Can LTX-Video Create Stunning Text-to-Video on Low VRAM (6/8 GB)? Find Out Now!

[@Crisisdarkness]

Thanks friend, I will try this, I hope it will run with 6 GB of GPU, I will tell you later if it was possible. Your channel is valuable, glad I found you

[@StableAIHub]

Thank you

[@ivoxx_]

OMFG this is fast!
I wish I could run in comfy!

[@StableAIHub]

Yes you can run on Comfyi, there is a workflow on their github page
github.com/Lightricks/LTX-Video

[@Parthi97]

can i run this on 1050 gtx 4gb VRAM??
with t5xxl_fp8_e4m3fn.safetensors text encoder and ltx-video-2b-v0.9.safetensors as checkpoint??

[@StableAIHub]

You will have to make code changes to load FP8 model.

[@slim420-e8v]

Does it work on AMD?

[@StableAIHub]

Unfortunately, No.

[@stable-ai]

hi there , how do do image to video with this , thanks

[@StableAIHub]

I have not tested if it is working.
try
1. Copy the image in LTXVideo folder
2. In config.yaml update the image file name input_image_path: ""
3. Update prompt and other settings.
Please let me know if it works. If not I will have a look

[@StableAIHub]

Tested, Verified I2V working.

[@bause6182]

Can it run on comfyui ?

[@StableAIHub]

This is for standalone installation. You can find Comfyi workflow in their github repo.

[@WileyHickok-sd6ov]

Yes

[@StableAIHub]

Yes you can run on Comfyi, there is a workflow on their github page
github.com/Lightricks/LTX-Video