Fix corner case where output size need to reduce by one in MaxPool #5741

liqunfu · 2023-11-08T05:38:57Z

Description

fix #5711 where output size shall not increase fully to padding part of the input region in ceil mode

Motivation and Context

bug fix

Signed-off-by: Liqun Fu <liqfu@microsoft.com>

onnx/defs/nn/defs.cc

In #106270, the solution managed to solve the [`ceil_model` corner issue](onnx/onnx#5711) with the usage of `get_pool_ceil_padding`. However, padding the ceil in converter side only works when we already know the input shapes, therefore, a regression happens when users want to do dynamic inputs. This PR provides (1) refactor codes with torchlib implementation, (2) add dynamic shapes test, and (3) disable the corner tests with comments saying re-enable it when the [real fix from ONNX](onnx/onnx#5741) is merged. [ghstack-poisoned]

Signed-off-by: Liqun Fu <liqfu@microsoft.com>

onnx/reference/ops/op_max_pool.py

codecov · 2023-11-09T00:53:34Z

Codecov Report

Attention: 6 lines in your changes are missing coverage. Please review.

Comparison is base (8a6b70f) 56.45% compared to head (d5ee359) 56.44%.

Files	Patch %	Lines
onnx/backend/test/case/node/maxpool.py	0.00%	6 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #5741      +/-   ##
==========================================
- Coverage   56.45%   56.44%   -0.01%     
==========================================
  Files         504      504              
  Lines       29872    29881       +9     
  Branches     4489     4491       +2     
==========================================
+ Hits        16863    16866       +3     
- Misses      12191    12197       +6     
  Partials      818      818

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

In #106270, the solution managed to solve the [`ceil_model` corner issue](onnx/onnx#5711) with the usage of `get_pool_ceil_padding`. However, padding the ceil in converter side only works when we already know the input shapes, therefore, a regression happens when users want to do dynamic inputs. This PR provides (1) refactor codes with torchlib implementation, (2) add dynamic shapes test, and (3) disable the corner tests with comments saying re-enable it when the [real fix from ONNX](onnx/onnx#5741) is merged. [ghstack-poisoned]

… inputs" In #106270, the solution managed to solve the [`ceil_model` corner issue](onnx/onnx#5711) with the usage of `get_pool_ceil_padding`. However, padding the ceil in converter side only works when we already know the input shapes, therefore, a regression happens when users want to do dynamic inputs. This PR provides (1) refactor codes with torchlib implementation, (2) add dynamic shapes test, and (3) disable the corner tests with comments saying re-enable it when the [real fix from ONNX](onnx/onnx#5741) is merged. [ghstack-poisoned]

In #106270, the solution managed to solve the [`ceil_model` corner issue](onnx/onnx#5711) with the usage of `get_pool_ceil_padding`. However, padding the ceil in converter side only works when we already know the input shapes, therefore, a regression happens when users want to do dynamic inputs. This PR provides (1) refactor codes with torchlib implementation, (2) add dynamic shapes test, and (3) disable the corner tests with comments saying re-enable it when the [real fix from ONNX](onnx/onnx#5741) is merged. [ghstack-poisoned]

In #106270, the solution managed to solve the [`ceil_model` corner issue](onnx/onnx#5711) with the usage of `get_pool_ceil_padding`. However, padding the ceil in converter side only works when we already know the input shapes, therefore, a regression happens when users want to do dynamic inputs. This PR provides (1) refactor codes with torchlib implementation, (2) add dynamic shapes test, and (3) disable the corner tests with comments saying re-enable it when the [real fix from ONNX](onnx/onnx#5741) is merged. Pull Request resolved: #113318 Approved by: https://github.com/thiagocrepaldi

In pytorch#106270, the solution managed to solve the [`ceil_model` corner issue](onnx/onnx#5711) with the usage of `get_pool_ceil_padding`. However, padding the ceil in converter side only works when we already know the input shapes, therefore, a regression happens when users want to do dynamic inputs. This PR provides (1) refactor codes with torchlib implementation, (2) add dynamic shapes test, and (3) disable the corner tests with comments saying re-enable it when the [real fix from ONNX](onnx/onnx#5741) is merged. Pull Request resolved: pytorch#113318 Approved by: https://github.com/thiagocrepaldi

Signed-off-by: Liqun Fu <liqfu@microsoft.com>

justinchuby · 2024-01-04T17:29:34Z

@fatcat-z do you know this is the expected behavior for tf2onnx?

fatcat-z · 2024-01-05T02:19:11Z

@fatcat-z do you know this is the expected behavior for tf2onnx?

In TF op MaxPool, there is no parameter to determine ceil_mode. At this moment, I don't see the impact to tf2onnx.

justinchuby · 2024-01-05T16:28:48Z

Probably need to skip the test_maxpool_2d_ceil_output_size_reduce_by_one test

…into liqun/max_pool_outsize

Signed-off-by: Liqun Fu <liqfu@microsoft.com>

onnx/backend/test/case/node/maxpool.py

onnx/reference/ops/op_max_pool.py

Signed-off-by: Liqun Fu <liqfu@microsoft.com>

### Description update with ONNX 1.16.0 branch according to https://github.com/microsoft/onnxruntime/blob/main/docs/How_To_Update_ONNX_Dev_Notes.md ONNX 1.16.0 release notes: https://github.com/onnx/onnx/releases/tag/v1.16.0 #### Updated ops for CPU EP: - DequantizeLinear(21) - Added int16 and uint16 support + various optimizer tests - Missing int4 and uint4 support - Missing block dequantization support - QuantizeLinear(21) - Added int16 and uint16 support + various optimizer tests - Missing int4 and uint4 support - Missing block quantization support - Cast(21) - Missing int4 and uint4 support - CastLike(21) - Missing int4 and uint4 support - ConstantOfShape(21) - Missing int4 and uint4 support - Identity(21) - Missing int4 and uint4 support - If(21) - Missing int4 and uint4 support - Loop(21) - Missing int4 and uint4 support - Reshape(21) - Missing int4 and uint4 support - Scan(21) - Missing int4 and uint4 support - Shape(21) - Missing int4 and uint4 support - Size(21) - Missing int4 and uint4 support - Flatten(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support - Pad(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support - Squeeze(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support - Transpose(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support - Unsqueeze(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support #### Unimplemented opset 21 features/ops - int4 and uint4 data type - QLinearMatMul(21) - GroupNormalization(21) - ai.onnx.ml.TreeEnsemble(5) ### Motivation and Context  ### Disabled tests #### ORT Training orttraining/orttraining/test/python/orttraining_test_ort_apis_py_bindings.py - test_ort_custom_ops: Potential shape inference bug for custom ops #### Python quantization unit tests test/onnx/python/quantization (shape inference bug) - test_op_conv_transpose.py: test_quantize_conv_transpose_u8u8_fp16 - test_op_conv_transpose.py: test_quantize_conv_transpose_s8s8_fp16 - test_op_gemm.py: test_quantize_qop_gemm_s8s8 - test_op_gemm.py: test_quantize_qop_gemm_e4m3fn_same - test_op_gemm.py: test_quantize_qop_gemm_e4m3fn_p3 - test_op_matmul.py: test_quantize_matmul_u8u8_f16 - test_op_matmul.py: test_quantize_matmul_s8s8_f16 - test_op_matmul.py: test_quantize_matmul_s8s8_f16_entropy - test_op_matmul.py: test_quantize_matmul_s8s8_f16_percentile - test_op_matmul.py: test_quantize_matmul_s8s8_f16_distribution - test_op_relu.py: test_quantize_qop_relu_s8s8 #### ONNX tests - test_maxpool_2d_ceil_output_size_reduce_by_one: ONNX 1.16.0 fixed a maxpool output size bug and added this test. Enable this test when [ORT PR](#18377) is merged. Refer to original [ONNX PR](onnx/onnx#5741). - test_ai_onnx_ml_tree_ensemble_set_membership_cpu: new unimplemented op ai.onnx.ml.TreeEnsemble - test_ai_onnx_ml_tree_ensemble_single_tree_cpu: same - test_ai_onnx_ml_tree_ensemble_set_membership_cuda: same - test_ai_onnx_ml_tree_ensemble_single_tree_cuda: same - test_cast_INT4_to_FLOAT_cpu: ORT Cast(21) impl doesn't support int4 yet - test_cast_INT4_to_INT8_cpu: same - test_cast_UINT4_to_FLOAT_cpu: same - test_cast_UINT4_to_UINT8_cpu: same - test_cast_INT4_to_FLOAT_cuda - test_cast_INT4_to_INT8_cuda - test_cast_UINT4_to_FLOAT_cuda - test_cast_UINT4_to_UINT8_cuda - test_constantofshape_float_ones_cuda: ConstantOfShape(21) not implemented for cuda - test_constantofshape_int_shape_zero_cuda: same - test_constantofshape_int_zeros_cuda: same - test_flatten_axis0_cuda: Flatten(21) not implemented for cuda - test_flatten_axis1_cuda: same - test_flatten_axis2_cuda: same - test_flatten_axis3_cuda: same - test_flatten_default_axis_cuda: same - test_flatten_negative_axis1_cuda: same - test_flatten_negative_axis2_cuda: same - test_flatten_negative_axis3_cuda: same - test_flatten_negative_axis4_cuda: same - test_qlinearmatmul_2D_int8_float16_cpu: QLinearMatMul(21) for onnx not implemented in ORT yet - test_qlinearmatmul_2D_int8_float32_cpu: same - test_qlinearmatmul_2D_uint8_float16_cpu: same - test_qlinearmatmul_2D_uint8_float32_cpu: same - test_qlinearmatmul_3D_int8_float16_cpu: same - test_qlinearmatmul_3D_int8_float32_cpu: same - test_qlinearmatmul_3D_uint8_float16_cpu: same - test_qlinearmatmul_3D_uint8_float32_cpu: same - test_qlinearmatmul_2D_int8_float16_cuda: same - test_qlinearmatmul_2D_int8_float32_cuda: same - test_qlinearmatmul_2D_uint8_float16_cuda: same - test_qlinearmatmul_2D_uint8_float32_cuda: same - test_qlinearmatmul_3D_int8_float16_cuda: same - test_qlinearmatmul_3D_int8_float32_cuda: same - test_qlinearmatmul_3D_uint8_float16_cuda: same - test_qlinearmatmul_3D_uint8_float32_cuda: same - test_size_cuda: Size(21) not implemented for cuda - test_size_example_cuda: same - test_dequantizelinear_blocked: Missing implementation for block dequant for DequantizeLinear(21) - test_quantizelinear_blocked_asymmetric: Missing implementation for block quant for QuantizeLinear(21) - test_quantizelinear_blocked_symmetric: Missing implementation for block quant for QuantizeLinear(21) --------- Signed-off-by: liqunfu <liqun.fu@microsoft.com> Signed-off-by: Ganesan Ramalingam <grama@microsoft.com> Co-authored-by: Ganesan Ramalingam <grama@microsoft.com> Co-authored-by: George Wu <jywu@microsoft.com> Co-authored-by: adrianlizarraga <adlizarraga@microsoft.com>

### Description update with ONNX 1.16.0 branch according to https://github.com/microsoft/onnxruntime/blob/main/docs/How_To_Update_ONNX_Dev_Notes.md ONNX 1.16.0 release notes: https://github.com/onnx/onnx/releases/tag/v1.16.0 #### Updated ops for CPU EP: - DequantizeLinear(21) - Added int16 and uint16 support + various optimizer tests - Missing int4 and uint4 support - Missing block dequantization support - QuantizeLinear(21) - Added int16 and uint16 support + various optimizer tests - Missing int4 and uint4 support - Missing block quantization support - Cast(21) - Missing int4 and uint4 support - CastLike(21) - Missing int4 and uint4 support - ConstantOfShape(21) - Missing int4 and uint4 support - Identity(21) - Missing int4 and uint4 support - If(21) - Missing int4 and uint4 support - Loop(21) - Missing int4 and uint4 support - Reshape(21) - Missing int4 and uint4 support - Scan(21) - Missing int4 and uint4 support - Shape(21) - Missing int4 and uint4 support - Size(21) - Missing int4 and uint4 support - Flatten(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support - Pad(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support - Squeeze(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support - Transpose(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support - Unsqueeze(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support #### Unimplemented opset 21 features/ops - int4 and uint4 data type - QLinearMatMul(21) - GroupNormalization(21) - ai.onnx.ml.TreeEnsemble(5) ### Motivation and Context  ### Disabled tests #### ORT Training orttraining/orttraining/test/python/orttraining_test_ort_apis_py_bindings.py - test_ort_custom_ops: Potential shape inference bug for custom ops #### Python quantization unit tests test/onnx/python/quantization (shape inference bug) - test_op_conv_transpose.py: test_quantize_conv_transpose_u8u8_fp16 - test_op_conv_transpose.py: test_quantize_conv_transpose_s8s8_fp16 - test_op_gemm.py: test_quantize_qop_gemm_s8s8 - test_op_gemm.py: test_quantize_qop_gemm_e4m3fn_same - test_op_gemm.py: test_quantize_qop_gemm_e4m3fn_p3 - test_op_matmul.py: test_quantize_matmul_u8u8_f16 - test_op_matmul.py: test_quantize_matmul_s8s8_f16 - test_op_matmul.py: test_quantize_matmul_s8s8_f16_entropy - test_op_matmul.py: test_quantize_matmul_s8s8_f16_percentile - test_op_matmul.py: test_quantize_matmul_s8s8_f16_distribution - test_op_relu.py: test_quantize_qop_relu_s8s8 #### ONNX tests - test_maxpool_2d_ceil_output_size_reduce_by_one: ONNX 1.16.0 fixed a maxpool output size bug and added this test. Enable this test when [ORT PR](microsoft#18377) is merged. Refer to original [ONNX PR](onnx/onnx#5741). - test_ai_onnx_ml_tree_ensemble_set_membership_cpu: new unimplemented op ai.onnx.ml.TreeEnsemble - test_ai_onnx_ml_tree_ensemble_single_tree_cpu: same - test_ai_onnx_ml_tree_ensemble_set_membership_cuda: same - test_ai_onnx_ml_tree_ensemble_single_tree_cuda: same - test_cast_INT4_to_FLOAT_cpu: ORT Cast(21) impl doesn't support int4 yet - test_cast_INT4_to_INT8_cpu: same - test_cast_UINT4_to_FLOAT_cpu: same - test_cast_UINT4_to_UINT8_cpu: same - test_cast_INT4_to_FLOAT_cuda - test_cast_INT4_to_INT8_cuda - test_cast_UINT4_to_FLOAT_cuda - test_cast_UINT4_to_UINT8_cuda - test_constantofshape_float_ones_cuda: ConstantOfShape(21) not implemented for cuda - test_constantofshape_int_shape_zero_cuda: same - test_constantofshape_int_zeros_cuda: same - test_flatten_axis0_cuda: Flatten(21) not implemented for cuda - test_flatten_axis1_cuda: same - test_flatten_axis2_cuda: same - test_flatten_axis3_cuda: same - test_flatten_default_axis_cuda: same - test_flatten_negative_axis1_cuda: same - test_flatten_negative_axis2_cuda: same - test_flatten_negative_axis3_cuda: same - test_flatten_negative_axis4_cuda: same - test_qlinearmatmul_2D_int8_float16_cpu: QLinearMatMul(21) for onnx not implemented in ORT yet - test_qlinearmatmul_2D_int8_float32_cpu: same - test_qlinearmatmul_2D_uint8_float16_cpu: same - test_qlinearmatmul_2D_uint8_float32_cpu: same - test_qlinearmatmul_3D_int8_float16_cpu: same - test_qlinearmatmul_3D_int8_float32_cpu: same - test_qlinearmatmul_3D_uint8_float16_cpu: same - test_qlinearmatmul_3D_uint8_float32_cpu: same - test_qlinearmatmul_2D_int8_float16_cuda: same - test_qlinearmatmul_2D_int8_float32_cuda: same - test_qlinearmatmul_2D_uint8_float16_cuda: same - test_qlinearmatmul_2D_uint8_float32_cuda: same - test_qlinearmatmul_3D_int8_float16_cuda: same - test_qlinearmatmul_3D_int8_float32_cuda: same - test_qlinearmatmul_3D_uint8_float16_cuda: same - test_qlinearmatmul_3D_uint8_float32_cuda: same - test_size_cuda: Size(21) not implemented for cuda - test_size_example_cuda: same - test_dequantizelinear_blocked: Missing implementation for block dequant for DequantizeLinear(21) - test_quantizelinear_blocked_asymmetric: Missing implementation for block quant for QuantizeLinear(21) - test_quantizelinear_blocked_symmetric: Missing implementation for block quant for QuantizeLinear(21) --------- Signed-off-by: liqunfu <liqun.fu@microsoft.com> Signed-off-by: Ganesan Ramalingam <grama@microsoft.com> Co-authored-by: Ganesan Ramalingam <grama@microsoft.com> Co-authored-by: George Wu <jywu@microsoft.com> Co-authored-by: adrianlizarraga <adlizarraga@microsoft.com>

…nnx#5741) ### Description fix onnx#5711 where output size shall not increase fully to padding part of the input region in ceil mode ### Motivation and Context bug fix --------- Signed-off-by: Liqun Fu <liqfu@microsoft.com> Co-authored-by: Justin Chu <justinchuby@users.noreply.github.com> Co-authored-by: G. Ramalingam <grama@microsoft.com> Signed-off-by: Linsho Kaku <linsho@preferred.jp>

### Description Pooling in reference currently has two bugs: (1) it has ["pads required"](https://github.com/onnx/onnx/blob/e292b4ae6d016c3231a801bfeb26f802ba95d82a/onnx/reference/ops/op_pool_common.py#L53) to make sure sliding window does not go out-of-bound, but it does not exclude pads required from pooling caculations. This causes microsoft/onnxruntime#16203 (2) In `ceil_mode`, the reference pooling pads the input image, but does not check if sliding window starts on pads. When we reach the end of the image, the window should stop working. This causes pytorch/pytorch#131272. Not directly, but related fix on MaxPooling: #5741. Detail: pytorch/pytorch#116420 (comment) ### Motivation and Context This PR fixes the two bugs and update their tests accordingly. I also drafted a PR to onnxruntime to have this update in CPU provider: microsoft/onnxruntime#16752 --------- Signed-off-by: titaiwangms <titaiwang@microsoft.com> Signed-off-by: Justin Chu <justinchuby@users.noreply.github.com> Signed-off-by: Andreas Fehlner <fehlner@arcor.de> Co-authored-by: Justin Chu <justinchuby@users.noreply.github.com> Co-authored-by: Andreas Fehlner <fehlner@arcor.de>

### Description Pooling in reference currently has two bugs: (1) it has ["pads required"](https://github.com/onnx/onnx/blob/e292b4ae6d016c3231a801bfeb26f802ba95d82a/onnx/reference/ops/op_pool_common.py#L53) to make sure sliding window does not go out-of-bound, but it does not exclude pads required from pooling caculations. This causes microsoft/onnxruntime#16203 (2) In `ceil_mode`, the reference pooling pads the input image, but does not check if sliding window starts on pads. When we reach the end of the image, the window should stop working. This causes pytorch/pytorch#131272. Not directly, but related fix on MaxPooling: #5741. Detail: pytorch/pytorch#116420 (comment) ### Motivation and Context This PR fixes the two bugs and update their tests accordingly. I also drafted a PR to onnxruntime to have this update in CPU provider: microsoft/onnxruntime#16752 --------- Signed-off-by: titaiwangms <titaiwang@microsoft.com> Signed-off-by: Justin Chu <justinchuby@users.noreply.github.com> Signed-off-by: Andreas Fehlner <fehlner@arcor.de> Co-authored-by: Justin Chu <justinchuby@users.noreply.github.com> Co-authored-by: Andreas Fehlner <fehlner@arcor.de> Signed-off-by: Andreas Fehlner <fehlner@arcor.de>

### Description Pooling in reference currently has two bugs: (1) it has ["pads required"](https://github.com/onnx/onnx/blob/e292b4ae6d016c3231a801bfeb26f802ba95d82a/onnx/reference/ops/op_pool_common.py#L53) to make sure sliding window does not go out-of-bound, but it does not exclude pads required from pooling caculations. This causes microsoft/onnxruntime#16203 (2) In `ceil_mode`, the reference pooling pads the input image, but does not check if sliding window starts on pads. When we reach the end of the image, the window should stop working. This causes pytorch/pytorch#131272. Not directly, but related fix on MaxPooling: onnx#5741. Detail: pytorch/pytorch#116420 (comment) ### Motivation and Context This PR fixes the two bugs and update their tests accordingly. I also drafted a PR to onnxruntime to have this update in CPU provider: microsoft/onnxruntime#16752 --------- Signed-off-by: titaiwangms <titaiwang@microsoft.com> Signed-off-by: Justin Chu <justinchuby@users.noreply.github.com> Signed-off-by: Andreas Fehlner <fehlner@arcor.de> Co-authored-by: Justin Chu <justinchuby@users.noreply.github.com> Co-authored-by: Andreas Fehlner <fehlner@arcor.de> Signed-off-by: seungwoo-ji <seungwoo.ji@nuvilab.com>

### Description update with ONNX 1.16.0 branch according to https://github.com/microsoft/onnxruntime/blob/main/docs/How_To_Update_ONNX_Dev_Notes.md ONNX 1.16.0 release notes: https://github.com/onnx/onnx/releases/tag/v1.16.0 #### Updated ops for CPU EP: - DequantizeLinear(21) - Added int16 and uint16 support + various optimizer tests - Missing int4 and uint4 support - Missing block dequantization support - QuantizeLinear(21) - Added int16 and uint16 support + various optimizer tests - Missing int4 and uint4 support - Missing block quantization support - Cast(21) - Missing int4 and uint4 support - CastLike(21) - Missing int4 and uint4 support - ConstantOfShape(21) - Missing int4 and uint4 support - Identity(21) - Missing int4 and uint4 support - If(21) - Missing int4 and uint4 support - Loop(21) - Missing int4 and uint4 support - Reshape(21) - Missing int4 and uint4 support - Scan(21) - Missing int4 and uint4 support - Shape(21) - Missing int4 and uint4 support - Size(21) - Missing int4 and uint4 support - Flatten(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support - Pad(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support - Squeeze(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support - Transpose(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support - Unsqueeze(21) - Missing float8e4m3fnuz, float8e5m2, float8e5m2fnuz, int4, and uint4 support #### Unimplemented opset 21 features/ops - int4 and uint4 data type - QLinearMatMul(21) - GroupNormalization(21) - ai.onnx.ml.TreeEnsemble(5) ### Motivation and Context  ### Disabled tests #### ORT Training orttraining/orttraining/test/python/orttraining_test_ort_apis_py_bindings.py - test_ort_custom_ops: Potential shape inference bug for custom ops #### Python quantization unit tests test/onnx/python/quantization (shape inference bug) - test_op_conv_transpose.py: test_quantize_conv_transpose_u8u8_fp16 - test_op_conv_transpose.py: test_quantize_conv_transpose_s8s8_fp16 - test_op_gemm.py: test_quantize_qop_gemm_s8s8 - test_op_gemm.py: test_quantize_qop_gemm_e4m3fn_same - test_op_gemm.py: test_quantize_qop_gemm_e4m3fn_p3 - test_op_matmul.py: test_quantize_matmul_u8u8_f16 - test_op_matmul.py: test_quantize_matmul_s8s8_f16 - test_op_matmul.py: test_quantize_matmul_s8s8_f16_entropy - test_op_matmul.py: test_quantize_matmul_s8s8_f16_percentile - test_op_matmul.py: test_quantize_matmul_s8s8_f16_distribution - test_op_relu.py: test_quantize_qop_relu_s8s8 #### ONNX tests - test_maxpool_2d_ceil_output_size_reduce_by_one: ONNX 1.16.0 fixed a maxpool output size bug and added this test. Enable this test when [ORT PR](microsoft/onnxruntime#18377) is merged. Refer to original [ONNX PR](onnx/onnx#5741). - test_ai_onnx_ml_tree_ensemble_set_membership_cpu: new unimplemented op ai.onnx.ml.TreeEnsemble - test_ai_onnx_ml_tree_ensemble_single_tree_cpu: same - test_ai_onnx_ml_tree_ensemble_set_membership_cuda: same - test_ai_onnx_ml_tree_ensemble_single_tree_cuda: same - test_cast_INT4_to_FLOAT_cpu: ORT Cast(21) impl doesn't support int4 yet - test_cast_INT4_to_INT8_cpu: same - test_cast_UINT4_to_FLOAT_cpu: same - test_cast_UINT4_to_UINT8_cpu: same - test_cast_INT4_to_FLOAT_cuda - test_cast_INT4_to_INT8_cuda - test_cast_UINT4_to_FLOAT_cuda - test_cast_UINT4_to_UINT8_cuda - test_constantofshape_float_ones_cuda: ConstantOfShape(21) not implemented for cuda - test_constantofshape_int_shape_zero_cuda: same - test_constantofshape_int_zeros_cuda: same - test_flatten_axis0_cuda: Flatten(21) not implemented for cuda - test_flatten_axis1_cuda: same - test_flatten_axis2_cuda: same - test_flatten_axis3_cuda: same - test_flatten_default_axis_cuda: same - test_flatten_negative_axis1_cuda: same - test_flatten_negative_axis2_cuda: same - test_flatten_negative_axis3_cuda: same - test_flatten_negative_axis4_cuda: same - test_qlinearmatmul_2D_int8_float16_cpu: QLinearMatMul(21) for onnx not implemented in ORT yet - test_qlinearmatmul_2D_int8_float32_cpu: same - test_qlinearmatmul_2D_uint8_float16_cpu: same - test_qlinearmatmul_2D_uint8_float32_cpu: same - test_qlinearmatmul_3D_int8_float16_cpu: same - test_qlinearmatmul_3D_int8_float32_cpu: same - test_qlinearmatmul_3D_uint8_float16_cpu: same - test_qlinearmatmul_3D_uint8_float32_cpu: same - test_qlinearmatmul_2D_int8_float16_cuda: same - test_qlinearmatmul_2D_int8_float32_cuda: same - test_qlinearmatmul_2D_uint8_float16_cuda: same - test_qlinearmatmul_2D_uint8_float32_cuda: same - test_qlinearmatmul_3D_int8_float16_cuda: same - test_qlinearmatmul_3D_int8_float32_cuda: same - test_qlinearmatmul_3D_uint8_float16_cuda: same - test_qlinearmatmul_3D_uint8_float32_cuda: same - test_size_cuda: Size(21) not implemented for cuda - test_size_example_cuda: same - test_dequantizelinear_blocked: Missing implementation for block dequant for DequantizeLinear(21) - test_quantizelinear_blocked_asymmetric: Missing implementation for block quant for QuantizeLinear(21) - test_quantizelinear_blocked_symmetric: Missing implementation for block quant for QuantizeLinear(21) --------- Signed-off-by: liqunfu <liqun.fu@microsoft.com> Signed-off-by: Ganesan Ramalingam <grama@microsoft.com> Co-authored-by: Ganesan Ramalingam <grama@microsoft.com> Co-authored-by: George Wu <jywu@microsoft.com> Co-authored-by: adrianlizarraga <adlizarraga@microsoft.com>

fix corner case where output size need to reduce by one

e0488c8

Signed-off-by: Liqun Fu <liqfu@microsoft.com>

liqunfu requested a review from a team as a code owner November 8, 2023 05:38

justinchuby reviewed Nov 8, 2023

View reviewed changes

onnx/defs/nn/defs.cc Show resolved Hide resolved

titaiwangms mentioned this pull request Nov 8, 2023

[onnx] Fix output shape mismatch issue of max_pool pytorch/pytorch#106270

Closed

justinchuby changed the title ~~fix corner case where output size need to reduce by one~~ Fix corner case where output size need to reduce by one Nov 8, 2023

justinchuby added the module: shape inference Issues related to shape inference label Nov 8, 2023

titaiwangms mentioned this pull request Nov 9, 2023

[ONNX] Refactor MaxPool to support dynamic inputs pytorch/pytorch#113318

Closed

fix op_max_pool, update md

10245ca

Signed-off-by: Liqun Fu <liqfu@microsoft.com>

github-advanced-security bot found potential problems Nov 9, 2023

View reviewed changes

onnx/reference/ops/op_max_pool.py Fixed Show fixed Hide fixed

onnx/reference/ops/op_max_pool.py Fixed Show fixed Hide fixed

liqunfu added 3 commits December 23, 2023 21:50

update doc

ce0a42c

Signed-off-by: Liqun Fu <liqfu@microsoft.com>

lint

256460e

Signed-off-by: Liqun Fu <liqfu@microsoft.com>

Merge branch 'main' into liqun/max_pool_outsize

3fb02e9

justinchuby added this to the 1.16 milestone Jan 4, 2024

justinchuby added the review needed: operators approvers Require reviews from members of operators-approvers label Jan 4, 2024

Merge branch 'main' into liqun/max_pool_outsize

2cd9ffc

justinchuby approved these changes Jan 5, 2024

View reviewed changes

justinchuby changed the title ~~Fix corner case where output size need to reduce by one~~ Fix corner case where output size need to reduce by one in MaxPool Jan 17, 2024

justinchuby mentioned this pull request Jan 17, 2024

Ambiguity in Pool spec (AveragePool and MaxPool) #1922

Open

justinchuby requested a review from xadupre January 17, 2024 04:18

Merge branch 'main' into liqun/max_pool_outsize

203808b

liqunfu added the run release CIs Use this label to trigger release tests in CI label Jan 18, 2024

liqunfu added 2 commits January 18, 2024 10:02

Merge branch 'liqun/max_pool_outsize' of https://github.com/onnx/onnx …

9b65b15

…into liqun/max_pool_outsize

skip ort test for the new test case until Ort has the corresponding fix

e1b5dc6

Signed-off-by: Liqun Fu <liqfu@microsoft.com>

liqunfu requested a review from a team as a code owner January 18, 2024 19:54

bumpup lintrunner-adapter version

ef609cb

Signed-off-by: Liqun Fu <liqfu@microsoft.com>

justinchuby approved these changes Jan 18, 2024

View reviewed changes

justinchuby reviewed Jan 18, 2024

View reviewed changes

onnx/backend/test/case/node/maxpool.py Outdated Show resolved Hide resolved

justinchuby reviewed Jan 18, 2024

View reviewed changes

onnx/reference/ops/op_max_pool.py Outdated Show resolved Hide resolved

update with reviewer's comments

b328fe4

Signed-off-by: Liqun Fu <liqfu@microsoft.com>

justinchuby enabled auto-merge January 19, 2024 00:25

justinchuby and others added 2 commits January 18, 2024 16:26

Merge branch 'main' into liqun/max_pool_outsize

c8ae0e2

Merge branch 'main' into liqun/max_pool_outsize

d5ee359

gramalingam approved these changes Jan 29, 2024

View reviewed changes

justinchuby added this pull request to the merge queue Jan 29, 2024

Merged via the queue into main with commit c5d91f6 Jan 29, 2024

justinchuby deleted the liqun/max_pool_outsize branch January 29, 2024 23:30

justinchuby mentioned this pull request Feb 8, 2024

operator MaxPool output tensor shape is ambiguously documented. #2927

Open

adrianlizarraga mentioned this pull request Apr 10, 2024

Integration with ONNX 1.16.0 microsoft/onnxruntime#19745

Merged

titaiwangms mentioned this pull request Jan 23, 2025

Fix pooling pads issues #6650

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Fix corner case where output size need to reduce by one in MaxPool #5741

Fix corner case where output size need to reduce by one in MaxPool #5741

Uh oh!

liqunfu commented Nov 8, 2023

Uh oh!

Uh oh!

Uh oh!

Uh oh!

codecov bot commented Nov 9, 2023 •

edited

Loading

Uh oh!

justinchuby commented Jan 4, 2024

Uh oh!

fatcat-z commented Jan 5, 2024

Uh oh!

justinchuby commented Jan 5, 2024

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Fix corner case where output size need to reduce by one in MaxPool #5741

Fix corner case where output size need to reduce by one in MaxPool #5741

Uh oh!

Conversation

liqunfu commented Nov 8, 2023

Description

Motivation and Context

Uh oh!

Uh oh!

Uh oh!

Uh oh!

codecov bot commented Nov 9, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

justinchuby commented Jan 4, 2024

Uh oh!

fatcat-z commented Jan 5, 2024

Uh oh!

justinchuby commented Jan 5, 2024

Uh oh!

Uh oh!

Uh oh!

Uh oh!

codecov bot commented Nov 9, 2023 •

edited

Loading