guidance
AGiXT
Our great sponsors
guidance | AGiXT | |
---|---|---|
23 | 26 | |
17,246 | 2,437 | |
5.1% | - | |
9.8 | 9.9 | |
4 days ago | 8 days ago | |
Jupyter Notebook | Python | |
MIT License | MIT License |
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.
guidance
-
Anthropic's Haiku Beats GPT-4 Turbo in Tool Use
[1]: https://github.com/guidance-ai/guidance/tree/main
-
Show HN: Prompts as (WASM) Programs
> The most obvious usage of this is forcing a model to output valid JSON
Isn't this something that Outlines [0], Guidance [1] and others [2] already solve much more elegantly?
0. https://github.com/outlines-dev/outlines
1. https://github.com/guidance-ai/guidance
2. https://github.com/sgl-project/sglang
- Show HN: Fructose, LLM calls as strongly typed functions
-
LiteLlama-460M-1T has 460M parameters trained with 1T tokens
Or combine it with something like llama.cpp's grammer or microsoft's guidance-ai[0] (which I prefer) which would allow adding some react-style prompting and external tools. As others have mentioned, instruct tuning would help too.
[0] https://github.com/guidance-ai/guidance
- Forcing AI to Follow a Specific Answer Pattern Using GBNF Grammar
-
Prompting LLMs to constrain output
have been experimenting with guidance and lmql. a bit too early to give any well formed opinions but really do like the idea of constraining llm output.
- Guidance is back π₯³
- New: LangChain templates β fastest way to build a production-ready LLM app
-
Is supervised learning dead for computer vision?
Thanks for your comment.
I did not know about "Betteridge's law of headlines", quite interesting. Thanks for sharing :)
You raise some interesting points.
1) Safety: It is true that LVMs and LLMs have unknown biases and could potentially create unsafe content. However, this is not necessarily unique to them, for example, Google had the same problem with their supervised learning model https://www.theverge.com/2018/1/12/16882408/google-racist-go.... It all depends on the original data. I believe we need systems on top of our models to ensure safety. It is also possible to restrict the output domain of our models (https://github.com/guidance-ai/guidance). Instead of allowing our LVMs to output any words, we could restrict it to only being able to answer "red, green, blue..." when giving the color of a car.
2) Cost: You are right right now LVMs are quite expensive to run. As you said are a great way to go to market faster but they cannot run on low-cost hardware for the moment. However, they could help with training those smaller models. Indeed, with see in the NLP domain that a lot of smaller models are trained on data created with GPT models. You can still distill the knowledge of your LVMs into a custom smaller model that can run on embedded devices. The advantage is that you can use your LVMs to generate data when it is scarce and use it as a fallback when your smaller device is uncertain of the answer.
3) Labelling data: I don't think labeling data is necessarily cheap. First, you have to collect the data, depending on the frequency of your events could take months of monitoring if you want to build a large-scale dataset. Lastly, not all labeling is necessarily cheap. I worked at a semiconductor company and labeled data was scarce as it required expert knowledge and could only be done by experienced employees. Indeed not all labelling can be done externally.
However, both approaches are indeed complementary and I think systems that will work the best will rely on both.
Thanks again for the thought-provoking discussion. I hope this answer some of the concerns you raised
-
Show HN: Elelem β TypeScript LLMs with tracing, retries, and type safety
I've had a bit of trouble getting function calling to work with cases that aren't just extracting some data from the input. The format is correct but it was harder to get the correct data if it wasn't a simple extraction.
Hopefully OpenAI and others will offer something like https://github.com/guidance-ai/guidance at some point to guarantee overall output structure.
Failed validations will retry, but from what I've seen JSONSchema + generated JSON examples are decently reliable in practice for gpt-3.5-turbo and extremely reliable on gpt-4.
AGiXT
-
Conversational "memory loss"?
If you are more interested in AI assistants check out AGiXT. It has some really cool features but it is under heavy development. Not everything works jet and updates break sometimes already working functions. But it is still far better than babyAGI and other proof of concepts.
-
Microsoft Research proposes new framework, LongMem, allowing for unlimited context length along with reduced GPU memory usage and faster inference speed. Code will be open-sourced
That's exactly my goal right now too! I have been trying to figure out how to use AGiXT agents to read and write to an "Adventurer's Log" text file to try to mimic a long term memory but honestly I'm not good enough with any of this to get it working yet. The idea I've got rn is that there'd be a DM agent which takes your input and then there'd be "memory" agents which would check text files such as "Adventurer's Log" and "Character Interactions/Relationships" to keep a contiguous understanding of what each character has done, who they've met, what they've been told/haven't been told by certain characters about their motivations. I'm sure there's someone *much* more talented than me working on this already, at this point I've sort of given up on the idea and I'm just waiting for someone to come out with a Tavern style interface where I can paste in world details and character details and just get going!
- AGiXT: A local automation platform with memories and SmartGPT-like prompting. Works with Ooba/LCPP/GPT4All, and more
-
What are the best AI tools you've ACTUALLY used?
AGiXT: A Python package for AGI research.
-
?Best LLM service for a tiny home server
Even if my (for example, privateGPT) LLM is glacially slow I'd still love to be able to say "Mr Holmes, have Mrs Doubtfire verb the data object in order to verb a product for me, please." (eg: analyse the wikipedia article on the peace of westfalia in order to ELI5 a short summary of it). Hopefully she'd crunch away at the data, and at my convenience, I could have her brief me on her conclusions. I'm sure folks here would do something more clever using AGiXT, or having the old girl prepare lesson-plans for Mycroft to deliver (I just think that sort of thing is world-changing-bonkers for anyone wanting to learn anything, perhaps for kids one day), but I'd have to work up to that.
-
LlamaCPP and LangChain Agent Quality
Keep an eye on this project as well. https://github.com/Josh-XT/AGiXT
- Using the right prompt format makes responses so much better
-
How big of a jump is 13B Vicuna Uncensored vs 30B Vicuna Uncensored?
File upload and automatic agents. It exists it is just buggy. They are working at an insane pace building it. It is practically broke 90% of the time. Maybe it's working better right now. I had success with v1.1.31 as well. https://github.com/Josh-xt/AGiXT
-
Langchain, Langchain.js, vs AutoGPT for local agent development
Maybe you want to check out josh-xt/AGiXT it has its roots in langchain so you can see what the prompts look like and the code. They have made a lot of tools as well although you are going to have issues getting it to work. The newest version kinda works and version 1.1.31 I had the fast API backend working. Maybe you can help them out. They need more people to show them bugs. https://github.com/Josh-XT/AGiXT
-
Is there an alternative to AgentGPT that I can run on my CPU with 32 GB of RAM?
https://github.com/Josh-XT/AGiXT I have tested this one and it is pretty much the same as AgentGPT, supports many providers + many local models (you can even make it work with oobabooga api which is pretty easy), donβt wait for insane results, the problem right now is context length with the local models, probably going to be an old issue in a few weeks we hope ;)
What are some alternatives?
lmql - A language for constraint-guided and efficient LLM programming.
AgentOoba - An autonomous AI agent extension for Oobabooga's web ui
semantic-kernel - Integrate cutting-edge LLM technology quickly and easily into your apps
private-gpt - Interact with your documents using the power of GPT, 100% privately, no data leaks
langchain - π¦π Build context-aware reasoning applications
AgentGPT - π€ Assemble, configure, and deploy autonomous AI Agents in your browser.
NeMo-Guardrails - NeMo Guardrails is an open-source toolkit for easily adding programmable guardrails to LLM-based conversational systems.
babyagi
text-generation-webui - A Gradio web UI for Large Language Models. Supports transformers, GPTQ, AWQ, EXL2, llama.cpp (GGUF), Llama models.
vault-ai - OP Vault ChatGPT: Give ChatGPT long-term memory using the OP Stack (OpenAI + Pinecone Vector Database). Upload your own custom knowledge base files (PDF, txt, epub, etc) using a simple React frontend.
outlines - Structured Text Generation
guidance - A guidance language for controlling large language models. [Moved to: https://github.com/guidance-ai/guidance]