rinna / text-generation updated 1 year ago

japanese-gpt-neox-small

Name: japanese-gpt-neox-small
Author: rinna

This repository provides a small-sized Japanese GPT-NeoX model. The model was trained using code based on EleutherAI/gpt-neox.

Params

200 M

Context

2,048

Downloads 30d

575 K

Likes

Commercial use: allowed mit Not gated SAFETENSORS 1 languages View on Hugging Face ↗

Download history

daily snapshots · 56 days

▲ 74 K in the last 30 days (14.7%)

584 K344 K

Jul 28Aug 15Sep 3Sep 21

Can you run it?

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

File	Quant	Size	Est. VRAM	Verdict on RTX 4090 · 24 GB
model.safetensors	f32	0.7 GB	1.3 GB	✅ Runs comfortably

Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Run it

copy-paste, exact tags checked against the Hub

~ · curl · api/v1

$ curl -s https://aimodelscomparison.com/api/v1/models/japanese-gpt-neox-small
{
  "hf_id": "rinna/japanese-gpt-neox-small",
  "params_b": 0.20,
  "context_length": 2048,
  "license": { "id": "mit", "commercial": "yes" },
  "downloads_30d": 575189,
  "vram_estimates": [
    { "quant": "f32", "gb": 1.3 }
  ],
  "updated_at": "2026-07-28T18:06:13Z"
}

est. VRAM —on RTX 4090 · 24 GBJSON API →

Specifications

Architecture: GPTNeoXForCausalLM
Parameters: 200 M
Tensor type: F32
Context length: 2,048
Vocabulary: 44,416
Layers / heads: 12 / 12
Licence: mit
First seen on the Hub: 2022-08-31
Training datasets: cc100, Wikipedia, mc4
Added to our catalog: 2026-07-28

Compare with any text-generation model

Popular comparisons

Qwen3-0.6B 23.7 M gpt2 15.1 M Qwen2.5-0.5B-Instruct 8.5 M opt-125m 7.0 M pythia-160m 3.5 M SmolLM2-135M 2.0 M