12 lines
374 B
YAML
12 lines
374 B
YAML
|
|
merge_method: breadcrumbs_ties
|
||
|
|
base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-7B
|
||
|
|
parameters:
|
||
|
|
density: 0.15 # Retain top 15% (Trims bottom 85% SFT noise)
|
||
|
|
gamma: 0.99 # Trim the top 1% (Destroys extreme RL outliers)
|
||
|
|
weight: 1.0
|
||
|
|
models:
|
||
|
|
- model: nvidia/AceReason-Nemotron-7B
|
||
|
|
- model: ./padded-maestro
|
||
|
|
- model: POLARIS-Project/Polaris-7B-Preview
|
||
|
|
dtype: bfloat16
|