9個技巧讓你的PyTorch模型訓(xùn)練變得飛快！
點(diǎn)擊上方“小白學(xué)視覺”，選擇加"星標(biāo)"或“置頂”
重磅干貨，第一時間送達(dá)
本文轉(zhuǎn)自|視覺算法
使用DataLoaders DataLoader中的workers數(shù)量 Batch size 梯度累計(jì) 保留的計(jì)算圖 移動到單個 16-bit 混合精度訓(xùn)練 移動到多個GPUs中（模型復(fù)制） 移動到多個GPU-nodes中 (8+GPUs) 思考模型加速的技巧
from pytorch_lightning import Trainer
model = LightningModule(…)
trainer = Trainer()
trainer.fit(model)1.DataLoaders
dataset = MNIST(root=self.hparams.data_root, train=train, download=True)
loader = DataLoader(dataset, batch_size=32, shuffle=True)
for batch in loader:
x, y = batch
model.training_step(x, y)
...2.DataLoaders中的workers的數(shù)量
# slow
loader = DataLoader(dataset, batch_size=32, shuffle=True)
# fast (use 10 workers)
loader = DataLoader(dataset, batch_size=32, shuffle=True, num_workers=10)3.Batch size
# clear last step
optimizer.zero_grad()
# 16 accumulated gradient steps
scaled_loss = 0
for accumulated_step_i in range(16):
out = model.forward()
loss = some_loss(out,y)
loss.backward()
scaled_loss += loss.item()
# update weights after 8 steps. effective batch = 8*16
optimizer.step()
# loss is now scaled up by the number of accumulated batches
actual_loss = scaled_loss / 16
accumulate_grad_batches=16：trainer = Trainer(accumulate_grad_batches=16)
trainer.fit(model)5.保留的計(jì)算圖
losses = []
...
losses.append(loss)
print(f'current loss: {torch.mean(losses)'})
# bad
losses.append(loss)
# good
losses.append(loss.item())
# put model on GPU
model.cuda(0)
# put data on gpu (cuda on a variable returns a cuda copy)
x = x.cuda(0)
# runs on GPU now
model(x)
Trainer(gpus=1)。# ask lightning to use gpu 0 for training
trainer = Trainer(gpus=[0])
trainer.fit(model)
# expensive
x = x.cuda(0)# very expensive
x = x.cpu()
x = x.cuda(0)
# really bad idea. Stops all the GPUs until they all catch up
torch.cuda.empty_cache()
# enable 16-bit on the model and the optimizer
model, optimizers = amp.initialize(model, optimizers, opt_level='O2')
# when doing .backward, let amp do it so it can scale the loss
with amp.scale_loss(loss, optimizer) as scaled_loss:
scaled_loss.backward()
Trainer(precision=16)就可以了。trainer = Trainer(amp_level='O2', use_amp=False)
trainer.fit(model)8.移動到多個GPUs中
# copy model on each GPU and give a fourth of the batch to each
model = DataParallel(model, devices=[0, 1, 2 ,3])
# out has 4 outputs (one for each gpu)
out = model(x.cuda(0))
# ask lightning to use 4 GPUs for training
trainer = Trainer(gpus=[0, 1, 2, 3])
trainer.fit(model)
# each model is sooo big we can't fit both in memory
encoder_rnn.cuda(0)
decoder_rnn.cuda(1)
# run input through encoder on GPU 0
encoder_out = encoder_rnn(x.cuda(0))
# run output through decoder on the next GPU
out = decoder_rnn(encoder_out.cuda(1))
# normally we want to bring all outputs back to GPU 0
out = out.cuda(0)
class MyModule(LightningModule):
def __init__():
self.encoder = RNN(...)
self.decoder = RNN(...)
def forward(x):
# models won't be moved after the first forward because
# they are already on the correct GPUs
self.encoder.cuda(0)
self.decoder.cuda(1)
out = self.encoder(x)
out = self.decoder(out.cuda(1))
# don't pass GPUs to trainer
model = MyModule()
trainer = Trainer()
trainer.fit(model)
# change these lines
self.encoder = RNN(...)
self.decoder = RNN(...)
# to these
# now each RNN is based on a different gpu set
self.encoder = DataParallel(self.encoder, devices=[0, 1, 2, 3])
self.decoder = DataParallel(self.encoder, devices=[4, 5, 6, 7])
# in forward...
out = self.encoder(x.cuda(0))
# notice inputs on first gpu in device
sout = self.decoder(out.cuda(4)) # <--- the 4 here
如果模型已經(jīng)在GPU上了，model.cuda()不會做任何事情。 總是把輸入放在設(shè)備列表中的第一個設(shè)備上。 在設(shè)備之間傳輸數(shù)據(jù)是昂貴的，把它作為最后的手段。 優(yōu)化器和梯度會被保存在GPU 0上，因此，GPU 0上使用的內(nèi)存可能會比其他GPU大得多。
在每個GPU上初始化一個模型的副本(確保設(shè)置種子，讓每個模型初始化到相同的權(quán)重，否則它會失敗)。 將數(shù)據(jù)集分割成子集(使用DistributedSampler)。每個GPU只在它自己的小子集上訓(xùn)練。 在.backward()上，所有副本都接收到所有模型的梯度副本。這是模型之間唯一一次的通信。
def tng_dataloader():
d = MNIST()
# 4: Add distributed sampler
# sampler sends a portion of tng data to each machine
dist_sampler = DistributedSampler(dataset)
dataloader = DataLoader(d, shuffle=False, sampler=dist_sampler)
def main_process_entrypoint(gpu_nb):
# 2: set up connections between all gpus across all machines
# all gpus connect to a single GPU "root"
# the default uses env://
world = nb_gpus * nb_nodes
dist.init_process_group("nccl", rank=gpu_nb, world_size=world)
# 3: wrap model in DPP
torch.cuda.set_device(gpu_nb)
model.cuda(gpu_nb)
model = DistributedDataParallel(model, device_ids=[gpu_nb])
# train your model now...
if __name__ == '__main__':
# 1: spawn number of processes
# your cluster will call main for each machine
mp.spawn(main_process_entrypoint, nprocs=8)
# train on 1024 gpus across 128 nodes
trainer = Trainer(nb_gpu_nodes=128, gpus=[0, 1, 2, 3, 4, 5, 6, 7])
# train on 4 gpus on the same machine MUCH faster than DataParallel
trainer = Trainer(distributed_backend='ddp', gpus=[0, 1, 2, 3])
好消息，小白學(xué)視覺團(tuán)隊(duì)的知識星球開通啦，為了感謝大家的支持與厚愛，團(tuán)隊(duì)決定將價(jià)值149元的知識星球現(xiàn)時免費(fèi)加入。各位小伙伴們要抓住機(jī)會哦！
交流群
歡迎加入公眾號讀者群一起和同行交流，目前有SLAM、三維視覺、傳感器、自動駕駛、計(jì)算攝影、檢測、分割、識別、醫(yī)學(xué)影像、GAN、算法競賽等微信群（以后會逐漸細(xì)分），請掃描下面微信號加群，備注：”昵稱+學(xué)校/公司+研究方向“，例如：”張三 + 上海交大 + 視覺SLAM“。請按照格式備注，否則不予通過。添加成功后會根據(jù)研究方向邀請進(jìn)入相關(guān)微信群。請勿在群內(nèi)發(fā)送廣告，否則會請出群，謝謝理解~
評論
圖片
表情