DeepSeek-OCR 2#
A document understanding model released by DeepSeek in 2026. Its core innovation is DeepEncoder V2, which uses cascaded causal reasoning to achieve 1D understanding of 2D images.
Core Innovation#
Visual Causal Flow#
Traditional vision encoders scan an image in a fixed order (top-left to bottom-right), while human reading follows semantically driven causal logic. DeepSeek-OCR 2 simulates this causal flow with DeepEncoder V2:
- Visual tokens: bidirectional attention (preserves a global receptive field)
- Causal flow queries: causal attention (learnable reordering tokens)
graph LR
A[Image] --> B[Vision Tokenizer]
B --> C[Visual Tokens]
C --> D[DeepEncoder V2]
D --> E[Causal Flow Queries]
E --> F[DeepSeek-MoE Decoder]
F --> G[Output]
Architecture#
| Component | Description |
|---|---|
| Vision Tokenizer | SAM-base + conv layers (80M params), 16x token compression |
| DeepEncoder V2 | Based on Qwen2-0.5B, replacing the original CLIP |
| Decoder | DeepSeek-MoE 3B (500M active) |
| Output token count | 256-1120 (on par with Gemini-1.5 Pro) |
Technical Details#
Attention Mask#
┌─────────────────────┬─────────────────────┐
│ Bidirectional attn │ Causal attn │
│ (visual tokens) │ (causal queries) │
│ all-ones matrix │ lower-triangular │
└─────────────────────┴─────────────────────┘
- Among visual tokens: fully visible in both directions (ViT-like)
- Causal flow queries: can only see preceding tokens (LLM-decoder-like)
- Each query can attend to all visual tokens plus preceding queries
Cascaded Causal Reasoning#
- Encoder: semantically reorders visual tokens via causal flow queries
- Decoder: performs autoregressive reasoning over the reordered sequence
This design bridges the gap between “2D spatial structure” and “1D causal language modeling”.
Performance#
- Significant improvement over DeepSeek-OCR on OmniDocBench v1.5
- Much stronger visual reading-order logic
- Retains high-ratio token compression efficiency