Update README.md

Add huggingface_hub dependency for compatibility with new router endpoint

Update README and requirements to reflect changes in API usage and dependencies

Refine user prompt for improved project description clarity and conciseness

Switch to router.huggingface.co OpenAI-compatible chat API for LLM completions

Remove base_url parameter - let library handle endpoint routing automatically

Pass model to InferenceClient initialization and remove from create call

Fix base_url for InferenceClient to use models endpoint

Update to use new InferenceClient API method

Pin specific compatible versions of gradio and huggingface_hub

Update gradio to fix HfFolder import error

Trigger Space rebuild

Update huggingface_hub to support new router endpoint

Add minijinja dependency

Trigger Space rebuild

Update README for clarity on Hugging Face token requirements and model details

Refactor: Remove rule-based fallback and improve error handling for LLM description improvement

Fix: Update model ID to meta-llama/Llama-3.2-3B-Instruct

Implement hybrid AI + rule-based fallback system

Fix: Improve error handling and switch to flan-t5-large

Refactor: Switch to google/flan-t5-xxl for reliability

Docs: Add instructions for setting HF token secret

Feat: Re-initialize with stable zero-shot Inference API project

Remove all project files

Fix HfFolder import by adding compatibility shim

Fix deps: pin gradio and huggingface_hub to compatible versions

Initial commit: Zero-shot technical description assistant