| merge_method: della_linear | |
| base_model: Qwen/Qwen2.5-1.5B-Instruct | |
| dtype: bfloat16 | |
| parameters: | |
| epsilon: 0.02 | |
| lambda: 1.65 | |
| normalize: true | |
| adaptive_merge_parameters: | |
| task_weights: | |
| tinyArc: 2.2 | |
| tinyHellaswag: 2.15 | |
| tinyMMLU: 1.9 | |
| tinyTruthfulQA: 2.0 | |
| tinyTruthfulQA_mc1: 1.95 | |
| tinyWinogrande: 1.85 | |
| IFEval: 2.5 # Best stat, for instruction-following | |
| BBH: 2.2 | |
| MATH: 2.3 | |
| GPQA: 2.0 | |
| MUSR: 2.0 | |
| MMLU-PRO: 2.15 | |
| smoothing_factor: 0.15 | |
| models: | |
| - model: Qwen/Qwen2.5-Coder-1.5B-Instruct | |
| parameters: | |
| weight: 1 | |
| density: 1 | |
| - model: Qwen/Qwen2.5-Math-1.5B-Instruct | |
| parameters: | |
| weight: 1 | |
| density: 1 | |
| - model: Qwen/Qwen2.5-1.5B-Instruct | |
| parameters: | |
| weight: 1 | |
| density: 1 | |
| - model: justinj92/Qwen2.5-1.5B-Thinking | |
| parameters: | |
| weight: 1 | |
| density: 1 | |
| - model: prithivMLmods/Bellatrix-Tiny-1.5B-R1 | |
| parameters: | |
| weight: 1 | |
| density: 1 |