7/26 Update - Just a quick tweak for Kijai's new lora options on the updated lora loaders. merge_loras set to True, low_mem_load set to False, upped block swap to 20 for 1280x720 to handle the increased memory requirements with how the loras now load.

7/30 update - replaced final frame selector with something more available.

My somewhat condensed workflow, meant for 24Gb+ VRAM cards. Mostly set up for a 5090, but can easily be used with a 3090 or 4090 by upping the block swap to fit. Lots of workflows are out there for folks with less powerful GPU's, but if you have the power, why not use it?

Also you should have at least 64Gb of system RAM as well, to handle model/lora offloading. ComfyUI's poor RAM clearing may require you to close the terminal and re-launch Comfy from time to time, if you're doing a lot of generations and swapping out loras. I've had it eventually run out my 96GB of RAM.

Skyreels V2 720p Model-

https://huggingface.co/Kijai/WanVideo_comfy/blob/main/Skyreels/Wan2_1-SkyReels-V2-T2V-14B-720P_fp8_e5m2.safetensors

Edit- Updated Self-Forcing Lightx2v Lora is out. You can either use 128 rank or 64 rank, they both use the same amount of VRAM and system RAM, and have the same speed. 128 will give you slightly better gens, but takes twice the hard drive space at 1.25 Gb vs 64 rank at 631 Mb.

128-

https://huggingface.co/Kijai/WanVideo_comfy/blob/main/Lightx2v/lightx2v_T2V_14B_cfg_step_distill_v2_lora_rank128_bf16.safetensors

64-

https://huggingface.co/Kijai/WanVideo_comfy/blob/main/Lightx2v/lightx2v_T2V_14B_cfg_step_distill_v2_lora_rank64_bf16_.safetensors

The Lightx2v lora should be kept around 0.7-1.0 strength. CFG must stay at 1 with this lora to keep the speed boost, and I like the quality at 6 steps.

Once you confirm models are loaded correctly on the left side of the workflow, and the VAE up top, all of the parameters you'd regularly want to change are in the center for ease of use.

Outputs will be in a dated SkyReelsT2V folder, and the final frame saved as well, in case you want to play around with extending in I2V.

On a 5090, puts out high quality gens at 24fps, 121 frame 1024x576 videos in around 1:30, 1280x720 in around 3:04.