Hi,
I think it would be good to mention more explicitly that the skill auto-downloads a model from HF (incl. size).
It was quite a surprise to see this message which also included file references to my source!
Warning: You are sending unauthenticated requests to the HF Hub. Please set a HF_TOKEN to enable higher rate limits and faster downloads.
{"query": "issue tracker", "results": [{"file_path": "src/main.rs",
...
...
Pi.dev explained it as follows.
### The 21.3s Wait
Why so slow? This was the very first search on this codebase. It had to:
1. Download the AI model from Hugging Face (~hundreds of MB)
2. Scan and index every file in the project (convert code into "meaning vectors")
3. Then answer your query
Next time? It will be nearly instant because the index is cached on disk. It only rebuilds when files change.
Hi,
I think it would be good to mention more explicitly that the skill auto-downloads a model from HF (incl. size).
It was quite a surprise to see this message which also included file references to my source!
Pi.dev explained it as follows.