|
--- |
|
tags: |
|
- gguf |
|
- quantized |
|
- roleplay |
|
- imatrix |
|
- mistral |
|
inference: false |
|
--- |
|
|
|
This repository hosts GGUF-Imatrix quantizations for [ResplendentAI/Datura_7B](https://huggingface.co/ResplendentAI/Datura_7B). |
|
``` |
|
Base⇢ GGUF(F16)⇢ Imatrix-Data(F16)⇢ GGUF(Imatrix-Quants) |
|
``` |
|
```python |
|
quantization_options = [ |
|
"Q4_K_M", "Q5_K_M", "Q6_K", "Q8_0" |
|
] |
|
``` |
|
|
|
**This is experimental.** |
|
|
|
For imatrix data generation, kalomaze's `groups_merged.txt` with added roleplay chats was used, you can find it [here](https://huggingface.co/Lewdiculous/Datura_7B-GGUF-Imatrix/blob/main/imatrix-with-rp-format-data.txt). |
|
|
|
The goal is to measure the (hopefully positive) impact of this data for consistent formatting in roleplay chatting scenarios. |
|
|
|
**Original model information:** |
|
|
|
# Datura 7B |
|
|
|
 |
|
|
|
Flora with a bit of toxicity. |
|
|
|
I've been making progress with my collection of tools, so I thought maybe I'd try something a little more toxic for this space. This should make for a more receptive model with fewer refusals. |
