NVIDIA
numpy reader error
The following error occurred while loading data to gpu using numpy reader.
Below is the source code and the printed error message.
from nvidia.dali import pipeline_def, fn
data_dir = os.path.join(data_dir_2d, 'SER00001')
@pipeline_def(batch_size=batch_size, num_threads=1, device_id=0)
def pipe_gds():
data = fn.readers.numpy(device='gpu', file_root=data_dir, file_filter='*.npy')
return data
p = pipe_gds()
p.build()
pipe_out = p.run()
data_gds = pipe_out[0].as_cpu().as_array() # as_cpu() to c|opy the data back to CPU memory
print(data_gds.shape)
RuntimeError: [/opt/dali/dali/python/backend_impl.cc:752] Assert on "tl.IsDenseTensor()" failed: Tensors in the list must have the same shape
Hi @ksm9413,
Please check the as_array function documentation. To call it, all tensors in the batch need to have the same shape, otherwise the bach cannot be represented as a single array.
Please resize/pad/crop the tensors in the batch or make sure that the input data is uniform in shape.
