Issues · Lightricks/LTX-Video-Trainer · GitHub

Labels Milestones New issue

[Bug]: Learning rate reaches end at `steps / num_processes` when using multi-GPU with default `split_batches=False`

#67

· Siqiao Huang (knightnemo) opened

on Feb 1, 2026

VRAM Requirements for LoRA Finetuning 13B model

#66

· 鸡蛋 (J1dan) opened

on Jan 26, 2026

Language fine-tune

#65

· LSI86 opened

on Jan 23, 2026

[Discussion/Question] Weak Controllability on Training Set with LoRA Finetuning (POV Motion Dataset)

#63

· 鸡蛋 (J1dan) opened

on Jan 2, 2026

Full finetuning error

#62

· Chi Tran (IceIce1ce) opened

on Dec 26, 2025

[Question] Guidance on training steps vs. dataset size for LoRA fine-tuning

#61

· wangxinliang (fyting) opened

on Dec 17, 2025

Preprocessing dataset with videos with too few frames can lead to incorrect labels

#59

· cj-gregory-kovelman opened

on Nov 27, 2025

Why model_source set for local path can not work?

#55

· redblack-cloud opened

on Oct 10, 2025

Higher GPU memory usage when setting load_text_encoder_in_8bit to true

#53

· Yi Zhu (bryanyzhu) opened

on Oct 1, 2025

Model is hard to be modified

#51

· Zihao Zou (zihaozou) opened

on Sep 24, 2025

i want to change the LTXVideoTransformer3DModel

#50

· ohad204 opened

on Sep 11, 2025

The quality of the VAE differs from the demo

#47

· julia-joanna opened

on Aug 4, 2025