choidm / lits-liver-tumor-segmentaton-challenge Goto Github PK

TypeError Traceback (most recent call last)
in <cell line: 1>()
----> 1 model.fit(total_patch, total_mask, batch_size = 512, epochs = 40)

2 frames
/tmp/__autograph_generated_filen463df7a.py in tf__weighted_binary_crossentropy(y_true, y_pred)
9 retval_ = ag__.UndefinedReturnValue()
10 y_pred = ag__.converted_call(ag__.ld(tf).clip_by_value, (ag__.ld(y_pred), 1e-07, 1.0 - 1e-07), None, fscope)
---> 11 loss = -(ag__.ld(y_true) * ag__.converted_call(ag__.ld(K).log, (ag__.ld(y_pred),), None, fscope) * 0.9 + (1 - ag__.ld(y_true)) * ag__.converted_call(ag__.ld(K).log, (1 - ag__.ld(y_pred),), None, fscope) * 0.1)
12 try:
13 do_return = True

TypeError: in user code:

File "/usr/local/lib/python3.10/dist-packages/keras/engine/training.py", line 1284, in train_function  *
    return step_function(self, iterator)
File "<ipython-input-41-259b4a3fa215>", line 3, in weighted_binary_crossentropy  *
    loss = - (y_true * K.log(y_pred) * 0.90 + (1 - y_true) * K.log(1 - y_pred) * 0.10)

TypeError: Input 'y' of 'Mul' Op has type float32 that does not match type uint8 of argument 'x'.

Hi.
In Training.ipynb has code like this:

model = model.get_layer("model_7")

and I run the note on the colab and I got ValueError like this.

ValueError Traceback (most recent call last)
in ()
----> 1 model = model.get_layer("model_7")

/usr/local/lib/python3.6/dist-packages/keras/engine/network.py in get_layer(self, name, index)
362 return layer
363
--> 364 raise ValueError('No such layer: ' + name)
365
366 @Property

ValueError: No such layer: model_7

I found "model_1" in the model.summary(). but

model = model.get_layer("model_1")

didn't work.

I also looked for the get_layer () function, but couldn't solve it.

I need help...
Thanks for reading.

Hi,
Thanks for your awesomework, I notice that your network output channel is 1 in train.ipynb, Is it have some problem when compute dice?
, It the sum of 0 and 1, not compute 0 and 1 Seperately .

Hi Choi!

Great work! Curious to know the quantitative results (loss and dice score) and training time. Could you please share some. Thank you!