incorrect increment of self.level

[tanjiarui]

Hi there,

I got the index error when I try to fit the model on a simple data set

Traceback (most recent call last):
  File "/home/ubuntu/.local/lib/python3.8/site-packages/keras/utils/traceback_utils.py", line 67, in error_handler
    raise e.with_traceback(filtered_tb) from None
  File "/home/ubuntu/.local/lib/python3.8/site-packages/tensorflow/python/framework/func_graph.py", line 1147, in autograph_handler
    raise e.ag_error_metadata.to_exception(e)
tensorflow.python.autograph.impl.api.StagingError: in user code

File "/home/ubuntu/.local/lib/python3.8/site-packages/keras/engine/training.py", line 1021, in train_function  *
    return step_function(self, iterator)
File "/home/ubuntu/.local/lib/python3.8/site-packages/keras/engine/training.py", line 1010, in step_function  **
    outputs = model.distribute_strategy.run(run_step, args=(data,))
File "/home/ubuntu/.local/lib/python3.8/site-packages/keras/engine/training.py", line 1000, in run_step  **
    outputs = model.train_step(data)
File "/home/ubuntu/.local/lib/python3.8/site-packages/keras/engine/training.py", line 859, in train_step
    y_pred = self(x, training=True)
File "/home/ubuntu/.local/lib/python3.8/site-packages/keras/utils/traceback_utils.py", line 67, in error_handler
    raise e.with_traceback(filtered_tb) from None
File "/home/ubuntu/桌面/face expression/fer/model/model.py", line 352, in call
    feature = self.bns[i][self.level](feature)

IndexError: Exception encountered when calling layer "box_net" (type BoxNet).

list index out of range

Call arguments received:
  • inputs=['tf.Tensor(shape=(None, 64, 64, 64), dtype=float32)', 'tf.Tensor(shape=(), dtype=int32)']
  • kwargs={'training': 'True'}

looks like the variable 'self.level' was over incremented to exceed the index range. this is the source code:

def call(self, inputs, **kwargs):
    feature, level = inputs
    for i in range(self.depth):
        feature = self.convs[i](feature)
        feature = self.bns[i][self.level](feature)
        feature = self.relu(feature)
    outputs = self.head(feature)
    outputs = self.reshape(outputs)
    self.level += 1
    return outputs

this is the configuration of self.bns

self.bns = [[layers.BatchNormalization(momentum=MOMENTUM, epsilon=EPSILON, name=f'{self.name}/box-{i}-bn-{j}') for j in range(3, 8)] for i in range(depth)]

obviously, there are only 'depth' groups of batch layers, and five batch layers within each group. thus, that code snippet should be revised as follows:

def call(self, inputs, **kwargs):
    feature, level = inputs
    for i in range(self.depth):
        feature = self.convs[i](feature)
        feature = self.bns[i][self.level](feature)
        feature = self.relu(feature)
    outputs = self.head(feature)
    outputs = self.reshape(outputs)
    if self.level < 4:
        self.level += 1
    else:
        self.level = 0
    return outputs

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[Jason-Young-AI]

Hi there,

I got the index error when I try to fit the model on a simple data set

Traceback (most recent call last):
  File "/home/ubuntu/.local/lib/python3.8/site-packages/keras/utils/traceback_utils.py", line 67, in error_handler
    raise e.with_traceback(filtered_tb) from None
  File "/home/ubuntu/.local/lib/python3.8/site-packages/tensorflow/python/framework/func_graph.py", line 1147, in autograph_handler
    raise e.ag_error_metadata.to_exception(e)
tensorflow.python.autograph.impl.api.StagingError: in user code

File "/home/ubuntu/.local/lib/python3.8/site-packages/keras/engine/training.py", line 1021, in train_function  *
    return step_function(self, iterator)
File "/home/ubuntu/.local/lib/python3.8/site-packages/keras/engine/training.py", line 1010, in step_function  **
    outputs = model.distribute_strategy.run(run_step, args=(data,))
File "/home/ubuntu/.local/lib/python3.8/site-packages/keras/engine/training.py", line 1000, in run_step  **
    outputs = model.train_step(data)
File "/home/ubuntu/.local/lib/python3.8/site-packages/keras/engine/training.py", line 859, in train_step
    y_pred = self(x, training=True)
File "/home/ubuntu/.local/lib/python3.8/site-packages/keras/utils/traceback_utils.py", line 67, in error_handler
    raise e.with_traceback(filtered_tb) from None
File "/home/ubuntu/桌面/face expression/fer/model/model.py", line 352, in call
    feature = self.bns[i][self.level](feature)

IndexError: Exception encountered when calling layer "box_net" (type BoxNet).

list index out of range

Call arguments received:
  • inputs=['tf.Tensor(shape=(None, 64, 64, 64), dtype=float32)', 'tf.Tensor(shape=(), dtype=int32)']
  • kwargs={'training': 'True'}

looks like the variable 'self.level' was over incremented to exceed the index range. this is the source code:

def call(self, inputs, **kwargs):
    feature, level = inputs
    for i in range(self.depth):
        feature = self.convs[i](feature)
        feature = self.bns[i][self.level](feature)
        feature = self.relu(feature)
    outputs = self.head(feature)
    outputs = self.reshape(outputs)
    self.level += 1
    return outputs

this is the configuration of self.bns

self.bns = [[layers.BatchNormalization(momentum=MOMENTUM, epsilon=EPSILON, name=f'{self.name}/box-{i}-bn-{j}') for j in range(3, 8)] for i in range(depth)]

obviously, there are only 'depth' groups of batch layers, and five batch layers within each group. thus, that code snippet should be revised as follows:

def call(self, inputs, **kwargs):
    feature, level = inputs
    for i in range(self.depth):
        feature = self.convs[i](feature)
        feature = self.bns[i][self.level](feature)
        feature = self.relu(feature)
    outputs = self.head(feature)
    outputs = self.reshape(outputs)
    if self.level < 4:
        self.level += 1
    else:
        self.level = 0
    return outputs

Your answer is right, while I changed my code based on yours, it is more adaptive:

def call(self, inputs, **kwargs):
    feature, level = inputs
    for i in range(self.depth):
        feature = self.convs[i](feature)
        feature = self.bns[i][self.level](feature)
        feature = self.relu(feature)
    outputs = self.head(feature)
    outputs = self.reshape(outputs)
    if self.level < self.depth:
        self.level += 1
    else:
        self.level = 0
    return outputs

See Another Solution in the Closed Issue

BTW, the author does not seem to maintain the repo...