Am I the only one who is really impressed by Granite4 from IBM?

It is small but still really good

11 points · 8 comments · view on lemmy.world

8 Comments

Smokeydope@lemmy.world · 5 pts · 327d (3 replies)

Havent heard of it till this post, what about it impressed you over something like llama, mistral, qwen?

for anyone who wants more info its a 7b Mixture of Experts model released under apache 2.0!

" Granite-4-Tiny-Preview is a 7B parameter fine-grained hybrid mixture-of-experts (MoE) instruct model fine-tuned from Granite-4.0-Tiny-Base-Preview using a combination of open source instruction datasets with permissive license and internally collected synthetic datasets tailored for solving long context problems. This model is developed using a diverse set of techniques with a structured chat format, including supervised fine-tuning, and model alignment using reinforcement learning."

Supported Languages: English, German, Spanish, French, Japanese, Portuguese, Arabic, Czech, Italian, Korean, Dutch, and Chinese. However, users may fine-tune this Granite model for languages beyond these 12 languages.

Intended Use: This model is designed to handle general instruction-following tasks and can be integrated into AI assistants across various domains, including business applications.

Capabilities

Thinking
Summarization
Text classification
Text extraction
Question-answering
Retrieval Augmented Generation (RAG)
Code related tasks
Function-calling tasks
Multilingual dialog use cases
Long-context tasks including long document/meeting summarization, long document QA, etc.

https://huggingface.co/ibm-granite/granite-4.0-tiny-preview

Xylight@lemdro.id · 3 pts · 327d (2 replies)

there's also a "small" and "micro" variant, which are 32b a6b MoE and 3b dense models respectively

basxto@discuss.tchncs.de · 1 pts · 326d (1 reply)

granite4:micro-h should be able to run on machines with 4GB RAM

Xylight@lemdro.id · 1 pts · 326d

You can run Qwen3 4b thinking at q4 quantization at 2.5GB, which is probably a better model too

Rhaxapopouetl@ttrpg.network · 4 pts · 328d (1 reply)

Do you mean IBM Granite 4?

possiblylinux127@lemmy.zip · 1 pts · 328d

Good catch

d0nkey@lemmy.zip · 2 pts · 316d

I have used the micro variant primarily with perplexica and I must say it is really good for summation and for answering further questions, especially when it comes to these tasks in my testing it has outclassed instruct models that are 2-3 times its size.

afk_strats@lemmy.world · 1 pts · 316d

You are not alone. It blew my mind at how good it is per billion parameters. As an example, I can't think of another model that will give you working code at 4B or less. I havent tried it on agentic tasks but that would be interesting