library_name, base_model, tags, datasets, model-index

library_name	base_model	tags	datasets	model-index
transformers	W-61/llama-3-8b-base-sft-hh-helpful-4xh200	alignment-handbook new-dpo generated_from_trainer	Anthropic/hh-rlhf	name results llama-3-8b-base-new-dpo-hh-helpful-4xh200-batch-64-q_t-0.5-s_star-0.85

llama-3-8b-base-new-dpo-hh-helpful-4xh200-batch-64-q_t-0.5-s_star-0.85

This model is a fine-tuned version of W-61/llama-3-8b-base-sft-hh-helpful-4xh200 on the Anthropic/hh-rlhf dataset. It achieves the following results on the evaluation set:

• Loss: 0.5312
• Fcm Dpo/beta: 0.0040
• Margin Dpo/margin Mean: 149.2117
• Margin Dpo/margin Std: 231.6038
• Logps/chosen: -579.7042
• Logps/rejected: -736.6628
• Logps/ref Chosen: -79.0510
• Logps/ref Rejected: -86.7979
• Logits/chosen: 0.3259
• Logits/rejected: 0.3477

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: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• total_eval_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Fcm Dpo/beta	Margin Dpo/margin Mean	Margin Dpo/margin Std	Logps/chosen	Logps/rejected	Logps/ref Chosen	Logps/ref Rejected	Logits/chosen	Logits/rejected
0.8901	0.2937	200	0.5793	0.0146	37.0480	73.5224	-146.9453	-191.7402	-79.0510	-86.7979	-0.4808	-0.4605
0.7153	0.5874	400	0.5444	0.0046	115.8499	187.1605	-440.1346	-563.7314	-79.0510	-86.7979	0.1299	0.1525
0.8902	0.8811	600	0.5312	0.0040	149.2117	231.6038	-579.7042	-736.6628	-79.0510	-86.7979	0.3259	0.3477

Framework versions

• Transformers 4.51.0
• Pytorch 2.3.1+cu121
• Datasets 2.21.0
• Tokenizers 0.21.4