VLM Benchmark
checking...
vLLM continuous batching - Qwen2.5-VL-7B-AWQ
Images (
0
)
Add Images
Clear All
Settings
Prompt
Describe this image briefly.
Benchmark Mode
Concurrent (all at once)
Sequential (one by one)
Simulate Real Traffic (random delays)
Min Delay (sec)
Max Delay (sec)
Run Benchmark
Running inference...
Results
Timeline (when each request was sent and completed)
-
Total Time (ms)
-
Avg Inference (ms)
-
Images/sec
-
Batching Benefit
-
Max Concurrent
Add images and run benchmark to see results