This is a reupload of <a target="_blank" rel="ugc" href="https://huggingface.co/alibaba-pai/Wan2.1-Fun-1.3B-InP">https://huggingface.co/alibaba-pai/Wan2.1-Fun-1.3B-InP</a> including a fp8 conversion for people who can't run the 1.3b model in 16 bit precision.Wan 2.1-Fun-1.3B-InP is an img2vid wan model at 1.3 billion parameters, it was trained by <a target="_blank" rel="ugc" href="https://huggingface.co/alibaba-pai">Alibaba-PAI</a>. Initialized from the 1.3b t2v model. The weights are similar to the 14b i2v models, but with the size of the 1.3b model. Making it an easy to run, but still good quality i2v model. It was trained for start and end frame inpainting. Setting just a start frame allows it to do i2v. Wan 14b workflows<h3 id="lora-training-rv57tqs2e">Lora training</h3><s>If you want to use diffusion-pipe for lora training, you can use </s><a target="_blank" rel="ugc" href="https://github.com/gitmylo/diffusion-pipe/tree/patch-1"><s>my fork</s></a><s>. Make sure you're on the patch-1 branch. There's also an open pull request for it to be merged into the main repository.</s><s>git clone --recurse-submodules </s><a target="_blank" rel="ugc" href="https://github.com/gitmylo/diffusion-pipe"><s>https://github.com/gitmylo/diffusion-pipe</s></a><s> -b patch-1</s>The pr has been merged, so regular diffusion pipe can be used now:git clone --recurse-submodules <a target="_blank" rel="ugc" href="https://github.com/tdrussell/diffusion-pipe">https://github.com/tdrussell/diffusion-pipe</a>

Wan2.1-Fun-1.3B-InP i2v fp16 and fp8e4m3fn

The fp16 model, original from huggingface <a target="_blank" rel="ugc" href="https://huggingface.co/alibaba-pai/Wan2.1-Fun-1.3B-InP/blob/main/diffusion_pytorch_model.safetensors">https://huggingface.co/alibaba-pai/Wan2.1-Fun-1.3B-InP/blob/main/diffusion_pytorch_model.safetensors</a>. And a converted fp8e4m3fn version (make sure you download the one you want.)

Files with SHA256: cdab0deae98394e017bf1e6c67c963e4efc0629b275ae439e92254d3506055f9