Instructions to use drbaph/MiniMax-H3-Turbo-Lora-ComfyUI with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Inference
- Notebooks
- Google Colab
- Kaggle
The v4 is the best lora currently
I just wanted to thank you. I've tried all the Turbo Loras currently available, and your V4 version is by far the best Loras of all. In the Lightx Loras, the women look like Flux1 with plastic skin, the movements are terrible, the quality is not good and the sound is terrible. With the V4 version everything is great, the sound, the picture quality everything. I hope you do more versions, your work is brilliant.
I was going to compliment on the 4-step 1.1 768p Rank 61 LoRA. The quality just pops and it is practically the same output as an 8-step LoRA.
if you test 4 steps you have to test it on fast motion
still slow motion vids don't really highlight its issue (if any)
if you test 4 steps you have to test it on fast motion
still slow motion vids don't really highlight its issue (if any)
I noticed that, despite the quality of the LoRA. 8-step was definitely more free-flowing with its movements.
if you test 4 steps you have to test it on fast motion
still slow motion vids don't really highlight its issue (if any)I noticed that, despite the quality of the LoRA. 8-step was definitely more free-flowing with its movements.
I wonder why we stopped seeing 8 steps. 4 step seems to give me a lot of body horror and character misspeaking issues.
if you test 4 steps you have to test it on fast motion
still slow motion vids don't really highlight its issue (if any)I noticed that, despite the quality of the LoRA. 8-step was definitely more free-flowing with its movements.
I wonder why we stopped seeing 8 steps. 4 step seems to give me a lot of body horror and character misspeaking issues.
Because everyone keeps pushing for speed and not quality.