Yaohui Liu
3db092b929
Support mutiple inputs of keras modek.predict.
2 years ago
Haiping Chen
8550dccc56
Add missing trackable class but not implemented.
2 years ago
Yaohui Liu
095bf33d7c
Refine the keras SavedModel unittest.
2 years ago
Yaohui Liu
8f6612d9e0
Reduce the time of keras unittest.
2 years ago
Yaohui Liu
ece36e616f
Automatically add KerasInterface to f.
2 years ago
Yaohui Liu
e5837dc8b3
Resolve the comment.
2 years ago
Yaohui Liu
875063abed
Add more explicit conversion for Tensors.
2 years ago
Haiping Chen
1767c3cbdb
GradientTest inherts from EagerModeTestBase.
2 years ago
Yaohui Liu
404c803ce2
Support the multiple inputs of keras model.fit.
2 years ago
Yaohui Liu
7465337b6e
Add default value of parameter of keras softmax.
2 years ago
Yaohui Liu
6b64d4be2f
Fix some typo of readme.
2 years ago
Haiping Chen
cbf2d81d6b
ICallback
2 years ago
Yaohui Liu
76a964f9f2
Refine some details of readme.
2 years ago
Yaohui Liu
4f794e8afa
Refine the resnet example.
2 years ago
Yaohui Liu
69889a3df1
Refine the readme structure.
2 years ago
Yaohui Liu
89c68cdc65
Add Chinese version readme.
2 years ago
Yaohui Liu
9458e838df
Refine some descriptions in readme.md.
2 years ago
Haiping Chen
f4450e9e53
output_names exeption if length is zero.
2 years ago
Haiping Chen
e5dfe90e9b
More SafeHandles.
2 years ago
Haiping Chen
016294d33f
Seperate SafeCheckpointReaderHandle.
2 years ago
Yaohui Liu
559d471407
Align keras.Input with tensorflow python.
2 years ago
Yaohui Liu
b8fd21c094
Fix issue #760
2 years ago
Haiping Chen
afa269699f
MultiThreads Failed on MacOS.
2 years ago
Haiping Chen
9aa8f758f4
Fix sparse_categorical_crossentropy.
2 years ago
Yaohui Liu
52b513d750
Support loading of SavedModel format ( #989 )
* Add CheckpointReader and corresponding C APIs.
* Add essential components of SavedModel format loading.
* Add checkpoint reading for SavedModel format loading.
* Revise customized json converters.
* Add support for loading models from python.
* Fix the duplicated weights in Keras.Model.
* Add alexnet loading test and check for loaded weights.
* Fix ci error caused by branch merge.
* Resolve the comments and errors.
* Fix the stucking of training when loading model.
* Fix the stucking of training when loading model.
* fix intptr.
---------
Co-authored-by: Haiping Chen <haiping008@gmail.com>
2 years ago
Haiping Chen
45f26269d8
Add SparseCategoricalCrossentropy and SparseCategoricalAccuracy.
2 years ago
Haiping Chen
2109063386
Fix FBetaScore data type issue.
2 years ago
Haiping Chen
3b0cdd878c
Fix double data type issue in hamming_loss_fn.
2 years ago
Haiping Chen
e5dc65a7d9
Support SigmoidFocalCrossEntropy, better for imbalanced multi-class task.
2 years ago
Haiping Chen
55cc4d0b78
Support save and load in numpy.lib.format.
2 years ago
Haiping Chen
d794576486
Add activations.mish.
2 years ago
Haiping Chen
c72f47990e
Add metric of HammingLoss.
2 years ago
Haiping Chen
32a3e48d19
All threshold as null in F1 Score.
2 years ago
Haiping Chen
9726fc6f81
Fix name of f1_score.
2 years ago
Haiping Chen
922139ff75
Set mean of all classes in F1 Score.
2 years ago
Haiping Chen
067c1ff92a
Add metrics of F1Score and FBetaScore.
2 years ago
Haiping Chen
77eb6f5e71
Add metrics of Accuracy and CosineSimilarity.
2 years ago
Haiping Chen
9e877d1c15
Add metrics of BinaryAccuracy, CategoricalAccuracy, CategoricalCrossentropy.
2 years ago
Haiping Chen
98919983b1
Add metrics of Precision.
2 years ago
Haiping Chen
217cfd2d41
Add mertic of Recall.
2 years ago
Haiping Chen
a5289b9bb3
Abstract IMetricFunc.
2 years ago
Haiping Chen
ca9f574fce
Add metric of top_k_categorical_accuracy.
2 years ago
Haiping Chen
b8645d3f83
Add keras.layers.CategoryEncoding.
2 years ago
AsakusaRinne
2ee08e8ce9
Fix the keras.sparse_categorical_crossentropy. ( #985 )
2 years ago
Superpiffer
5821275145
Reimplemented NDArray == and != operators, handling null values. Added unit tests.
2 years ago
Haiping Chen
ccda2c39ff
Add BinaryCrossentropy loss function.
2 years ago
Haiping Chen
8a21ad2816
Fix math_ops.max and min.
2 years ago
AsakusaRinne
a44028ed23
Fix ConvTraspose2D gradient and register rsqrt gradient.
2 years ago
Superpiffer
a7c9a75954
Use a local Status variable
Using a local reference ensure that the Status object cannot be disposed before the Dispose. This way it's also possible to use an external Status instance instead of the static one, if needed.
2 years ago
Haiping Chen
0ee50d319e
Add double to NDArrayConverter.
2 years ago