Update README.md
Browse files
README.md
CHANGED
|
@@ -25,8 +25,8 @@ As a workaround, to inject the <think> token in OpenWebUI, you can use the [inje
|
|
| 25 |
./build/bin/llama-server \
|
| 26 |
--alias anikifoss/MiniMax-M2-HQ4_K \
|
| 27 |
--model ~/Env/models/anikifoss/MiniMax-M2-HQ4_K/MiniMax-M2-HQ4_K-00001-of-00004.gguf \
|
| 28 |
-
--temp 1.0 --top-k 0 --top-p 1.0 --min-p 0.
|
| 29 |
-
--repeat-penalty 1.
|
| 30 |
--ctx-size 95000 \
|
| 31 |
-ctk q8_0 -ctv q8_0 \
|
| 32 |
-fa on \
|
|
|
|
| 25 |
./build/bin/llama-server \
|
| 26 |
--alias anikifoss/MiniMax-M2-HQ4_K \
|
| 27 |
--model ~/Env/models/anikifoss/MiniMax-M2-HQ4_K/MiniMax-M2-HQ4_K-00001-of-00004.gguf \
|
| 28 |
+
--temp 1.0 --top-k 0 --top-p 1.0 --min-p 0.02 \
|
| 29 |
+
--repeat-penalty 1.04 --repeat-last-n 256 \
|
| 30 |
--ctx-size 95000 \
|
| 31 |
-ctk q8_0 -ctv q8_0 \
|
| 32 |
-fa on \
|