Exporting Nemotron-3-nano-4B LoRa as GGUF failure | Unsloth AI | Page 1

storm dawn Mar 29, 2026, 6:54 AM

#

GGUF export failed: Failed to save/merge model: Unsloth: Saving LoRA finetune failed since # of LoRAs = 50 does not match # of saved modules = 0. Please file a bug report!

untold finch Mar 29, 2026, 10:46 AM

#

are you able to export to GGUF using just llama-cpp

#

?

storm dawn Mar 29, 2026, 2:17 PM

#

untold finch are you able to export to GGUF using just llama-cpp

Hmm, I suppose I am not, simply from my own inexperience

untold finch Mar 29, 2026, 2:43 PM

#

is your lora anywhere on HF?

storm dawn Mar 30, 2026, 3:22 AM

#

untold finch is your lora anywhere on HF?

I have not posted it. Would it help?

#

I have uploaded stuff to HF before, but only when trying to use a cloud GPU on a headless training task, automatically as part of the script. Unsloth Studio doesn't automatically upload, does it?

#Exporting Nemotron-3-nano-4B LoRa as GGUF failure