1.0.8

• Additional indication of size and number of parameters

Kazam_screencast_00002_.mp4

1.0.7

Added display of parameter size for each OP.

Kazam_screencast_00001_.mp4

What's Changed

• Added display of parameter size for each OP by @PINTO0309 in #4

Full Changelog: 1.0.6...1.0.7

1.0.6

• thousands separator
  

What's Changed

• thousands separator by @PINTO0309 in #3

Full Changelog: 1.0.5...1.0.6

1.0.5

• Extended supported data types.

  ONNX_DTYPES_TO_NUMPY_DTYPES: dict = {
      f'{onnx.TensorProto.FLOAT16}': np.float16,
      f'{onnx.TensorProto.FLOAT}': np.float32,
      f'{onnx.TensorProto.DOUBLE}': np.float64,
      f'{onnx.TensorProto.INT8}': np.int8,
      f'{onnx.TensorProto.INT16}': np.int16,
      f'{onnx.TensorProto.INT32}': np.int32,
      f'{onnx.TensorProto.INT64}': np.int64,
      f'{onnx.TensorProto.UINT8}': np.uint8,
      f'{onnx.TensorProto.UINT16}': np.uint16,
      f'{onnx.TensorProto.UINT32}': np.uint32,
      f'{onnx.TensorProto.UINT64}': np.uint64,
  }

What's Changed

• Extended supported data types by @PINTO0309 in #2

New Contributors

• @PINTO0309 made their first contribution in #2

Full Changelog: 1.0.4...1.0.5

1.0.4

• Add short form

  $ ssc4onnx -h
  
  usage:
      ssc4onnx [-h]
      -if INPUT_ONNX_FILE_PATH
  
  optional arguments:
    -h, --help
          show this help message and exit.
  
    -if INPUT_ONNX_FILE_PATH, --input_onnx_file_path INPUT_ONNX_FILE_PATH
          Input onnx file path.
  

1.0.3

bugfix and add return value #1 @fateshelled

1. Fixed input_onnx_file_path must be specified.
2. Added return values.
   • op_nums
   • model_size

1.0.2

• producer
• opset

  ┏━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
  ┃ OP Type                ┃ OPs        ┃
  ┡━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
  │ Add                    │ 3907       │
  │ AveragePool            │ 3          │
  │ Cast                   │ 2652       │
  │ Concat                 │ 1983       │
  │ Constant               │ 14992      │
  │ ConstantOfShape        │ 1350       │
  │ Conv                   │ 710        │
  │ Div                    │ 1107       │
  │ Einsum                 │ 353        │
  │ Equal                  │ 1240       │
  │ Expand                 │ 1662       │
  │ Floor                  │ 416        │
  │ Gather                 │ 1411       │
  │ GatherElements         │ 832        │
  │ Greater                │ 832        │
  │ InstanceNormalization  │ 15         │
  │ Less                   │ 832        │
  │ MatMul                 │ 1          │
  │ Mul                    │ 5267       │
  │ Neg                    │ 206        │
  │ Not                    │ 102        │
  │ Pad                    │ 212        │
  │ Range                  │ 206        │
  │ ReduceSum              │ 14         │
  │ Relu                   │ 352        │
  │ Reshape                │ 2410       │
  │ ScatterND              │ 102        │
  │ Shape                  │ 1556       │
  │ Sigmoid                │ 208        │
  │ Slice                  │ 620        │
  │ Softmax                │ 1          │
  │ Split                  │ 208        │
  │ Sqrt                   │ 13         │
  │ Sub                    │ 2446       │
  │ Tanh                   │ 104        │
  │ Tile                   │ 2          │
  │ Transpose              │ 317        │
  │ Unsqueeze              │ 3866       │
  │ Where                  │ 2904       │
  │ ---------------------- │ ---------- │
  │ Total number of OPs    │ 55414      │
  │ ====================== │ ========== │
  │ Model Size             │ 37.2MiB    │
  └────────────────────────┴────────────┘
  INFO: file: deqflow_b_things_opset12_192x320.onnx
  INFO: producer: pytorch 1.11.0
  INFO: opset: 12
  INFO: input_name.1: input1 shape: [1, 3, 192, 320] dtype: float32
  INFO: input_name.2: input2 shape: [1, 3, 192, 320] dtype: float32
  INFO: output_name.1: flow_up shape: [1, 2, 192, 320] dtype: float32
  

1.0.1

• Total number of OPs

  ┏━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓
  ┃ OP Type                ┃ OPs        ┃
  ┡━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩
  │ Add                    │ 3907       │
  │ AveragePool            │ 3          │
  │ Cast                   │ 2652       │
  │ Concat                 │ 1983       │
  │ Constant               │ 14992      │
  │ ConstantOfShape        │ 1350       │
  │ Conv                   │ 710        │
  │ Div                    │ 1107       │
  │ Einsum                 │ 353        │
  │ Equal                  │ 1240       │
  │ Expand                 │ 1662       │
  │ Floor                  │ 416        │
  │ Gather                 │ 1411       │
  │ GatherElements         │ 832        │
  │ Greater                │ 832        │
  │ InstanceNormalization  │ 15         │
  │ Less                   │ 832        │
  │ MatMul                 │ 1          │
  │ Mul                    │ 5267       │
  │ Neg                    │ 206        │
  │ Not                    │ 102        │
  │ Pad                    │ 212        │
  │ Range                  │ 206        │
  │ ReduceSum              │ 14         │
  │ Relu                   │ 352        │
  │ Reshape                │ 2410       │
  │ ScatterND              │ 102        │
  │ Shape                  │ 1556       │
  │ Sigmoid                │ 208        │
  │ Slice                  │ 620        │
  │ Softmax                │ 1          │
  │ Split                  │ 208        │
  │ Sqrt                   │ 13         │
  │ Sub                    │ 2446       │
  │ Tanh                   │ 104        │
  │ Tile                   │ 2          │
  │ Transpose              │ 317        │
  │ Unsqueeze              │ 3866       │
  │ Where                  │ 2904       │
  │ ---------------------- │ ---------- │
  │ Total number of OPs    │ 55414      │
  │ ====================== │ ========== │
  │ Model Size             │ 37.2MiB    │
  └────────────────────────┴────────────┘
  INFO: file: deqflow_b_things_opset12_192x320.onnx
  INFO: input_name.1: input1 shape: [1, 3, 192, 320] dtype: float32
  INFO: input_name.2: input2 shape: [1, 3, 192, 320] dtype: float32
  INFO: output_name.1: flow_up shape: [1, 2, 192, 320] dtype: float32
  

1.0.0

• Initial release