This is a merge of pre-trained language models created using
mergekit.
This model was merged using the
TIES merge method using
unsloth/Meta-Llama-3.1-8B as a base.
1base_model: unsloth/Meta-Llama-3.1-8B
2dtype: bfloat16
3merge_method: ties
4parameters:
5 density: 1.0
6 weight: 1.0
7slices:
8- sources:
9 - layer_range: [0, 32]
10 model: T145/KRONOS-8B-V1-P1
11 parameters:
12 density: 1.0
13 weight: 1.0
14 - layer_range: [0, 32]
15 model: T145/KRONOS-8B-V1-P2
16 parameters:
17 density: 1.0
18 weight: 1.0
19 - layer_range: [0, 32]
20 model: T145/KRONOS-8B-V1-P3
21 parameters:
22 density: 1.0
23 weight: 1.0
24 - layer_range: [0, 32]
25 model: unsloth/Meta-Llama-3.1-8B-Instruct
26 parameters:
27 density: 1.0
28 weight: 1.0
29 - layer_range: [0, 32]
30 model: unsloth/Meta-Llama-3.1-8B
31tokenizer_source: unsloth/Meta-Llama-3.1-8B-Instruct
Detailed results can be found
here!
Summarized results can be found
here!