A model is still running

A model is still running or downloading. Leaving will stop the current operation. Stay on this page to let it finish.

Skip to content
LocalAI.toolsBeta

Join the waitlist

Get an email when the new apps and model runs launch.

No account needed.

PP-OCR 6 Desktop

Locate printed text regions in a still image. Returns axis-aligned pixel boxes and confidence scores; it does not read their text.

Desktop inferencePrivate · Runs on your computer
About this model

Locate printed text regions in a still image. Returns axis-aligned pixel boxes and confidence scores; it does not read their text.

Created by
PaddlePaddle
License
Apache-2.0
Model released
Unknown
Source repository created
Jun 10, 2026
Model weights · FP32
9.9 MB

Best for

  • Text-region localization before recognition.
  • Reviewable local document-processing prototypes.

Keep in mind

  • Local benchmark scope is printed English fixtures only; upstream Chinese,Japanese,handwriting,curved and vertical text capabilities are not locally certified.
  • Poor contrast,blur,unusual fonts,small or rotated text can fail; review important results.
  • Exactly uniform decoded RGB regions return empty without neural inference; textured no-text regions can still produce false positives or hallucinations. Recognition applies per-region linear contrast stretching; detection caps its canvas long side at736pixels.
  • Input is one still image URL with optional in-bounds pixel crop. Images are limited to8192 pixels per side,16megapixels and64MiB encoded.
  • Axis-aligned connected-component postprocessing is adapter-local,not a claim of exact official polygon/unclip parity.
  • The pinned ONNX graph produces a complete tensor; incremental streaming is unavailable.
  • Publication,complete dependency backup and forced-source recovery remain separate release gates.

Processed on your deviceYour input is processed locally in the desktop app. Model and runtime files download from external hosts.

Target platforms · verification required

  • macos · aarch64

Recommended configuration

FP32 · cpu (verified platform only)

Sources & technical details
Original model
PaddlePaddle/PP-OCRv6_small_det
Release verification
Unknown
Release evidence
Official exact checkpoint card and repository inspected; no unambiguous original public checkpoint release day established. Repository creation/conversion dates are not treated as release dates.
Source updated
Jun 12, 2026
Native model files
PaddlePaddle/PP-OCRv6_small_det_onnx
Files updated
Jun 18, 2026
Artifact revision
28fe5895c24fd108c19eb3e8479f4ab385fbfc62

Desktop app

Sample result

PP-OCR 6 Small Text Detector Desktop

Run PP-OCR 6 Desktop on your computer

Download the desktop app to load this model, provide input and view results locally.

Public desktop downloads are coming soon.

Device requirements

macOS Apple SiliconCPU / FP32TestedNot establishedRAM estimate
If your device falls short

MemoryAdvisory only: close other apps, reduce image/tile/token/batch settings or choose a smaller verified model/precision. CPU can need more RAM than GPU.

GPUIf the selected GPU runtime fails, existing CPU fallback may recover when enabled, but CPU has its own memory/runtime needs.

PlatformUse a compatible OS/architecture/runtime or another model. These estimates do not add hardware detection or block runs.

Benchmark evidence

Saved desktop measurements apply to the recorded hardware and runtime. Fallback cases report the device that actually ran; they do not prove GPU support on the requested device.

PP-OCR 6 Small Text Detector Desktop

1 configurations tested
Passed1
Failed0
TestedNov 14, 2026
Runtimeonnxruntime 1.24.4

Quality criteria

{ "detectorMinimumMatchedIoU": 0.4, "detectorMinimumRecall": 0.9, "detectorMaximumExtraBoxes": 1, "uniformNoTextMustBeEmpty": true, "recognizerMaximumPerCaseCER": 0.05, "maximumInferenceMs": 5000 }

PrecisionDeviceResultPlatformInference / load
FP32cpuPassedVery fastmacos · aarch6473 ms / 181 ms
Test setup & sample details
Quality gate
{ "detectorMinimumMatchedIoU": 0.4, "detectorMinimumRecall": 0.9, "detectorMaximumExtraBoxes": 1, "uniformNoTextMustBeEmpty": true, "recognizerMaximumPerCaseCER": 0.05, "maximumInferenceMs": 5000 }
Tested
2026-11-14T11:08:30.669684+00:00
Environment
Apple M3 Max macOS 27.2; saved still-image OCR fixtures; CPU FP32.
Runtime
onnxruntime 1.24.4
Artifact revision
28fe5895c24fd108c19eb3e8479f4ab385fbfc62
Fixture
{
  "file": "model-samples/ppocrv5-menu.png",
  "url": "https://f005.backblazeb2.com/file/hubters-webai-public/assets/benchmarks/ppocrv5-menu.png",
  "size": 16719,
  "sha1": "a238414c0632eb69535a9b5307b61cfea4da5421"
}
Rerun when
artifact revision; runtime version; OCR fixture or preprocessing; host device
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "recall": 1,
      "extra": 0
    },
    "runtime": {
      "acceleration": "cpu",
      "artifactRevision": "28fe5895c24fd108c19eb3e8479f4ab385fbfc62",
      "cpuThreads": 4,
      "device": "cpu",
      "durationMs": 85.013666,
      "engine": "onnxruntime",
      "fallback": false,
      "fallbackReason": null,
      "family": {
        "catalogId": "pp-ocrv6-native",
        "defaultModelId": "ppocrv6-small-det-native",
        "displayName": "PP-OCR 6 Desktop",
        "id": "pp-ocrv6",
        "variantId": "small-det",
        "variantName": "Small detector"
      },
      "gpuPlanOperations": 0,
      "imagesPerSecond": 13.72095017543147,
      "inferenceMs": 72.88125000195578,
      "loadMs": 180.782167,
      "modelId": "ppocrv6-small-det-native",
      "neuralExecuted": true,
      "peakRssBytes": 248365056,
      "placement": null,
      "precision": "fp32",
      "provider": "CPUExecutionProvider",
      "requestedDevice": "cpu",
      "requestedPrecision": "fp32",
      "reused": false,
      "selectionReason": "Explicit device selection; actual placement is reported after inference",
      "version": "1.24.4",
      "totalMs": 253.66341700195576,
      "grade": "good"
    }
  }
]

PP-OCR 6 Medium Text Detector Desktop

2 configurations tested
Passed2
Failed0
TestedNov 14, 2026
Runtimeonnxruntime 1.24.4

Quality criteria

{ "detectorMinimumMatchedIoU": 0.4, "detectorMinimumRecall": 0.9, "detectorMaximumExtraBoxes": 1, "uniformNoTextMustBeEmpty": true, "recognizerMaximumPerCaseCER": 0.05, "maximumInferenceMs": 5000 }

PrecisionDeviceResultPlatformInference / load
FP32cpuPassedFastmacos · aarch64211 ms / 205 ms
FP32coremlPassedFastmacos · aarch64306 ms / 2880 ms
Test setup & sample details
Quality gate
{ "detectorMinimumMatchedIoU": 0.4, "detectorMinimumRecall": 0.9, "detectorMaximumExtraBoxes": 1, "uniformNoTextMustBeEmpty": true, "recognizerMaximumPerCaseCER": 0.05, "maximumInferenceMs": 5000 }
Tested
2026-11-14T11:08:31.216446+00:00
Environment
Apple M3 Max macOS 27.2; saved still-image OCR fixtures; CPU FP32 and partial Core ML GPU FP32.
Runtime
onnxruntime 1.24.4
Artifact revision
61323801669c338b7891481ec7bac61ce31b576a
Fixture
{
  "file": "model-samples/ppocrv5-menu.png",
  "url": "https://f005.backblazeb2.com/file/hubters-webai-public/assets/benchmarks/ppocrv5-menu.png",
  "size": 16719,
  "sha1": "a238414c0632eb69535a9b5307b61cfea4da5421"
}
Rerun when
artifact revision; runtime version; OCR fixture or preprocessing; host device
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "recall": 1,
      "extra": 0
    },
    "runtime": {
      "acceleration": "cpu",
      "artifactRevision": "61323801669c338b7891481ec7bac61ce31b576a",
      "cpuThreads": 4,
      "device": "cpu",
      "durationMs": 213.372458,
      "engine": "onnxruntime",
      "fallback": false,
      "fallbackReason": null,
      "family": {
        "catalogId": "pp-ocrv6-native",
        "defaultModelId": "ppocrv6-small-det-native",
        "displayName": "PP-OCR 6 Desktop",
        "id": "pp-ocrv6",
        "variantId": "medium-det",
        "variantName": "Medium detector"
      },
      "gpuPlanOperations": 0,
      "imagesPerSecond": 4.735144610681135,
      "inferenceMs": 211.18679200299084,
      "loadMs": 205.214042,
      "modelId": "ppocrv6-medium-det-native",
      "neuralExecuted": true,
      "peakRssBytes": 371343360,
      "placement": null,
      "precision": "fp32",
      "provider": "CPUExecutionProvider",
      "requestedDevice": "cpu",
      "requestedPrecision": "fp32",
      "reused": false,
      "selectionReason": "Explicit device selection; actual placement is reported after inference",
      "version": "1.24.4",
      "totalMs": 416.40083400299085,
      "grade": "good"
    }
  },
  {
    "precision": "fp32",
    "device": "coreml",
    "status": "pass",
    "quality": {
      "passed": true,
      "recall": 1,
      "extra": 0
    },
    "runtime": {
      "acceleration": "partial",
      "artifactRevision": "61323801669c338b7891481ec7bac61ce31b576a",
      "cpuThreads": 4,
      "device": "coreml",
      "durationMs": 307.0805,
      "engine": "onnxruntime",
      "fallback": false,
      "fallbackReason": null,
      "family": {
        "catalogId": "pp-ocrv6-native",
        "defaultModelId": "ppocrv6-small-det-native",
        "displayName": "PP-OCR 6 Desktop",
        "id": "pp-ocrv6",
        "variantId": "medium-det",
        "variantName": "Medium detector"
      },
      "gpuPlanOperations": 74,
      "imagesPerSecond": 3.2726939578414487,
      "inferenceMs": 305.5586660047993,
      "loadMs": 2879.8072079999997,
      "modelId": "ppocrv6-medium-det-native",
      "neuralExecuted": true,
      "peakRssBytes": 2202075136,
      "placement": {
        "CPUExecutionProvider": 11,
        "CoreMLExecutionProvider": 10
      },
      "precision": "fp32",
      "provider": "CoreMLExecutionProvider",
      "requestedDevice": "coreml",
      "requestedPrecision": "fp32",
      "reused": false,
      "selectionReason": "Explicit device selection; actual placement is reported after inference",
      "version": "1.24.4",
      "visualizationMs": 0.166584,
      "totalMs": 3185.365874004799,
      "grade": "good"
    }
  }
]

PP-OCR 6 Tiny Text Detector Desktop

1 configurations tested
Passed1
Failed0
TestedNov 14, 2026
Runtimeonnxruntime 1.24.4

Quality criteria

{ "detectorMinimumMatchedIoU": 0.4, "detectorMinimumRecall": 0.9, "detectorMaximumExtraBoxes": 1, "uniformNoTextMustBeEmpty": true, "recognizerMaximumPerCaseCER": 0.05, "maximumInferenceMs": 5000 }

PrecisionDeviceResultPlatformInference / load
FP32cpuPassedVery fastmacos · aarch6443 ms / 169 ms
Test setup & sample details
Quality gate
{ "detectorMinimumMatchedIoU": 0.4, "detectorMinimumRecall": 0.9, "detectorMaximumExtraBoxes": 1, "uniformNoTextMustBeEmpty": true, "recognizerMaximumPerCaseCER": 0.05, "maximumInferenceMs": 5000 }
Tested
2026-11-14T11:08:30.597477+00:00
Environment
Apple M3 Max macOS 27.2; saved still-image OCR fixtures; CPU FP32.
Runtime
onnxruntime 1.24.4
Artifact revision
2ba1506c0380b8f0b03dd142459aac66d4421f6c
Fixture
{
  "file": "model-samples/ppocrv5-menu.png",
  "url": "https://f005.backblazeb2.com/file/hubters-webai-public/assets/benchmarks/ppocrv5-menu.png",
  "size": 16719,
  "sha1": "a238414c0632eb69535a9b5307b61cfea4da5421"
}
Rerun when
artifact revision; runtime version; OCR fixture or preprocessing; host device
Per-case quality & failure details
[
  {
    "precision": "fp32",
    "device": "cpu",
    "status": "pass",
    "quality": {
      "passed": true,
      "recall": 1,
      "extra": 0
    },
    "runtime": {
      "acceleration": "cpu",
      "artifactRevision": "2ba1506c0380b8f0b03dd142459aac66d4421f6c",
      "cpuThreads": 4,
      "device": "cpu",
      "durationMs": 48.883125,
      "engine": "onnxruntime",
      "fallback": false,
      "fallbackReason": null,
      "family": {
        "catalogId": "pp-ocrv6-native",
        "defaultModelId": "ppocrv6-small-det-native",
        "displayName": "PP-OCR 6 Desktop",
        "id": "pp-ocrv6",
        "variantId": "tiny-det",
        "variantName": "Tiny detector"
      },
      "gpuPlanOperations": 0,
      "imagesPerSecond": 23.49085912572734,
      "inferenceMs": 42.56974998861551,
      "loadMs": 168.718125,
      "modelId": "ppocrv6-tiny-det-native",
      "neuralExecuted": true,
      "peakRssBytes": 157646848,
      "placement": null,
      "precision": "fp32",
      "provider": "CPUExecutionProvider",
      "requestedDevice": "cpu",
      "requestedPrecision": "fp32",
      "reused": false,
      "selectionReason": "Explicit device selection; actual placement is reported after inference",
      "version": "1.24.4",
      "totalMs": 211.2878749886155,
      "grade": "good"
    }
  }
]
Release validation still required
  • Public dependency backup, sample publication and forced-outage acceptance are deferred.
  • Windows runtime and platform verification are deferred.

Something not quite right?

Tell us what happened with PP-OCR 6 Desktop. Loading, results, or the page itself—we’re here for all of it.

Attachments Optional

Up to 3 files · 50 MB each

Browser, OS, CPU/GPU, memory, screen, storage, and network details are included automatically. Model inputs, results, and browsing history are not collected.