Leveraging Large language models (LLM), extract information from the documents.
This project utilizes the Mistral-7B quantized model, a powerful large language model (LLM), to effectively extract key information from diverse documents.
!pip3 install --upgrade auto-gptq
!pip3 install --upgrade transformers optimum
!pip install paddlepaddle-gpu
!pip install paddleocr
Quantized gptq model runs only in GPU environment . Use ggml or gguf versions for cpu based inference.
{ "invoice": { "invoice_number": "4172", "client_id": "335", "date": "03/27/2003" } }