| models: | |
| - model: mistralai/Mistral-7B-v0.1 | |
| # No parameters necessary for base model | |
| - model: samir-fama/SamirGPT-v1 | |
| parameters: | |
| density: 0.53 | |
| weight: 0.4 | |
| - model: abacusai/Slerp-CM-mist-dpo | |
| parameters: | |
| density: 0.53 | |
| weight: 0.3 | |
| - model: EmbeddedLLM/Mistral-7B-Merge-14-v0.2 | |
| parameters: | |
| density: 0.53 | |
| weight: 0.3 | |
| merge_method: dare_ties | |
| base_model: mistralai/Mistral-7B-v0.1 | |
| parameters: | |
| int8_mask: true | |
| dtype: bfloat16 | |