Python profile示例

示例#1

文件： _norm.py 项目： vlamprinidis/nn-estimation

model = nn.Sequential(
    layer, nn.Flatten(),
    nn.Linear(in_features=args.channels * args.numf**DIM, out_features=10))

train_dataset = give(DIM, args.numf, args.channels)

if args.nodes > 1:
    model, train_loader = lib_torch.distribute(model, train_dataset,
                                               args.nodes, args.batch)
else:
    train_loader = torch.utils.data.DataLoader(dataset=train_dataset,
                                               batch_size=args.batch,
                                               shuffle=True)

time = lib_torch.profile(['batch_norm', 'NativeBatchNormBackward'], model,
                         train_loader, args.epochs)

import numpy as np

data = np.array([[
    args.epochs,
    tor_data.ds_size,  # dataset size
    args.numf,
    args.channels,
    args.batch,
    args.nodes,
    time
]])
with open('norm{}d.ptorch'.format(DIM), 'a') as file:
    np.savetxt(file, data, delimiter=",", fmt="%s")

示例#2

文件： _final_dense.py 项目： vlamprinidis/nn-estimation

layer = nn.Linear(in_features=args.numf, out_features=args.units)

model = nn.Sequential(
    nn.Flatten(), nn.Linear(in_features=args.numf, out_features=args.units))

train_dataset = give(1, args.numf, 1, out_size=args.units)

if args.nodes > 1:
    model, train_loader = lib_torch.distribute(model, train_dataset,
                                               args.nodes, args.batch)
else:
    train_loader = torch.utils.data.DataLoader(dataset=train_dataset,
                                               batch_size=args.batch,
                                               shuffle=True)

time = lib_torch.profile(['addmm', 'AddmmBackward'], model, train_loader,
                         args.epochs)

import numpy as np

data = np.array([[
    args.epochs,
    tor_data.ds_size,  # dataset size
    args.numf,
    args.batch,
    args.nodes,
    args.units,
    time
]])
with open('final_dense.ptorch', 'a') as file:
    np.savetxt(file, data, delimiter=",", fmt="%s")

示例#3

文件： _flatten.py 项目： vlamprinidis/nn-estimation

model = nn.Sequential(
    nn.Flatten(),
    nn.Linear(in_features=args.channels * args.numf**DIM, out_features=10))

train_dataset = give(DIM, args.numf, args.channels)

if args.nodes > 1:
    model, train_loader = lib_torch.distribute(model, train_dataset,
                                               args.nodes, args.batch)
else:
    train_loader = torch.utils.data.DataLoader(dataset=train_dataset,
                                               batch_size=args.batch,
                                               shuffle=True)

time = lib_torch.profile(['flatten'], model, train_loader, args.epochs)

import numpy as np

data = np.array([[
    args.epochs,
    tor_data.ds_size,  # dataset size
    args.numf,
    args.channels,
    args.batch,
    args.nodes,
    time
]])
with open('flatten{}d.ptorch'.format(DIM), 'a') as file:
    np.savetxt(file, data, delimiter=",", fmt="%s")

示例#4

文件： _max.py 项目： vlamprinidis/nn-estimation

        out_features = 10
    )
)

train_dataset = give(DIM, args.numf, args.channels)

if args.nodes > 1:
    model, train_loader = lib_torch.distribute(model, train_dataset, args.nodes, args.batch)
else:
    train_loader = torch.utils.data.DataLoader(
        dataset = train_dataset,
        batch_size = args.batch,
        shuffle = True
    )

time = lib_torch.profile(['max_pool{}d'.format(DIM)], 
                         model, train_loader, args.epochs)

import numpy as np

data = np.array([[
    args.epochs, tor_data.ds_size, # dataset size
    args.numf,
    args.channels,
    args.batch,
    args.nodes,
    args.pool,
    args.stride,
    time
]])
with open('max{}d.ptorch'.format(DIM),'a') as file:
    np.savetxt(file, data, delimiter=",", fmt="%s")

示例#5

model = nn.Sequential(
    layer, nn.Flatten(),
    nn.Linear(in_features=args.channels * args.numf**DIM, out_features=10))

train_dataset = give(DIM, args.numf, args.channels)

if args.nodes > 1:
    model, train_loader = lib_torch.distribute(model, train_dataset,
                                               args.nodes, args.batch)
else:
    train_loader = torch.utils.data.DataLoader(dataset=train_dataset,
                                               batch_size=args.batch,
                                               shuffle=True)

time = lib_torch.profile(['relu'], model, train_loader, args.epochs)

import numpy as np

data = np.array([[
    args.epochs,
    tor_data.ds_size,  # dataset size
    args.numf,
    args.channels,
    args.batch,
    args.nodes,
    time
]])
with open('relu{}d.ptorch'.format(DIM), 'a') as file:
    np.savetxt(file, data, delimiter=",", fmt="%s")

示例#6

    nn.Linear(
        in_features=args.filters *
        lib_torch.conv_size_out(args.numf, args.kernel, args.stride)**DIM,
        out_features=10))

train_dataset = give(DIM, args.numf, args.channels)

if args.nodes > 1:
    model, train_loader = lib_torch.distribute(model, train_dataset,
                                               args.nodes, args.batch)
else:
    train_loader = torch.utils.data.DataLoader(dataset=train_dataset,
                                               batch_size=args.batch,
                                               shuffle=True)

time = lib_torch.profile(['conv{}d'.format(DIM), 'MkldnnConvolutionBackward'],
                         model, train_loader, args.epochs)

import numpy as np

data = np.array([[
    args.epochs,
    tor_data.ds_size,  # dataset size
    args.numf,
    args.channels,
    args.batch,
    args.nodes,
    args.kernel,
    args.stride,
    args.filters,
    time
]])

示例#7

文件： _drop.py 项目： vlamprinidis/nn-estimation

model = nn.Sequential(
    layer, nn.Flatten(),
    nn.Linear(in_features=args.channels * args.numf**DIM, out_features=10))

train_dataset = give(DIM, args.numf, args.channels)

if args.nodes > 1:
    model, train_loader = lib_torch.distribute(model, train_dataset,
                                               args.nodes, args.batch)
else:
    train_loader = torch.utils.data.DataLoader(dataset=train_dataset,
                                               batch_size=args.batch,
                                               shuffle=True)

time = lib_torch.profile(['dropout', 'feature_dropout'], model, train_loader,
                         args.epochs)

import numpy as np

data = np.array([[
    args.epochs,
    tor_data.ds_size,  # dataset size
    args.numf,
    args.channels,
    args.batch,
    args.nodes,
    args.drop,
    time
]])
with open('drop{}d.ptorch'.format(DIM), 'a') as file:
    np.savetxt(file, data, delimiter=",", fmt="%s")