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.

558 B
Raw Blame
{
  "zero_optimization": {
    "stage": 3,
    "offload_optimizer": {"device": "cpu", "pin_memory": true},
    "offload_param": {"device": "cpu", "pin_memory": true},
    "overlap_comm": true,
    "contiguous_gradients": true,
    "stage3_max_live_parameters": 1000000000,
    "stage3_max_reuse_distance": 1000000000,
    "stage3_gather_16bit_weights_on_model_save": true
  },
  "fp16": {"enabled": "auto"},
  "bf16": {"enabled": "auto"},
  "gradient_accumulation_steps": "auto",
  "train_micro_batch_size_per_gpu": "auto",
  "wall_clock_breakdown": false
}