TrueOCR
OCR tiếng Việt tốc độ cao trên CPU. Đọc ảnh nhiều dòng hoặc crop một dòng; trả về text, bbox và confidence. Model khoảng 8 MB kèm trong package, chạy offline, không cần GPU hay PyTorch.
Tốc độ CPU
Xeon Platinum 8358P, 1 thread, batch 1, cùng 128 crop × 3 lượt/model @TrueOCR 0.1.0.
| Model | p50, ms/crop ↓ | p95, ms/crop ↓ | p50 / TrueOCR |
|---|---|---|---|
| TrueOCR 0.1.0 | 1.54 | 2.35 | 1.0× |
| Tesseract fast | 5.89 | 15.29 | 3.8× |
| PaddleOCR v6 tiny | 10.76 | 12.29 | 7.0× |
| Tesseract best | 20.24 | 49.06 | 13.2× |
| PaddleOCR v5 Latin mobile | 29.43 | 30.79 | 19.1× |
| EasyOCR latin_g2 | 30.03 | 68.90 | 19.5× |
| VietOCR vgg_seq2seq | 71.77 | 98.01 | 46.7× |
| VietOCR vgg_transformer | 137.85 | 231.45 | 89.7× |
Đo latency của crop recognition, gồm preprocessing, inference và text decoding. Không gồm model loading, image I/O hay text detection cho cả trang. Benchmark setup, baseline config và số liệu từng request.
Quick Start
Python ≥3.10.
pip install trueocr
from trueocr import load_model
model = load_model() # Load một lần, dùng lại cho nhiều ảnh
result = model.predict("anh.jpg")
print(result.text)
Mặc định đọc ảnh nhiều dòng. Với crop một dòng:
model.predict("crop.png", mode="line"). Lấy bbox và confidence qua
result.lines; lưu text, JSON và ảnh bbox bằng result.save("output").
CLI: trueocr anh.jpg --out output/.
Chất lượng và giới hạn
Model dùng cho chữ in tiếng Việt. Handwriting, scene text và font lạ còn yếu; crop quá rộng hoặc chứa nhiều dòng dễ sai. Kết quả theo dataset.
1 maintainer
Metadata
1 maintainer
Release files for trueocr 0.1.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| trueocr-0.1.1.tar.gz | 7.4 MB | With provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| trueocr-0.1.1-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 14.8 MB
Release files / trueocr-0.1.1.tar.gz
| Download URL | trueocr-0.1.1.tar.gz |
|---|---|
| Size | 7.4 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
c25b4b2e486b9434f600fb083db70022b6b01f55708b868df4373950cd7e7d3e
|
|
BLAKE2b-256 checksum How to use checksums |
695fd537adcab7ab74fc705f419a4be0571d6e4022fb0aaaf478c9911171d43d
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 24, 2026.
Transparency logRelease files / trueocr-0.1.1-py3-none-any.whl
| Download URL | trueocr-0.1.1-py3-none-any.whl |
|---|---|
| Size | 7.4 MB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
8a7e4141d5f06f602684ea5dbafd775d11787d7319d39baaabf5e41158597374
|
|
BLAKE2b-256 checksum How to use checksums |
dbdbf32d13bb09f875b8f8b301c83ea4f8dfe92fcb543f49ca11daed5a0657b5
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 24, 2026.
Transparency logSecurity for trueocr 0.1.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
All 2 files in this release have been uploaded with provenance metadata.
Attestation for trueocr-0.1.1.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 24, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.
