Update README.md
#1
by eugr-nv - opened
README.md
CHANGED
|
@@ -216,8 +216,6 @@ vllm serve nvidia/GLM-5.3-Flash-FP4-Local \
|
|
| 216 |
--host 0.0.0.0 \
|
| 217 |
--port 8000 \
|
| 218 |
--tensor-parallel-size 2 \
|
| 219 |
-
--pipeline-parallel-size 1 \
|
| 220 |
-
--decode-context-parallel-size 1 \
|
| 221 |
--mamba-cache-mode align \
|
| 222 |
--enable-prefix-caching \
|
| 223 |
--enable-chunked-prefill \
|
|
@@ -251,8 +249,6 @@ vllm serve nvidia/GLM-5.3-Flash-FP4-Local \
|
|
| 251 |
--host 0.0.0.0 \
|
| 252 |
--port 8000 \
|
| 253 |
--tensor-parallel-size 2 \
|
| 254 |
-
--pipeline-parallel-size 1 \
|
| 255 |
-
--decode-context-parallel-size 1 \
|
| 256 |
--mamba-cache-mode align \
|
| 257 |
--enable-prefix-caching \
|
| 258 |
--enable-chunked-prefill \
|
|
@@ -398,7 +394,6 @@ vllm serve nvidia/GLM-5.3-Flash-FP4-Local \
|
|
| 398 |
--max-parallel-prefills 1 \
|
| 399 |
--moe-backend b12x \
|
| 400 |
--override-generation-config '{"temperature":1.0,"top_p":0.95}' \
|
| 401 |
-
--pipeline-parallel-size 1 \
|
| 402 |
--prefill-compute-share 0.4 \
|
| 403 |
--prefill-schedule-interval 1 \
|
| 404 |
--quantization modelopt_mixed \
|
|
|
|
| 216 |
--host 0.0.0.0 \
|
| 217 |
--port 8000 \
|
| 218 |
--tensor-parallel-size 2 \
|
|
|
|
|
|
|
| 219 |
--mamba-cache-mode align \
|
| 220 |
--enable-prefix-caching \
|
| 221 |
--enable-chunked-prefill \
|
|
|
|
| 249 |
--host 0.0.0.0 \
|
| 250 |
--port 8000 \
|
| 251 |
--tensor-parallel-size 2 \
|
|
|
|
|
|
|
| 252 |
--mamba-cache-mode align \
|
| 253 |
--enable-prefix-caching \
|
| 254 |
--enable-chunked-prefill \
|
|
|
|
| 394 |
--max-parallel-prefills 1 \
|
| 395 |
--moe-backend b12x \
|
| 396 |
--override-generation-config '{"temperature":1.0,"top_p":0.95}' \
|
|
|
|
| 397 |
--prefill-compute-share 0.4 \
|
| 398 |
--prefill-schedule-interval 1 \
|
| 399 |
--quantization modelopt_mixed \
|