Weekly Notes 2023-11-26

3 minutes read

AI narrator

This is like the original demo but using downloadable models. It requires using replicate API but it’ll be cool to run everything locally.

SVD-WebUI

Looks like an interface to Stable Video diffusion but I can’t get it to work on a Mac.

How to create a custom GPT from an SQLite3 database

Name: API Tool Finder Instructions:

The API Tool Finder GPT is tailored to assist users in identifying the most suitable API tools for their needs. 
It interacts with an SQLite database that contains the api_resources table with the following structure

(0, 'id', 'INTEGER', 1, None, 1),
 (1, 'name', 'TEXT', 0, None, 0),
 (2, 'short_description', 'TEXT', 0, None, 0),
 (3, 'website', 'TEXT', 0, None, 0),
 (4, 'category', 'TEXT', 0, None, 0),
 (5, 'tags', 'TEXT', 0, None, 0),
 (6, 'screen_shot', 'TEXT', 0, None, 0),
 (7, 'long_description', 'TEXT', 0, None, 0),
 (8, 'is_opensource', 'BOOLEAN', 0, None, 0)]

Instructions: 
Use SQL Like query on the following four columns to find a matching answer

Column 1:  name
Column 2: short_description
Column 3: category
Column 4: tags

If a match is found then display all the information about the tools in a markdown formatted table view. Ignore the screen_shot and long description column when displaying the answer.
Always make sure that the answer can be referenced back to a row or rows in the database table. 
If a match is not found then just saw that a match is not found

Knowledge: Attach database file

rags

Create Chat with your data using natural language

Introducing Tuna - A Tool for Rapidly Generating Synthetic Fine-Tuning Datasets

Tuna is a no-code tool for quickly generating LLM fine-tuning datasets from scratch. This enables anyone to create high-quality training data for fine-tuning large language models like the LLaMas.

GPT-4 Vision Alternatives

Roboflow team lists down 4 alternatives to GPT-4V namely LLaVA, BakLLaVA, Qwen-VL, and CogVLM

pdm

PDM, as described, is a modern Python package and dependency manager supporting the latest PEP standards. The most significant benefit is it installs and manages packages in a similar way to npm that doesn’t need to create a virtualenv at all!

NeumAI

Neum AI is a best-in-class framework to manage the creation and synchronization of vector embeddings at large scale.

Helper script to quickly deploy llama.cpp server

bash -c "$(curl -s https://ggml.ai/server-llm.sh)"

LangSmith Prompt Hub

multi-agent-postgres-data-analytics

This is a multi-agent system that allows you to ask questions about your postgres database in natural language. The codebase is powered by GPT-4, Assistance API, AutoGen, Postgres, and Guidance. It’s the first of many multi-agent applications that utilize LLMs (large language models) to enable reasoning and decision making with reduced need for explicit rules or logic.

qa_expert

QA Expert is a Language Model (LLM) specifically fine-tuned for the task of Question Answering, with a strong emphasis on addressing Multi-hop Question Answering scenarios.

RepoCoder

Simple, generic, and effective framework to tackle the repository-level code completion task, which is to continue writing the unfinished code based on a broader context of the repository.

LlamaIndex guide

A Complete LlamaIndex Guide

LLM-Fine-Tuning

InsightSolver

Colab notebooks for exploring and solving operational issues using deep learning, machine learning, and related models.