FileGPT: Local cognitive search over a pdf

Steps to run the model locally

clone this repository

   git clone https://github.com/allthatido/FileGPT.git

Navigate to the models folder
Download the LLama family model (GGML V3 only) you want to test. Refer to README.md in the models folder. Huggingface is a good source for these.
install dependencies. If you get an error for the below command, install each library indivisually using pip install. Windows users may have to install latest Visual C++ libraries from here and CMake from here. If you have a Nvidia GPU use cuBLAS to speed up inference. Install llama-cpp-python as described here
```
   pip install -r requirements.txt
```
Start streamlit app
```
   streamlit run app.py
```
Open browser and go to http://localhost:8501/

Name		Name	Last commit message	Last commit date
Latest commit History 10 Commits
.streamlit		.streamlit
data		data
media		media
models		models
.gitignore		.gitignore
README.md		README.md
app.py		app.py
css.py		css.py
db.py		db.py
llm.py		llm.py
requiremnts.txt		requiremnts.txt