File size: 2,190 Bytes
bfe59bb |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 |
---
base_model:
- meta-llama/Meta-Llama-3-70B-Instruct
license: llama3
language:
- en
pipeline_tag: text-generation
tags:
- merge
- frankenmerge
- 96b
---
# BigWeave v33 105b
<img src="https://cdn-uploads.huggingface.co/production/uploads/65a6db055c58475cf9e6def1/4CbbAN-X7ZWj702JrcCGH.png" width=600>
The BigWeave models aim to experimentally identify merge settings for increasing model performance. The version number merely tracks various attempts and is not a quality indicator. Only results demonstrating good performance are retained and shared.
# Prompting Format
llamav3
# Merge process
This is a self-merge of meta-llama/Meta-Llama-3-70B-Instruct. Middle layers are duplicated and various matrices are scaled according to the template by jukofyork as shown here: https://github.com/arcee-ai/mergekit/issues/198#issuecomment-2079950009
Merge configuration:
```
const_tag: &MODEL meta-llama/Meta-Llama-3-70B-Instruct
const_tag: &RESIDUAL_SCALE_FACTOR 0.5
const_tag: &QK_ATTENUATION_FACTOR 0.7071067812
const_tag: &OUT_FACTOR 0.9
scale-filter-env: &scale_filter_env
parameters:
scale:
- filter: o_proj
value: *RESIDUAL_SCALE_FACTOR
- filter: down_proj
value: *RESIDUAL_SCALE_FACTOR
- filter: q_proj
value: *QK_ATTENUATION_FACTOR
- filter: k_proj
value: *QK_ATTENUATION_FACTOR
- filter: v_proj
value: *OUT_FACTOR
- filter: up_proj
value: *OUT_FACTOR
- value: 1.0
slices:
- sources:
- model: *MODEL
layer_range: [0, 19]
- sources:
- model: *MODEL
layer_range: [19, 20]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [10, 29]
- sources:
- model: *MODEL
layer_range: [29, 30]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [20, 39]
- sources:
- model: *MODEL
layer_range: [39, 40]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [30, 49]
- sources:
- model: *MODEL
layer_range: [49, 50]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [40, 80]
merge_method: passthrough
dtype: float16
``` |