Fix incorrect model information in MANIFEST.json and add test data for version-RFB (#553)

jcwchen · web-flow · commit e4aa511d91cc · 2022-10-10T10:30:43.000-07:00
* Add missing version-RFB

Signed-off-by: jcwchen &lt;jacky82226@gmail.com&gt;

* updated MAINIFEST

Signed-off-by: jcwchen &lt;jacky82226@gmail.com&gt;

* nit

Signed-off-by: jcwchen &lt;jacky82226@gmail.com&gt;

* update

Signed-off-by: jcwchen &lt;jacky82226@gmail.com&gt;

* update .json

Signed-off-by: jcwchen &lt;jacky82226@gmail.com&gt;

* for example

Signed-off-by: jcwchen &lt;jacky82226@gmail.com&gt;

* use string Python 3.10

Signed-off-by: jcwchen &lt;jacky82226@gmail.com&gt;

* add yolov2.tar.gz

Signed-off-by: jcwchen &lt;jacky82226@gmail.com&gt;

Signed-off-by: jcwchen &lt;jacky82226@gmail.com&gt;
diff --git a/.github/workflows/linux_ci.yml b/.github/workflows/linux_ci.yml
@@ -17,7 +17,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [3.8]
+        python-version: ['3.10']
         architecture: ['x64']
 
     steps:
diff --git a/.github/workflows/windows_ci.yml b/.github/workflows/windows_ci.yml
@@ -17,7 +17,7 @@ jobs:
     runs-on: windows-latest
     strategy:
       matrix:
-        python-version: [3.8]
+        python-version: ['3.10']
         architecture: ['x64']
 
     steps:
@@ -33,8 +33,6 @@ jobs:
       - name: Install dependencies
         run: |
           python -m pip install --upgrade pip
-          # TODO: now ONNX only supports Protobuf <= 3.20.1
-          python -m pip install protobuf==3.20.1
           python -m pip install onnx onnxruntime requests py-cpuinfo
           # Print CPU info for debugging ONNX Runtime inference difference
           python -m cpuinfo
diff --git a/ONNX_HUB_MANIFEST.json b/ONNX_HUB_MANIFEST.json
@@ -753,7 +753,7 @@
     {
         "model": "T5-decoder-with-lm-head",
         "model_path": "text/machine_comprehension/t5/model/t5-decoder-with-lm-head-12.onnx",
-        "onnx_version": "1.6",
+        "onnx_version": "1.7",
         "opset_version": 12,
         "metadata": {
             "model_sha": "235afca35d3c47a5fd7209844ac11f3157fe7263fd074197a45b7f536e40ea56",
@@ -794,13 +794,16 @@
                         "type": "tensor(float)"
                     }
                 ]
-            }
+            },
+            "model_with_data_path": "text/machine_comprehension/t5/model/t5-decoder-with-lm-head-12.tar.gz",
+            "model_with_data_sha": "fa6edefa5951479f7b68bea693950a4ebc188f54c7971bf73f2b036b91bc8066",
+            "model_with_data_bytes": 287840759
         }
     },
     {
         "model": "T5-encoder",
         "model_path": "text/machine_comprehension/t5/model/t5-encoder-12.onnx",
-        "onnx_version": "1.6",
+        "onnx_version": "1.7",
         "opset_version": 12,
         "metadata": {
             "model_sha": "4523122d7cf0f50905694d84995633c8a0cc223da762d3eb2aaffa17251a6f60",
@@ -832,7 +835,10 @@
                         "type": "tensor(float)"
                     }
                 ]
-            }
+            },
+            "model_with_data_path": "text/machine_comprehension/t5/model/t5-encoder-12.tar.gz",
+            "model_with_data_sha": "434e2f691bd71c838bd4b68be47bdf32c899313fdebfe4b77c7bea0b2f52e831",
+            "model_with_data_bytes": 194535656
         }
     },
     {
@@ -1025,7 +1031,10 @@
                         "type": "tensor(float)"
                     }
                 ]
-            }
+            },
+            "model_with_data_path": "vision/body_analysis/ultraface/models/version-RFB-320.tar.gz",
+            "model_with_data_sha": "628d0dd3e0288adb821f211e13d4e97f6d6f4527237339606732dffa6f19d381",
+            "model_with_data_bytes": 2015397
         }
     },
     {
@@ -1074,7 +1083,10 @@
                         "type": "tensor(float)"
                     }
                 ]
-            }
+            },
+            "model_with_data_path": "vision/body_analysis/ultraface/models/version-RFB-640.tar.gz",
+            "model_with_data_sha": "07cc0e284b7924bd89f2ec103254ef3ab4b673fd12341153faaff07f3c1137e3",
+            "model_with_data_bytes": 4818743
         }
     },
     {
@@ -7446,7 +7458,10 @@
                         "type": "tensor(float)"
                     }
                 ]
-            }
+            },
+            "model_with_data_path": "vision/object_detection_segmentation/yolov2-coco/model/yolov2-coco-9.tar.gz",
+            "model_with_data_sha": "c34966e4e96165f8db6f2549f509b6a8bdfc01ddd978e6fd07daad4e665d5383",
+            "model_with_data_bytes": 191439022
         }
     },
     {
diff --git a/text/machine_comprehension/t5/README.md b/text/machine_comprehension/t5/README.md
@@ -12,10 +12,10 @@ understanding through the training of multiple tasks at once.
 
 ## Model
 
- |Model        |Download  | Compressed |ONNX version|Opset version|
-|-------------|:--------------|:--------------|:--------------|:--------------|
-|T5-encoder       |[650.6 MB](model/t5-encoder-12.onnx) | [205.0 MB](model/t5-encoder-12.tar.gz)| 1.6 | 12
-|T5-decoder-with-lm-head |[304.9 MB](model/t5-decoder-with-lm-head-12.onnx) | [304.9 MB](model/t5-decoder-with-lm-head-12.tar.gz)| 1.6 | 12
+| Model | Download  | Download (with sample test data) | ONNX version | Opset version |
+| ----------- | ---------- |--------------| -------------- | -------------- |
+|T5-encoder       |[650.6 MB](model/t5-encoder-12.onnx) | [205.0 MB](model/t5-encoder-12.tar.gz)| 1.7 | 12
+|T5-decoder-with-lm-head |[304.9 MB](model/t5-decoder-with-lm-head-12.onnx) | [304.9 MB](model/t5-decoder-with-lm-head-12.tar.gz)| 1.7 | 12
 
 
 ### Source
diff --git a/vision/body_analysis/age_gender/README.md b/vision/body_analysis/age_gender/README.md
@@ -6,7 +6,7 @@
 Automatic age and gender classification has become relevant to an increasing amount of applications, particularly since the rise of social platforms and social media. Nevertheless, performance of existing methods on real-world images is still significantly lacking, especially when compared to the tremendous leaps in performance recently reported for the related task of face recognition.
 
 ## Models
-| Model (Caffe) | Download | ONNX version | Opset version | Dataset |
+| Model | Download | ONNX version | Opset version | Dataset |
 |:-------------|:--------------|:--------------|:--------------|:--------------|
 | [googlenet_age_adience](https://drive.google.com/drive/folders/1GeLTHzHALgTYFj2Q9o5aWdztA9WzoErx?usp=sharing) | [23 MB](models/age_googlenet.onnx) | 1.6 | 11 | Adience |
 | [googlenet_gender_adience](https://drive.google.com/drive/folders/1r0GroTfsF7VpLhcS3IxU-LmAh6rI6vbQ?usp=sharing) | [23 MB](models/gender_googlenet.onnx)| 1.6 | 11 | Adience |
diff --git a/vision/body_analysis/ultraface/README.md b/vision/body_analysis/ultraface/README.md
@@ -6,10 +6,10 @@
 This model is a lightweight facedetection model designed for edge computing devices.
 
 ## Model
-| Model| Download | ONNX version | Opset version |
-|----------------|:-----------|:-----------|:--------|
-|version-RFB-320|[1.21 MB](models/version-RFB-320.onnx)|1.4|9|
-|version-RFB-640|[1.51 MB](models/version-RFB-640.onnx)|1.4|9|
+| Model | Download | Download (with sample test data) | ONNX version | Opset version |
+| ------------- | ------------- | ------------- | ------------- | ------------- |
+|version-RFB-320| [1.21 MB](models/version-RFB-320.onnx) | [1.92 MB](models/version-RFB-320.tar.gz) | 1.4 | 9 |
+|version-RFB-640| [1.51 MB](models/version-RFB-640.onnx) | [4.59 MB](models/version-RFB-640.tar.gz) | 1.4 | 9 |
 
 ### Dataset
 The training set is the VOC format data set generated by using the cleaned widerface labels provided by [Retinaface](https://arxiv.org/pdf/1905.00641.pdf) in conjunction with the widerface [dataset](http://shuoyang1213.me/WIDERFACE/).
diff --git a/vision/body_analysis/ultraface/models/version-RFB-320.tar.gz b/vision/body_analysis/ultraface/models/version-RFB-320.tar.gz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:628d0dd3e0288adb821f211e13d4e97f6d6f4527237339606732dffa6f19d381
+size 2015397
diff --git a/vision/body_analysis/ultraface/models/version-RFB-640.tar.gz b/vision/body_analysis/ultraface/models/version-RFB-640.tar.gz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07cc0e284b7924bd89f2ec103254ef3ab4b673fd12341153faaff07f3c1137e3
+size 4818743
diff --git a/vision/object_detection_segmentation/yolov2-coco/README.md b/vision/object_detection_segmentation/yolov2-coco/README.md
@@ -8,9 +8,9 @@ This model aims to detect objects in real time. It detects 80 different classes
 ## Model
 The model was converted to ONNX from PyTorch version of YOLOv2 using [PyTorch-Yolo2](https://github.com/marvis/pytorch-yolo2). The output is fully verified by generating bounding boxes under PyTorch and onnxruntime.
 
-|Model|Download| ONNX version |Opset version|
-|-----|:--------------|:-------------|:------------|
-|YOLOv2|[203.9 MB](model/yolov2-coco-9.onnx) |1.5  |9 |
+| Model | Download | Download (with sample test data) | ONNX version | Opset version |
+| ----- | -------- | -------------------------------- | ------------ | ------------- |
+| YOLOv2 | [203.9 MB](model/yolov2-coco-9.onnx) | [182.6 MB](model/yolov2-coco-9.tar.gz) | 1.5  | 9 |
 
 ## Inference
 ### Input to model
diff --git a/vision/object_detection_segmentation/yolov2-coco/model/yolov2-coco-9.tar.gz b/vision/object_detection_segmentation/yolov2-coco/model/yolov2-coco-9.tar.gz
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c34966e4e96165f8db6f2549f509b6a8bdfc01ddd978e6fd07daad4e665d5383
+size 191439022
diff --git a/workflow_scripts/generate_onnx_hub_manifest.py b/workflow_scripts/generate_onnx_hub_manifest.py
@@ -43,7 +43,7 @@ def parse_html(table):
 
 
 def parse_readme(filename):
-    with open(filename, "r") as f:
+    with open(filename, "r", encoding="utf8") as f:
         parsed = markdown.markdown(f.read(), extensions=["markdown.extensions.tables"])
         soup = bs4.BeautifulSoup(parsed, "html.parser")
         return [parse_html(table) for table in soup.find_all("table")]
@@ -65,7 +65,11 @@ def parse_readme(filename):
 all_tables = []
 for markdown_file in markdown_files:
     with open(markdown_file, "r") as f:
-        for parsed in parse_readme(markdown_file):
+        parsed_readme = parse_readme(markdown_file)
+        if not parsed_readme:
+            print(f"{markdown_file} needs to be updated to include a table.")
+            continue
+        for parsed in parsed_readme:
             parsed = parsed.rename(columns={"Opset Version": "Opset version"})
             if all(col in parsed.columns.values for col in ["Model", "Download", "Opset version", "ONNX version"]):
                 parsed["source_file"] = markdown_file
@@ -212,7 +216,7 @@ def get_model_ports(source_file, metadata, model_name):
                     help="Update target? (all, diff, single)",
                     choices=["all", "diff", "single"])
 parser.add_argument("--path", required=False, default=None, type=str,
-                    help="The model path which you want to update.")
+                    help="The model path which you want to update. e.g., vision/classification/resnet/model/resnet50.onnx")
 parser.add_argument("--drop", required=False, default=False, action="store_true",
                     help="Drop downloaded models after verification. (For space limitation in CIs)")
 args = parser.parse_args()
@@ -244,7 +248,7 @@ def get_model_ports(source_file, metadata, model_name):
         elif args.target == "single":
             if args.path is None:
                 raise ValueError("Please specify --path if you want to update by single model.")
-            target_models = set(args.path)
+            target_models = set([args.path.replace("\\", "/")])
         model_info = get_file_info(row, "model_path", target_models)
         if model_info is None:
             continue
@@ -261,7 +265,7 @@ def get_model_ports(source_file, metadata, model_name):
             for k, v in get_file_info(row, "model_with_data_path").items():
                 metadata[k] = v
         except (AttributeError, FileNotFoundError) as e:
-            print("no model_with_data in file {}: {}".format(row["source_file"], e))
+            print(f"no model_with_data in file {row['source_file']}: {e}")
 
         try:
             opset = int(row["opset_version"].contents[0])
@@ -292,6 +296,7 @@ def get_model_ports(source_file, metadata, model_name):
     else:
         print("Missing model in {}".format(row["source_file"]))
 output.sort(key=lambda x: x["model_path"])
+
 with open("ONNX_HUB_MANIFEST.json", "w+") as f:
     print("Found {} models".format(len(output)))
     json.dump(output, f, indent=4)

Original file line number	Diff line number	Diff line change
`@@ -753,7 +753,7 @@`
`753`	`753`	`{`
`754`	`754`	`"model": "T5-decoder-with-lm-head",`
`755`	`755`	`"model_path": "text/machine_comprehension/t5/model/t5-decoder-with-lm-head-12.onnx",`
`756`		`- "onnx_version": "1.6",`
	`756`	`+ "onnx_version": "1.7",`
`757`	`757`	`"opset_version": 12,`
`758`	`758`	`"metadata": {`
`759`	`759`	`"model_sha": "235afca35d3c47a5fd7209844ac11f3157fe7263fd074197a45b7f536e40ea56",`
`@@ -794,13 +794,16 @@`
`794`	`794`	`"type": "tensor(float)"`
`795`	`795`	`}`
`796`	`796`	`]`
`797`		`- }`
	`797`	`+ },`
	`798`	`+ "model_with_data_path": "text/machine_comprehension/t5/model/t5-decoder-with-lm-head-12.tar.gz",`
	`799`	`+ "model_with_data_sha": "fa6edefa5951479f7b68bea693950a4ebc188f54c7971bf73f2b036b91bc8066",`
	`800`	`+ "model_with_data_bytes": 287840759`
`798`	`801`	`}`
`799`	`802`	`},`
`800`	`803`	`{`
`801`	`804`	`"model": "T5-encoder",`
`802`	`805`	`"model_path": "text/machine_comprehension/t5/model/t5-encoder-12.onnx",`
`803`		`- "onnx_version": "1.6",`
	`806`	`+ "onnx_version": "1.7",`
`804`	`807`	`"opset_version": 12,`
`805`	`808`	`"metadata": {`
`806`	`809`	`"model_sha": "4523122d7cf0f50905694d84995633c8a0cc223da762d3eb2aaffa17251a6f60",`
`@@ -832,7 +835,10 @@`
`832`	`835`	`"type": "tensor(float)"`
`833`	`836`	`}`
`834`	`837`	`]`
`835`		`- }`
	`838`	`+ },`
	`839`	`+ "model_with_data_path": "text/machine_comprehension/t5/model/t5-encoder-12.tar.gz",`
	`840`	`+ "model_with_data_sha": "434e2f691bd71c838bd4b68be47bdf32c899313fdebfe4b77c7bea0b2f52e831",`
	`841`	`+ "model_with_data_bytes": 194535656`
`836`	`842`	`}`
`837`	`843`	`},`
`838`	`844`	`{`
`@@ -1025,7 +1031,10 @@`
`1025`	`1031`	`"type": "tensor(float)"`
`1026`	`1032`	`}`
`1027`	`1033`	`]`
`1028`		`- }`
	`1034`	`+ },`
	`1035`	`+ "model_with_data_path": "vision/body_analysis/ultraface/models/version-RFB-320.tar.gz",`
	`1036`	`+ "model_with_data_sha": "628d0dd3e0288adb821f211e13d4e97f6d6f4527237339606732dffa6f19d381",`
	`1037`	`+ "model_with_data_bytes": 2015397`
`1029`	`1038`	`}`
`1030`	`1039`	`},`
`1031`	`1040`	`{`
`@@ -1074,7 +1083,10 @@`
`1074`	`1083`	`"type": "tensor(float)"`
`1075`	`1084`	`}`
`1076`	`1085`	`]`
`1077`		`- }`
	`1086`	`+ },`
	`1087`	`+ "model_with_data_path": "vision/body_analysis/ultraface/models/version-RFB-640.tar.gz",`
	`1088`	`+ "model_with_data_sha": "07cc0e284b7924bd89f2ec103254ef3ab4b673fd12341153faaff07f3c1137e3",`
	`1089`	`+ "model_with_data_bytes": 4818743`
`1078`	`1090`	`}`
`1079`	`1091`	`},`
`1080`	`1092`	`{`
`@@ -7446,7 +7458,10 @@`
`7446`	`7458`	`"type": "tensor(float)"`
`7447`	`7459`	`}`
`7448`	`7460`	`]`
`7449`		`- }`
	`7461`	`+ },`
	`7462`	`+ "model_with_data_path": "vision/object_detection_segmentation/yolov2-coco/model/yolov2-coco-9.tar.gz",`
	`7463`	`+ "model_with_data_sha": "c34966e4e96165f8db6f2549f509b6a8bdfc01ddd978e6fd07daad4e665d5383",`
	`7464`	`+ "model_with_data_bytes": 191439022`
`7450`	`7465`	`}`
`7451`	`7466`	`},`
`7452`	`7467`	`{`
Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,3 @@`
	`1`	`+version https://git-lfs.github.com/spec/v1`
	`2`	`+oid sha256:628d0dd3e0288adb821f211e13d4e97f6d6f4527237339606732dffa6f19d381`
	`3`	`+size 2015397`
Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,3 @@`
	`1`	`+version https://git-lfs.github.com/spec/v1`
	`2`	`+oid sha256:07cc0e284b7924bd89f2ec103254ef3ab4b673fd12341153faaff07f3c1137e3`
	`3`	`+size 4818743`
Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,3 @@`
	`1`	`+version https://git-lfs.github.com/spec/v1`
	`2`	`+oid sha256:c34966e4e96165f8db6f2549f509b6a8bdfc01ddd978e6fd07daad4e665d5383`
	`3`	`+size 191439022`