File size: 1,940 Bytes
711d2d7 941fd3f 711d2d7 4ad0372 711d2d7 941fd3f a4d40e8 941fd3f a4d40e8 941fd3f 4ad0372 941fd3f |
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 |
---
base_model:
- grimjim/Mistral-7B-Instruct-demi-merge-v0.2-7B
- grimjim/kukulemon-7B
library_name: transformers
tags:
- mergekit
- merge
license: cc-by-nc-4.0
pipeline_tag: text-generation
---
# kukulemon-32K-7B
This is a proof of concept a merge capable of functional 32K context length while being derived from [kukulemon-7B](https://huggingface.co/grimjim/kukulemon-7B).
The functioning 32K context window has been folded in via a merger of Mistral 7B v0.2 models.
SLERP merge appears to be viable, but DARE-TIES merge risks producing a damaged model and is therefore not recommended.
Although the resulting model natively supports Alpaca prompt, I've tested with ChatML prompts successfuly. Medium temperature (around 1) with low minP (e.g., 0.01) works with ChatML prompts in my most recent testing.
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
- GGUF quants: [grimjim/kukulemon-32K-7B-GGUF](https://huggingface.co/grimjim/kukulemon-32K-7B-GGUF)
## Merge Details
### Merge Method
This model was merged using the SLERP merge method.
### Models Merged
The following models were included in the merge:
* [grimjim/Mistral-7B-Instruct-demi-merge-v0.2-7B](https://huggingface.co/grimjim/Mistral-7B-Instruct-demi-merge-v0.2-7B)
* [grimjim/kukulemon-7B](https://huggingface.co/grimjim/kukulemon-7B)
### Configuration
The following YAML configuration was used to produce this model:
```yaml
slices:
- sources:
- model: grimjim/kukulemon-7B
layer_range: [0, 32]
- model: grimjim/Mistral-7B-Instruct-demi-merge-v0.2-7B
layer_range: [0, 32]
# or, the equivalent models: syntax:
# models:
merge_method: slerp
base_model: grimjim/kukulemon-7B
parameters:
t:
- filter: self_attn
value: [0, 0.5, 0.3, 0.7, 1]
- filter: mlp
value: [1, 0.5, 0.7, 0.3, 0]
- value: 0.5 # fallback for rest of tensors
dtype: bfloat16
```
|