Загрузка видео...
Не удалось загрузить видео
Most AI video tools stop at prompting, but LTX goes a layer deeper. Introducing the new LTX Trainer. Train LoRAs and IC-LoRAs across video, audio, cross-modal, and reference-conditioned workflows from a single framework. Plus: • New agentic skill • Flexible conditioning • Free IC-LoRAs • Fully open source Most... show more
6,487,225 просмотров • 3 месяцев назад •via X (Twitter)
Комментарии: 35
So, I can train a personalized model for each of my clients? Using their products, brand guides etc? 👀
Yes exactly!
Ltx + open source = Paradise!
imagine building a custom model for each business (yummieee)
Ahhh Requirements: Nvidia GPU with 80GB+ VRAM - Recommended for the standard config. For GPUs with 32GB VRAM (e.g., RTX 5090), use the low VRAM config which enables INT8 quantization and other memory optimizations Can this be run with 8gb vram?
oh wow!! now i'm excited 😍
So epic!!!!
You're the best Congrats on the launch
Looking forward to digging into this soon.
All in LTX 2.3, Open Source FTW!
This is great, excited to give it a try!
nice suff
nice, thank you LTX team
Thanks ! Hope there will be more guides for non technical artists!
nice ad
The future isn't using generic AI models. It's training models that understand your style.
This is awesome. I integrated it into my open source Automatic training Eval tool called "bracket". No modifications to the LTX trainer though to respect the requirements listed. Bracket let's you auto run across 30 training configs in sample or full datasets to get best score runs.
Will the Lora be supported in LTX Desktop? Or only comfyui?
This video is 🔥
Why is none of this stuff in the Web and Desktop Apps?
This is awesome, man. Finally a way to make the best AI models yours by training custom LoRAs for video and audio, all open source with LTX Trainer.
This seems really powerful and it's far more realistic to train LoRAs for LTX than it was for Wan 2.2
Already building something with it
LTX Trainer sounds lit for training custom LoRAs on video, audio and more, dude! Being fully open source is a huge plus.
This is great
You guys really need to team up with BMD, LTX and Davinci is the complete package
LTX Trainer go beyond prompts, make the model yours. Train video, audio, LoRAs free + open source
LTX Trainer: go beyond prompts, make the model yours train video, audio, and LoRAs, free and open source.
love this going beyond prompting into training LoRAs for video feels like a real shift toward control not just generation.
Bro, most models are generic but with LTX Trainer you can make them yours by training LoRAs and IC-LoRAs for video and audio. Open source all the way!
what do i do with these color problems?
seems awesome
banger
"Not borrowed" No, just stolen 🤣
Maybe you can use it to train your management in what dealing in good faith is? Is continuing to charge a subscription to someone who hasn’t touched your software in 3 years ethical?
![Video Model Comparison: Text to Video → Veo 3.1 → Sora 2 Pro → Kling 2.6 → Luma Ray3 → Wan 2.5 → Hailuo 2.3 → LTX-2 Pro → Seedance Pro Many of the new models automatically generate sound with video results including Veo 3.1, Sora 2 Pro, Kling 2.6, Wan 2.5, and LTX-2 Pro. I used Mirelo AI SFX and ElevenLabs to add sound to the others. I used the same prompt on each model 4x and chose the best result: Meteor Shower Panic Subject: [Father holding daughter's hand, looking up] Action: [Running through a suburban street, looking up as meteors streak across the sky, a small impact occurs nearby blowing wind and dust at them] Scene: [Suburban street, night, streetlights, burning trails in sky] Style: [Cinematic disaster, low angle tracking shot, high contrast, lens flares from meteors] Audio: Audio: sonic booms, wind rushing, car alarms triggered, shouting. Technical: [daytime, calm, sunny, happy] Which one is your favorite?](https://image.24vids.com/tw-1997700238828568730/ext_tw_video_thumb/1997700182079574016/pu/img/3WgOmHUz5ujEFJiF.jpg)
