Stuck for 10 mins at a Node
I am stuck at Prompt enhancer when I tried to generate video using LTX-2.3_-_I2V_T2V_Basic, released in your repo 3days ago. Prior to this new workflows the earlier workflow which I was using on 06/03/2026, worked smoothly. I am running ComfyUI of RTX 4070 12GB of VRAM, and 64GB of RAM.
Could you help me. Before it took me less than 5mins to create the video.
I seen that happen once or twice as well. You can always stop the generation (click the red X at top right corner in comfy).
And run again.
If it happens a lot, you can disable the prompt enhancer node. There is a toggle to turn it off at bottom of the workflow
Thank you it did work I toggled Off the prompt enhancer node, and then bypassed LTX2 Sampling Preview Override as this node gave me OOM errors instantly. But still I get error my process just gets killed silently. Any help would be appreciated. Even reduced the width and height as per given in the workflow. But process gets killed quietly.
How much VRAM you got? and what size are you trying?
If you got very low vram, try a test run (just to check if all works) at a low resolution. For example 832x480, or 768x512
And if you do have very low VRAM, and want to get higher res, a good option then is to use GGUF models (tiny bit slower, but can be a LOT smaller models).
For example Gemma 12B guff at Q2, and LTX at Q4
https://huggingface.co/unsloth/gemma-3-12b-it-GGUF/tree/main
https://huggingface.co/unsloth/LTX-2.3-GGUF/tree/main/distilled
Or alternatively https://huggingface.co/QuantStack/LTX-2.3-GGUF/tree/main/LTX-2.3-distilled
That being said, I did notice ComfyUI was quite slow and resource hungry after the updates today / yesterday.
So hopefully they will make another update
( a few updates ago things ran super smooth, so maybe a new update will come to fix that part)
I have an RTX 5070 12GB and 48GB of RAM. I updated to ComfyUI 0.16.4 yesterday or the day before, used LTX-2.3_-_I2V_T2V_Basic and ltx-2.3-22b-distilled_transformer_only_fp8_scaled. I disabled the Prompt Enhancer. When generating a 10s 720*1280 video, it takes about 2 to 4 minutes, and it almost never crashes.
Thank you.
I solve it at last, I was low on RAM. Placed some extra RAM increased it to 64GB. The entire workflow runs like butter. Thank you for pointing out the RAM issue.
Lastly are there any discord/social media channels where I can stay in touch with the Owner for making contributions?
I guess i could always create a Discord channel.
Created a new one for things related to this https://discord.gg/dyrq3ybF