Gemma2 errors after merging :

[Gomez12]

If I create a gguff from unsloth, if I then try loading it with llama.cpp it errors on : llama_model_load: error loading model: check_tensor_dims: tensor 'blk.0.ffn_norm.weight' not found

Does this with 9b and 27b

[danielhanchen]

@Gomez12 Apologies - my bad - please reinstall and update unsloth! For local machines, use:

pip uninstall unsloth -y
pip install --upgrade --force-reinstall --no-cache-dir git+https://github.com/unslothai/unsloth.git

For Colab & Kaggle, just refresh

Apologies on the issue!

[Gomez12]

Thnx, works perfectly now for me