skillsafe-ai/pp-ocrv6-tiny-rec

PP-OCRv6 tiny text recognition

Browser-ready import artifacts for text-recognition, produced by SkillSafe's reproducible converter (models/ in skillsafe.ai) from a pinned upstream source. Every byte here is derivable from that source plus the recipe below; nothing was edited by hand.

Provenance

Upstream	https://huggingface.co/PaddlePaddle/PP-OCRv6_tiny_rec_onnx/tree/2612ab37152ae0a677521bae4e1e3d4fb4cf7c30
Upstream SHA-256 / commit	2612ab37152ae0a677521bae4e1e3d4fb4cf7c30
Recipe	recipes/pp-ocrv6-tiny-rec.yaml — sha256 95bdbfd15e694d66d1e10276ce7a493e9d6d5c28af2bf2d5869dd914a465a53a
Toolchain	Python 3.12.13, torch 2.10.0, onnx 1.23.0, onnxruntime 1.30.0 on Darwin 25.6.0 arm64
Converted	2026-09-22T21:39:21+00:00

Files

file	class	size	SHA-256
inference.onnx	registry	4.26 MB	9ef676d6ed3c88256a2d92c640c44f25b0c40947e111b14b8be8f594091563e6

registry files are parameter files served from models.skillsafe.ai once vetted; bundle files ship inside an app; registry-shared is a runtime library reused by every model of the same architecture.

Verification

Imported as published upstream (no conversion). Each file is pinned by SHA-256 to its source; every ONNX file passed onnx.checker and a CPU smoke run under onnxruntime with zero-filled inputs at the declared shapes:

file	inputs	outputs	ms
inference.onnx	x[1, 3, 48, 320]	fetch_name_0[1, 40, 6906]	2.4

Use in the browser

import * as ort from "onnxruntime-web";
const session = await ort.InferenceSession.create("https://huggingface.co/skillsafe-ai/pp-ocrv6-tiny-rec/resolve/main/inference.onnx", { executionProviders: ["webgpu", "wasm"] });

Contract (inference.onnx): input x float32 ['DynamicDimension.0', 3, 48, 'DynamicDimension.1'] → output fetch_name_0 float32 ['DynamicDimension.0', 'floor(floor(floor(DynamicDimension.1/2 - 1/2)/2)/2 - 1/2) + 1', 6906]. Opset 11.

Licence and attribution

PP-OCRv6 tiny (rec): PaddlePaddle / PaddleOCR, Apache License 2.0. https://huggingface.co/PaddlePaddle/PP-OCRv6_tiny_rec_onnx

Licence: Apache-2.0 — notice: https://github.com/PaddlePaddle/PaddleOCR/blob/main/LICENSE. The conversion recipe and this model card are part of the SkillSafe repository and carry its licence; the weights remain under the upstream licence above.

The full manifest.json in this repo records the recipe, sources, toolchain (including the uv.lock hash) and per-file verification numbers.