instructor-embedding
vlite
instructor-embedding | vlite | |
---|---|---|
4 | 7 | |
1,703 | 690 | |
3.1% | - | |
5.9 | 9.3 | |
10 days ago | 5 days ago | |
Python | Python | |
Apache License 2.0 | GNU Affero General Public License v3.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
instructor-embedding
-
My experience on starting with fine tuning LLMs with custom data
If you li embeddings and vector DB, you should look into this: https://github.com/HKUNLP/instructor-embedding
-
Build Personal ChatGPT Using Your Data
If you look at a embeddings leaderboard [1], one of the top competitors called InstructorXL [2] is just a pip install away. It's neck and neck with Ada v2 except for a shorter input length and half the dimensions, with the added benefit that you'll always have the model available.
Most of the other options just work with the transformers library.
[1] https://huggingface.co/spaces/mteb/leaderboard
[2] https://github.com/HKUNLP/instructor-embedding
-
I've made a customisable SMS personal assistant which has infinite and persistent semantic memory.
Use instructor-embedding to to make it 100% local and even maybe quick relationship lookup (embed relationship info with sentiment analysis instruction)
-
Whisper Transcription Formatting
First.I believe having srt subtitles as whisper result would be better.Essentially you don't need just a list of words like YouTube does.You need something more structured.I don't remember what whisper outputs so I might be wrong.There is whisperx for that as example. And then maybe use gpt index over it.Or something like instructor model That can work.
vlite
- Fastest vector database made in NumPy
- FLaNK Stack Weekly for 27 November 2023
- Vlite – simple vector database written in less than 200 lines of code
-
Build Personal ChatGPT Using Your Data
I am working on a simple vector db just with numpy: https://github.com/sdan/vlite
I think milvus, quickwit, and pinecone are geared more towards enterprise and are hard to use.
- Vlite: Simple Open-source project for vector embeddings
- Vlite: Simple Vector Database with NumPy
- VLite: Fast vector db written in NumPy
What are some alternatives?
h2ogpt - Private chat with local GPT with document, images, video, etc. 100% private, Apache 2.0. Supports oLLaMa, Mixtral, llama.cpp, and more. Demo: https://gpt.h2o.ai/ https://codellama.h2o.ai/
gpt-2 - Code for the paper "Language Models are Unsupervised Multitask Learners"
openai-cookbook - Examples and guides for using the OpenAI API
PdfGptIndexer - RAG based tool for indexing and searching PDF text data using OpenAI API and FAISS (Facebook AI Similarity Search) index, designed for rapid information retrieval and superior search accuracy.
Nuggt - An Autonomous LLM Agent that runs on Wizcoder-15B
OpenLLM - Run any open-source LLMs, such as Llama 2, Mistral, as OpenAI compatible API endpoint in the cloud.
easydiffusion - Easiest 1-click way to create beautiful artwork on your PC using AI, with no tech knowledge. Provides a browser UI for generating images from text prompts and images. Just enter your text prompt, and see the generated image.
lit-gpt - Hackable implementation of state-of-the-art open-source LLMs based on nanoGPT. Supports flash attention, 4-bit and 8-bit quantization, LoRA and LLaMA-Adapter fine-tuning, pre-training. Apache 2.0-licensed. [Moved to: https://github.com/Lightning-AI/litgpt]
FLaNK-ContinuousSQL
haystack - :mag: LLM orchestration framework to build customizable, production-ready LLM applications. Connect components (models, vector DBs, file converters) to pipelines or agents that can interact with your data. With advanced retrieval methods, it's best suited for building RAG, question answering, semantic search or conversational agent chatbots.
private-gpt - Interact with your documents using the power of GPT, 100% privately, no data leaks