Views
No views yet
1merge_method: linear
2parameters:
3 weight: 1.0
4slices:
5- sources:
6 - model: nvidia/Llama-3.1-Nemotron-70B-Instruct-HF
7 layer_range: [0, 16]
8- sources:
9 - model: cyberagent/Llama-3.1-70B-Japanese-Instruct-2407
10 layer_range: [8, 24]
11- sources:
12 - model: nvidia/Llama-3.1-Nemotron-70B-Instruct-HF
13 layer_range: [17, 32]
14- sources:
15 - model: cyberagent/Llama-3.1-70B-Japanese-Instruct-2407
16 layer_range: [25, 40]
17- sources:
18 - model: nvidia/Llama-3.1-Nemotron-70B-Instruct-HF
19 layer_range: [33, 48]
20- sources:
21 - model: cyberagent/Llama-3.1-70B-Japanese-Instruct-2407
22 layer_range: [41, 56]
23- sources:
24 - model: nvidia/Llama-3.1-Nemotron-70B-Instruct-HF
25 layer_range: [49, 64]
26- sources:
27 - model: cyberagent/Llama-3.1-70B-Japanese-Instruct-2407
28 layer_range: [57, 72]
29- sources:
30 - model: nvidia/Llama-3.1-Nemotron-70B-Instruct-HF
31 layer_range: [65, 80]
32dtype: float16
33tokenizer_source: nvidia/Llama-3.1-Nemotron-70B-Instruct-HF
34