convnet_benchmarks.py

"""
Benchmark for common convnets.

Speed on Titan X, with 10 warmup steps and 10 main steps and with different
versions of cudnn, are as follows (time reported below is per-batch time,
forward / forward+backward):

                    CuDNN V3        CuDNN v4
AlexNet         32.5 / 108.0    27.4 /  90.1
OverFeat       113.0 / 342.3    91.7 / 276.5
Inception      134.5 / 485.8   125.7 / 450.6
VGG (batch 64) 200.8 / 650.0   164.1 / 551.7

Speed on Inception with varied batch sizes and CuDNN v4 is as follows:

Batch Size   Speed per batch     Speed per image
 16             22.8 /  72.7         1.43 / 4.54
 32             38.0 / 127.5         1.19 / 3.98
 64             67.2 / 233.6         1.05 / 3.65
128            125.7 / 450.6         0.98 / 3.52

Speed on Tesla M40, which 10 warmup steps and 10 main steps and with cudnn
v4, is as follows:

AlexNet         68.4 / 218.1
OverFeat       210.5 / 630.3
Inception      300.2 / 1122.2
VGG (batch 64) 405.8 / 1327.7

(Note that these numbers involve a "full" backprop, i.e. the gradient
with respect to the input image is also computed.)

To get the numbers, simply run:

for MODEL in AlexNet OverFeat Inception; do
  PYTHONPATH=../gen:$PYTHONPATH python convnet_benchmarks.py \
    --batch_size 128 --model $MODEL --forward_only True
done
for MODEL in AlexNet OverFeat Inception; do
  PYTHONPATH=../gen:$PYTHONPATH python convnet_benchmarks.py \
    --batch_size 128 --model $MODEL
done
PYTHONPATH=../gen:$PYTHONPATH python convnet_benchmarks.py \
  --batch_size 64 --model VGGA --forward_only True
PYTHONPATH=../gen:$PYTHONPATH python convnet_benchmarks.py \
  --batch_size 64 --model VGGA

for BS in 16 32 64 128; do
  PYTHONPATH=../gen:$PYTHONPATH python convnet_benchmarks.py \
    --batch_size $BS --model Inception --forward_only True
  PYTHONPATH=../gen:$PYTHONPATH python convnet_benchmarks.py \
    --batch_size $BS --model Inception
done

Note that VGG needs to be run at batch 64 due to memory limit on the backward
pass.
"""

import argparse

from caffe2.python import cnn, workspace


def MLP(order, cudnn_ws):
    model = cnn.CNNModelHelper()
    d = 256
    depth = 20
    width = 3
    for i in range(depth):
        for j in range(width):
            current = "fc_{}_{}".format(i, j) if i > 0 else "data"
            next_ = "fc_{}_{}".format(i + 1, j)
            model.FC(
                current, next_,
                dim_in=d, dim_out=d,
                weight_init=model.XavierInit,
                bias_init=model.XavierInit)
    model.Sum(["fc_{}_{}".format(depth, j) for j in range(width)], ["sum"])
    model.FC("sum", "last",
             dim_in=d, dim_out=1000,
             weight_init=model.XavierInit,
             bias_init=model.XavierInit)
    xent = model.LabelCrossEntropy(["last", "label"], "xent")
    model.AveragedLoss(xent, "loss")
    return model, d


def AlexNet(order, cudnn_ws):
    model = cnn.CNNModelHelper(
        order, name="alexnet",
        use_cudnn=True, cudnn_exhaustive_search=True,
        ws_nbytes_limit=cudnn_ws)
    conv1 = model.Conv(
        "data",
        "conv1",
        3,
        64,
        11,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        stride=4,
        pad=2
    )
    relu1 = model.Relu(conv1, "conv1")
    pool1 = model.MaxPool(relu1, "pool1", kernel=3, stride=2)
    conv2 = model.Conv(
        pool1,
        "conv2",
        64,
        192,
        5,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=2
    )
    relu2 = model.Relu(conv2, "conv2")
    pool2 = model.MaxPool(relu2, "pool2", kernel=3, stride=2)
    conv3 = model.Conv(
        pool2,
        "conv3",
        192,
        384,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu3 = model.Relu(conv3, "conv3")
    conv4 = model.Conv(
        relu3,
        "conv4",
        384,
        256,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu4 = model.Relu(conv4, "conv4")
    conv5 = model.Conv(
        relu4,
        "conv5",
        256,
        256,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu5 = model.Relu(conv5, "conv5")
    pool5 = model.MaxPool(relu5, "pool5", kernel=3, stride=2)
    fc6 = model.FC(
        pool5, "fc6", 256 * 6 * 6, 4096, ('XavierFill', {}),
        ('ConstantFill', {})
    )
    relu6 = model.Relu(fc6, "fc6")
    fc7 = model.FC(
        relu6, "fc7", 4096, 4096, ('XavierFill', {}), ('ConstantFill', {})
    )
    relu7 = model.Relu(fc7, "fc7")
    fc8 = model.FC(
        relu7, "fc8", 4096, 1000, ('XavierFill', {}), ('ConstantFill', {})
    )
    pred = model.Softmax(fc8, "pred")
    xent = model.LabelCrossEntropy([pred, "label"], "xent")
    loss = model.AveragedLoss(xent, "loss")
    return model, 224


def OverFeat(order, cudnn_ws):
    model = cnn.CNNModelHelper(
        order, name="overfeat",
        use_cudnn=True, cudnn_exhaustive_search=True,
        ws_nbytes_limit=cudnn_ws)
    conv1 = model.Conv(
        "data",
        "conv1",
        3,
        96,
        11,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        stride=4
    )
    relu1 = model.Relu(conv1, "conv1")
    pool1 = model.MaxPool(relu1, "pool1", kernel=2, stride=2)
    conv2 = model.Conv(
        pool1, "conv2", 96, 256, 5, ('XavierFill', {}), ('ConstantFill', {})
    )
    relu2 = model.Relu(conv2, "conv2")
    pool2 = model.MaxPool(relu2, "pool2", kernel=2, stride=2)
    conv3 = model.Conv(
        pool2,
        "conv3",
        256,
        512,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu3 = model.Relu(conv3, "conv3")
    conv4 = model.Conv(
        relu3,
        "conv4",
        512,
        1024,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu4 = model.Relu(conv4, "conv4")
    conv5 = model.Conv(
        relu4,
        "conv5",
        1024,
        1024,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu5 = model.Relu(conv5, "conv5")
    pool5 = model.MaxPool(relu5, "pool5", kernel=2, stride=2)
    fc6 = model.FC(
        pool5, "fc6", 1024 * 6 * 6, 3072, ('XavierFill', {}),
        ('ConstantFill', {})
    )
    relu6 = model.Relu(fc6, "fc6")
    fc7 = model.FC(
        relu6, "fc7", 3072, 4096, ('XavierFill', {}), ('ConstantFill', {})
    )
    relu7 = model.Relu(fc7, "fc7")
    fc8 = model.FC(
        relu7, "fc8", 4096, 1000, ('XavierFill', {}), ('ConstantFill', {})
    )
    pred = model.Softmax(fc8, "pred")
    xent = model.LabelCrossEntropy([pred, "label"], "xent")
    loss = model.AveragedLoss(xent, "loss")
    return model, 231


def VGGA(order, cudnn_ws):
    model = cnn.CNNModelHelper(
        order, name='vgg-a',
        use_cudnn=True, cudnn_exhaustive_search=True,
        ws_nbytes_limit=cudnn_ws)
    conv1 = model.Conv(
        "data",
        "conv1",
        3,
        64,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu1 = model.Relu(conv1, "conv1")
    pool1 = model.MaxPool(relu1, "pool1", kernel=2, stride=2)
    conv2 = model.Conv(
        pool1,
        "conv2",
        64,
        128,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu2 = model.Relu(conv2, "conv2")
    pool2 = model.MaxPool(relu2, "pool2", kernel=2, stride=2)
    conv3 = model.Conv(
        pool2,
        "conv3",
        128,
        256,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu3 = model.Relu(conv3, "conv3")
    conv4 = model.Conv(
        relu3,
        "conv4",
        256,
        256,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu4 = model.Relu(conv4, "conv4")
    pool4 = model.MaxPool(relu4, "pool4", kernel=2, stride=2)
    conv5 = model.Conv(
        pool4,
        "conv5",
        256,
        512,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu5 = model.Relu(conv5, "conv5")
    conv6 = model.Conv(
        relu5,
        "conv6",
        512,
        512,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu6 = model.Relu(conv6, "conv6")
    pool6 = model.MaxPool(relu6, "pool6", kernel=2, stride=2)
    conv7 = model.Conv(
        pool6,
        "conv7",
        512,
        512,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu7 = model.Relu(conv7, "conv7")
    conv8 = model.Conv(
        relu7,
        "conv8",
        512,
        512,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu8 = model.Relu(conv8, "conv8")
    pool8 = model.MaxPool(relu8, "pool8", kernel=2, stride=2)

    fcix = model.FC(
        pool8, "fcix", 512 * 7 * 7, 4096, ('XavierFill', {}),
        ('ConstantFill', {})
    )
    reluix = model.Relu(fcix, "fcix")
    fcx = model.FC(
        reluix, "fcx", 4096, 4096, ('XavierFill', {}), ('ConstantFill', {})
    )
    relux = model.Relu(fcx, "fcx")
    fcxi = model.FC(
        relux, "fcxi", 4096, 1000, ('XavierFill', {}), ('ConstantFill', {})
    )
    pred = model.Softmax(fcxi, "pred")
    xent = model.LabelCrossEntropy([pred, "label"], "xent")
    loss = model.AveragedLoss(xent, "loss")
    return model, 231


def _InceptionModule(
    model, input_blob, input_depth, output_name, conv1_depth, conv3_depths,
    conv5_depths, pool_depth
):
    # path 1: 1x1 conv
    conv1 = model.Conv(
        input_blob, output_name + ":conv1", input_depth, conv1_depth, 1,
        ('XavierFill', {}), ('ConstantFill', {})
    )
    conv1 = model.Relu(conv1, conv1)
    # path 2: 1x1 conv + 3x3 conv
    conv3_reduce = model.Conv(
        input_blob, output_name + ":conv3_reduce", input_depth, conv3_depths[0],
        1, ('XavierFill', {}), ('ConstantFill', {})
    )
    conv3_reduce = model.Relu(conv3_reduce, conv3_reduce)
    conv3 = model.Conv(
        conv3_reduce,
        output_name + ":conv3",
        conv3_depths[0],
        conv3_depths[1],
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    conv3 = model.Relu(conv3, conv3)
    # path 3: 1x1 conv + 5x5 conv
    conv5_reduce = model.Conv(
        input_blob, output_name + ":conv5_reduce", input_depth, conv5_depths[0],
        1, ('XavierFill', {}), ('ConstantFill', {})
    )
    conv5_reduce = model.Relu(conv5_reduce, conv5_reduce)
    conv5 = model.Conv(
        conv5_reduce,
        output_name + ":conv5",
        conv5_depths[0],
        conv5_depths[1],
        5,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=2
    )
    conv5 = model.Relu(conv5, conv5)
    # path 4: pool + 1x1 conv
    pool = model.MaxPool(
        input_blob,
        output_name + ":pool",
        kernel=3,
        stride=1,
        pad=1
    )
    pool_proj = model.Conv(
        pool, output_name + ":pool_proj", input_depth, pool_depth, 1,
        ('XavierFill', {}), ('ConstantFill', {})
    )
    pool_proj = model.Relu(pool_proj, pool_proj)
    output = model.Concat([conv1, conv3, conv5, pool_proj], output_name)
    return output


def Inception(order, cudnn_ws):
    model = cnn.CNNModelHelper(
        order, name="inception",
        use_cudnn=True, cudnn_exhaustive_search=True,
        ws_nbytes_limit=cudnn_ws)
    conv1 = model.Conv(
        "data",
        "conv1",
        3,
        64,
        7,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        stride=2,
        pad=3
    )
    relu1 = model.Relu(conv1, "conv1")
    pool1 = model.MaxPool(relu1, "pool1", kernel=3, stride=2, pad=1)
    conv2a = model.Conv(
        pool1, "conv2a", 64, 64, 1, ('XavierFill', {}), ('ConstantFill', {})
    )
    conv2a = model.Relu(conv2a, conv2a)
    conv2 = model.Conv(
        conv2a,
        "conv2",
        64,
        192,
        3,
        ('XavierFill', {}),
        ('ConstantFill', {}),
        pad=1
    )
    relu2 = model.Relu(conv2, "conv2")
    pool2 = model.MaxPool(relu2, "pool2", kernel=3, stride=2, pad=1)
    # Inception modules
    inc3 = _InceptionModule(
        model, pool2, 192, "inc3", 64, [96, 128], [16, 32], 32
    )
    inc4 = _InceptionModule(
        model, inc3, 256, "inc4", 128, [128, 192], [32, 96], 64
    )
    pool5 = model.MaxPool(inc4, "pool5", kernel=3, stride=2, pad=1)
    inc5 = _InceptionModule(
        model, pool5, 480, "inc5", 192, [96, 208], [16, 48], 64
    )
    inc6 = _InceptionModule(
        model, inc5, 512, "inc6", 160, [112, 224], [24, 64], 64
    )
    inc7 = _InceptionModule(
        model, inc6, 512, "inc7", 128, [128, 256], [24, 64], 64
    )
    inc8 = _InceptionModule(
        model, inc7, 512, "inc8", 112, [144, 288], [32, 64], 64
    )
    inc9 = _InceptionModule(
        model, inc8, 528, "inc9", 256, [160, 320], [32, 128], 128
    )
    pool9 = model.MaxPool(inc9, "pool9", kernel=3, stride=2, pad=1)
    inc10 = _InceptionModule(
        model, pool9, 832, "inc10", 256, [160, 320], [32, 128], 128
    )
    inc11 = _InceptionModule(
        model, inc10, 832, "inc11", 384, [192, 384], [48, 128], 128
    )
    pool11 = model.AveragePool(inc11, "pool11", kernel=7, stride=1)
    fc = model.FC(
        pool11, "fc", 1024, 1000, ('XavierFill', {}), ('ConstantFill', {})
    )
    # It seems that Soumith's benchmark does not have softmax on top
    # for Inception. We will add it anyway so we can have a proper
    # backward pass.
    pred = model.Softmax(fc, "pred")
    xent = model.LabelCrossEntropy([pred, "label"], "xent")
    loss = model.AveragedLoss(xent, "loss")
    return model, 224


def AddParameterUpdate(model):
    """ Simple plain SGD update -- not tuned to actually train the models """
    ITER = model.Iter("iter")
    LR = model.LearningRate(
        ITER, "LR", base_lr=-1e-8, policy="step", stepsize=10000, gamma=0.999)
    ONE = model.param_init_net.ConstantFill([], "ONE", shape=[1], value=1.0)
    for param in model.params:
        param_grad = model.param_to_grad[param]
        model.WeightedSum([param, ONE, param_grad, LR], param)


def Benchmark(model_gen, arg):
    model, input_size = model_gen(arg.order, arg.cudnn_ws)
    model.Proto().type = arg.net_type
    model.Proto().num_workers = arg.num_workers

    # In order to be able to run everything without feeding more stuff, let's
    # add the data and label blobs to the parameter initialization net as well.
    if arg.order == "NCHW":
        input_shape = [arg.batch_size, 3, input_size, input_size]
    else:
        input_shape = [arg.batch_size, input_size, input_size, 3]
    if arg.model == "MLP":
        input_shape = [arg.batch_size, input_size]

    model.param_init_net.GaussianFill(
        [],
        "data",
        shape=input_shape,
        mean=0.0,
        std=1.0
    )
    model.param_init_net.UniformIntFill(
        [],
        "label",
        shape=[arg.batch_size, ],
        min=0,
        max=999
    )

    if arg.forward_only:
        print('{}: running forward only.'.format(arg.model))
    else:
        print('{}: running forward-backward.'.format(arg.model))
        model.AddGradientOperators(["loss"])
        AddParameterUpdate(model)
        if arg.order == 'NHWC':
            print(
                '==WARNING==\n'
                'NHWC order with CuDNN may not be supported yet, so I might\n'
                'exit suddenly.'
            )

    if not arg.cpu:
        model.param_init_net.RunAllOnGPU()
        model.net.RunAllOnGPU()

    if arg.engine:
        for op in model.net.Proto().op:
            op.engine = arg.engine

    if arg.dump_model:
        # Writes out the pbtxt for benchmarks on e.g. Android
        with open(
            "{0}_init_batch_{1}.pbtxt".format(arg.model, arg.batch_size), "w"
        ) as fid:
            fid.write(str(model.param_init_net.Proto()))
        with open("{0}.pbtxt".format(arg.model, arg.batch_size), "w") as fid:
            fid.write(str(model.net.Proto()))

    workspace.RunNetOnce(model.param_init_net)
    workspace.CreateNet(model.net)
    workspace.BenchmarkNet(
        model.net.Proto().name, arg.warmup_iterations, arg.iterations,
        arg.layer_wise_benchmark)


def GetArgumentParser():
    parser = argparse.ArgumentParser(description="Caffe2 benchmark.")
    parser.add_argument(
        "--batch_size",
        type=int,
        default=128,
        help="The batch size."
    )
    parser.add_argument("--model", type=str, help="The model to benchmark.")
    parser.add_argument(
        "--order",
        type=str,
        default="NCHW",
        help="The order to evaluate."
    )
    parser.add_argument(
        "--cudnn_ws",
        type=int,
        help="The cudnn workspace size."
    )
    parser.add_argument(
        "--iterations",
        type=int,
        default=10,
        help="Number of iterations to run the network."
    )
    parser.add_argument(
        "--warmup_iterations",
        type=int,
        default=10,
        help="Number of warm-up iterations before benchmarking."
    )
    parser.add_argument(
        "--forward_only",
        action='store_true',
        help="If set, only run the forward pass."
    )
    parser.add_argument(
        "--layer_wise_benchmark",
        action='store_true',
        help="If True, run the layer-wise benchmark as well."
    )
    parser.add_argument(
        "--cpu",
        action='store_true',
        help="If True, run testing on CPU instead of GPU."
    )
    parser.add_argument(
        "--engine",
        type=str,
        default="",
        help="If set, blindly prefer the given engine(s) for every op.")
    parser.add_argument(
        "--dump_model",
        action='store_true',
        help="If True, dump the model prototxts to disk."
    )
    parser.add_argument("--net_type", type=str, default="dag")
    parser.add_argument("--num_workers", type=int, default=2)
    parser.add_argument("--use-nvtx", default=False, action='store_true')
    parser.add_argument("--htrace_span_log_path", type=str)
    return parser


if __name__ == '__main__':
    args = GetArgumentParser().parse_args()
    if (
        not args.batch_size or not args.model or not args.order
    ):
        GetArgumentParser().print_help()
    else:
        workspace.GlobalInit(
            ['caffe2', '--caffe2_log_level=0'] +
            (['--caffe2_use_nvtx'] if args.use_nvtx else []) +
            (['--caffe2_htrace_span_log_path=' + args.htrace_span_log_path]
                if args.htrace_span_log_path else []))

        model_map = {
            'AlexNet': AlexNet,
            'OverFeat': OverFeat,
            'VGGA': VGGA,
            'Inception': Inception,
            'MLP': MLP,
        }
        Benchmark(model_map[args.model], args)