Update README.md
Browse files
README.md
CHANGED
|
@@ -30,9 +30,9 @@ tags:
|
|
| 30 |
- mixture of experts
|
| 31 |
- 128 experts
|
| 32 |
- 8 active experts
|
| 33 |
-
-
|
| 34 |
- qwen3
|
| 35 |
-
-
|
| 36 |
- brainstorm 20x
|
| 37 |
- brainstorm
|
| 38 |
- optional thinking
|
|
@@ -128,7 +128,10 @@ QUANTS
|
|
| 128 |
|
| 129 |
---
|
| 130 |
|
| 131 |
-
Quants are build using NEO Imatrix dataset, and optimized specifically for thinking block models
|
|
|
|
|
|
|
|
|
|
| 132 |
|
| 133 |
---
|
| 134 |
|
|
|
|
| 30 |
- mixture of experts
|
| 31 |
- 128 experts
|
| 32 |
- 8 active experts
|
| 33 |
+
- 256k context
|
| 34 |
- qwen3
|
| 35 |
+
- gguf
|
| 36 |
- brainstorm 20x
|
| 37 |
- brainstorm
|
| 38 |
- optional thinking
|
|
|
|
| 128 |
|
| 129 |
---
|
| 130 |
|
| 131 |
+
Quants are build using NEO Imatrix dataset (by DavidAU), and optimized specifically for thinking block models
|
| 132 |
+
with the output tensor at 16 bits.
|
| 133 |
+
|
| 134 |
+
I have found this specific combo to be unbeatable even at IQ2_M.
|
| 135 |
|
| 136 |
---
|
| 137 |
|