README.md

metadata

library_name: transformers
license: llama3
base_model: Dynosaur/llama3-8b-math-sft
tags:
  - alignment-handbook
  - trl
  - sft
  - generated_from_trainer
  - trl
  - sft
  - generated_from_trainer
datasets:
  - Dynosaur/math-sft-full-command-r
model-index:
  - name: llama3-8b-math-sft-full-command-r
    results: []

llama3-8b-math-sft-full-command-r

This model is a fine-tuned version of Dynosaur/llama3-8b-math-sft on the Dynosaur/math-sft-full-command-r dataset.

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: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 8
• total_train_batch_size: 128
• total_eval_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 2

Training results

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1