Skip learn.predict and run manual inference

Updated transform pipeline

fix(runtime): enforce python 3.10 for fastai compatibility on Hugging Face Spaces

fix(app): correct gradio Blocks setup and inference function for flower classifier

fix(gradio): return dict instead of list for Label output to resolve runtime TypeError

Switch to Gradio Blocks with submit button to fix dict output error

Remove num_top_classes to fix Gradio dict output error

Fix Gradio interface: use Blocks API for top-5 label display

Specified Gradio version in requirements.txt file to >=3.0

Sorted dictionary for Gradio to load 5 probs

Used Gr.Label with probabilities instead of dictionaries

Re-exported model using fastai 2.8.4

Specified fastai version to 2.8.4

Removed pytorch to let Hugging Face install it's own compatible version

Specified fastai and torch versions for compatability

Added ipython to requirements.txt file

Changed how I called the model to just the file name

Cleaned up python file- removed Kaggle-specific code

Add fastai dependency

Gradio UI issue- changed old py script name to app.ppy

Initial commit- python script and model for HuggingFace Spaces

Add application file

initial commit