我已經(jīng)手動(dòng)將整個(gè)事情間隔開。盡管如此，它還是行不通。第一行之后直接出現(xiàn)縮進(jìn)錯(cuò)誤。代碼：def train_epoch(model, data_loader, loss_fn, optimizer, device, scheduler, n_examples):""" docstring? """losses = []for d in data_loader: input_ids = d['input_ids'].to(device) targets = d['targets'].to(device) outputs = model(input_ids = input_ids, labels = targets) loss = loss_fn(outputs, targets) losses.append( loss.item() ) loss.backward() optimizer.step() scheduler.step() optimizer.zero_grad()return np.mean(losses)錯(cuò)誤：def train_epoch(模型、data_loader、loss_fn、優(yōu)化器、設(shè)備、調(diào)度器、n_examples): ...文件“”，第 2 行 ^ IndentationError：需要縮進(jìn)塊到底是怎么回事？我沒有看到任何地方有問題。
3 回答
ibeautiful
TA貢獻(xiàn)1993條經(jīng)驗(yàn) 獲得超6個(gè)贊
def train_epoch(model, data_loader, loss_fn, optimizer, device, scheduler, n_examples):
"""
docstring?
"""
losses = []
for d in data_loader:
input_ids = d['input_ids'].to(device)
targets = d['targets'].to(device)
outputs = model(input_ids = input_ids, labels = targets)
loss = loss_fn(outputs, targets)
losses.append( loss.item() )
loss.backward()
optimizer.step()
scheduler.step()
optimizer.zero_grad()
return np.mean(losses)
請像這樣格式化。您的代碼不在您的train_epoch()方法的縮進(jìn)塊下。
慕田峪4524236
TA貢獻(xiàn)1875條經(jīng)驗(yàn) 獲得超5個(gè)贊
我修改了它：for循環(huán)必須向右縮進(jìn)才能使其函數(shù)代碼運(yùn)行或在函數(shù)內(nèi)部，否則它不會(huì)被視為函數(shù)代碼。此外，在這種情況下，損失列表必須位于 for 循環(huán)內(nèi)部，而不是在其之前，也不能位于同一 for 循環(huán)級(jí)別。嘗試一下然后告訴我。如果成功投票并回答:-)
def train_epoch(model, data_loader, loss_fn, optimizer, device, scheduler, n_examples):
for d in data_loader:
losses = []
input_ids = d['input_ids'].to(device)
targets = d['targets'].to(device)
outputs = model(input_ids = input_ids, labels = targets)
loss = loss_fn(outputs, targets)
losses.append( loss.item() )
loss.backward()
optimizer.step()
scheduler.step()
optimizer.zero_grad()
return np.mean(losses)
翻閱古今
TA貢獻(xiàn)1780條經(jīng)驗(yàn) 獲得超5個(gè)贊
這是因?yàn)榈谝恍?。函?shù)體必須縮進(jìn)。
def train_epoch(model, data_loader, loss_fn, optimizer, device, scheduler, n_examples):
""" docstring? """
losses = []
for d in data_loader:
input_ids = d['input_ids'].to(device)
targets = d['targets'].to(device)
outputs = model(input_ids = input_ids, labels = targets)
loss = loss_fn(outputs, targets)
losses.append( loss.item() )
loss.backward()
optimizer.step()
scheduler.step()
optimizer.zero_grad()
return np.mean(losses)
添加回答
舉報(bào)
0/150
提交
取消