DeepSeek-OCR-2
🌟 Github 📥 Model Download 📄 Paper Link 📄 Arxiv Paper Link DeepSeek-OCR 2: Visual Causal Flow Explore more human-like visual encoding.
Params
3.4 B
Context
8,192
Downloads 30d
900 K
Likes
1,101
Download history
daily snapshots · 56 days
▲ 437 K in the last 30 days (32.7%)
1.3 M889 K
Aug 23Sep 2Sep 12Sep 21
2.5 M889 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 | bf16 | 6.8 GB | 8.5 GB | ✅ Runs comfortably |
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.
Specifications
- Architecture
- DeepseekOCR2ForCausalLM
- Parameters
- 3.4 B
- Tensor type
- BF16
- Context length
- 8,192
- Vocabulary
- 129,280
- Layers / heads
- 12 / 10
- Licence
- apache-2.0
- First seen on the Hub
- 2026-01-27
- Training datasets
- undisclosed
- Added to our catalog
- 2026-07-28
Compare with any image-text-to-text model