Frieren フリーレン - 葬送のフリーレン
This LoRA is intended to generate images of the character Frieren in a high quality aesthetic illustration style.
ℹ️ LoRA work best when applied to the base models on which they are trained. Please read the About This Version on the appropriate base models and workflow/training information.
Metadata is included in all uploaded files, you can drag the generated videos into ComfyUI to use the embedded workflows.
For other characters, see: Sousou no Frieren Collection
License
Description
[WAN 14B] LoRA
Trained with diffusion-pipe on Wan2.1-T2V-14B with a dataset including image and video
22 images, 9 videos
Video previews generated with ComfyUI_examples/wan/#text-to-video
Loading the LoRA with LoraLoaderModelOnly node and using the fp8 14B wan2.1_t2v_14B_fp8_e4m3fn.safetensors
Image previews generated with modified ComfyUI_examples/wan/#text-to-video
Setting the frame length to 1
Adding Upscaling
Image to Video previews generated with ComfyUI_examples/wan/#image-to-video
FAQ
Comments (4)
So you can train a lora for wan t2v and later use it in the i2v too? That has great potential.
Yes it seems to work well enough for guidance on image-to-video. The text-to-image capabilities of wan are also very impressive. I think text-to-image, upscale, then image-to-video is an effective workflow - all with the same LoRA
@motimalu Nice! Thank you for answering :D May i ask you something more? I'm interesting in training Wan2.1 too. I have tagged images and videos, but idk yet how to train this model. What software did you use? I would like to train it with my 3090 but perhaps 24 gb vram is not enough for videos of 480p and 5secs.
@VCominos No problem - it should be possible I think with a 3090. For wan I trained with diffusion-pipe on a 4090, with videos of around 33 frames at 16fps (~2 seconds) at 244p.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.