mmaudio will slow down cuda usage on the consequent generations.. i already git pulled the custom node... nodiff1.3 is now 1.3.1, i removed GIMMVIF and swapped the sharpen node for another one since those were causing issues with ramConsider this workflow as a control panel, if you don't like the design, just don't download. Here a screenshot:<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/1fcbb0a1-52c2-4118-9c98-d03743e537e3/width=525/1fcbb0a1-52c2-4118-9c98-d03743e537e3.jpeg" />v1.3.1<ul><li>Updated resize image node</li><li>Added batch loader for start image input</li><li>Added Video To Extras Option</li><li>Added isolated face refiner step</li><li>Added MMaudio section (it will load the model to vram)</li><li>Added visual resolution/aspect ratio selector for T2V</li><li>Changed frame trimming nodes</li><li>Removed GIMMVIF interpolationf for film_net_fp32</li><li>Sharpen image swaped for sharpen MTB</li></ul>v1.2.1<ul><li>added GIMMVIF interpolationn (ds_factor (downscaling) at 0.25 can increase interpolation speed buy with a slightly quality trade-off)</li><li>added wildcard process node</li></ul>Added v1.2 <ul><li>cleaned prompt inputs and model loaders,</li><li>added CLIP VISION for helping image details in positive prompt (<a target="_blank" rel="ugc" href="https://huggingface.co/EQX55/VideoTest/blob/main/clip-vit-large-patch14-OPENAI.safetensors">download</a>)</li><li>added vram purge management nodes,</li><li>also made external common inputs for length and latent_window_size</li></ul>Used kijai's workflow, customized and added<ul><li>upscaler with model and rescaler</li><li>interpolation with sharpening</li><li>switch from samplers (default or F1)</li><li>switch for mode (i2v or t2v)</li><li>remove 5 frames (from the start for the non pingpong VHS, and from the start and from the end in the pingponged VSH) this is default at 40 frames for t2v</li><li>Reactor nodes for face swap/refine but those will slow down subsequent generations due to some kind of bug, feel free to delete reactor nodes if you feel like it...</li></ul>Text to video mode with F1 Sampler:<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/caf1a11e-3b09-4a7b-9d93-666e9842eb2c/width=525/caf1a11e-3b09-4a7b-9d93-666e9842eb2c.jpeg" />Hover the FramePack Text Encode (Timestamped) (prompt for F1) for seeing more info about the timestamped patterns for prompts:<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/1c92d2ac-1af1-4780-b7a7-a900b6c9c28b/width=525/1c92d2ac-1af1-4780-b7a7-a900b6c9c28b.jpeg" />* Please let me know if im missing any other custom nodes links so i can add them Also, theres a bug with current wrappers, if you incrase CFG avobe 1 (default) inference time will by x2 slower.. the only way to fix the workflow, even restarting comfy, is to add a fresh sampler and move the connections.. sad but it is what it is..Nodes:Latest Kijai Wrapper version: <a target="_blank" rel="ugc" href="https://github.com/kijai/ComfyUI-FramePackWrapper.git">https://github.com/kijai/ComfyUI-FramePackWrapper.git</a>F1 Sampler <a target="_blank" rel="ugc" href="https://github.com/kijai/ComfyUI-FramePackWrapper/pull/14/files">https://github.com/kijai/ComfyUI-FramePackWrapper/pull/14/files</a>You can also use:<a target="_blank" rel="ugc" href="https://github.com/ShmuelRonen/ComfyUI-FramePackWrapper_Plus">https://github.com/ShmuelRonen/ComfyUI-FramePackWrapper_Plus</a>GIMMVF:[<a target="_blank" rel="ugc" href="https://github.com/kijai/ComfyUI-GIMM-VFI">https://github.com/kijai/ComfyUI-GIMM-VFI</a>] (model with self download)Model links:<a target="_blank" rel="ugc" href="https://huggingface.co/Kijai/HunyuanVideo_comfy/blob/main/FramePackI2V_HY_fp8_e4m3fn.safetensors">https://huggingface.co/Kijai/HunyuanVideo_comfy/blob/main/FramePackI2V_HY_fp8_e4m3fn.safetensors</a><a target="_blank" rel="ugc" href="https://huggingface.co/Kijai/HunyuanVideo_comfy/blob/main/FramePackI2V_HY_bf16.safetensors">https://huggingface.co/Kijai/HunyuanVideo_comfy/blob/main/FramePackI2V_HY_bf16.safetensors</a>sigclip:<a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/sigclip_vision_384/tree/main">https://huggingface.co/Comfy-Org/sigclip_vision_384/tree/main</a>text encoder and VAE:<a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/HunyuanVideo_repackaged/tree/main/split_files">https://huggingface.co/Comfy-Org/HunyuanVideo_repackaged/tree/main/split_files</a>----

FramePack 💎 img2vid / txt2vid with LoRa Workflow - K3NK

cleaned prompt inputs and model loaders, added clip vision for helping positive prompt, added vram purge management nodes, also made external common inputs for length and latent_window_size, positions reorganized and tagged sections with a title

Files with SHA256: 782a97ae25ae182eb2c2ed02f6a1b55d4e0c34d9a32bb4a1cdcf1f4d019c74bb