| slices: | |
| - sources: | |
| - model: mlabonne/NeuralHermes-2.5-Mistral-7B | |
| layer_range: [0, 32] | |
| - model: teknium/OpenHermes-2.5-Mistral-7B | |
| layer_range: [0, 32] | |
| merge_method: slerp | |
| base_model: mlabonne/NeuralHermes-2.5-Mistral-7B | |
| parameters: | |
| t: | |
| - filter: self_attn | |
| value: [0, 0.3, 0.6, 0.9, 1] # Progressive fusion des couches d'attention | |
| - filter: mlp | |
| value: [1, 0.7, 0.4, 0.1, 0] # Transition inverse pour les MLP | |
| - value: 0.45 # Ratio de fusion global | |
| dtype: bfloat16 | |