[Bug] "Unsloth: Your model needs to call `.get_peft_model` first!" in gemma-3 Colab Notebook Example

[Coder3333]

This notebook is referenced in the unsloth docs, but I keep getting stuck at one step with an exception. I swear I have run all of the previous steps in order properly. Please, help me get through this. Thank you.

Error:
"Unsloth: Your model needs to call .get_peft_model first!"

Step: <- Have to change the False to True on this step
if True:
from unsloth import FastLanguageModel
model, tokenizer = FastLanguageModel.from_pretrained(
model_name = "gemma-3", # YOUR MODEL YOU USED FOR TRAINING
max_seq_length = 2048,
load_in_4bit = False,
)

Notebook:
https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Gemma3_(270M).ipynb

Document reference:
https://docs.unsloth.ai/models/gemma-3-how-to-run-and-fine-tune

[danielhanchen]

Sorry we just fixed the issue - also I fixed the model name mismatch since it was "lora_model" before.

For Colab / Kaggle, please disconnect and delete runtime, or to use saved work, update Unsloth (also update on local machines as well) via:

pip install --upgrade --force-reinstall --no-cache-dir --no-deps unsloth unsloth_zoo

[Coder3333]

@danielhanchen , I just got a new error on the same step. This is after starting a clean notebook and also using your force install recommendation.

==((====))==  Unsloth 2025.9.7: Fast Gemma patching. Transformers: 4.55.4.
   \\   /|    Tesla T4. Num GPUs = 1. Max memory: 14.741 GB. Platform: Linux.
O^O/ \_/ \    Torch: 2.8.0+cu126. CUDA: 7.5. CUDA Toolkit: 12.6. Triton: 3.4.0
\        /    Bfloat16 = FALSE. FA [Xformers = 0.0.32.post2. FA2 = False]
 "-____-"     Free license: http://github.com/unslothai/unsloth
Unsloth: Fast downloading is enabled - ignore downloading bars which are red colored!
tokenizer_config.json: 
 1.16M/? [00:00<00:00, 26.2MB/s]
tokenizer.model: 100%
 4.69M/4.69M [00:00<00:00, 9.88MB/s]
added_tokens.json: 100%
 35.0/35.0 [00:00<00:00, 1.92kB/s]
special_tokens_map.json: 100%
 670/670 [00:00<00:00, 14.3kB/s]
tokenizer.json: 100%
 33.4M/33.4M [00:00<00:00, 74.9MB/s]
chat_template.jinja: 
 1.65k/? [00:00<00:00, 49.7kB/s]
Unsloth: Tokenizer is most likely buggy, and Unsloth failed to repair it.
It will still work, but beware of out of bounds memory accesses.
Please file an issue on the model owner's repo about this issue.
---------------------------------------------------------------------------
NotImplementedError                       Traceback (most recent call last)
[/tmp/ipython-input-4142586749.py](https://localhost:8080/#) in <cell line: 0>()
      1 if True:
      2     from unsloth import FastLanguageModel
----> 3     model, tokenizer = FastLanguageModel.from_pretrained(
      4         model_name = "gemma-3", # YOUR MODEL YOU USED FOR TRAINING
      5         max_seq_length = 2048,

1 frames
[/usr/local/lib/python3.12/dist-packages/unsloth/models/llama.py](https://localhost:8080/#) in patch_peft_model(model, use_gradient_checkpointing)
   2740         elif model_type == "qwen3moe":  apply_lora_mlp = apply_lora_mlp_swiglu
   2741         else:
-> 2742             raise NotImplementedError(f"Unsloth: {model_type} is not yet implemented!")
   2743         pass
   2744 

NotImplementedError: Unsloth: gemma3_text is not yet implemented!

[Coder3333]

I will mark this as closed. The new build has cleared up the originally reported error. There is now a new error in the next step of the notebook, but I will create a new issue for that.