section4.ipynb

Kernel: Unknown Kernel

调试训练管道

Install the Transformers, Datasets, and Evaluate libraries to run this notebook.

!pip install datasets evaluate transformers[sentencepiece]

from datasets import load_dataset, load_metric
from transformers import (
    AutoTokenizer,
    AutoModelForSequenceClassification,
    TrainingArguments,
    Trainer,
)

raw_datasets = load_dataset("glue", "mnli")

model_checkpoint = "distilbert-base-uncased"
tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)


def preprocess_function(examples):
    return tokenizer(examples["premise"], examples["hypothesis"], truncation=True)


tokenized_datasets = raw_datasets.map(preprocess_function, batched=True)
model = AutoModelForSequenceClassification.from_pretrained(model_checkpoint)

args = TrainingArguments(
    f"distilbert-finetuned-mnli",
    evaluation_strategy="epoch",
    save_strategy="epoch",
    learning_rate=2e-5,
    num_train_epochs=3,
    weight_decay=0.01,
)

metric = load_metric("glue", "mnli")


def compute_metrics(eval_pred):
    predictions, labels = eval_pred
    return metric.compute(predictions=predictions, references=labels)


trainer = Trainer(
    model,
    args,
    train_dataset=raw_datasets["train"],
    eval_dataset=raw_datasets["validation_matched"],
    compute_metrics=compute_metrics,
)
trainer.train()

'ValueError: You have to specify either input_ids or inputs_embeds'

trainer.train_dataset[0]

{'hypothesis': 'Product and geography are what make cream skimming work. ',
 'idx': 0,
 'label': 1,
 'premise': 'Conceptually cream skimming has two basic dimensions - product and geography.'}

from datasets import load_dataset, load_metric
from transformers import (
    AutoTokenizer,
    AutoModelForSequenceClassification,
    TrainingArguments,
    Trainer,
)

raw_datasets = load_dataset("glue", "mnli")

model_checkpoint = "distilbert-base-uncased"
tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)


def preprocess_function(examples):
    return tokenizer(examples["premise"], examples["hypothesis"], truncation=True)


tokenized_datasets = raw_datasets.map(preprocess_function, batched=True)
model = AutoModelForSequenceClassification.from_pretrained(model_checkpoint)

args = TrainingArguments(
    f"distilbert-finetuned-mnli",
    evaluation_strategy="epoch",
    save_strategy="epoch",
    learning_rate=2e-5,
    num_train_epochs=3,
    weight_decay=0.01,
)

metric = load_metric("glue", "mnli")


def compute_metrics(eval_pred):
    predictions, labels = eval_pred
    return metric.compute(predictions=predictions, references=labels)


trainer = Trainer(
    model,
    args,
    train_dataset=tokenized_datasets["train"],
    eval_dataset=tokenized_datasets["validation_matched"],
    compute_metrics=compute_metrics,
)
trainer.train()

'ValueError: expected sequence of length 43 at dim 1 (got 37)'

tokenizer.decode(trainer.train_dataset[0]["input_ids"])

'[CLS] conceptually cream skimming has two basic dimensions - product and geography. [SEP] product and geography are what make cream skimming work. [SEP]'

trainer.train_dataset[0].keys()

dict_keys(['attention_mask', 'hypothesis', 'idx', 'input_ids', 'label', 'premise'])

type(trainer.model)

transformers.models.distilbert.modeling_distilbert.DistilBertForSequenceClassification

tokenizer.decode(trainer.train_dataset[0]["attention_mask"])

[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]

len(trainer.train_dataset[0]["attention_mask"]) == len(
    trainer.train_dataset[0]["input_ids"]
)

True

trainer.train_dataset[0]["label"]

1

trainer.train_dataset.features["label"].names

['entailment', 'neutral', 'contradiction']

for batch in trainer.get_train_dataloader():
    break

~/git/transformers/src/transformers/data/data_collator.py in torch_default_data_collator(features)
    105                 batch[k] = torch.stack([f[k] for f in features])
    106             else:
--> 107                 batch[k] = torch.tensor([f[k] for f in features])
    108 
    109     return batch

ValueError: expected sequence of length 45 at dim 1 (got 76)

data_collator = trainer.get_train_dataloader().collate_fn
data_collator

<function transformers.data.data_collator.default_data_collator(features: List[InputDataClass], return_tensors='pt') -> Dict[str, Any]>

from datasets import load_dataset, load_metric
from transformers import (
    AutoTokenizer,
    AutoModelForSequenceClassification,
    DataCollatorWithPadding,
    TrainingArguments,
    Trainer,
)

raw_datasets = load_dataset("glue", "mnli")

model_checkpoint = "distilbert-base-uncased"
tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)


def preprocess_function(examples):
    return tokenizer(examples["premise"], examples["hypothesis"], truncation=True)


tokenized_datasets = raw_datasets.map(preprocess_function, batched=True)
model = AutoModelForSequenceClassification.from_pretrained(model_checkpoint)

args = TrainingArguments(
    f"distilbert-finetuned-mnli",
    evaluation_strategy="epoch",
    save_strategy="epoch",
    learning_rate=2e-5,
    num_train_epochs=3,
    weight_decay=0.01,
)

metric = load_metric("glue", "mnli")


def compute_metrics(eval_pred):
    predictions, labels = eval_pred
    return metric.compute(predictions=predictions, references=labels)


data_collator = DataCollatorWithPadding(tokenizer=tokenizer)

trainer = Trainer(
    model,
    args,
    train_dataset=tokenized_datasets["train"],
    eval_dataset=tokenized_datasets["validation_matched"],
    compute_metrics=compute_metrics,
    data_collator=data_collator,
    tokenizer=tokenizer,
)
trainer.train()

RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)`

data_collator = trainer.get_train_dataloader().collate_fn
batch = data_collator([trainer.train_dataset[i] for i in range(4)])

data_collator = trainer.get_train_dataloader().collate_fn
actual_train_set = trainer._remove_unused_columns(trainer.train_dataset)
batch = data_collator([actual_train_set[i] for i in range(4)])

for batch in trainer.get_train_dataloader():
    break

outputs = trainer.model.cpu()(**batch)

~/.pyenv/versions/3.7.9/envs/base/lib/python3.7/site-packages/torch/nn/functional.py in nll_loss(input, target, weight, size_average, ignore_index, reduce, reduction)
   2386         )
   2387     if dim == 2:
-> 2388         ret = torch._C._nn.nll_loss(input, target, weight, _Reduction.get_enum(reduction), ignore_index)
   2389     elif dim == 4:
   2390         ret = torch._C._nn.nll_loss2d(input, target, weight, _Reduction.get_enum(reduction), ignore_index)

IndexError: Target 2 is out of bounds.

trainer.model.config.num_labels

2

from datasets import load_dataset, load_metric
from transformers import (
    AutoTokenizer,
    AutoModelForSequenceClassification,
    DataCollatorWithPadding,
    TrainingArguments,
    Trainer,
)

raw_datasets = load_dataset("glue", "mnli")

model_checkpoint = "distilbert-base-uncased"
tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)


def preprocess_function(examples):
    return tokenizer(examples["premise"], examples["hypothesis"], truncation=True)


tokenized_datasets = raw_datasets.map(preprocess_function, batched=True)
model = AutoModelForSequenceClassification.from_pretrained(model_checkpoint, num_labels=3)

args = TrainingArguments(
    f"distilbert-finetuned-mnli",
    evaluation_strategy="epoch",
    save_strategy="epoch",
    learning_rate=2e-5,
    num_train_epochs=3,
    weight_decay=0.01,
)

metric = load_metric("glue", "mnli")


def compute_metrics(eval_pred):
    predictions, labels = eval_pred
    return metric.compute(predictions=predictions, references=labels)


data_collator = DataCollatorWithPadding(tokenizer=tokenizer)

trainer = Trainer(
    model,
    args,
    train_dataset=tokenized_datasets["train"],
    eval_dataset=tokenized_datasets["validation_matched"],
    compute_metrics=compute_metrics,
    data_collator=data_collator,
    tokenizer=tokenizer,
)

for batch in trainer.get_train_dataloader():
    break

outputs = trainer.model.cpu()(**batch)

import torch

device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
batch = {k: v.to(device) for k, v in batch.items()}

outputs = trainer.model.to(device)(**batch)

loss = outputs.loss
loss.backward()

trainer.create_optimizer()
trainer.optimizer.step()

# This will take a long time and error out, so you shouldn't run this cell
trainer.train()

TypeError: only size-1 arrays can be converted to Python scalars

trainer.evaluate()

TypeError: only size-1 arrays can be converted to Python scalars

for batch in trainer.get_eval_dataloader():
    break

batch = {k: v.to(device) for k, v in batch.items()}

with torch.no_grad():
    outputs = trainer.model(**batch)

predictions = outputs.logits.cpu().numpy()
labels = batch["labels"].cpu().numpy()

compute_metrics((predictions, labels))

TypeError: only size-1 arrays can be converted to Python scalars

predictions.shape, labels.shape

((8, 3), (8,))

import numpy as np


def compute_metrics(eval_pred):
    predictions, labels = eval_pred
    predictions = np.argmax(predictions, axis=1)
    return metric.compute(predictions=predictions, references=labels)


compute_metrics((predictions, labels))

{'accuracy': 0.625}

import numpy as np
from datasets import load_dataset, load_metric
from transformers import (
    AutoTokenizer,
    AutoModelForSequenceClassification,
    DataCollatorWithPadding,
    TrainingArguments,
    Trainer,
)

raw_datasets = load_dataset("glue", "mnli")

model_checkpoint = "distilbert-base-uncased"
tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)


def preprocess_function(examples):
    return tokenizer(examples["premise"], examples["hypothesis"], truncation=True)


tokenized_datasets = raw_datasets.map(preprocess_function, batched=True)
model = AutoModelForSequenceClassification.from_pretrained(model_checkpoint, num_labels=3)

args = TrainingArguments(
    f"distilbert-finetuned-mnli",
    evaluation_strategy="epoch",
    save_strategy="epoch",
    learning_rate=2e-5,
    num_train_epochs=3,
    weight_decay=0.01,
)

metric = load_metric("glue", "mnli")


def compute_metrics(eval_pred):
    predictions, labels = eval_pred
    predictions = np.argmax(predictions, axis=1)
    return metric.compute(predictions=predictions, references=labels)


data_collator = DataCollatorWithPadding(tokenizer=tokenizer)

trainer = Trainer(
    model,
    args,
    train_dataset=tokenized_datasets["train"],
    eval_dataset=tokenized_datasets["validation_matched"],
    compute_metrics=compute_metrics,
    data_collator=data_collator,
    tokenizer=tokenizer,
)
trainer.train()

for batch in trainer.get_train_dataloader():
    break

batch = {k: v.to(device) for k, v in batch.items()}
trainer.create_optimizer()

for _ in range(20):
    outputs = trainer.model(**batch)
    loss = outputs.loss
    loss.backward()
    trainer.optimizer.step()
    trainer.optimizer.zero_grad()

with torch.no_grad():
    outputs = trainer.model(**batch)
preds = outputs.logits
labels = batch["labels"]

compute_metrics((preds.cpu().numpy(), labels.cpu().numpy()))

{'accuracy': 1.0}