Update README.md
Browse files
README.md
CHANGED
|
@@ -34,6 +34,8 @@ Full offload possible on 16GB VRAM with a decent context size.
|
|
| 34 |
|
| 35 |
Bonus : a Kobold.CPP Frankenstein which reads IQ3_XXS models and is not affected by the Kobold.CPP 1.56/1.57 slowdown at the cost of an absent Mixtral fix.
|
| 36 |
https://github.com/Nexesenex/kobold.cpp/releases/tag/v1.57_b2030
|
|
|
|
|
|
|
| 37 |
|
| 38 |
---
|
| 39 |
|
|
|
|
| 34 |
|
| 35 |
Bonus : a Kobold.CPP Frankenstein which reads IQ3_XXS models and is not affected by the Kobold.CPP 1.56/1.57 slowdown at the cost of an absent Mixtral fix.
|
| 36 |
https://github.com/Nexesenex/kobold.cpp/releases/tag/v1.57_b2030
|
| 37 |
+
Now supperseded with another KCPP-F, with 13 different KV cache quantization lebel to chose from :
|
| 38 |
+
https://github.com/Nexesenex/kobold.cpp/releases
|
| 39 |
|
| 40 |
---
|
| 41 |
|