phanviethoang1512/llama3.2-1b-deita-dpo-ref_teacher

ref_teacher

This model is a fine-tuned version of meta-llama/Llama-3.1-8B on the HuggingFaceH4/deita-10k-v0-sft dataset. It achieves the following results on the evaluation set:

• Loss: 1.0297

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: 2e-05
• train_batch_size: 4
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• total_eval_batch_size: 4
• 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: 3

Training results

Training Loss	Epoch	Step	Validation Loss
0.8886	1.0	571	0.9567
0.5882	2.0	1142	0.9569
0.3662	3.0	1713	1.0297

Framework versions

• Transformers 4.52.4
• Pytorch 2.10.0+cu126
• Datasets 4.8.4
• Tokenizers 0.21.4