microsoft / onnxconverter-common

Common utilities for ONNX converters

microsoft/onnxconverter-common Issues

convert to FP16 generate orphan and self-recurring nodes
Updated 9 days ago2
[Error] Load fp16
Updated 11 days ago2
Performance degrade after sess_options.enable_profiling = True
Updated 12 days ago1
convert float32 model to float16, but memory usage has not decreased
Updated 12 days ago5
protobuf version
Closed a month ago4
onnxconverter_common.auto_mixed_precision.auto_convert_mixed_precision never ends
Updated a month ago1
convert_float_to_float16() produces a model that causes ValidationError with onnx.checker.check_model()
Updated a month ago6
Converting model fp32 to fp16 with auto_mixed_precision_model_path from gets NaN
Updated a month ago1
`auto_convert_mixed_precision` Error: two nodes with same node name error occurred during
Closed 9 months ago5
FP16 conversion yields an unusable model
Updated a month ago2
Is there any upgrades on onnxconverter-common?
Updated a month ago1
resize op convert to FP16 fail
Updated a month ago4
#Onnx Quantisation
Updated a month ago2
Integrate with ONNX 1.16.0 release branch
Closed a month ago1
Redundant dependencies in requirements.txt
Closed a month ago1
Documentation
Closed a month ago1
Extra Cast nodes causes overflow in onnxruntime 1.17
Closed a month ago1
How to convert a model to mixed precision?
Closed 2 months ago21
FP16 model can not get acceleration on GPU with ONNXRuntime-GPU
Closed 2 months ago2
Add `convert_float_to_bfloat16` function to avoid overflow
Updated 6 months ago
Failing tests
Closed 2 years ago3
onnx.onnx_cpp2py_export.checker.ValidationError: Nodes in a graph must be topologically sorted, however input 'Resize__139_input_cast_1' of node: name: Resize__139 OpType: Resize is not output of any previous nodes.
Updated 9 months ago2
Issues when converting model to float16
Closed 2 years ago3
Inference issue after convert_float_to_float16
Closed 2 years ago2
fp32 to fp16
Closed 2 years ago2
Fp32 to Fp16 conversion failure for CSNet
Closed 2 years ago2
Publish source distribution to pypi
Closed 2 years ago2
1.12.2 version released？
Closed 2 years ago1
Fp32-->fp16: original fp32 model works well with input data, but converted fp16 model failed with the same input data
Closed 2 years ago
Source tarball for v1.9.0 republished?
Closed 2 years ago
Add bounds warning to FP16 conversion script
Closed 2 years ago1
Error: onnx.onnx_cpp2py_export.checker.ValidationError: Nodes in a graph must be topologically sorted, however input 'TopK_111_input_cast_0' of node:
Closed 2 years ago1
publish onnxconverter package in anaconda
Closed 2 years ago1
support sizes for Resize op
Closed 2 years ago4
`Unsupported shape calculation for operator mlProgram` while using `onnxmltools.convert_coreml`
Closed 2 years ago1
F16 file does not convert correctly
Closed 2 years ago3
Problem of converting model from 32-bit to 16-bit
Closed 2 years ago1
StrictVersion is deprecated
Closed 2 years ago1
Security Development Lifecycle review for 2022-06
Closed 2 years ago2
"No space left on device" issue on auto_convert_mixed_precision_model_path()
Closed 2 years ago1
add NOTICE file to onnxconverter-common
Closed 2 years ago1
verify onnx 1.12.0 rc
Closed 2 years ago3
Fp16 model runs slower than fp32 model
Closed 2 years ago4
auto_convert_mixed_precision() doesn't support >2GB model
Closed 2 years ago2
regarding the keep_io_types in float16 converter
Closed 3 years ago1
convert to .onnx fail with tensorflow2.4
Closed 3 years ago1
tensorflow max_pool_with_argmax op does not return indices
Closed 3 years ago1
Bad opitmization for MergePadConvOptimizer
Closed 3 years ago2
apply_relu6 was deprecated and remove.
Closed 4 years ago3
Test test_to_onnx_type fails in version 1.6.0
Closed 4 years ago1