This is a Wan2.2 Bernini-R video workflow that supports:<ol><li>Local Gemma 4 gguf file for LLM prompt enhancing.</li><li>Input of both images and videos for LLM.</li></ol>Example use case: Video editing with reference images.Custom node used:<a target="_blank" rel="ugc" href="https://github.com/KLL535/ComfyUI_Simple_Qwen3-VL-gguf">ComfyUI_Simple_Qwen3-VL-gguf</a><ul><li>Set "cpu_moe": true to speed up LLM with limited vram. (Q8 model needs ~6GB vram.)</li><li>To set image_min_tokens for Gemma 4, follow n_ubatch &gt; image_max_tokens &gt; image_min_tokens. (e.g. 2240, 2240, 560)</li><li>Qwen 3.6 can also be used instead of Gemma 4. Just set image_min_tokens to 1024 and n_ctx to a larger value (e.g. 16384).</li></ul>

Files with SHA256: 823a0bc1e9485f571310fa5ca21c8357e2ef603b4a7a5b958c41776e7243640b