mirror of
https://github.com/SakanaAI/doc-to-lora.git
synced 2026-07-23 17:01:04 +02:00
d99fa8732b
- work with vllm self-gen data (everything is tokenized during self-gen) - fix chat template! - not use liger kernel for distillation (for both training and eval)
20 lines
No EOL
485 B
JSON
20 lines
No EOL
485 B
JSON
{
|
|
"bf16": {
|
|
"enabled": "auto"
|
|
},
|
|
"zero_optimization": {
|
|
"stage": 0,
|
|
"stage3_gather_16bit_weights_on_model_save": "no",
|
|
"offload_optimizer": {
|
|
"device": "auto"
|
|
},
|
|
"offload_param": {
|
|
"device": "auto"
|
|
}
|
|
},
|
|
"gradient_clipping": "auto",
|
|
"train_batch_size": "auto",
|
|
"train_micro_batch_size_per_gpu": "auto",
|
|
"gradient_accumulation_steps": "auto",
|
|
"steps_per_print": 2000000
|
|
} |