Train with PyTorch

tp new my-cnn --method pytorch/image-classification
tp dataset push my-images v1 <your image directory>
tp submit my-cnn --profile <card>:1 --train-dataset <you>/my-images@v1

train/scratch/0001.png      this image is labelled scratch
train/dent/0002.png         this one is labelled dent
train/ok/0003.png
val/scratch/0101.png        validation set, same class names
val/ok/0102.png

tp new my-cnn --method pytorch/image-classification

input_pipeline: imagefolder   # reads <data_dir>/{train,val}/<class>/*
num_classes: 3
arch: simple_cnn              # or any torchvision.models constructor, e.g. resnet18
epochs: 5
batch_size: 64                # per process; the global batch is this times the GPU count
learning_rate: 1.0e-3
optimizer: adamw              # adamw | sgd
lr_scheduler: cosine          # constant | step | cosine
amp: true

tp submit my-cnn --profile <card>:1 --set epochs=20 --set arch=resnet18

tp dataset push my-images v1 <directory>
tp submit my-cnn --profile <card>:1 --train-dataset <you>/my-images@v1

tp volume push <directory> --name my-images
tp submit my-cnn --profile <card>:1 --volume <you>/my-images

data_dir: ${VOLUMES_DIR}/my-images   # the platform chooses the path; name it, never hard-code it

tp submit my-cnn --profile <card>:1     # one card
tp submit my-cnn --profile <card>:8     # eight on one node

def build_model(config, in_channels):
    """Return any nn.Module."""

def build_datasets(config, data_dir, *, may_download=True):
    """Return (train_set, val_set, in_channels)."""

./deploy/docker/build-runtimes.sh pytorch --push