File size: 564 Bytes
a57dcb6 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 |
base_model: NousResearch/Meta-Llama-3-8B
dtype: bfloat16
merge_method: dare_ties
parameters:
int8_mask: 1.0
slices:
- sources:
- layer_range: [0, 32]
model: Weyaxi/Einstein-v6.1-Llama3-8B
parameters:
density: 0.1
weight: 1.0
- layer_range: [0, 32]
model: asiansoul/Versatile-Llama-3-8B-1m
parameters:
density: 0.2
weight: 0.35
- layer_range: [0, 32]
model: NousResearch/Hermes-2-Pro-Llama-3-8B
parameters:
density: 0.5
weight: 0.23
- layer_range: [0, 32]
model: NousResearch/Meta-Llama-3-8B |