RHab's comments

On: Ask HN: What are you working on? (October 2024)

[RHab]

A llm backend fantasy game. It uses structured output and supports Openai, Anthropic and LM Studio. Gemini support is ending, at the moment it is not working reliable. https://github.com/HabermannR/Fantasy-Tribe-Game

[Aeolun → RHab]

That’s a fun idea. I may try this one.

On: Structured Output for LM Studio 0.3.5, OpenAI, Ant...

[RHab]

I am trying to understand structured output from LLMs, and as training I am building a text game.

OpenAI is using pydantic to create objects directly in its beta branch, it's quite nice. Anthropic is a bit more involved, as you need to involve tool calling.

LM Studio 0.3.5 introduced a bug for structured output when using pydantic and enums or literals, I created a workaround.

Also Gemini is not using Pydantic at all, instead TypedDict to create the json scheme.

And Gemini does not have a system message.

I am quite proud to have everything working, if you want to check out my code please take a look: https://github.com/HabermannR/Fantasy-Tribe-Game

Here is the backend: https://github.com/HabermannR/Fantasy-Tribe-Game/blob/main/L...

For example, this is how I call Gemini:

completion = model.generate_content(

messages[0]['content'] + ": " + messages[1]['content'],

generation_config=genai.GenerationConfig(

response_mime_type="application/json", response_schema=response_types.typed_dict

),

)

result = response_types.pydantic_model.model_validate_json(completion.text)

Happy for any feedback!

On: Show HN: Fantasy Tribe Game: LLM backed interactiv...

[RHab]

Hi, I am programming as a hobby since I am seven, and I always wanted to program a game, but never finished anything. I am super fascinated by LLMs, and made a small interactive text-based strategy game, where the player can choice a tribe and then gets presented 3 choices, which will shape the development of the tribe from humble beginnings to god like ascension. Everything is done procedurally by an LLM using structured output. I support OpenAI, Anthropic, and local LLM, tested with LM Studio. Please take a look and give me some feedback, thanks! https://github.com/HabermannR/Fantasy-Tribe-Game

On: AMD's Turin: 5th Gen EPYC Launched

[RHab]

Abaqus for example is by core, I am severly limited, for me this makes totally sense.

On: Ask HN: Who is hiring? (October 2024)

[RHab]

Würselen! The world is small :D

On: Replit's new Code LLM: Open Source, 77% smaller th...

[RHab]

No, I could only get 2.7B to run on 8GB VRam unfortunatly.

[amasad → RHab]

it is 2.7B

[tarruda → amasad]

Loading seems to have worked on my laptop's RTX 3070, `nvidia-smi` shows `5188MiB / 8192MiB` in memory usage.

On: FEA Solver that doesn't need defeaturing

[RHab]

No plastic material, only basic contact, no advanced connectors... This is my problem with all these fancy FEA codes, nice in theory, not fit for industrial models.

[franckl → RHab]

if you are interested in mechanical contact, we developed a symmetrical algorithm (no master/slave) and can make it available for you on https://tanatloc.com