FP8-Dynamic quant created with llm-compressor, can run on 16 VRAM cards. Update vLLM and Transformers:

pip install vllm>=0.7.2
pip install git+https://github.com/huggingface/transformers

Then run with:

vllm serve leon-se/Qwen2.5-VL-7B-Instruct-FP8-Dynamic --trust-remote-code
Downloads last month
369
Safetensors
Model size
8.29B params
Tensor type
BF16
·
F8_E4M3
·
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no library tag.

Model tree for leon-se/Qwen2.5-VL-7B-Instruct-FP8-Dynamic

Quantized
(8)
this model