distilgpt2
DistilGPT2 (short for Distilled-GPT2) is an English-language model pre-trained with the supervision of the smallest version of Generative Pre-trained Transformer 2 (GPT-2). Like GPT-2, DistilGPT2 can be used to generate text. Users of this model card should also consider information about the design, training, and limitations of GPT-2.
Params
90 M
Context
—
Downloads 30d
1.5 M
Likes
667
Download history
daily snapshots · 106 days
▲ 464 K in the last 30 days (23.9%)
1.9 M1.5 M
Oct 12Oct 22Nov 1Nov 10
2.5 M1.5 M
Aug 13Sep 12Oct 12Nov 10
2.7 M1.5 M
Jul 28Sep 1Oct 6Nov 10
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 | 0.4 GB | 0.9 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 -s https://aimodelscomparison.com/api/v1/models/distilgpt2
{
"hf_id": "distilbert/distilgpt2",
"params_b": 0.09,
"context_length": null,
"license": { "id": "apache-2.0", "commercial": "yes" },
"downloads_30d": 1481788,
"vram_estimates": [
{ "quant": "f32", "gb": 0.9 }
],
"updated_at": "2026-07-28T17:07:36Z"
}
Specifications
- Architecture
- GPT2LMHeadModel
- Parameters
- 90 M
- Tensor type
- F32
- Vocabulary
- 50,257
- Licence
- apache-2.0
- First seen on the Hub
- 2022-03-02
- Training datasets
- openwebtext
- WikiText-103 (reported)
- 21.1
- Added to our catalog
- 2026-07-28
Compare with any text-generation model