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1