Image-Text-to-Text
PaddleOCR
Safetensors
English
Chinese
multilingual
paddleocr_vl
ERNIE4.5
PaddlePaddle
image-to-text
ocr
document-parse
layout
table
formula
chart
conversational
custom_code
Eval Results
Instructions to use PaddlePaddle/PaddleOCR-VL with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PaddleOCR
How to use PaddlePaddle/PaddleOCR-VL with PaddleOCR:
# See https://www.paddleocr.ai/latest/version3.x/pipeline_usage/PaddleOCR-VL.html to installation from paddleocr import PaddleOCRVL pipeline = PaddleOCRVL(pipeline_version="v1") output = pipeline.predict("path/to/document_image.png") for res in output: res.print() res.save_to_json(save_path="output") res.save_to_markdown(save_path="output") - Notebooks
- Google Colab
- Kaggle
Record full-checkpoint compatibility regression results
Browse files- tests/README.md +0 -2
- tests/compatibility_results.json +44 -11
tests/README.md
CHANGED
|
@@ -18,5 +18,3 @@ Fixtures:
|
|
| 18 |
- `paddle_general_ocr_001.png`: official synthetic Chinese OCR test image from https://huggingface.co/datasets/hf-internal-testing/transformers-synthetic-assets/resolve/main/images/paddle_general_ocr_001.png .
|
| 19 |
- `invoice_en.png`: synthetic three-line English invoice with known expected text.
|
| 20 |
- `table_en.png`: synthetic Item/Qty/Price table containing Apple and Orange rows.
|
| 21 |
-
|
| 22 |
-
Draft progress: only processor loading and meta-device construction have passed for this model so far. Full-checkpoint regression is pending; do not interpret the summary as a generation test.
|
|
|
|
| 18 |
- `paddle_general_ocr_001.png`: official synthetic Chinese OCR test image from https://huggingface.co/datasets/hf-internal-testing/transformers-synthetic-assets/resolve/main/images/paddle_general_ocr_001.png .
|
| 19 |
- `invoice_en.png`: synthetic three-line English invoice with known expected text.
|
| 20 |
- `table_en.png`: synthetic Item/Qty/Price table containing Apple and Orange rows.
|
|
|
|
|
|
tests/compatibility_results.json
CHANGED
|
@@ -2,11 +2,11 @@
|
|
| 2 |
"model": "PaddlePaddle/PaddleOCR-VL",
|
| 3 |
"revision": "7fa00a8c55b735ba51ba49a9058f3f9c57a99a11",
|
| 4 |
"trust_remote_code": true,
|
| 5 |
-
"validation_scope": "
|
| 6 |
-
"full_checkpoint_pending":
|
| 7 |
-
"device": "
|
| 8 |
-
"dtype":
|
| 9 |
-
"checkpoint_sha256":
|
| 10 |
"tested_versions": [
|
| 11 |
"4.57.6",
|
| 12 |
"5.0.0",
|
|
@@ -51,8 +51,19 @@
|
|
| 51 |
"stages": [
|
| 52 |
"config",
|
| 53 |
"processor",
|
| 54 |
-
"
|
| 55 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 56 |
},
|
| 57 |
{
|
| 58 |
"version": "5.0.0",
|
|
@@ -60,8 +71,19 @@
|
|
| 60 |
"stages": [
|
| 61 |
"config",
|
| 62 |
"processor",
|
| 63 |
-
"
|
| 64 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 65 |
},
|
| 66 |
{
|
| 67 |
"version": "5.21.1",
|
|
@@ -69,8 +91,19 @@
|
|
| 69 |
"stages": [
|
| 70 |
"config",
|
| 71 |
"processor",
|
| 72 |
-
"
|
| 73 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 74 |
}
|
| 75 |
]
|
| 76 |
}
|
|
|
|
| 2 |
"model": "PaddlePaddle/PaddleOCR-VL",
|
| 3 |
"revision": "7fa00a8c55b735ba51ba49a9058f3f9c57a99a11",
|
| 4 |
"trust_remote_code": true,
|
| 5 |
+
"validation_scope": "full checkpoint and generation",
|
| 6 |
+
"full_checkpoint_pending": false,
|
| 7 |
+
"device": "cpu",
|
| 8 |
+
"dtype": "float32",
|
| 9 |
+
"checkpoint_sha256": "3085f1042e184f68f8a412aa0f64f2c4b8562989598bbfba326aaa11fc685de8",
|
| 10 |
"tested_versions": [
|
| 11 |
"4.57.6",
|
| 12 |
"5.0.0",
|
|
|
|
| 51 |
"stages": [
|
| 52 |
"config",
|
| 53 |
"processor",
|
| 54 |
+
"model",
|
| 55 |
+
"weights_and_buffers",
|
| 56 |
+
"mask_semantics",
|
| 57 |
+
"english_generate",
|
| 58 |
+
"uncached_prefix",
|
| 59 |
+
"batch_prefill_isolation",
|
| 60 |
+
"uncached_full_english",
|
| 61 |
+
"full_chinese_batch_table"
|
| 62 |
+
],
|
| 63 |
+
"parameter_mismatches": 0,
|
| 64 |
+
"batch_logits_max_abs_difference": 4.3392181396484375e-05,
|
| 65 |
+
"uncached_full_english_matches_cached": true,
|
| 66 |
+
"full_batch_matches_individual": true
|
| 67 |
},
|
| 68 |
{
|
| 69 |
"version": "5.0.0",
|
|
|
|
| 71 |
"stages": [
|
| 72 |
"config",
|
| 73 |
"processor",
|
| 74 |
+
"model",
|
| 75 |
+
"weights_and_buffers",
|
| 76 |
+
"mask_semantics",
|
| 77 |
+
"english_generate",
|
| 78 |
+
"uncached_prefix",
|
| 79 |
+
"batch_prefill_isolation",
|
| 80 |
+
"uncached_full_english",
|
| 81 |
+
"full_chinese_batch_table"
|
| 82 |
+
],
|
| 83 |
+
"parameter_mismatches": 0,
|
| 84 |
+
"batch_logits_max_abs_difference": 4.3392181396484375e-05,
|
| 85 |
+
"uncached_full_english_matches_cached": true,
|
| 86 |
+
"full_batch_matches_individual": true
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"version": "5.21.1",
|
|
|
|
| 91 |
"stages": [
|
| 92 |
"config",
|
| 93 |
"processor",
|
| 94 |
+
"model",
|
| 95 |
+
"weights_and_buffers",
|
| 96 |
+
"mask_semantics",
|
| 97 |
+
"english_generate",
|
| 98 |
+
"uncached_prefix",
|
| 99 |
+
"batch_prefill_isolation",
|
| 100 |
+
"uncached_full_english",
|
| 101 |
+
"full_chinese_batch_table"
|
| 102 |
+
],
|
| 103 |
+
"parameter_mismatches": 0,
|
| 104 |
+
"batch_logits_max_abs_difference": 4.3392181396484375e-05,
|
| 105 |
+
"uncached_full_english_matches_cached": true,
|
| 106 |
+
"full_batch_matches_individual": true
|
| 107 |
}
|
| 108 |
]
|
| 109 |
}
|