Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/patw/sumbot

A local API for using a Mistral 7b LLM to summarize JSON data into semantically rich plain text.
https://github.com/patw/sumbot

fastapi json mistral-7b

Last synced: 29 days ago
JSON representation

A local API for using a Mistral 7b LLM to summarize JSON data into semantically rich plain text.

Awesome Lists containing this project

README

        

# SumBot
Summarize structured JSON data into LLM produced, semantically rich paragraphs. Great for embedding!

## Local Installation

```
pip install -r requirements.txt
```

Rename the mode.json.sample to model.json. This file is used to set the prompt format and ban tokens, the default
is ChatML format so it should work with most recent models. Set the llama_endpoint to point to your llama.cpp running
in server mode, if it's not on the same container/server as your SumBot service (see below!)

## Downloading an LLM model

We highly recommend OpenHermes 2.5 Mistral-7b fine tune for this task, as it's currently the best (Nov 2023) that
we've tested personally. You can find different quantized versions of the model here:

https://huggingface.co/TheBloke/OpenHermes-2.5-Mistral-7B-GGUF/tree/main

I'd suggest the Q6 quant for GPU and Q4_K_M for CPU

## Running a model on llama.cpp in API mode

### Windows

Go to the llama.cpp releases and download either the win-avx2 package for CPU or the cublas for nvidia cards:

https://github.com/ggerganov/llama.cpp/releases

Extract the files out and run the following for nvidia GPUs:
```
server.exe -m .gguf -t 4 -c 2048 -ngl 33 --host 0.0.0.0 --port 8086
```

For CPU only:
```
server.exe -m .gguf -c 2048 --host 0.0.0.0 --port 8086
```

Replace with whatever model you downloaded and put into the llama.cpp directory

### Linux, MacOS or WSL2

Follow the install instructions for llama.cpp at https://github.com/ggerganov/llama.cpp

Git clone, compile and run the following for GPU:
```
./server -m models/.gguf -t 4 -c 2048 -ngl 33 --host 0.0.0.0 --port 8086
```

For CPU only:
```
./server -m models/.gguf -c 2048 --host 0.0.0.0 --port 8086
```

Replace with whatever model you downloaded and put into the llama.cpp/models directory

## Running Sumbot API

```
uvicorn main:app --host 0.0.0.0 --port 3002 --reload
```

## Accessing API

http://localhost:3002/docs

## Example output

The Godfather is a 1972 crime drama film directed by Francis Ford Coppola and starring Marlon Brando, Al Pacino, James Caan, and Richard S. Castellano. The movie has an IMDb rating of 9.2 out of 10 based on 1,038,358 votes and a Metacritic score of 100. It was released on March 24, 1972, and won 3 Oscars with a total of 33 wins and 19 nominations. The film's plot revolves around the aging patriarch of an organized crime dynasty transferring control of his empire to his reluctant son, leading to a series of unfortunate events and conflicts within the family. The movie is available in English, Italian, and Latin languages with subtitles. It was produced by Paramount Pictures and has a runtime of 175 minutes. The film's writers include Mario Puzo (screenplay), Francis Ford Coppola (screenplay), and Mario Puzo (novel).

## Example Input

Set entity to "movies"

Use the following JSON:

```
{
"_id": {
"$oid": "573a1396f29313caabce4a9a"
},
"fullplot": "When the aging head of a famous crime family decides to transfer his position to one of his subalterns, a series of unfortunate events start happening to the family, and a war begins between all the well-known families leading to insolence, deportation, murder and revenge, and ends with the favorable successor being finally chosen.",
"imdb": {
"rating": 9.2,
"votes": 1038358,
"id": 68646
},
"year": 1972,
"plot": "The aging patriarch of an organized crime dynasty transfers control of his clandestine empire to his reluctant son.",
"genres": [
"Crime",
"Drama"
],
"rated": "R",
"metacritic": 100,
"title": "The Godfather",
"lastupdated": "2015-09-02 00:08:23.680000000",
"languages": [
"English",
"Italian",
"Latin"
],
"writers": [
"Mario Puzo (screenplay)",
"Francis Ford Coppola (screenplay)",
"Mario Puzo (novel)"
],
"type": "movie",
"tomatoes": {
"website": "http://www.thegodfather.com",
"viewer": {
"rating": 4.4,
"numReviews": 725773,
"meter": 98
},
"dvd": {
"$date": "2001-10-09T00:00:00.000Z"
},
"critic": {
"rating": 9.2,
"numReviews": 84,
"meter": 99
},
"lastUpdated": {
"$date": "2015-09-12T17:15:13.000Z"
},
"consensus": "One of Hollywood's greatest critical and commercial successes, The Godfather gets everything right; not only did the movie transcend expectations, it established new benchmarks for American cinema.",
"rotten": 1,
"production": "Paramount Pictures",
"fresh": 83
},
"poster": "https://m.media-amazon.com/images/M/MV5BM2MyNjYxNmUtYTAwNi00MTYxLWJmNWYtYzZlODY3ZTk3OTFlXkEyXkFqcGdeQXVyNzkwMjQ5NzM@._V1_SY1000_SX677_AL_.jpg",
"num_mflix_comments": 131,
"released": {
"$date": "1972-03-24T00:00:00.000Z"
},
"awards": {
"wins": 33,
"nominations": 19,
"text": "Won 3 Oscars. Another 30 wins & 19 nominations."
},
"countries": [
"USA"
],
"cast": [
"Marlon Brando",
"Al Pacino",
"James Caan",
"Richard S. Castellano"
],
"directors": [
"Francis Ford Coppola"
],
"runtime": 175
}
```