issues
search
PINTO0309
/
onnx2tf
Self-Created Tools to convert ONNX files (NCHW) to TensorFlow/TFLite/Keras format (NHWC). The purpose of this tool is to solve the massive Transpose extrapolation problem in onnx-tensorflow (onnx-tf). I don't need a Star, but give me a pull request.
MIT License
712
stars
73
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Added optimization process for `Resize` with undefined dimensions
#523
PINTO0309
closed
1 year ago
0
Fix `Gather` parameter substitution logic
#522
PINTO0309
closed
1 year ago
0
Cannot use converted model with dynamic input shape
#521
sklum
closed
1 year ago
6
Slicing model into parts and separately quantizing each of them.
#520
stupidcucumber
closed
1 year ago
4
Improved stability of `Float16 Pow` conversion
#519
PINTO0309
closed
1 year ago
0
tensorflow==2.14.0, onnx==1.14.1, onnxruntime==1.16.0
#518
PINTO0309
closed
1 year ago
0
Improved conversion stability for tensors with `None` geometry
#517
PINTO0309
closed
1 year ago
0
Fixed a conversion bug in the `Dropout`, `GRU`, `LSTM`, `RNN` OP for a very old opset
#516
PINTO0309
closed
1 year ago
0
`converter._experimental_disable_batchmatmul_unfold` -> `converter.unfold_batchmatmul`
#514
PINTO0309
closed
1 year ago
0
Fixed reference paths for `Lambda` functions.
#513
PINTO0309
closed
1 year ago
0
Significantly faster `Softmax`, `InstanceNormalization`, `Add`, `Sub`, `Div`, `Mul` and `Mod` conversions when converting models too large to perform dummy inference
#512
PINTO0309
closed
1 year ago
0
Significantly improved the efficiency of `MatMul` error correction process when dummy inference is not possible.
#511
PINTO0309
closed
1 year ago
0
Added display of log messages
#510
PINTO0309
closed
1 year ago
0
`np.dot` -> `np.matmul`, `np.float64` -> `np.float32`
#509
PINTO0309
closed
1 year ago
0
Fixed to output all TF, h5, and keras formats when outputting .h5
#508
PINTO0309
closed
1 year ago
0
Cannot make the .pb predict at all
#507
eladmeir
closed
1 year ago
12
Ubuntu22.04, tensorflow v2.14.0rc1, ml_dtypes, flatc v23.5.26
#506
PINTO0309
closed
1 year ago
0
Added conversion progress indicator
#505
PINTO0309
closed
1 year ago
0
Improved stability of `MatMul` transformations when the output tensor is too large to allow dummy inference
#504
PINTO0309
closed
1 year ago
0
Improved `ScatterND` transposition determination algorithm
#503
PINTO0309
closed
1 year ago
0
Improved optimization of non-optimizable `OneHot` with onnx-optimizer and onnxsim
#502
PINTO0309
closed
1 year ago
0
Added NHWC information propagation processing to `DepthToSparse` and improved conversion stability
#501
PINTO0309
closed
1 year ago
0
Added correction process for `Softmax` and `GridSample` accuracy degradation
#500
PINTO0309
closed
1 year ago
0
Improved model conversion stability by adding an initialization process for useless output shapes.
#499
PINTO0309
closed
1 year ago
0
Optimization of `Selu`
#498
PINTO0309
closed
1 year ago
0
Significantly improved conversion success rate for models containing `Einsum`
#497
PINTO0309
closed
1 year ago
0
Implemented workaround for shape correction when final output shapes do not match
#496
PINTO0309
closed
1 year ago
0
Add handling of cases where the first or second tensor is 1-dimensional
#495
PINTO0309
closed
1 year ago
0
Fix `Slice` degrades
#494
PINTO0309
closed
1 year ago
0
Improved processing stability for `Where` and `GatherND`
#493
PINTO0309
closed
1 year ago
0
Fixed a problem in which tensors were automatically transposed when the input tensor consisted of multiple values in one dimension
#492
PINTO0309
closed
1 year ago
0
Improve accuracy of NHWC decision in `Concat` and support edge cases in `Slice`
#491
PINTO0309
closed
1 year ago
0
Improved `Einsum` conversion stability
#490
PINTO0309
closed
1 year ago
0
Major overhaul of `ScatterND` process
#489
PINTO0309
closed
1 year ago
0
Improved stability of the process of calculating expected RAM consumption
#488
PINTO0309
closed
1 year ago
0
[experimental] Revised the method of determining the position of the axis of the Unsqueeze target.
#487
PINTO0309
closed
1 year ago
0
tf.nn.softmax doesn't support Dim(logits) > 5
#486
babiking
closed
1 year ago
2
Improved stability of `ScatterND` and `Where` conversions
#485
PINTO0309
closed
1 year ago
0
Correction of missing judgment conditions for parameter replacement
#484
PINTO0309
closed
1 year ago
0
Reduces post-concat accuracy degradation when all axes except batch size have the same value
#483
PINTO0309
closed
1 year ago
0
Reduce error frequency of `Concat` of two tensors containing constants and disable tensor swapping during `Sub`, `Div` and `Mod`.
#482
PINTO0309
closed
1 year ago
0
Eliminates unnecessary narrowing of the search for the correct answer axis
#481
PINTO0309
closed
1 year ago
0
Feature/babiking/grid sampler
#480
babiking
closed
1 year ago
0
Reduction of transposition errors when all axis values except batch size are the same
#479
PINTO0309
closed
1 year ago
0
Reshape Unmatched max abs error when doing ONNX and TF output value validation
#478
jlamperez
closed
1 year ago
4
[experimental] `ConvInteger`
#477
PINTO0309
closed
1 year ago
0
Bug fixes in `LSTM`, `GRU`, `RNN`, and `MatMul`
#476
PINTO0309
closed
1 year ago
0
Attempts to force axis correction when the number of axes in the combined tensor do not exactly match
#475
PINTO0309
closed
1 year ago
0
Negative padding supported
#474
PINTO0309
closed
1 year ago
0
[D3Net] Clarification regarding a specific input dimension type
#473
ajithvcoder
closed
1 year ago
8
Previous
Next