LSI

[LSI @SidaLukas]

@MercuryBarbora @OndejStary Je to AI. Takhle dobře nevypadá. Obyčejný FaceSwap :)

[LSI @SidaLukas]

@Alibaba_Wan There was a lot of talk about focusing on the open-source community, yet you're sticking with a closed model and won't release version 2.5

[LSI @SidaLukas]

@Alibaba_Wan The main thing is that at the conference you talked about a return to a community-focused and open-source approach. Boom—the first model is closed-source. :(

[LSI @SidaLukas]

@stefan_fee How do I fine-tune other languages? The Czech language is important to me. THANKS

[LSI @SidaLukas]

@stefan_fee How do I fine-tune other languages? The Czech language is important to me.

[LSI @SidaLukas]

@ltx_model Please also focus on supporting smaller languages or give us a powerful tool to train them. The model understands Czech, but it makes many mistakes in interpreting it and doesn’t recognize some words. Give us the ability to fine-tune the languages used in dialogues.

[LSI @SidaLukas]

@ComfyUI In comfyui, but definitely via API nodes 😁

[LSI @SidaLukas]

@PurzBeats @ComfyUI I did the update all directly from the manager in comfyui. 🤷🏼‍♂️

[LSI @SidaLukas]

@ComfyUI WAN 2.5 local please 😁

[LSI @SidaLukas]

@yoavhacohen @ltx_model Thank you for your help, I have sent you a private message.

[LSI @SidaLukas]

@yoavhacohen @ltx_model I tried https://t.co/xMFODI2AbJ, but I don't think it has much of an effect. Maybe a small dataset of 30 videos with descriptions. Does it have to be a special audio Lora node in comfyui, or is a regular Lora node used? Thank you very much.

[LSI @SidaLukas]

@yoavhacohen @ltx_model @yoavhacohen Hello, I don't know exactly who, but someone from your team replied that Lora can be used to fine-tune the language. I am referring to Czech. It is in the model, but it has an accent. Would it be possible to train the intonation with a Czech dataset using Lora?

[LSI @SidaLukas]

@BrentLynch @FCG_Studio @ltx_model @higgsfield_ai Thank you for your prompt reply.

[LSI @SidaLukas]

@BrentLynch @FCG_Studio @ltx_model @higgsfield_ai ...Will there be any possibility of using Czech/other audio dubbing in the future? Thank you very much

[LSI @SidaLukas]

@BrentLynch @FCG_Studio @ltx_model @higgsfield_ai It all sounds like a fairy tale that I might have a chance even with 8GB VRAM (RTX3070) and 64GB RAM for the VA model :) It's a terrible shame for small countries with minority languages that the model does not generate in smaller languages (Czech) like the veo3 model and Sora 2

[LSI @SidaLukas]

@ltx_model This month? 🥹

[LSI @SidaLukas]

@ComfyUI I'll wait for the local implementation next month :)

[LSI @SidaLukas]

@P_Fiala Děkuji za vysvětlení