Update README.md
Browse files
README.md
CHANGED
@@ -33,7 +33,8 @@ Two sets of models are provided:
|
|
33 |
* Groupsize = 1024
|
34 |
* Should work reliably in 24GB VRAM
|
35 |
* Groupsize = 128
|
36 |
-
*
|
|
|
37 |
* In my testing it ran out of VRAM on a 24GB card around 1500 tokens returned.
|
38 |
|
39 |
For each model, two versions are available:
|
|
|
33 |
* Groupsize = 1024
|
34 |
* Should work reliably in 24GB VRAM
|
35 |
* Groupsize = 128
|
36 |
+
* Optimal setting for highest inference quality
|
37 |
+
* But may require more than 24GB VRAM, depending on response length
|
38 |
* In my testing it ran out of VRAM on a 24GB card around 1500 tokens returned.
|
39 |
|
40 |
For each model, two versions are available:
|