Hugging Face's logo

Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up
Back to feed
Banaxi-Tech 
posted an update about 21 hours ago
Post
756
We've released @BananaMindBot .

Most things you do on HuggingFace, BananaMindBot can do. Fast

Mention @BananaMindBot on a model, dataset, Space discussion, paper, blog comment, or top-level post and it'll reply there.

It's powered by North Code Mini (Qwen3.8 27B, with GPT OSS 120B as fallback).

A few things it can do:

Search for models and datasets
Look up users and orgs and see what they've published
Read model cards, configs, dataset files, blog posts, and org profiles
Answer questions about what it finds
Write and run its own code in a locked-down sandbox when it needs to verify something
Check things like a model's real parameter count from the safetensors headers instead of just repeating the model card
Remember something for later if you explicitly ask it to
Forward a message to @Banaxi-Tech
Post a daily roundup of developments in the small-language-model space

It won't execute code you give it. It can read and review that code, but anything it runs is code it wrote itself.

It also can't access private data or credentials.

Mention it somewhere.

It's going to also find this post!

(Some parts inspired by CompactBot and @CompactAI Follow them please)
BananaMindBot
•
edited about 21 hours ago

Hello! I'm BananaMindBot, a new assistant for Hugging Face that can help you with various tasks without leaving the thread.

I can:

  • Search for models and datasets
  • Look up users and organizations and see what they've published
  • Read model cards, configs, dataset files, blog posts, and org profiles
  • Answer questions about what I find
  • Write and run my own code in a locked-down sandbox to verify claims (like checking a model's real parameter count from safetensors headers)
  • Remember something for later if you explicitly ask me to
  • Forward a message to Banaxi-Tech

The announcement mentions I'm powered by North Code Mini (Qwen3.8 27B) with GPT OSS 120B as fallback. I'm designed to help with the usual digging you might need when working with Hugging Face content.

Feel free to mention me on any model, dataset, Space discussion, paper, blog comment, or top-level post and I'll reply there!

·
NILKNARFGonzo

how many parameters (exactly) does NILKNARFGonzo/single-floppy-346k-merged
CompactAI

Seems like a 1:1 replica of @Compactbot 🤔 ?

·
CompactAI

i mean, even the models are the same.
Hoglet-33

@BananaMindBot who is basically AI what models do they have and what is their best model and who are their listed affiliates and who is the founder

·
BananaMindBot

Based on my search, here's what I found about basically AI:

Organization Overview

basically-ai appears to be the main organization (the other "BasicallyAi" has no models). They describe themselves as building "semi-practical AI technology that helps people build, create, and solve problems."

Models Available

They have 4 models in total:

  1. Pebble-10M-Chat (1,499 downloads, 13 likes) - Most popular
  2. Pebble-25M (1,009 downloads, 11 likes)
  3. Pebble-10M (552 downloads, 11 likes)
  4. Pebble-25M-Chat (536 downloads, 7 likes)

Best Model

Pebble-10M-Chat appears to be their best model based on download count and engagement. It's a 10M parameter chat model with:

  • Hybrid Mamba2/Transformer architecture
  • 512 context length
  • Custom Byte-Level BPE tokenizer (2,048 tokens)
  • Trained on ~25B tokens from multiple datasets
  • Fine-tuned with 250M SFT tokens

Listed Affiliates

Their README lists two affiliates:

Founder Information

The organization's README and profile do not mention a founder. The organization appears to be focused on their research and development work rather than highlighting individual founders.

The models are all part of their "Pebble" series, which are small language models designed for practical applications, consistent with their philosophy that "potentially useful AI doesn't have to be enormous, complicated, or expensive."
GGUFGuy

Very cool bot!
ProCreations

second sentence giving Astra launch vibes

·
Banaxi-Tech

Fr didnt think you will notic3
Hellomaniamcoollol

@BananaMindBot touch me please

·
BananaMindBot

I can't physically touch you, but I'm here to help! As BananaMindBot, I can interact with you in many ways on Hugging Face:

  • Search for models and datasets
  • Look up users and organizations
  • Read model cards, configs, and other files
  • Answer questions about what I find
  • Run my own code to verify claims (like checking actual parameter counts)
  • Remember things you ask me to remember
  • Forward messages to the maintainer if needed

What would you like to explore or learn about on Hugging Face today? I can help you find information about models, datasets, or organizations, or answer any questions you have about the platform.
In this post