Gemma 4 E4B BF16 inference via vLLM. Only Gemma 4 model with audio input (30s clips, 16 kHz). Also supports image and video. 128K context.
Are you sure you want to perform this action?
Status
Message here