keras-team/tf-keras issues and pull requests

#170 - Add max false positive rate option to tf.keras.metrics.AUC

Issue - State: open - Opened by 9las over 1 year ago - 12 comments
Labels: stat:contributions welcome, type:feature

#159 - AdamW does not load for tensorflow 2.11.0

Issue - State: closed - Opened by maerory about 1 year ago - 15 comments
Labels: stat:awaiting response from contributor, stale

#159 - AdamW does not load for tensorflow 2.11.0

Issue - State: closed - Opened by maerory about 1 year ago - 15 comments
Labels: stat:awaiting response from contributor, stale

#151 - Keras.callbacks.ModelCheckpoint doesn't work

Issue - State: closed - Opened by maxkskhor about 1 year ago - 11 comments
Labels: stat:awaiting response from contributor

#151 - Keras.callbacks.ModelCheckpoint doesn't work

Issue - State: closed - Opened by maxkskhor about 1 year ago - 11 comments
Labels: stat:awaiting response from contributor

#141 - Tensorflow (model.predict) still prints progress bars when verbose=0

Issue - State: closed - Opened by sadegh6383 about 1 year ago - 4 comments
Labels: stat:awaiting response from contributor, stale

#141 - Tensorflow (model.predict) still prints progress bars when verbose=0

Issue - State: closed - Opened by sadegh6383 about 1 year ago - 7 comments
Labels: stat:awaiting response from contributor, stale

#140 - Apple Silicon M2 Loss Goes Up After 1-2 (Few) Epochs but not on Google Colab

Issue - State: closed - Opened by felixm3 about 1 year ago - 12 comments
Labels: stat:awaiting response from contributor

#135 - Calling a model and using model.predict yields different results for the same input

Issue - State: closed - Opened by tobiasarndt about 1 year ago - 5 comments
Labels: stat:awaiting response from contributor, stale

#128 - Support input of temporal sample_weights for model training on ragged tensors

Issue - State: open - Opened by yashk2810 about 3 years ago - 2 comments
Labels: stat:contributions welcome, stale

#127 - Cannot load back model with no-op Concatenate layer

Issue - State: open - Opened by stefanistrate almost 3 years ago - 15 comments
Labels: stat:awaiting keras-eng, type:bug

#124 - tensorflow/core/util/gpu_launch_config.h:129] Check failed: work_element_count > 0 (-1018167296 vs. 0)

Issue - State: open - Opened by csteel45 almost 3 years ago - 32 comments

#123 - Scoring for metrics and losses,

Issue - State: open - Opened by EngrStudent almost 3 years ago - 8 comments
Labels: stat:contributions welcome, type:feature, stale

#115 - model.save fails with ValueError __inference_conv2d_transpose_layer_call_fn_4530 when Conv2DTranspose is quantization aware

Issue - State: open - Opened by ypsilon-elle over 2 years ago - 8 comments

#107 - Gradient accumulation support?

Issue - State: open - Opened by andreped over 2 years ago - 28 comments
Labels: type:feature

#103 - Type inference failed. This indicates an invalid graph that escaped type checking. Error message: INVALID_ARGUMENT

Issue - State: open - Opened by sky712345678 about 2 years ago - 21 comments

#98 - Support gradient checkpointing for functional models via the standard fit method

Issue - State: open - Opened by rivershah over 1 year ago - 8 comments
Labels: type:feature, type:performance

#98 - Support gradient checkpointing for functional models via the standard fit method

Issue - State: open - Opened by rivershah over 1 year ago - 8 comments
Labels: type:feature, type:performance

#83 - [BUG] Model state is not correctly after saving and loading

Issue - State: closed - Opened by yangw1234 over 2 years ago - 10 comments
Labels: stat:awaiting response from contributor, stale

#81 - Dropout during prediction phase

Issue - State: closed - Opened by ProfessorHT over 1 year ago - 7 comments
Labels: stat:awaiting response from contributor, stale

#77 - Using StringLookup as 1st layer in a Sequential model raises UnimplementedError since TF 2.8.0

Issue - State: closed - Opened by ageron over 2 years ago - 8 comments
Labels: stat:awaiting response from contributor, stale

#77 - Using StringLookup as 1st layer in a Sequential model raises UnimplementedError since TF 2.8.0

Issue - State: closed - Opened by ageron over 2 years ago - 8 comments
Labels: stat:awaiting response from contributor, stale

#75 - layers.pop no longer removes the last layer

Issue - State: open - Opened by PolarBean over 1 year ago - 13 comments

#74 - Add Rotary Position Embedding, or RoPE, layer in Keras or Keras NLP layers

Issue - State: closed - Opened by anish9 over 1 year ago - 3 comments
Labels: stat:awaiting keras-eng, type:feature

#66 - InvalidArgumentError: Graph execution error:

Issue - State: open - Opened by David-hosting almost 2 years ago - 52 comments
Labels: stat:awaiting keras-eng, type:bug

#62 - New optimizers fail to load CUDA installed through conda

Issue - State: open - Opened by drasmuss over 1 year ago - 14 comments
Labels: type:bug

#56 - MultiHeadAttention quietly ignores masking when annotated with tf.function

Issue - State: open - Opened by tsdeng over 1 year ago - 4 comments
Labels: type:bug

#55 - Support for `MelSpectrogram` layer

Issue - State: closed - Opened by awsaf49 over 1 year ago - 6 comments
Labels: stat:awaiting keras-eng, type:feature

#52 - [Feature Request]: Warn user directly when custom loss is not differentiable

Issue - State: closed - Opened by Frightera over 1 year ago - 13 comments
Labels: stat:awaiting response from contributor, stat:awaiting keras-eng, type:feature, stale

#46 - keras create/load model fails on Apple M2 cpu. Same code works on Ryzen

Issue - State: closed - Opened by drmatthewclark about 1 year ago - 31 comments
Labels: type:bug

#46 - keras create/load model fails on Apple M2 cpu. Same code works on Ryzen

Issue - State: closed - Opened by drmatthewclark about 1 year ago - 30 comments
Labels: type:bug

#45 - Slow training for tf.keras.layers.Embedding with the new optimizers

Issue - State: open - Opened by vysarge about 1 year ago - 5 comments
Labels: type:performance

#44 - tf.keras.models.model_from_json() missing a safe_mode parameter

Issue - State: closed - Opened by epetrovski about 1 year ago - 13 comments
Labels: stat:awaiting response from contributor, type:bug

#42 - Serialization/deserialization of h5 model failed with custom initializer (TF 2.13)

Issue - State: closed - Opened by cofri about 1 year ago - 8 comments
Labels: stat:awaiting response from contributor, type:bug, stale

#42 - Serialization/deserialization of h5 model failed with custom initializer (TF 2.13)

Issue - State: closed - Opened by cofri about 1 year ago - 8 comments
Labels: stat:awaiting response from contributor, type:bug, stale

#41 - TensorFlow 'NoneType' object is not subscriptable during fit()

Issue - State: open - Opened by vrunm about 1 year ago - 4 comments

#39 - TF 2.13 error load h5 with tf.keras.initializers.Constant of numpy

Issue - State: closed - Opened by Chizkiyahu about 1 year ago - 8 comments
Labels: stat:awaiting response from contributor

#39 - TF 2.13 error load h5 with tf.keras.initializers.Constant of numpy

Issue - State: closed - Opened by Chizkiyahu about 1 year ago - 8 comments
Labels: stat:awaiting response from contributor

#37 - Unable to deserialize Sequential model from config

Issue - State: open - Opened by shkarupa-alex about 1 year ago - 2 comments
Labels: type:bug

#37 - Unable to deserialize Sequential model from config

Issue - State: open - Opened by shkarupa-alex about 1 year ago - 2 comments
Labels: type:bug

#36 - Wrong weight names after deserializing model from config

Issue - State: open - Opened by shkarupa-alex about 1 year ago - 5 comments
Labels: type:bug

#33 - tf.keras.metrics.F1Score produces ValueError: Tensor conversion requested dtype float32 for Tensor with dtype int32

Issue - State: open - Opened by meekus-fischer about 1 year ago - 7 comments

#33 - tf.keras.metrics.F1Score produces ValueError: Tensor conversion requested dtype float32 for Tensor with dtype int32

Issue - State: open - Opened by meekus-fischer about 1 year ago - 8 comments

GitHub / keras-team/tf-keras issues and pull requests