Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model:
|
| 3 |
+
- Doctor-Shotgun/Mixtral-8x7B-Instruct-v0.1-LimaRP-ZLoss
|
| 4 |
+
- jondurbin/bagel-dpo-8x7b-v0.2
|
| 5 |
+
- NeverSleep/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss
|
| 6 |
+
license: cc-by-nc-4.0
|
| 7 |
+
tags:
|
| 8 |
+
- not-for-all-audiences
|
| 9 |
+
- nsfw
|
| 10 |
+
- mergekit
|
| 11 |
+
- merge
|
| 12 |
+
- exllamav2
|
| 13 |
+
---
|
| 14 |
+
|
| 15 |
+
# Undi95/BagelMix-8x7B 3.5bpw
|
| 16 |
+
Exllama quant of [Undi95/BagelMix-8x7B](https://huggingface.co/Undi95/BagelMix-8x7B)
|
| 17 |
+
|
| 18 |
+
You will need 24gb of vram to run this model at about half context (16k, you can probably go a bit higher too)
|
| 19 |
+
|
| 20 |
+
|
| 21 |
+
### Prompt format:
|
| 22 |
+
ChatML? maybe? Unclear
|
| 23 |
+
|
| 24 |
+
```
|
| 25 |
+
<|im_start|>system
|
| 26 |
+
{sysprompt}<|im_end|>
|
| 27 |
+
<|im_start|>user
|
| 28 |
+
{input}<|im_end|>
|
| 29 |
+
<|im_start|>assistant
|
| 30 |
+
{output}<|im_end|>
|
| 31 |
+
```
|
| 32 |
+
|
| 33 |
+
### Contact
|
| 34 |
+
Kooten on discord.
|
| 35 |
+
|