sglang.0.4.8.post1/sglang/benchmark/mmmu

Run evaluation

Evaluate sglang

Host the VLM:

python -m sglang.launch_server --model-path Qwen/Qwen2-VL-7B-Instruct --port 30000

It's recommended to reduce the memory usage by appending something like --mem-fraction-static 0.6 to the command above.

Benchmark:

python benchmark/mmmu/bench_sglang.py --port 30000 --concurrency 16

You can adjust the --concurrency to control the number of concurrent OpenAI calls.

You can use --lora-path to specify the LoRA adapter to apply during benchmarking. E.g.,

# Launch server with LoRA enabled
python -m sglang.launch_server --model-path microsoft/Phi-4-multimodal-instruct --port 30000 --trust-remote-code --disable-radix-cache --lora-paths vision=<LoRA path>

# Apply LoRA adapter during inferencing
python -m benchmark/mmmu/bench_sglang.py --concurrency 8 --lora-path vision

Evaluate hf

python benchmark/mmmu/bench_hf.py --model-path Qwen/Qwen2-VL-7B-Instruct