tengxiao1
TX
ae60326
metadata
license: llama3
base_model: meta-llama/Meta-Llama-3-8B-Instruct
tags:
  - alignment_handbook-handbook
  - generated_from_trainer
datasets:
  - princeton-nlp/llama3-ultrafeedback
model-index:
  - name: LLama-8B-Instruct-v0.1-MI-2e-5
    results: []

Visualize in Weights & Biases

LLama-8B-Instruct-v0.1-MI-2e-5

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the princeton-nlp/llama3-ultrafeedback dataset. It achieves the following results on the evaluation set:

  • Loss: 1.2552
  • Rewards/chosen: -0.5792
  • Rewards/rejected: -0.7270
  • Rewards/accuracies: 0.6626
  • Rewards/margins: 0.1478
  • Logps/rejected: -0.7270
  • Logps/chosen: -0.5792
  • Logits/rejected: -0.3939
  • Logits/chosen: -0.3830

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: 2
  • eval_batch_size: 4
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 4
  • gradient_accumulation_steps: 16
  • total_train_batch_size: 128
  • total_eval_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: 1

Training results

Training Loss Epoch Step Validation Loss Rewards/chosen Rewards/rejected Rewards/accuracies Rewards/margins Logps/rejected Logps/chosen Logits/rejected Logits/chosen
1.2571 0.8550 400 1.2552 -0.5792 -0.7270 0.6626 0.1478 -0.7270 -0.5792 -0.3939 -0.3830

Framework versions

  • Transformers 4.42.0
  • Pytorch 2.1.2+cu121
  • Datasets 2.14.6
  • Tokenizers 0.19.1