BigWeave-v32-96b / README.md
llmixer's picture
Create README.md
4c8cf31 verified
---
base_model:
- meta-llama/Meta-Llama-3-70B-Instruct
license: llama3
language:
- en
pipeline_tag: text-generation
tags:
- merge
- frankenmerge
- 96b
---
# BigWeave v32 96b
<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, 25]
- sources:
- model: *MODEL
layer_range: [25, 26]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [25, 26]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [26, 27]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [26, 27]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [27, 28]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [27, 28]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [28, 29]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [28, 29]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [29, 30]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [29, 30]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [30, 31]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [30, 31]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [31, 32]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [31, 32]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [32, 33]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [32, 33]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [33, 34]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [33, 34]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [34, 35]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [34, 35]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [35, 36]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [35, 36]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [36, 37]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [36, 37]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [37, 38]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [37, 38]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [38, 39]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [38, 39]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [39, 40]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [39, 40]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [40, 41]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [40, 41]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [41, 42]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [41, 42]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [42, 43]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [42, 43]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [43, 44]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [43, 44]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [44, 45]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [44, 45]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [45, 46]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [45, 46]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [46, 47]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [46, 47]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [47, 48]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [47, 48]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [48, 49]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [48, 49]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [49, 50]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [49, 50]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [50, 51]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [50, 51]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [51, 52]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [51, 52]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [52, 53]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [52, 53]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [53, 54]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [53, 54]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [54, 55]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [54, 55]
<<: *scale_filter_env
- sources:
- model: *MODEL
layer_range: [55, 80]
merge_method: passthrough
dtype: float16
```