Terjemahan disediakan oleh mesin penerjemah. Jika konten terjemahan yang diberikan bertentangan dengan versi bahasa Inggris aslinya, utamakan versi bahasa Inggris.

Konfigurasi Umum

defaults:
  - _self_
  - cluster: slurm 
  - recipes: training/llama/hf_llama3_8b_seq8192_gpu
instance_type: p5.48xlarge
git:
  repo_url_or_path: null
  branch: null
  commit: null
  entry_script: null
  token: null
env_vars:
  NCCL_DEBUG: WARN

run

run:
  name: llama-8b
  results_dir: ${base_results_dir}/${.name}
  time_limit: "6-00:00:00"
  model_type: hf
    

exp_manager:
  exp_dir: null
  name: experiment
  create_tensorboard_logger: True
    

exp_manager
  checkpoint_dir: ${recipes.exp_manager.exp_dir}/checkpoints/
  auto_checkpoint:
    enabled: True

exp_manager:
  checkpoint_callback_params:
    # Set save_top_k = 0 to disable sharded checkpointing
    save_top_k: 3
    every_n_train_steps: 10
    monitor: "step"
    mode: "max"
    save_last: False
  resume_from_checkpoint: ${recipes.exp_manager.exp_dir}/checkpoints/

exp_manager:
  export_full_model:
    # Set every_n_train_steps = 0 to disable full checkpointing
    every_n_train_steps: 0
    save_last: True
    final_export_dir : null

model:
  model_type: llama_v3
  # Base configs
  train_batch_size: 4
  val_batch_size: 1
  seed: 12345
  grad_clip: 1.0

  # Model parallelism
  tensor_model_parallel_degree: 4
  expert_model_parallel_degree: 1
  context_parallel_degree: 2

FP8

model:
  # FP8 config
  fp8: True
  fp8_amax_history_len: 1024
  fp8_amax_compute_algo: max

data

model:
  data:
    train_dir: /path/to/your/train/data
    val_dir: /path/to/your/val/data
    dataset_type: hf
    use_synthetic_data: False

model:
  # Fine tuning config
  do_finetune: True
  # The path to resume from, needs to be HF compatible
  hf_model_name_or_path: null
  hf_access_token: null
  # PEFT config
  peft:
    peft_type: lora
    rank: 32
    alpha: 16
    dropout: 0.1