List of pip requirements
kavi6024 opened this issue · 0 comments
kavi6024 commented
Can someone share the requirements.txt file since the packages I use seems to not have some functions that are used here ?
Also, I get Reshape errors when converting from CNN to RNN (using keras.backend.set_image_data_format("channels_first")).
Shape must be at least rank 3 but is rank 2 for '{{node BiasAdd}} = BiasAdd[T=DT_FLOAT, data_format="NCHW"](MatMul, unstack)' with input shap input shapes: [?,384], [384].
Call arguments received by layer 'forward_gru' (type GRU):
• inputs=tf.Tensor(shape=(None, 60, 128), dtype=float32)
• mask=None
• training=None
• initial_state=None
Using keras.backend.set_image_data_format("channels_last") gives me this
total size of new array must be unchanged, input_shape = [9, 2, 64], output_shape = [60, -1]
Call arguments received by layer "reshape" (type Reshape):
• inputs=tf.Tensor(shape=(None, 9, 2, 64), dtype=float32)