22810c97b7
* use warmup_ratio as a better default than warmup steps since it's data dependent * replace remainder of warmup_steps
81 lines
1.5 KiB
YAML
81 lines
1.5 KiB
YAML
base_model: Qwen/Qwen3-8B
|
|
# Automatically upload checkpoint and final model to HF
|
|
# hub_model_id: username/custom_model_name
|
|
|
|
load_in_8bit: false
|
|
load_in_4bit: false
|
|
strict: false
|
|
|
|
plugins:
|
|
- axolotl.integrations.liger.LigerPlugin
|
|
|
|
liger_rope: true
|
|
liger_rms_norm: true
|
|
liger_glu_activation: true
|
|
liger_layer_norm: true
|
|
liger_fused_linear_cross_entropy: true
|
|
|
|
datasets:
|
|
- path: tatsu-lab/alpaca
|
|
type: alpaca
|
|
|
|
output_dir: ./outputs/qat_out/
|
|
|
|
sequence_len: 2048
|
|
sample_packing: true
|
|
flex_attention: true
|
|
|
|
|
|
flex_attn_compile_kwargs:
|
|
dynamic: false
|
|
mode: max-autotune-no-cudagraphs
|
|
|
|
qat:
|
|
activation_dtype: int8
|
|
weight_dtype: int4
|
|
group_size: 256
|
|
fake_quant_after_n_steps: 1000
|
|
|
|
wandb_project:
|
|
wandb_entity:
|
|
wandb_watch:
|
|
wandb_name:
|
|
wandb_log_model:
|
|
|
|
gradient_accumulation_steps: 1
|
|
micro_batch_size: 2
|
|
max_steps: 2000
|
|
optimizer: adamw_torch_fused
|
|
lr_scheduler: cosine
|
|
learning_rate: 2e-5
|
|
|
|
bf16: true
|
|
tf32: true
|
|
|
|
resume_from_checkpoint:
|
|
logging_steps: 1
|
|
|
|
evals_per_epoch: 1
|
|
saves_per_epoch: 1
|
|
|
|
warmup_ratio: 0.1
|
|
weight_decay: 0.0
|
|
fsdp:
|
|
- full_shard
|
|
- auto_wrap
|
|
|
|
fsdp_config:
|
|
fsdp_version: 2
|
|
fsdp_offload_params: false
|
|
fsdp_cpu_ram_efficient_loading: true
|
|
fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP
|
|
fsdp_transformer_layer_cls_to_wrap: Qwen3DecoderLayer
|
|
fsdp_state_dict_type: FULL_STATE_DICT
|
|
fsdp_sharding_strategy: FULL_SHARD
|
|
fsdp_reshard_after_forward: true
|
|
fsdp_activation_checkpointing: true
|
|
|
|
special_tokens:
|
|
|
|
# save_first_step: true # uncomment this to validate checkpoint saving works with your config
|