CatalogWantedMirrorsLedgerOperating StatementDocsStudio statusCard system
alt.freeweights.deepseek / DeepSeek-OCR-2
From: deepseek-ai <huggingface.co/deepseek-ai>Newsgroups: alt.freeweights.deepseekSubject:

DeepSeek-OCR-2

Date: 2026-11-14Commit: aaa02f3Signed: 2DC28CC7B26A283B ✓

Image text to text model with 3.4B parameters. Safetensors weights, 6.79 GB in all. Published by deepseek-ai under Apache-2.0.

3.4B params6.79GBDeepSeek-OCR-2safetensorspermissive
FactsParams3.4BSize6.79 GBFormatssafetensorsTaskimage text to textLicenseApache-2.0From sourcepermissive
$ fw get deepseek-ai/DeepSeek-OCR-2
# manifest sha256 5f6db77082ed… · 6.79 GB · 16 files
$ fw verify ./DeepSeek-OCR-2 deepseek-ai/DeepSeek-OCR-2
From sourceHosted (signed)deepseek-ai/DeepSeek-OCR-2Last check: 2026-11-14 21:14 UTC
Run at homeGet help

Get help: custom fine-tune or private deploy. We review your setup before we quote.

Works with

Run23Fine-tune29Convert15Evaluate13Scan13Apps62
OverviewModel cardFilesUse itLicenseProvenanceModel tree

Overview

Size
6.79 GB
Weights
6.78 GB (safetensors)
Parameters
3.4B
Formats
safetensors
Task
image text to text
License
Apache-2.0 (permissive)
Files
16, each with a sha256
Publisher
deepseek-ai
Added
2026-11-14
Section
alt.freeweights.deepseek
Lineage

Original release from the publisher.

Where these bytes come from
Publisher at the pinned commit→You

Sources are tried in order. A file counts only when its size and sha256 match the manifest. Sources are checked when you ask for them, not on a timer.

Sources
BadgeHostLast checkResult
From sourcehuggingface.co2026-11-14 21:14 UTCpassed

Model card

From the publisher at commit aaa02f3. Shown as plain text with links; images open at the publisher.

Image: DeepSeek AI

Homepage Hugging Face

Discord Twitter Follow

🌟 Github | 📥 Model Download | 📄 Paper Link | 📄 Arxiv Paper Link |

DeepSeek-OCR 2: Visual Causal Flow

Image: image

Explore more human-like visual encoding.

Usage#

Inference using Huggingface transformers on NVIDIA GPUs. Requirements tested on python 3.12.9 + CUDA11.8：

torch==2.6.0
transformers==4.46.3
tokenizers==0.20.3
einops
addict 
easydict
pip install flash-attn==2.7.3 --no-build-isolation
from transformers import AutoModel, AutoTokenizer
import torch
import os
os.environ["CUDA_VISIBLE_DEVICES"] = '0'
model_name = 'deepseek-ai/DeepSeek-OCR-2'

tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
model = AutoModel.from_pretrained(model_name, _attn_implementation='flash_attention_2', trust_remote_code=True, use_safetensors=True)
model = model.eval().cuda().to(torch.bfloat16)

# prompt = "<image>\nFree OCR. "
prompt = "<image>\n<|grounding|>Convert the document to markdown. "
image_file = 'your_image.jpg'
output_path = 'your/output/dir'


res = model.infer(tokenizer, prompt=prompt, image_file=image_file, output_path = output_path, base_size = 1024, image_size = 768, crop_mode=True, save_results = True)

vLLM#

Refer to 🌟GitHub for guidance on model inference acceleration and PDF processing, etc.

Support-Modes#

  • Dynamic resolution
    • Default: (0-6)×768×768 + 1×1024×1024 — (0-6)×144 + 256 visual tokens ✅

Main Prompts#

# document: <image>\n<|grounding|>Convert the document to markdown.
# without layouts: <image>\nFree OCR.

Acknowledgement#

We would like to thank DeepSeek-OCR, Vary, GOT-OCR2.0, MinerU, PaddleOCR for their valuable models and ideas.

We also appreciate the benchmark OmniDocBench.

Citation#

@article{wei2025deepseek,
  title={DeepSeek-OCR: Contexts Optical Compression},
  author={Wei, Haoran and Sun, Yaofeng and Li, Yukun},
  journal={arXiv preprint arXiv:2510.18234},
  year={2025}
}
@article{wei2026deepseek,
  title={DeepSeek-OCR 2: Visual Causal Flow},
  author={Wei, Haoran and Sun, Yaofeng and Li, Yukun},
  journal={arXiv preprint arXiv:2601.20552},
  year={2026}
}

Files

16 files · 6.79 GBWeights 6.78 GB (safetensors). Sizes and sha256 from the signed manifest.
Hide all 16 files with their sha256
FileSizesha256
.gitattributes1.57 KBd8c58cd65e84cd2e9639504e924a7b7e9319b5db04ef5eb6257da76c4769088c
LICENSE.txt11.3 KB9debfe045c2a2755a52d0ae1087605d15770fe6636798c09efd93c1dc732ef28
README.md4.14 KB507812c8e4620f72625305377daa697faf7135d7400ac25ab4b7d6299a92d60b
assets/fig1.png141 KBbfa3301b397df240c5eba7206c7865cae3d9aaf6674a6c8a68904a9c98bcbc77
config.json2.59 KBe2bb23b49b445f0bfd6eeb978bac9f19a2df82cb38472754e4ae83662d18f845
configuration_deepseek_v2.py10.6 KB6ab21f29a4722e26fa28c8e0d4277591689a598df17cf6c712330e8f62b3fc7c
conversation.py9.25 KBec7b6ce89bcda643de1f43269ffa66a7b2e65dc3ed30e427958f776546b4ba03
deepencoderv2.py36.3 KB562e4da383b7ba5ef5db79b5ddfb705c649a80044b99122fe4d5d42c16284193
model-00001-of-000001.safetensors6.78 GBd8ff67a424ba6f4dd077885eb9d6a05d2537e76fe5491f0e2a9b712f8c8870fa
model.safetensors.index.json247 KBf97aa45bbff64ccb2d645789303b1e2be17f45990812e5f0c2f80d169587aed7
modeling_deepseekocr2.py39.2 KB4166e8f014250143e257f48fb4eaa57a2ef49fa6ad375ecc016bc7d262fc7160
modeling_deepseekv2.py82.2 KB69184a0493d1fdac21c360f9f9eabc5b5af3188e0265df44cfaeafc44f9095e3
processor_config.json460 B0fe7ba9aa6b967a90e4af40d43c8030cbdd3dbcfbb387b8907625d5e54f0dbbe
special_tokens_map.json801 Bab4bd57ce17d62e39e0a39e739de1e407484f090f0b2c7e391312bca7a5b061a
tokenizer.json9.98 MBa02f8fd5228c90256bb4f6554c34a579d48f909e5beb232dc4afad870b55a8b4
tokenizer_config.json166 KBa0cbe8464049da1f891b7a12676de06af4cb54c130995d42f71adc1c30c6e9f3
VersionsEach time the pinned upstream commit or file list changed in this repo. Newest first.
Upstream commitDateContents
aaa02f3811 pinned now manifest updated, same commit2026-11-14 21:14 UTC16 files · 6.79 GB
aaa02f38112026-11-14 21:14 UTC16 files · 6.79 GB

Use it

$ fw get deepseek-ai/DeepSeek-OCR-2
# manifest sha256 5f6db77082ed… · 6.79 GB · 16 files
$ fw verify ./DeepSeek-OCR-2 deepseek-ai/DeepSeek-OCR-2
fw.pyChecks the signature and every sha256, tries each source in order, resumes.
$ python3 fw.py get https://freeweights.allternit.com/manifests/deepseek-ai/DeepSeek-OCR-2.json --dir ./DeepSeek-OCR-2 $ python3 fw.py verify ./DeepSeek-OCR-2 ./deepseek-ai__DeepSeek-OCR-2.json
hfThe Hugging Face CLI with one setting. Same repo id.
$ HF_ENDPOINT=https://freeweights.allternit.com hf download deepseek-ai/DeepSeek-OCR-2 --revision aaa02f3811945a91062062994c5c4a3f4c0af2b0 --local-dir ./DeepSeek-OCR-2
transformersLoad the weights in Python.
import os os.environ["HF_ENDPOINT"] = "https://freeweights.allternit.com" from transformers import AutoModelForImageTextToText, AutoProcessor processor = AutoProcessor.from_pretrained("deepseek-ai/DeepSeek-OCR-2", revision="aaa02f3811945a91062062994c5c4a3f4c0af2b0") model = AutoModelForImageTextToText.from_pretrained("deepseek-ai/DeepSeek-OCR-2", revision="aaa02f3811945a91062062994c5c4a3f4c0af2b0", device_map="auto")
vLLMServe an OpenAI-compatible API.
$ HF_ENDPOINT=https://freeweights.allternit.com vllm serve deepseek-ai/DeepSeek-OCR-2 --revision aaa02f3811945a91062062994c5c4a3f4c0af2b0
Run at home

License

Name
Apache License 2.0
Id
Apache-2.0
Class
Permissive
Redistribution
yes
Listing
Hosted (signed)
Decided
License as declared by the publisher. No person read it. 2026-11-14
Text in
LICENSE.txt
Text sha256
9debfe045c2a2755a52d0ae1087605d15770fe6636798c09efd93c1dc732ef28
Terms gate
No extra terms beyond the license.

Read the license text

Licenses are checked against the license table. A custom license passes only after a person reads it.

Provenance

From: deepseek-ai <huggingface.co/deepseek-ai>Newsgroups: alt.freeweights.deepseekSubject: DeepSeek-OCR-2Date: 2026-11-14Commit: aaa02f3Signed: 2DC28CC7B26A283B ✓
Upstream
deepseek-ai/DeepSeek-OCR-2
Pinned commit
aaa02f3811945a91062062994c5c4a3f4c0af2b0
Manifest
DeepSeek-OCR-2.json · .minisig
Manifest sha256
5f6db77082ed025810fa4946fc1af9c9056cb3afb8ec67255e8d3e5decc469ee
Signature
minisign, Ed25519, key 2DC28CC7B26A283B
Created
2026-11-14 21:14 UTC
Generator
freeweights-manifest/0.1.0
Takedowns
None.

How to verify

$ curl -O https://freeweights.allternit.com/manifests/deepseek-ai/DeepSeek-OCR-2.json $ curl -O https://freeweights.allternit.com/manifests/deepseek-ai/DeepSeek-OCR-2.json.minisig $ minisign -Vm DeepSeek-OCR-2.json -P RWQ7KGqyx4zCLXtUFOqvfv8L3g+8fpKJ1FsuxPyOYMF+4KufbUxKq1Oq $ shasum -a 256 <file> # compare with the Files tab

A manifest whose signature fails must be rejected. A file whose size or sha256 does not match must be discarded. Verify a download

Model tree

No related models in this catalog.Nothing here lists this model as its parent, and it names no parent.
Get it
Download
$ HF_ENDPOINT=https://freeweights.allternit.com hf download deepseek-ai/DeepSeek-OCR-2
Open in Freeweights Studio

Pull, verify and run it in the Allternit app on Desktop, web or phone.

Signed manifest
minisign, key 2DC28CC7B26A283B. DeepSeek-OCR-2.json. Provenance. Takedown history: none.
Source
huggingface.co/deepseek-ai/DeepSeek-OCR-2

Related models

All of alt.freeweights.deepseek →
From: timm <huggingface.co/timm>Subject: deepseek_vit_412m.deepseek_v4_1_flash
From: bartowski <huggingface.co/bartowski>Subject: deepseek-ai_DeepSeek-R1-0528-GGUF
From: bartowski <huggingface.co/bartowski>Subject: deepseek-ai_DeepSeek-R1-0528-Qwen3-8B-GGUF
From: bartowski <huggingface.co/bartowski>Subject: deepseek-ai_DeepSeek-V3-0324-GGUF
From: bartowski <huggingface.co/bartowski>Subject: deepseek-ai_DeepSeek-V3.1-GGUF
From: mlx-community <huggingface.co/mlx-community>Subject: deepseek-ai-DeepSeek-V4-Flash-8bit