Converting pytorch convolutional neural nets and certain transformers to Timeloop workload files.
After cloning this repository, run python setup.py install to finish the installation. Note that this converter has been developed and tested with:
- python 3.7
- pytorch 1.7.1
- torchvision 0.8.2
- numpy 1.21.2
- pyyaml 5.3.1
import torchvision.models as models
import pytorch2timeloop
# Define a pytorch-based neural network model, for example, a pre-defined alexnet from torchvision.
net = models.alexnet()
# Define the shape of a single input sample, in the following format:
# (# of channels, height, width)
# For example, the above alexnet will get a 224x224 RGB image:
input_shape = (3, 224, 224)
# Define the number of batches that will be used for the inference
batch_size = 1
# Define the directory names where the timeloop workload yaml files will be stored.
# The yaml files will be stored in ./workloads/alexnet/ in this example.
top_dir = 'workloads'
sub_dir = 'alexnet'
# By default, nn.Conv2d modules will be automatically converted, but nn.Linear modules will be ignored.
# If you want to convert nn.Linear, set the option to be true.
# The converter will change the description of nn.Linear into Convolution-like layer.
# (e.g., in_channel=in_features, out_channel=out_features, input_height=1, input_width=1, filter size = 1x1, stride = 1x1, padding = 0x0)
# If you want to ignore nn.Linear layers, set this option to be false.
convert_fc = True
# Finally, in case there exists a layer that is only used during the training phase, define an identifier for a such layer.
# For example, in torchvision.models.inception_v3, auxiliary classification layers are not used during the inference (e.g., InceptionAux).
# In this case, include a string that can serve as an identifier for such layers (e.g., 'Aux') in exception_module_names.
# But for the above alexnet, there is no necessity to define this.
exception_module_names = []
# Now, convert!
pytorch2timeloop.convert_model(net, input_shape, batch_size, sub_dir, top_dir, convert_fc, exception_module_names)This code is licensed with MIT License. This code has been modified from the works of Anurag Golla and Alex Moser.