I’m training an IC-LoRA / video-to-video LoRA setup and I’m seeing frequent non-finite gradients during training.
The loss itself is still finite, but after backward / grad clipping the trainer reports non-finite gradients and skips the optimizer step.
Model: LTX-2.3 22B
Training mode: LoRA
LoRA rank / alpha: 128 / 128
Precision: bf16
Distributed: FSDP2
Learning rate: 1e-4
max_grad_norm: 1.0
In my logs, the issue seems to happen during backward rather than forward, because the loss remains finite. The skipped steps are also often associated with high sigma values.
I’m training an IC-LoRA / video-to-video LoRA setup and I’m seeing frequent non-finite gradients during training.
The loss itself is still finite, but after backward / grad clipping the trainer reports non-finite gradients and skips the optimizer step.
In my logs, the issue seems to happen during backward rather than forward, because the loss remains finite. The skipped steps are also often associated with high sigma values.