OpenAI Plugin Store and Generative AI Projects
OpenAI Plugin Store:
- Adept.ai got copied as an in the OpenAI Plugin Store, called Multi-on
Experience needed:
- People with experience in CLIP, BLIP-2, and VQA are needed for a project
- Someone with solid experience in multimodal vector similarity search is also needed
Meeting invite:
- A meeting invite was shared in the WhatsApp group for the project
Projects:
- Multimodal vector similarity search
- Personal search
- Generative AI for solving problems and building multi-modal systems
- Domain-specific feedback loop and fine-tuning
- Context-based search with cohere embeddings and OpenAI GPT3.5 for non-English languages
- Text to video
- Applications of generative AI in finance
Resources:
- GitHub repository for the project: https://github.com/marqo-ai/marqo
- Deepset’s Haystack and Milvus for modern DL infra
- Character.ai for personal search
- OpenAI’s Point-e for converting image/text to 3D model
- Segment Anything Model for creating an object mask from an image
- Langdetect for language detection in text
- ControlNet models for image-to-image generation and control
Links:
- Meetup details: https://hasgeek.com/generativeAI/april-meetup/
- Wired article on giant AI models: https://www.wired.com/story/openai-ceo-sam-altman-the-age-of-giant-ai-models-is-already-over/
- GitHub repository for OpenBBTerminal: https://github.com/OpenBB-finance/OpenBBTerminal/releases/tag/v3.0.0rc2
- Gradio space for ControlNet: https://huggingface.co/spaces/hysts/ControlNet
Social:
- Instagram account for making videos via SD: https://www.instagram.com/reel/Cq5eWq4rRQC/?utm_source=ig_web_copy_link
Links
The description and link can be mismatched because of extraction errors.
- https://github.com/marqo-ai/marqo - a potential database to use for multimodal vector similarity search, which involves indexing image and text embeddings of useful information. The message also mentions the need for someone with solid experience in this area.
- https://yaofu.notion.site/How-does-GPT-Obtain-its-Ability-Tracing-Emergent-Abilities-of-Language-Models-to-their-Sources-b9a57ac0fcf74f30a1ab9e3e36fa1dc1: Found this in the eval PRs. Are there folks in this group interested in working on these problems?
- https://www.together.xyz/blog/redpajama (related to reducing unknown unknowns)
- https://t.co/sVSHuljpwe: A demo for a program that is US or US remote, with a message indicating that the person might DM (direct message) for more information.
- Personal search related: https://github.com/KnowledgeCanvas/knowledge - interested in working on PKM adjacent problems (personal search etc), mostly solving for myself…building on top of my Readwise data. Both Gh and hf are facing issues.
- Very good read: https://www.wired.com/story/openai-ceo-sam-altman-the-age-of-giant-ai-models-is-already-over/ Title is a bit misleading though. The point is that models will become better, but not by adding more base training data.
- https://hasgeek.com/generativeAI/april-meetup/ - I heard there’s a meet up this sat. Where can I find details about this and how do I register?
- https://github.com/Mimino666/langdetect: The message discusses the difficulty of detecting language in audio and references the open-source project langdetect on GitHub for text language detection.
- https://www.youtube.com/watch?v=30xueN12guw: The video demonstrates how specifying the language can make language detection and translation easier for a generic question. The message in the same link as the video is related to the topic.
- https://www.causal.app/ - A neat offering related to finance, but not specifically Generative AI.
- https://www.causal.app/ - A neat offering that is not GenerativeAI.
- https://github.com/OpenBB-finance/OpenBBTerminal/releases/tag/v3.0.0rc2 - Open Source Bloomberg GPT.
- https://www.instagram.com/reel/Cq5eWq4rRQC/?utm_source=ig_web_copy_link - This link leads to a video reel on Instagram that the sender thinks is very good.
- https://www.instagram.com/reel/Cq5eWq4rRQC/?utm_source=ig_web_copy_link - This link leads to a video made via SD and the message mentions that the guy’s last couple of reels are very good.
- https://huggingface.co/liuhaotian/LLaVA-13b-delta-v0: This link is mentioned in the context of a limited version of logojoy.com and is being appreciated for being open. The message also mentions Midjourney for logo creation.
- https://huggingface.co/spaces/hysts/ControlNet: A Gradio space for segmenting various attributes of UI elements, suggested to someone who mentioned having existing designs and asked for elaboration on segmenting UI elements. The message also discusses the potential of control net models and T2I adapters for achieving maximum control.
- Controlnet (https://github.com/lllyasviel/ControlNet) is being discussed as a useful tool for maximum control, with community momentum behind it.