Skip to content

armgit / Soup

Public Imported from GitHub
0

Fine-tune LLMs from one YAML. Layer streaming trains an 8B model on a 4 GB laptop GPU.

583 B
Raw Blame
# Soup template: Reasoning / GRPO
# Fine-tune a model for chain-of-thought reasoning with GRPO

base: meta-llama/Llama-3.1-8B-Instruct
task: grpo
# backend: unsloth  # 2-5x faster, pip install "soup-cli[fast]"

data:
  train: ./data/reasoning_train.jsonl
  format: sharegpt
  val_split: 0.1
  max_length: 4096

training:
  epochs: 3
  lr: 1e-5
  batch_size: auto
  gradient_accumulation_steps: 8
  lora:
    r: 64
    alpha: 16
    target_modules: auto
  quantization: 4bit
  gradient_checkpointing: true
  grpo_beta: 0.1
  num_generations: 4
  reward_fn: accuracy

output: ./output