calmilovesai / basic_cnns_tensorflow2 Goto Github PK

Which version of CUDA should tensorflow-gpu2.2.0rc3 correspond to? I use cuda10.0 under Ubuntu 18.04.4, and tensorflow-gpu2.2.0rc3 cannot use GPU

在这里感谢楼主的工作，很多地方值得借鉴、、
同时我在训练的过程中，完全没法识别。。应该是楼主没有用什么增强操作吧 直接硬上了

I tried all models but this error happened every time after some steps in the first epoch.

Traceback (most recent call last):
File "D:\Py\Basic_CNNs_TensorFlow2-master\Basic_CNNs_TensorFlow2-master\train.py", line 150, in
images, labels = process_features(features, data_augmentation=True)
File "D:\Py\Basic_CNNs_TensorFlow2-master\Basic_CNNs_TensorFlow2-master\train.py", line 94, in process_features
image_tensor = load_and_preprocess_image(image, data_augmentation=data_augmentation)
File "D:\Py\Basic_CNNs_TensorFlow2-master\Basic_CNNs_TensorFlow2-master\prepare_data.py", line 17, in load_and_preprocess_image
image = tf.image.random_crop(value=image, size=[IMAGE_HEIGHT, IMAGE_WIDTH, CHANNELS])
File "C:\Users\Viking\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow_core\python\ops\random_ops.py", line 348, in random_crop
math_ops.reduce_all(shape >= size),
File "C:\Users\Viking\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow_core\python\ops\gen_math_ops.py", line 4077, in greater_equal
_ops.raise_from_not_ok_status(e, name)
File "C:\Users\Viking\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow_core\python\framework\ops.py", line 6606, in raise_from_not_ok_status
six.raise_from(core._status_to_exception(e.code, message), None)
File "", line 3, in raise_from
tensorflow.python.framework.errors_impl.InvalidArgumentError: Incompatible shapes: [4] vs. [3] [Op:GreaterEqual]

Thanks.

Do you do ./255 each image before training?

训练数据为什么不需要给出坐标，而是直接一张图呀

new to TF2

It confuses me that there is no parameters in all denseblocks. Is there any problem in the program?

How to write the code to assign different gradient weights to different classes of SparseCategoricalCrossentropy

我在使用GPU进行RegNet训练的时候，报错了

UnimplementedError:  Fused conv implementation does not support grouped convolutions for now.
	 [[node reg_net/any_stage/res_bottleneck_block/bottleneck_transform/conv2d_3/BiasAdd (defined at /filepath/anynet.py:119) ]] [Op:__inference_train_function_10336]

Function call stack:
train_function

但是CPU下是正常的，我检查了一下代码，GPU和CPU模式下，get_group_conv是不一样的，但我不知道怎么debug
希望能得到您的回复

RuntimeError: tensorflow/lite/kernels/transpose.cc Transpose op only supports 1D-4D input arrays.Node number 9 (TRANSPOSE) failed to prepare

I'm sure that the model had been save by train.py.

` /home/du/Desk/my_project/Basic_CNNs_TensorFlow2/evaluate.py:28 test_step *
predictions = model(images, training=False)

TypeError: '_UserObject' object is not callable

`

My validation data is the same as training data but the validation result is not good as training in training process. Anyone met this issue?

你好，我在运行了博主的to_tfrecord代码以后生成的datasettest.tfrecord、datasettrain.tfrecord、datasetvaild.tfrecord都是空的，这是为什么？博主的数据集应该是什么格式？是不是有格式要求？还有运行了train以后出来的训练和测试的准确率都是0?源代码里是还有要改的地方吗？
Epoch: 0/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 1/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 2/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 3/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 4/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 5/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 6/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 7/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 8/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 9/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 10/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 11/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 12/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 13/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 14/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 15/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 16/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 17/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 18/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000
Epoch: 19/20, train loss: 0.00000, train accuracy: 0.00000, valid loss: 0.00000, valid accuracy: 0.00000

I follow the instructions in READMD.md.But ‘
def get_the_length_of_dataset(dataset):
count = 0
for i in dataset:
count += 1
return count
’ looks like a dead cycle.

My Python version is 3.7 and tensorflow version is 2.0.0.
Windows10

Have you ever met this situation and how can I solve it
Thanks.

I am a beginner in model training. I train the model using MobileNetV2 and I need to test on a smart phone so I need a .tflite file and label.txt but the "saved_model" folder only has:

都不用迁移学习的吗 硬训练啊 结果不理想

Hello,I found a performance issue in the definition of get_parsed_dataset ,
parse_tfrecord.py,
parsed_dataset = raw_dataset.map was called without num_parallel_calls.
I think it will increase the efficiency of your program if you add this.

Here is the documemtation of tensorflow to support this thing.

Looking forward to your reply. Btw, I am very glad to create a PR to fix it if you are too busy.

2023-02-21 19:53:25.950975: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: AVX2
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
2023-02-21 19:53:25.953773: I tensorflow/core/common_runtime/process_util.cc:146] Creating new thread pool with default inter op setting: 2. Tune using inter_op_parallelism_threads for best performance.
Tensor("args_0:0", shape=(), dtype=float32)
Traceback (most recent call last):
File "k:/resnet/TensorFlow2.0_ResNet-master/train.py", line 33, in
train_dataset, valid_dataset, test_dataset, train_count, valid_count, test_count = generate_datasets()
File "k:\resnet\TensorFlow2.0_ResNet-master\prepare_data.py", line 49, in generate_datasets
train_dataset, train_count = get_dataset(dataset_root_dir=config.train_dir)
File "k:\resnet\TensorFlow2.0_ResNet-master\prepare_data.py", line 40, in get_dataset
image_dataset = tf.data.Dataset.from_tensor_slices(all_image_path).map(load_and_preprocess_image)
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\data\ops\dataset_ops.py", line 2048, in map
return MapDataset(self, map_func, preserve_cardinality=True, name=name)
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\data\ops\dataset_ops.py", line 5243, in init
self._map_func = structured_function.StructuredFunctionWrapper(
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\data\ops\structured_function.py", line 271, in init
self._function = fn_factory()
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\eager\function.py", line 2567, in get_concrete_function
graph_function = self._get_concrete_function_garbage_collected(
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\eager\function.py", line 2533, in _get_concrete_function_garbage_collected
graph_function, _ = self.maybe_define_function(args, kwargs)
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\eager\function.py", line 2711, in maybe_define_function
graph_function = self.create_graph_function(args, kwargs)
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\eager\function.py", line 2627, in create_graph_function
func_graph_module.func_graph_from_py_func(
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\framework\func_graph.py", line 1141, in func_graph_from_py_func
func_outputs = python_func(*func_args, **func_kwargs)
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\data\ops\structured_function.py", line 248, in wrapped_fn
ret = wrapper_helper(*args)
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\data\ops\structured_function.py", line 177, in wrapper_helper
ret = autograph.tf_convert(self.func, ag_ctx)(*nested_args)
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\autograph\impl\api.py", line 692, in wrapper
raise e.ag_error_metadata.to_exception(e)
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\autograph\impl\api.py", line 689, in wrapper
return converted_call(f, args, kwargs, options=options)
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\autograph\impl\api.py", line 439, in converted_call
result = converted_f(*effective_args, **kwargs)
File "C:\Users\16388\AppData\Local\Temp_autograph_generated_filen3u7t458.py", line 11, in tf__load_and_preprocess_image
img_raw = ag.converted_call(ag.ld(tf).io.read_file, (ag.ld(img_path),), None, fscope)
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\autograph\impl\api.py", line 377, in converted_call
return _call_unconverted(f, args, kwargs, options)
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\autograph\impl\api.py", line 459, in _call_unconverted
return f(*args)
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\ops\io_ops.py", line 133, in read_file
return gen_io_ops.read_file(filename, name)
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\ops\gen_io_ops.py", line 570, in read_file
_, _, _op, _outputs = _op_def_library._apply_op_helper(
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\framework\op_def_library.py", line 779, in _apply_op_helper
_ExtractInputsAndAttrs(op_type_name, op_def, allowed_list_attr_map,
File "G:\study\anaconda\envs\supercjq_copy1128\lib\site-packages\tensorflow\python\framework\op_def_library.py", line 579, in _ExtractInputsAndAttrs
raise TypeError(f"{prefix} expected type of "
TypeError: in user code:

File "k:\resnet\TensorFlow2.0_ResNet-master\prepare_data.py", line 11, in load_and_preprocess_image  *
    img_raw = tf.io.read_file(img_path)

TypeError: Input 'filename' of 'ReadFile' Op has type float32 that does not match expected type of string.

Did you shuffle the training data after every training epoch?

When I tried to execute evaluate.py, I've got error "Unsuccessful TensorSliceReader constructor: Failed to find any matching files for saved_model/" Why is this happening?

Hello!

Sorry to bother you.

It seems that creating layers inside 'call' function in a custom layer inheriting from tf.keras.layers.Layer is not proper.

The problem happens when I use your codes as libraries and then I do model.fit with model checkpoint call-back function.

Also, custom layers inheriting from tf.keras.layers.Layer should include get_config as well, or some errors will happen while saving, too.

I've fixed them like below:

class DenseBlock(tf.keras.layers.Layer):
    def __init__(self, num_layers, growth_rate, drop_rate, name=None):
        super(DenseBlock, self).__init__(name=name)
        self.num_layers = num_layers
        self.growth_rate = growth_rate
        self.drop_rate = drop_rate
        self.features_list = []
        self.model_list = []
        for i in range(0, self.num_layers):
            tmp = BottleNeck(growth_rate=self.growth_rate, drop_rate=self.drop_rate)
            self.model_list.append(tmp)

    # def _make_layer(self, x, training):
    #     y = BottleNeck(growth_rate=self.growth_rate, drop_rate=self.drop_rate)(x, training=training)
    #     self.features_list.append(y)
    #     y = tf.concat(self.features_list, axis=-1)
    #     return y

    def call(self, inputs, training=None, **kwargs):
        self.features_list.append(inputs)
        x = inputs
        # x = self._make_layer(inputs, training=training)
        for i in range(0, self.num_layers):
            # x = self._make_layer(x, training=training)
            x = self.model_list[i](x, training=training)
            self.features_list.append(x)
            x = tf.concat(self.features_list, axis=-1)
        self.features_list.clear()
        return x

    def get_config(self):
        config = super(DenseBlock, self).get_config()
        config.update({'growth_rate': self.growth_rate,
                       'drop_rate': self.drop_rate,
                       'num_layers': self.num_layers})
        return config

Please fix them if you are free!
Thank you!

If I need to know which testing data is predicted right or wrong, what can I do to get the image file name information?

Hi brother, well-done.
Can you create confusion_matrix function in evaluate.py if possible?

加油

Where to download pretrained model about EfficientNet？ Thanks！

I'm a newcomer to Machine Learning. It's my pleasure to run your code. When I run split_dataset.py, no data is copied into folders train and test. I copy the data from folder original_dataset into folders test and train, and run train.py will report an error:
`Traceback (most recent call last):
File "C:/Python/Tensorflow/TensorFlow2.0_ResNet-master/train.py", line 33, in
train_dataset, valid_dataset, test_dataset, train_count, valid_count, test_count = generate_datasets()
File "C:\Python\Tensorflow\TensorFlow2.0_ResNet-master\prepare_data.py", line 47, in generate_datasets
train_dataset, train_count = get_dataset(dataset_root_dir=config.train_dir)
File "C:\Python\Tensorflow\TensorFlow2.0_ResNet-master\prepare_data.py", line 38, in get_dataset
image_dataset = tf.data.Dataset.from_tensor_slices(all_image_path).map(load_and_preprocess_image)
File "C:\Python\Tensorflow\test\venv\lib\site-packages\tensorflow_core\python\data\ops\dataset_ops.py", line 1588, in map
return MapDataset(self, map_func, preserve_cardinality=True)
File "C:\Python\Tensorflow\test\venv\lib\site-packages\tensorflow_core\python\data\ops\dataset_ops.py", line 3888, in init
use_legacy_function=use_legacy_function)
File "C:\Python\Tensorflow\test\venv\lib\site-packages\tensorflow_core\python\data\ops\dataset_ops.py", line 3147, in init
self._function = wrapper_fn._get_concrete_function_internal()
File "C:\Python\Tensorflow\test\venv\lib\site-packages\tensorflow_core\python\eager\function.py", line 2395, in _get_concrete_function_internal
*args, **kwargs)
File "C:\Python\Tensorflow\test\venv\lib\site-packages\tensorflow_core\python\eager\function.py", line 2389, in _get_concrete_function_internal_garbage_collected
graph_function, _, _ = self._maybe_define_function(args, kwargs)
File "C:\Python\Tensorflow\test\venv\lib\site-packages\tensorflow_core\python\eager\function.py", line 2703, in _maybe_define_function
graph_function = self._create_graph_function(args, kwargs)
File "C:\Python\Tensorflow\test\venv\lib\site-packages\tensorflow_core\python\eager\function.py", line 2593, in _create_graph_function
capture_by_value=self._capture_by_value),
File "C:\Python\Tensorflow\test\venv\lib\site-packages\tensorflow_core\python\framework\func_graph.py", line 978, in func_graph_from_py_func
func_outputs = python_func(*func_args, **func_kwargs)
File "C:\Python\Tensorflow\test\venv\lib\site-packages\tensorflow_core\python\data\ops\dataset_ops.py", line 3140, in wrapper_fn
ret = _wrapper_helper(*args)
File "C:\Python\Tensorflow\test\venv\lib\site-packages\tensorflow_core\python\data\ops\dataset_ops.py", line 3082, in _wrapper_helper
ret = autograph.tf_convert(func, ag_ctx)(*nested_args)
File "C:\Python\Tensorflow\test\venv\lib\site-packages\tensorflow_core\python\autograph\impl\api.py", line 237, in wrapper
raise e.ag_error_metadata.to_exception(e)
TypeError: in converted code:

C:\Python\Tensorflow\TensorFlow2.0_ResNet-master\prepare_data.py:9 load_and_preprocess_image  *
    img_raw = tf.io.read_file(img_path)
C:\Python\Tensorflow\test\venv\lib\site-packages\tensorflow_core\python\ops\gen_io_ops.py:568 read_file
    "ReadFile", filename=filename, name=name)
C:\Python\Tensorflow\test\venv\lib\site-packages\tensorflow_core\python\framework\op_def_library.py:491 _apply_op_helper
    (prefix, dtypes.as_dtype(input_arg.type).name))

TypeError: Input 'filename' of 'ReadFile' Op has type float32 that does not match expected type of string.

Process finished with exit code 1
`
I'm also a newcomer to tensorflow. How can I solve this problem?

After each epoch I have te same values:
Epoch: 0/2, train loss: nan, train accuracy: 0.10688, valid loss: nan, valid accuracy: 0.10680
regardless of the model.

tensorflow 2.3.1 (install via pip)
Python 3.7.9
What is wrong?

run train.py report errors
Traceback (most recent call last):
File "E:/work/Basic_CNNs_TensorFlow2-master/train_test.py", line 100, in
model.save_weights(filepath=save_model_dir+"epoch-{}".format(epoch), save_format='tf')
File "E:\anaconda3\envs\tf2\lib\site-packages\tensorflow_core\python\keras\engine\network.py", line 1123, in save_weights
self._trackable_saver.save(filepath, session=session)
File "E:\anaconda3\envs\tf2\lib\site-packages\tensorflow_core\python\training\tracking\util.py", line 1168, in save
file_prefix=file_prefix_tensor, object_graph_tensor=object_graph_tensor)
File "E:\anaconda3\envs\tf2\lib\site-packages\tensorflow_core\python\training\tracking\util.py", line 1108, in _save_cached_when_graph_building
object_graph_tensor=object_graph_tensor)
File "E:\anaconda3\envs\tf2\lib\site-packages\tensorflow_core\python\training\tracking\util.py", line 1076, in _gather_saveables
feed_additions) = self._graph_view.serialize_object_graph()
File "E:\anaconda3\envs\tf2\lib\site-packages\tensorflow_core\python\training\tracking\graph_view.py", line 379, in serialize_object_graph
trackable_objects, path_to_root = self._breadth_first_traversal()
File "E:\anaconda3\envs\tf2\lib\site-packages\tensorflow_core\python\training\tracking\graph_view.py", line 199, in _breadth_first_traversal
for name, dependency in self.list_dependencies(current_trackable):
File "E:\anaconda3\envs\tf2\lib\site-packages\tensorflow_core\python\training\tracking\graph_view.py", line 159, in list_dependencies
return obj._checkpoint_dependencies
File "E:\anaconda3\envs\tf2\lib\site-packages\tensorflow_core\python\training\tracking\data_structures.py", line 509, in _checkpoint_dependencies
"automatically un-wrapped and subsequently ignored." % (self,)))
ValueError: Unable to save the object ListWrapper([]) (a list wrapper constructed to track trackable TensorFlow objects). A list element was replaced (setitem, setslice), deleted (delitem, delslice), or moved (sort). In order to support restoration on object creation, tracking is exclusively for append-only data structures.

If you don't need this list checkpointed, wrap it in a tf.contrib.checkpoint.NoDependency object; it will be automatically un-wrapped and subsequently ignored.

How to solve

Tensorflow-gpu版本为2.0.0 ，使用的flower_photos数据集
您好，我在运行您的代码时发现：运行mobilenet、densenet、effcientnet等模型时，模型效果不收敛。准确率一直在0.2左右徘徊。请问这个是什么原因呢？

pip insatll tensorflow==2.0.0 or pip install tensorflow==2.0.0-beta1，tensorflow2.0.0or tensorflow2.0.0-beta1,but when I run train.py, the error of "ValueError: This converter can only convert a single ConcreteFunction. Converting multiple functions is under development" appears,why?How to do?

why Input Image Size is the same? does it depend on my dataset?

Hello all
I am trying to use this repo., but I dont know how I can edit the codes so as (1) To plot the training and validation loss graphy

(2) To make prediction to get the predictions probability

Any help please

Thanks

could you supply the sample of dataset format for mobilenet

It may sound like a naive question, but before starting training, which path should I put my images in?