|
--- |
|
license: cc-by-nc-4.0 |
|
tags: |
|
- conversational |
|
- mixtral |
|
- merge |
|
- mergekit |
|
--- |
|
|
|
``` |
|
e88 88e d8 |
|
d888 888b 8888 8888 ,"Y88b 888 8e d88 |
|
C8888 8888D 8888 8888 "8" 888 888 88b d88888 |
|
Y888 888P Y888 888P ,ee 888 888 888 888 |
|
"88 88" "88 88" "88 888 888 888 888 |
|
b |
|
8b, |
|
|
|
e88'Y88 d8 888 |
|
d888 'Y ,"Y88b 888,8, d88 ,e e, 888 |
|
C8888 "8" 888 888 " d88888 d88 88b 888 |
|
Y888 ,d ,ee 888 888 888 888 , 888 |
|
"88,d88 "88 888 888 888 "YeeP" 888 |
|
|
|
PROUDLY PRESENTS |
|
``` |
|
|
|
<img src="https://files.catbox.moe/zdxyzv.png" width="400"/> |
|
|
|
## TeTO-MS-8x7b-iMat-GGUF |
|
<i>Weighted quants were made using the full precision fp16 model and groups_merged_enhancedV3.</i> |
|
|
|
<u><b>Te</b></u>soro + <u><b>T</b></u>yphon + <u><b>O</b></u>penGPT |
|
|
|
Presenting a Model Stock experiment combining the unique strengths from the following 8x7b Mixtral models: |
|
* Tess-2.0-Mixtral-8x7B-v0.2 / [migtissera](https://huggingface.co/migtissera) / General Purpose |
|
* Typhon-Mixtral-v1 / [Sao10K](https://huggingface.co/Sao10K) / Creative & Story Completion |
|
* Open_Gpt4_8x7B_v0.2 / [rombodawg](https://huggingface.co/rombodawg) / Conversational |
|
|
|
# Recommended Template |
|
* Basic: Alpaca Format |
|
* Advanced: See context/instruct/sampler settings in [our new Recommended Settings repo](https://huggingface.co/Quant-Cartel/Recommended-Settings/tree/main/Teto-MS-8x7b). |
|
* Huge shout out to [rAIfle](https://huggingface.co/rAIfle) for his original work on the Wizard 8x22b templates which were modified for this model. |
|
|
|
<H2>Methodology</H2> |
|
|
|
> [I]nnovative layer-wise weight averaging technique surpasses state-of-the-art model methods such as Model Soup, utilizing only two fine-tuned models. This strategy can be aptly coined Model Stock, highlighting its reliance on selecting a minimal number of models to draw a more optimized-averaged model |
|
<i> (From [arXiv:2403.19522](https://arxiv.org/pdf/2403.19522))</i> |
|
|
|
|
|
* Methodology and merging process was based on the following paper - [Model Stock: All we need is just a few fine-tuned models](https://arxiv.org/abs/2403.19522) |
|
* Initial model selection was based on top performing models of Mixtral architecture covering a variety of use cases and skills |
|
* Base model (Mixtral Instruct 8x7b v0.1) was chosen after outperforming two other potential base models in terms of MMLU benchmark performance. |
|
|
|
# Output |
|
|
|
<img src="https://files.catbox.moe/bw97yg.PNG" width="400"/> |
|
|
|
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit). |
|
|
|
## Merge Details |
|
### Merge Method |
|
|
|
This model was merged using the [Model Stock](https://arxiv.org/abs/2403.19522) merge method using Mixtral-8x7B-v0.1-Instruct as a base. |
|
|
|
### Models Merged |
|
|
|
The following models were included in the merge: |
|
* migtissera_Tess-2.0-Mixtral-8x7B-v0.2 |
|
* rombodawg_Open_Gpt4_8x7B_v0.2 |
|
* Sao10K_Typhon-Mixtral-v1 |
|
|
|
### Configuration |
|
|
|
The following YAML configuration was used to produce this model: |
|
|
|
```yaml |
|
models: |
|
- model: models/migtissera_Tess-2.0-Mixtral-8x7B-v0.2 |
|
- model: models/Sao10K_Typhon-Mixtral-v1 |
|
- model: models/rombodawg_Open_Gpt4_8x7B_v0.2 |
|
merge_method: model_stock |
|
base_model: models/Mixtral-8x7B-v0.1-Instruct |
|
dtype: float16 |
|
``` |
|
|
|
|
|
## Appendix - Llama.cpp MMLU Benchmark Results* |
|
|
|
<i>These results were calculated via perplexity.exe from llama.cpp using the following params:</i> |
|
|
|
`.\perplexity -m .\models\TeTO-8x7b-MS-v0.03\TeTO-MS-8x7b-Q6_K.gguf -bf .\evaluations\mmlu-test.bin --multiple-choice -c 8192 -t 23 -ngl 200` |
|
|
|
|
|
``` |
|
* V0.01 (4 model / Mixtral Base): |
|
Final result: 43.3049 +/- 0.4196 |
|
Random chance: 25.0000 +/- 0.3667 |
|
|
|
|
|
* V0.02 (3 model / Tess Mixtral Base): |
|
Final result: 43.8356 +/- 0.4202 |
|
Random chance: 25.0000 +/- 0.3667 |
|
|
|
|
|
* V0.03 (4 model / Mixtral Instruct Base): |
|
Final result: 45.7004 +/- 0.4219 |
|
Random chance: 25.0000 +/- 0.3667 |
|
``` |
|
|
|
*Please be advised metrics above are not representative of final HF benchmark scores for reasons given [here](https://github.com/ggerganov/llama.cpp/pull/5047) |
