SONG Ge comments

Results 164 comments of


                                            SONG Ge

Setting `class_weight` in `fit()` causes error: `AttributeError: 'DistributedDatasetsFromFunction' object has no attribute 'map'`

Hi sorry for the later reply. We have reproduced and looked into this issue. ### Reason This is a Tensorflow limit. Orca tf2 estimator will use `experimental_distribute_dataset ` strategy when...

Setting `class_weight` in `fit()` causes error: `AttributeError: 'DistributedDatasetsFromFunction' object has no attribute 'map'`

> > Hi sorry for the later reply. We have reproduced and looked into this issue. > > ### Reason > > This is a Tensorflow limit. Orca tf2 estimator...

Setting `class_weight` in `fit()` causes error: `AttributeError: 'DistributedDatasetsFromFunction' object has no attribute 'map'`

> > > > Hi sorry for the later reply. We have reproduced and looked into this issue. > > > > ### Reason > > > > This is...

Setting `class_weight` in `fit()` causes error: `AttributeError: 'DistributedDatasetsFromFunction' object has no attribute 'map'`

> > > > > > Hi sorry for the later reply. We have reproduced and looked into this issue. > > > > > > ### Reason > >...

Use spark backend got error:NotImplementedError: Saving the model to HDF5 format requires the model to be a Functional model or a Sequential model. It does not work for subclassed models, because such models are defined via the body of a Python method, which isn't safely serializable. Consider saving to the Tensorflow SavedModel format (by setting save_format="tf") or using `save_weights`.

@Alxe1 Hey sorry for the late reply, would you mind providing the code you build the model (`model_creator`) and save it (`est.save()`)?

Use spark backend got error:NotImplementedError: Saving the model to HDF5 format requires the model to be a Functional model or a Sequential model. It does not work for subclassed models, because such models are defined via the body of a Python method, which isn't safely serializable. Consider saving to the Tensorflow SavedModel format (by setting save_format="tf") or using `save_weights`.

> @sgwhat Code: > > ``` > def model_creator(config): > deep_cross = DeepCross(user_num=config["uid_num"], > item_num=config["item_num"], > user_item_dim=16, > sparse_num=config["sparse_num"], > feature_embed_dim=16, > embed_norm=0.001, > dnn_hidden_units=[int(e) for e in [128, 64,...

Use spark backend got error:NotImplementedError: Saving the model to HDF5 format requires the model to be a Functional model or a Sequential model. It does not work for subclassed models, because such models are defined via the body of a Python method, which isn't safely serializable. Consider saving to the Tensorflow SavedModel format (by setting save_format="tf") or using `save_weights`.

> the DeepCross model is a user defined subclassed model, in tf2, it cannot save as h5 file. Does there any arguments in `__init__` of `DeepCross`？

Use spark backend got error:NotImplementedError: Saving the model to HDF5 format requires the model to be a Functional model or a Sequential model. It does not work for subclassed models, because such models are defined via the body of a Python method, which isn't safely serializable. Consider saving to the Tensorflow SavedModel format (by setting save_format="tf") or using `save_weights`.

> > > the DeepCross model is a user defined subclassed model, in tf2, it cannot save as h5 file. > > > > > > Does there any arguments...

Use spark backend got error:NotImplementedError: Saving the model to HDF5 format requires the model to be a Functional model or a Sequential model. It does not work for subclassed models, because such models are defined via the body of a Python method, which isn't safely serializable. Consider saving to the Tensorflow SavedModel format (by setting save_format="tf") or using `save_weights`.

> > > > > > the DeepCross model is a user defined subclassed model, in tf2, it cannot save as h5 file. > > > > > > >...

Use spark backend got error:NotImplementedError: Saving the model to HDF5 format requires the model to be a Functional model or a Sequential model. It does not work for subclassed models, because such models are defined via the body of a Python method, which isn't safely serializable. Consider saving to the Tensorflow SavedModel format (by setting save_format="tf") or using `save_weights`.

> > > > > > > > the DeepCross model is a user defined subclassed model, in tf2, it cannot save as h5 file. > > > > >...