mirror of
https://github.com/hiyouga/LLaMA-Factory.git
synced 2026-07-28 11:46:09 +08:00
b61140db3e
Co-authored-by: frozenleaves <frozenleaves@users.noreply.github.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com> Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
28 lines
496 B
YAML
28 lines
496 B
YAML
model: Qwen/Qwen3-0.6B
|
|
model_class: llm
|
|
|
|
|
|
kernel_config:
|
|
name: auto
|
|
include_kernels: auto # choice: null/true/false/auto/kernel_id1,kernel_id2,kernel_id3, default is null
|
|
|
|
quant_config: null
|
|
|
|
dist_config:
|
|
name: fsdp2
|
|
dcp_path: null # /mnt/f/pretrain_models/Qwen3-0.6B-dcp
|
|
|
|
### data
|
|
train_dataset: data/v1_sft_demo.yaml
|
|
|
|
### training
|
|
output_dir: outputs/test_fsdp2
|
|
micro_batch_size: 1
|
|
cutoff_len: 2048
|
|
learning_rate: 1.0e-4
|
|
max_steps: 10
|
|
|
|
### sample
|
|
sample_backend: hf
|
|
max_new_tokens: 128
|