Commit
•
5ecc992
1
Parent(s):
946d23d
Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- gguf
|
| 4 |
+
- mistral
|
| 5 |
+
- conversational
|
| 6 |
+
---
|
| 7 |
+
<img src="https://i.imgur.com/P68dXux.png" width="400"/>
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
# Mistral 7B v0.2 iMat GGUF
|
| 11 |
+
|
| 12 |
+
<h1>Not to be confused with Mistral 7B Instruct v0.2 (this is the latest release from 3/23) </h1>
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
Mistral 7B v0.2 iMat GGUF quantized from fp16 with love.
|
| 17 |
+
* iMat dat file created using groups_merged.txt
|
| 18 |
+
* Not sure what to expect from this model by itself but uploading to repo incase anyone is curious like me
|
| 19 |
+
|
| 20 |
+
<b>Legacy quants (i.e. Q8, Q5_K_M) in this repo have all been enhanced with importance matrix calculation. These quants show improved KL-Divergence over their static counterparts.</b>
|
| 21 |
+
|
| 22 |
+
All files have been tested for your safety and convenience. No need to clone the entire repo, just pick the quant that's right for you.
|
| 23 |
+
|
| 24 |
+
For more information on latest iMatrix quants see this PR - https://github.com/ggerganov/llama.cpp/pull/5747
|
| 25 |
+
|