gemma-4-31B-it-scotoma-2

gemma-4-31B-it-scotoma-2

Text generation model of 31.3B parameters from ReadyArt, built on Google's gemma-4-31b-it and stored as Safetensors. Its page gives no statement of purpose, so it suits people who want to inspect a Gemma 4 variant.

Made by ReadyArt on Hugging Face. These notes are Mindket's, written from its model page and model card; the model's own files and license are on Hugging Face.

About

What it is: gemma-4-31B-it-scotoma-2 is a text generation model from ReadyArt with 31.3 billion parameters. It is based on Google's gemma-4-31b-it and stored in the Safetensors format. The readable part of the page is only page styling, so it does not say what the model was tuned for.

Who it suits: People who work with large open models and are happy to test an unexplained variant on their own prompts before relying on it.

Industry | Who uses it | How they use it | Example
Research | Model evaluators | Compare a Gemma 4 variant with its base model | The same prompts run on both models
Software | Developers | Try a 31B text model on their own hardware | A chat session with a local copy
Home labs | Hobbyists | Test how a community variant behaves | A set of favorite prompts

Skip it if you need documented behavior, because the page gives no description, examples or results.

Where to use it:

• Python with the transformers library, which the page lists
• Your own computer, on Windows, macOS or Linux, for the download
• A machine with large GPU memory, since the weights are about 62.5 GB in BF16

Use case | Who | What they get back
Chat with a 31B model | Developers | Text replies from a Gemma 4 based model
Base comparison | Evaluators | Outputs to set beside the original gemma-4-31b-it

How to set it up: You need Python with pip, the transformers library and a lot of disk space. The BF16 weights are about 62.5 GB, and a 4-bit copy would be about 15.6 GB by the record's estimate.

• Python and pip
• transformers
• About 62.5 GB of disk for BF16

1. Install the Hugging Face hub tool.
2. Download the repository's files.
3. Load them with transformers on a machine with enough memory.

pip install -U huggingface_hub
hf download ReadyArt/gemma-4-31B-it-scotoma-2

How to use it: The page gives no run steps that can be read. Load the downloaded files in transformers as you would a Gemma chat model, and check the model page for its own guidance.

hf download ReadyArt/gemma-4-31B-it-scotoma-2

Tool | What it does
hf download | Copies this repository's files to your computer
transformers | Loads the weights for text generation

Example: You download the files, load the model and send it a short question. Because the page gives no sample, judge the reply against what the base Gemma model would say.

For newcomers and experts: If you are new to this, know that a 31B model needs a strong GPU. Start with a smaller model first, and come back to this one when you can test it properly.

If you know the field, this is a derivative of a 31B instruction-tuned Gemma 4 model. The record shows 21 quantized and 9 fine-tuned versions made by others. The tuning data, method and goals are not stated in the readable text.

Start from gemma-4-31b-it
Apply the ReadyArt tuning
Publish Safetensors weights
Load and test on your prompts

Good to know: The page does not explain the model, so test it on your own prompts before you rely on it.

Title: Approximate weight size in GB
BF16: 62.5 GB
4-bit estimate: 15.6 GB

• License: Apache 2.0, as recorded
• Cost to run: none named
• Data it reaches: the files you download from Hugging Face, plus the text you give your local runner
• Not for: use without testing, since its purpose is not described

On Hugging Face, 13 November 2026

• Task: text generation
• Library: transformers
• Parameters: 31.3B
• Weights: about 62.5 GB in BF16; about 15.6 GB at 4-bit (estimate)
• Formats: Safetensors
• Downloads in the last 30 days: 1,643
• Likes: 60
• License: apache-2.0
• Based on: google/gemma-4-31b-it
• Quantized versions: 21
• Fine-tunes: 9
• Tags: transformers, safetensors, gemma4, image-text-to-text, text-generation, conversational

Links

• Model: https://huggingface.co/ReadyArt/gemma-4-31B-it-scotoma-2
• Files: https://huggingface.co/ReadyArt/gemma-4-31B-it-scotoma-2/tree/main

Get the model

Download its files with Hugging Face's command line:

pip install -U huggingface_hub
hf download ReadyArt/gemma-4-31B-it-scotoma-2

Then follow the usage in its model card. Nothing in this kit runs by itself.