| slices: | |
| - sources: | |
| - model: IntervitensInc/Llama-3.1-Minitron-4B-Width-Base-chatml | |
| layer_range: [0, 24] | |
| - sources: # add middle layers with residuals scaled to zero | |
| - model: IntervitensInc/Llama-3.1-Minitron-4B-Width-Base-chatml | |
| layer_range: [8, 24] | |
| parameters: | |
| scale: | |
| - filter: o_proj | |
| value: 0.0 | |
| - filter: down_proj | |
| value: 0.0 | |
| - value: 1.0 | |
| - sources: # add middle layers with residuals scaled to zero | |
| - model: IntervitensInc/Llama-3.1-Minitron-4B-Width-Base-chatml | |
| layer_range: [8, 24] | |
| parameters: | |
| scale: | |
| - filter: o_proj | |
| value: 0.0 | |
| - filter: down_proj | |
| value: 0.0 | |
| - value: 1.0 | |
| - sources: | |
| - model: IntervitensInc/Llama-3.1-Minitron-4B-Width-Base-chatml | |
| layer_range: [24, 32] | |
| merge_method: passthrough | |
| dtype: bfloat16 | |