having a slight issue with template workflow on comfy
so 5 seconds runs fine everytime. i get those outputs they work and are clearly good gens.
the moment you add the required frame length, audio frame length and seconds for 10+ it will finish its global run but the local fails to a shape error:
cannot reshape tensor of 0 elements into shape [-1, 0] because the unspecified dimension size -1 can be any value and is ambiguous
im not sure if its the nodes audio encoding because thats where it fails everytime or the way the flow is set up currently but i cant get past 5 seconds.
heres my workflow i adjusted for gguf use, its legit just the unet s replacing the diffusion model nodes.
https://huggingface.co/realrebelai/Wan_Dancer_GGUFs/blob/main/WAN%20DANCER%20GGUF%20(Workflow).json
i had to remove the dance style and motion amplitude from the subgraph chain as the workflow would NaN the drop down if i tried to change the style.
this could completely be a ME issue but i cant get past it currently and claude cant see inside the subgraph for some reason
Uodate: issue was that i strung the local audio encoder into the total length node and that must remain seperate.