A list of resources around generative content and machine learning - Compiled by Phil Carlisle (Generative Machines/University of Lincoln) Note: inclusion in the list isn't any form of endorsement. Some of these resources are commercial and require money, I'll try and highlight Open Source where possible. Also note: I'll keep a list of academic papers in another document so that this is more about resources to play with.
- HuggingFace - Basically one of THE primary sources for new models, often with demonstrations
- FutureTools.io - Aggregator of new AI services and applications
- SadTalker - Generates moving head videos, with animated mouth shapes and some relatively unnatural head movement
- Runway Gen 2 - Runway have a number of video-based tools, like automatic removal etc.
- WonderShare - Spielberg's company doing generative characters from example videos (character replacement).
- Huggingface - Transformers library does a wide range of language tasks, with a good model hub that auto-downloads etc.
- [NVidia Nemo Toolkit] (https://www.example.com) - Language and voice libraries from NVIDIA, very good speech transcription, limited TTS, has NLP
- [Spacy] (https://www.example.com) - A "traditional" NLP library. Gives parts of speech info (i.e. is a word a verb, adverb, noun etc)
- [Langchain] (https://www.example.com) - Python library to develop large language model applications.
- [MindMeld] (https://www.example.com) - Older python library for NLP used for non LLM chatbot style interactions. Seems a bit abandoned now.
- [Rasa] (https://www.example.com) - Older python library for NLP used for non LLM chatbot style interactions. A bit too strong on the upsell, chatbots are old.
- [OpenAssistant] (https://www.example.com) - Recent LLM + application for chat interactions. Language model not so great yet, but usable.
- Dolly
- OpenAssistant
- Vicuna
- Koala
- Llama
- Alpaca
- Eleven Labs - Voice generation and cloning cloud service - One of the better generators for Text to speech (TTS) with Cloud API
- MotionFusion
- SimMDM