Kijai/MiniMax-H3-experimental

Alibaba - H3 3-Step LoRA and Long-Form Video

#44 opened 1 day ago by

sunnyboxs

Upload 2 files

#43 opened 1 day ago by

kuro522

Upload 2 files

#42 opened 1 day ago by

Subzero784

Works much better with 8 steps

#41 opened 3 days ago by

Geffers

minimax_h3_fastvideo_vsa_datafree_1300step_4step_int8_convrot.safetensors

❤️👍 7

#40 opened 4 days ago by

AlinBig

PDD ACC LORA After updating Comfy UI to version 0.34.5, it started giving an error about tensors.

👍 1

#39 opened 4 days ago by

AIBOB1199

MiniMax-H3-exper

#38 opened 5 days ago by

willowoods

H3 fastvideo VSA acceleration becomes much slower after loading a LoRA in ComfyUI

#37 opened 11 days ago by

Ree0123

W4A8 version of the new 4step FastVideo model?

1

#36 opened 12 days ago by

IxMxAMAR

has anyone tried this AIO modal yet its awesome i ditched all others and loras and it works great standalone

👍 1

#35 opened 12 days ago by

dennismc1974

An interesting LoRa..

#34 opened 12 days ago by

anr2me

Please resolve the errors and warnings related to LoRA.

7

#33 opened 13 days ago by

makisekurisu-jp

Controlnet lora? Please always provide some text file in dir explaining what workflow, how to use it, source of idea etc

4

#31 opened 15 days ago by

MavenDE

MiniMax H3 Acc LoRas support?

👍 5

26

#30 opened 16 days ago by

anr2me

Cannot use with Native Comfy Minimax H3 I2V

👍 1

#29 opened 19 days ago by

Fatcraig

Unpruned W4A8 Diffusion Model Request

2

#27 opened 28 days ago by

makisekurisu-jp

The smaller models slow my generation down

#26 opened 29 days ago by

leonnn1

Reference Video

#25 opened 30 days ago by

Pentox

works well - NO audio?

#24 opened 30 days ago by

daspin335

dynamic vram not working in comfy

1

#23 opened about 1 month ago by

inykokso

please - do the same to TE

7

#22 opened about 1 month ago by

OliviaRossi

Why is int8_convrot VAE slower than fp16 VAE?

➕ 1

15

#21 opened about 1 month ago by

starmapking

What is this used for?【minimax_h3_ref_lora_rank_256_bf16】

➕ 1

7

#20 opened about 1 month ago by

RedHn

adding Lora freezes "Model Initializing ... "?

➕ 3

8

#18 opened about 1 month ago by

Dalvogalbo2

Unprunned model versions w4a8 pLZ :)

🚀🔥 3

1

#17 opened about 1 month ago by

ass5002

w4a8 slower than int8??

4

#15 opened about 1 month ago by

Fronyax

question about format and running in comfy

2

#14 opened about 1 month ago by

realrebelai

minimax_h3_video_vae_int8_convrot.safetensors

17

#13 opened about 1 month ago by

Portland01

Custom node required?

👀 1

3

#9 opened about 1 month ago by

MJS255

Update README.md

#8 opened about 1 month ago by

ryanlee-dev

KeyError: 'asym_w4a8_int8'

2

#7 opened about 1 month ago by

Danzelus

VRAM savings?

#6 opened about 1 month ago by

pizzawookiee

I found MiniMax-H3-Turbo-Lora ~ 4-step

❤️👍 2

2

#5 opened about 1 month ago by

sunnyboxs

Great work! Please also make one for the ref2va version. Thanks!

👍 8

6

#4 opened about 1 month ago by

lucas-ai26

finally!

1

#3 opened about 1 month ago by

Kutches