search

huyquoctrinh / MetaPolyp-CBMS2023

Meta-Polyp: a baseline for efficient Polyp segmentation (CBMS 2023)
MIT License
68 stars 9 forks source link

facing issue while training #5

Closed kunalpurkayastha closed 10 months ago

kunalpurkayastha commented 11 months ago

Hello there, I am facing the below error while I try to train with the given dataset only

Traceback (most recent call last): File "train.py", line 35, in model = build_model(img_size) File "/home/kunal/Documents/MetaPolyp-CBMS2023/model.py", line 21, in build_model upscale_feature = decode(x, scale = 4, filters = x.shape[channel_axis]) File "/home/kunal/Documents/MetaPolyp-CBMS2023/layers/upsampling.py", line 14, in decode x1 = tf.keras.layers.Conv2D(filters, (1, 1), activation=activation, use_bias=False, File "/home/kunal/.conda/envs/scorpio/lib/python3.8/site-packages/keras/src/utils/traceback_utils.py", line 70, in error_handler raise e.with_traceback(filtered_tb) from None File "/home/kunal/.conda/envs/scorpio/lib/python3.8/site-packages/keras/src/engine/input_spec.py", line 253,in assert_input_compatibility raise ValueError( ValueError: Input 0 of layer "conv2d" is incompatible with the layer: expected min_ndim=4, found ndim=3. Full shape received: (None, 64, 512)

please help me to fix this

huyquoctrinh commented 11 months ago

Hi @kunalpurkayastha , thank you so much for this issue. This issue can be fixed via the installation keras-cv-attention-models==1.3.9 package, I updated in the requirements.txt file. I hope you can fix it.

huyquoctrinh commented 10 months ago

Hi everyone, this issue solution is updated in the README, you guys can follow in that, I will close this issue