Superman37891/PostgreSQL_Documentation_AI_Assistant

Fixed bug when typing same query twice

Removed debugging mode

Fixing query embedding latency time measurement

Measuring performance of model

Added temporary debugging for performance measurement

Modified README to reflect using combined semantic/keyword search

Removed useless function

Removed useless notebook and implemented RRF for combining vector search and text search similarities

Fixed README for HuggingFace compatibility

Added README

Switching to a much cheaper model

Switching model for long-term support

Updating model

Debugging

Updated model to use Gwen Instruct

Fixing tiny typo

Removing unexpected keyword argument

Made a dummy GPU function to run everything on CPU

Changing model because Gwen took too long

Switched from Mistral to Qwen

Fixed typo

Trying one last time with Mistral before switching models

Fixing API Key Error

Attempting to use HuggingFace router again

Modified API Key variables

Used OpenAI Key instead of HF Key

Switched to openai Python SDK to remove routing issues

Resolving provider API issue in get_answer

Fixing provider in InferenceClient

Use text_generation inside of get_answer instead of completion

Updated model to Mistral-7B-Instruct-v0.3 for compatibility with HuggingFace

Further debugging

Decorating answer function with spaces.GPU instead of sub-function

Fixing gpu annotation

Re-added code to use spaces gpu

Attempted debugging

Attempted debugging again

Attempted debugging

Attempted debugging again

Attempted debugging

Attempted debugging

Debugging

Attempting to fix app

Modified app.py to work with HuggingFace

Updated requirements

Initial commit

initial commit