trouble in running dataset_cheaker.py

[FelixZhang7]

tensorflow.python.framework.errors_impl.InvalidArgumentError: {{function_node __inference_Dataset_map_parse_tfrecord_98}} Feature: image/encoded (data type: string) is required but could not be found. [[{{node ParseSingleExample/ParseSingleExample}}]] [Op:IteratorGetNextSync]

[peteryuX]

Hi @FelixZhang7 ~ You already downloaded MS-Celeb-1M datasets, then extract and convert them to tfrecord as traning data right?

[FelixZhang7]

yes,i have convert mxrecord to tfrecord

[peteryuX]

That's weird. Do you mind provide me the tf.record you produced.

[FelixZhang7]

i have solved this problem but i got a new one

python3.7/site-packages/tensorflow_core/python/keras/engine/training_utils.py", line 144, in finalize raise ValueError('Empty training data.') ValueError: Empty training data.

[peteryuX]

Had you set your dataset path in the config.file?

[FelixZhang7]

After cheaking the data,i think i got worong dataset,but the ms1m112*112 in face.evoLVe.PyTorch is not available,could you provide baidu drive or google drive?

[peteryuX]

Check this temporal backup.

[FelixZhang7]

I solved the dataset problem,but... "/home/anaconda3/lib/python3.7/site-packages/tensorflow_core/python/keras/engine/network.py", line 708, in call convert_kwargs_to_constants=base_layer_utils.call_context().saving) File "/home/anaconda3/lib/python3.7/site-packages/tensorflow_core/python/keras/engine/network.py", line 870, in _run_internal_graph assert str(id(x)) in tensor_dict, 'Could not compute output ' + str(x) AssertionError: Could not compute output Tensor("ArcHead/Identity:0", shape=(None, 85742), dtype=float32)`

[peteryuX]

So the dataset_checker.py is running well right now? Where the last error massage come from? May you please provide me more information about this? Thanks :)

[FelixZhang7]

Traceback (most recent call last): File "train.py", line 135, in app.run(main) File "/home/anaconda3/lib/python3.7/site-packages/absl/app.py", line 299, in run _run_main(main, args) File "/home/anaconda3/lib/python3.7/site-packages/absl/app.py", line 250, in _run_main sys.exit(main(argv)) File "train.py", line 78, in main logist = model(inputs, training=True) File "/home/anaconda3/lib/python3.7/site-packages/tensorflow_core/python/keras/engine/base_layer.py", line 891, in call outputs = self.call(cast_inputs, *args, **kwargs) File "/home/anaconda3/lib/python3.7/site-packages/tensorflow_core/python/keras/engine/network.py", line 708, in call convert_kwargs_to_constants=base_layer_utils.call_context().saving) File "/home/anaconda3/lib/python3.7/site-packages/tensorflow_core/python/keras/engine/network.py", line 870, in _run_internal_graph assert str(id(x)) in tensor_dict, 'Could not compute output ' + str(x) AssertionError: Could not compute output Tensor("ArcHead/Identity:0", shape=(None, 85742), dtype=float32)

[peteryuX]

Hi @FelixZhang7~ I don't know why the problem present... I training with both binary and online image version tf.record dataset well. But the error message you provided is similar to https://github.com/peteryuX/arcface-tf2/issues/7. Hope this information can help you. Is there any possible still have any difference between with the steps in my repo?

[FelixZhang7]

I solved that according to the link, Thank you so much! Sorry to disturb you for a long time

[LackesLab]

i have solved this problem but i got a new one

python3.7/site-packages/tensorflow_core/python/keras/engine/training_utils.py", line 144, in finalize raise ValueError('Empty training data.') ValueError: Empty training data.

Could you please give additional information how you solved your problem?

Thanks