search

calico / basenji

Sequential regulatory activity predictions with deep convolutional neural networks.
Apache License 2.0
386 stars 120 forks source link

AssertionError, Basenji_test.py 'Conv1D' object has no attribute 'input_shape' #199

Open PouriaAzadehR opened 1 month ago

PouriaAzadehR commented 1 month ago

Hi,

I follow up the installation instructions and test_train tutorial. I used the pre-trained model and did not train a new one.

when running the following command

../bin/basenji_test.py --ai 0,1,2 -o output/heart_test --rc --shifts "1,0,-1" models/params_small.json models/heart/model_best.h5 heart_l131k

I get error

Total params: 111,011 (433.64 KB) Trainable params: 109,235 (426.70 KB) Non-trainable params: 1,776 (6.94 KB) None Traceback (most recent call last): File "/project/6014832/pouria81/basenji/tutorials/../bin/basenji_test.py", line 433, in main() File "/project/6014832/pouria81/basenji/tutorials/../bin/basenji_test.py", line 143, in main seqnn_model = seqnn.SeqNN(params_model) ^^^^^^^^^^^^^^^^^^^^^^^^^ File "/project/6014832/pouria81/basenji/basenji/seqnn.py", line 35, in init self.build_model() File "/project/6014832/pouria81/basenji/basenji/seqnn.py", line 190, in build_model stride_factor = layer.input_shape[1] / layer.output_shape[1]

davek44 commented 1 month ago

We're not maintaining this repository anymore. Please move to https://github.com/calico/baskerville