outlines VS TypeChat

> [2]: https://github.com/outlines-dev/outlines?tab=readme-ov-file#...

It's interesting as speech recognition has become more popular than ever through services like Alexa, and other iot devices support for OS speech recognition

Unfortunately most implementations (especially those that are iot focused) don't have very important features for robust speech recognition.

1. Ability to enable and disable a grammar

[4] https://github.com/outlines-dev/outlines/tree/main

The next step is defining how to guide generation. For this step, we'll use the Outlines library. Outlines is a library for controlling how tokens are generated. It applies logic to enforce schemas, regular expressions and/or specific output formats such as JSON.

No benchmarks, just my anecdotal experience trying to get local LLM's to respond with JSON. The method above works for my use case nearly 100% of the time. Other things I've tried (e.g. `outlines`[0]) are really slow or don't work at all. Would love to hear what others have tried!

0 - https://github.com/outlines-dev/outlines

As I was playing with the Outlines library (https://outlines-dev.github.io/outlines/), I discussed with my friend Maxime how funny it would be if we set up a way to pair LLMs in chess matches till one wins. The first time I tried it, it required substantial prompt engineering to get some of those LLMs to propose valid moves. Large language models can mostly stay focused and even play rather well; see https://news.ycombinator.com/item?id=37616170 for example. However small language models aren't as easy to convince.

Some of those LLMs have seen very little chess notation and so after the first few opening moves there aren't any valid tactics, let alone strategy, so they would end up either repeating the same move, or hallucinate moves that are not valid (Kxe5, but there would be a queen on e5!)

Then Outlines came along and we could force them to pick valid moves with little cost! Maxime worked super fast and got a first version of this idea as a gradio space.

I think it is pretty fun to see the (mostly terrible, but otherwise valid) chess that those LLMs play. Maybe it will even be instructive to how we can create small LLMs that can play much better than the ones on the leaderboard.

Anyway, you can check it out here:

https://huggingface.co/spaces/mlabonne/chessllm

What is interactive about it: you can pick the LLMs from available models on HuggingFace (within reason, small LLMs are preferable so that the space does not crash) or push one of your own small models to HF and have it fight with others. At the end of the game the leaderboard is updated.

Hope you find it fun!

> 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

And I think “just” Markdown doesn’t quite cut it for safe guidance. For example: directly generating content for your components. But I’m really excited about tooling like outlines appearing, with a greater focus on guided generation for structured data. Because this is often what we actually need!

It's pretty easy to force a locally running model to always output valid JSON: when it gives you probabilities for the next tokens, discard all tokens that would result in invalid JSON at that point (basically reverse parsing), and then apply the usual techniques to pick the completion only from the remaining tokens. You can even validate against a JSON schema that way, so long as it is simple enough.

There are a bunch of libraries for this already, e.g.: https://github.com/outlines-dev/outlines

https://github.com/outlines-dev/outlines/blob/7fae436345e621... squares with my experience using LLMs for anything real

  sequence = generator("Alice had 4 apples and Bob ate 2. Write an expression for Alice's apples:")

Not sure it's related to function calling. GPT4 can do function calling without using the specific function-calling API just by injecting the schema you want into the prompt with directions and asking it to return JSON. It works like >99% of the time. Same with 3.5-turbo.

The problem is these libraries convert pydantic models into json schemas and inject them into the prompt, which uses up like 80% more tokens than just describing the schema using typescript type syntax for example. See https://microsoft.github.io/TypeChat/, where they prompt using typescript type descriptions to get json data from LLMs. It's similar to what we built but with more boilerplate.

Semantic Memory (renamed to Kernel Memory - https://github.com/microsoft/kernel-memory) complements SK. Guidance's features are being absorbed into SK, following the departure of that team from Microsoft. Additionally, we have TypeChat (https://github.com/microsoft/TypeChat), which aims to ensure type-safe responses from LLMs. Most features of Autogen are also being integrated into SK, along with Assistants. SK serves as the orchestration engine powering Microsoft Copilots.

I tried TypeChat for my use case and ended up defining functions as typescript data types. This approach sounds much better, and leverages the newer OpenAI function calling, which should be more reliable I would think. Thanks for creating+sharing.

https://microsoft.github.io/TypeChat/

That re-prompting error on is what this new Microsoft library does, too: https://github.com/microsoft/TypeChat

Here's their prompt for that: https://github.com/microsoft/TypeChat/blob/c45460f4030938da3...

I think the approach using grammars (seen here, but also in things like https://github.com/ggerganov/llama.cpp/pull/1773 ) is a much more elegant solution.

I'm very surprised that they're not using `guidance` [0] here.

It not only would allow them to suggest that required fields be completed (avoiding the need for validation [1]) and probably save them GPU time in the end.

There must be a reason and I'm dying to know what it is! :)

[0] https://github.com/microsoft/guidance

[1] https://github.com/microsoft/TypeChat/blob/main/src/typechat...