|
--- |
|
base_model: |
|
- NousResearch/Yarn-Mistral-7b-128k |
|
- Test157t/Kunocchini-1.1-7b |
|
library_name: transformers |
|
tags: |
|
- mistral |
|
- quantized |
|
- text-generation-inference |
|
- merge |
|
- mergekit |
|
pipeline_tag: text-generation |
|
inference: false |
|
--- |
|
|
|
## "NOTES: This model seems to be overtly confident leading to hallucinations, normalization has seemed to also break the long context chaining. I do not recommend this model." Use the previous, Kunoccini-7B-128k-test, in [**this collection**](https://huggingface.co/collections/Lewdiculous/quantized-models-gguf-65d8399913d8129659604664). |
|
|
|
# **GGUF-Imatrix quantizations for [Test157t/Kunocchini-1.2-7b-longtext](https://huggingface.co/Test157t/Kunocchini-1.2-7b-longtext/).** |
|
|
|
# What does "Imatrix" mean? |
|
|
|
It stands for **Importance Matrix**, a technique used to improve the quality of quantized models. |
|
|
|
The **Imatrix** is calculated based on calibration data, and it helps determine the importance of different model activations during the quantization process. The idea is to preserve the most important information during quantization, which can help reduce the loss of model performance. |
|
|
|
One of the benefits of using an Imatrix is that it can lead to better model performance, especially when the calibration data is diverse. |
|
|
|
More information: [[1]](https://github.com/ggerganov/llama.cpp/discussions/5006) [[2]](https://github.com/ggerganov/llama.cpp/discussions/5263#discussioncomment-8395384) |
|
|
|
SillyTavern preset files for the previous version are located [here](https://huggingface.co/Test157t/Kunocchini-7b-128k-test/tree/main/ST%20presets). |
|
|
|
*If you want any specific quantization to be added, feel free to ask.* |
|
|
|
All credits belong to the [creator](https://huggingface.co/Test157t/). |
|
|
|
`Base⇢ GGUF(F16)⇢ Imatrix(F16)⇢ GGUF-Imatrix(Quants)` |
|
|
|
The new **IQ3_S** merged today has shown to be better than the old Q3_K_S, but will only be supported in `koboldcpp-1.60` or newer. |
|
|
|
Using [llama.cpp](https://github.com/ggerganov/llama.cpp/)-[b2254](https://github.com/ggerganov/llama.cpp/releases/tag/b2254). |
|
|
|
For --imatrix data, `imatrix-Kunocchini-1.2-7b-longtext-F16.dat` was used. |
|
|
|
# Original model information: |
|
|
|
Thanks to @Epiculous for the dope model/ help with llm backends and support overall. |
|
|
|
Id like to also thank @kalomaze for the dope sampler additions to ST. |
|
|
|
@SanjiWatsuki Thank you very much for the help, and the model! |
|
 |
|
|
|
This model was merged using the [DARE](https://arxiv.org/abs/2311.03099) [TIES](https://arxiv.org/abs/2306.01708). |
|
|
|
### Models Merged |
|
|
|
The following models were included in the merge: |
|
* [NousResearch/Yarn-Mistral-7b-128k](https://huggingface.co/NousResearch/Yarn-Mistral-7b-128k) + [Test157t/Kunocchini-1.1-7b](https://huggingface.co/Test157t/Kunocchini-1.1-7b) |
|
|
|
### Configuration |
|
|
|
The following YAML configuration was used to produce this model: |
|
|
|
```yaml |
|
merge_method: dare_ties |
|
base_model: Test157t/Kunocchini-1.1-7b |
|
parameters: |
|
normalize: true |
|
models: |
|
- model: NousResearch/Yarn-Mistral-7b-128k |
|
parameters: |
|
weight: 1 |
|
- model: Test157t/Kunocchini-1.1-7b |
|
parameters: |
|
weight: 1 |
|
dtype: float16 |
|
``` |
