kyush/vllm-recipe
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Update device/nvidia-agx-thor/nvidia-diffusiongemma-26B-A4B-it-NVFP4.md

Browse source
This commit is contained in:
Kyush 2026-06-22 11:09:56 +00:00
commit ae5fe10662
1 changed files with 5 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

6
device/nvidia-agx-thor/nvidia-diffusiongemma-26B-A4B-it-NVFP4.md
View file

@ -48,6 +48,7 @@ vllm bench serve \
--num-prompts 5 \
--max-concurrency 1
```
![x1](nvidia-diffusiongemma-26B-A4B-it-NVFP4-x1.png)
```bash
vllm bench serve \
@ -60,4 +61,7 @@ vllm bench serve \
--random-output-len 1024 \
--num-prompts 32 \
--max-concurrency 8
```
```
![x81](nvidia-diffusiongemma-26B-A4B-it-NVFP4-x8.png)
![tj](nvidia-diffusiongemma-26B-A4B-it-NVFP4-tj.png)