Model tags wrong | Unsloth AI | Page 1

frigid frost May 12, 2024, 2:48 PM

#

Hello everyone, I'm trying to upload to HuggingFace via code as follows:

model_llama3.push_to_hub_merged("Fallperino/llama3-alpacca-sloth", tokenizer_llama3, save_method = "merged_16bit", token = userdata.get('hg-write'))

But for some reason I'm getting following error message:

/usr/local/lib/python3.10/dist-packages/transformers/utils/hub.py in create_and_tag_model_card(repo_id, tags, token, ignore_metadata_errors)
   1140     if tags is not None:
   1141         for model_tag in tags:
-> 1142             if model_tag not in model_card.data.tags:
   1143                 model_card.data.tags.append(model_tag)
   1144 

TypeError: argument of type 'NoneType' is not iterable

The repo does exist, it is only private, I don't have gatting enabled, names of model and tokenizer are correct. Could anyone, please, point me in the right direction?

#

@dire spire adding u here so i dont forget

#

so sorry this slipped past me

#

i know u solved it

#

but its still a bug 🙂

dire spire May 12, 2024, 2:49 PM

#

Thanks and no worries, atleast this now works as work around 😄

frigid frost May 12, 2024, 2:49 PM

#

🙂

kindred pine Jul 20, 2024, 5:12 AM

#

Hello, I was about to ask this question but found this post. Is any troubleshooting available for this issue?

Initially, I had fine-tuned Gemma 2 with a small dataset and successfully pushed the model to HF. However, subsequent attempts to push to HF are failing. Now I tried using SmolLM as the base model -- in case there is any problem with Gemma 2 -- but I still get the same error. Interestingly, the initial model push to HF did not create a README file with the Unsloth logo...I think that used to be the case earlier, no?

#Model tags wrong