因果语言模型¶

pip install transformers datasets evaluate

from huggingface_hub import notebook_login
notebook_login()

from datasets import load_dataset
eli5 = load_dataset("eli5_category", split="train[:5000]")

eli5 = eli5.train_test_split(test_size=0.2)

eli5["train"][0]

from transformers import AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("distilbert/distilgpt2")

eli5 = eli5.flatten()
eli5["train"][0]

def preprocess_function(examples):
    return tokenizer([" ".join(x) for x in examples["answers.text"]])

tokenized_eli5 = eli5.map(
    preprocess_function,
    batched=True,
    num_proc=4,
    remove_columns=eli5["train"].column_names,
)

block_size = 128

def group_texts(examples):
    # 连接所有文本。
    concatenated_examples = {k: sum(examples[k], []) for k in examples.keys()}
    total_length = len(concatenated_examples[list(examples.keys())[0]])
    # 我们丢弃小的余数，如果模型支持，我们可以添加填充而不是丢弃，您可以自定义这部分以满足您的需求。
    if total_length >= block_size:
        total_length = (total_length // block_size) * block_size
    # 按 block_size 分块。
    result = {
        k: [t[i:i + block_size] for i in range(0, total_length, block_size)]
        for k, t in concatenated_examples.items()
    }
    result["labels"] = result["input_ids"].copy()
    return result

lm_dataset = tokenized_eli5.map(group_texts, batched=True, num_proc=4)

from transformers import DataCollatorForLanguageModeling
tokenizer.pad_token = tokenizer.eos_token
data_collator = DataCollatorForLanguageModeling(tokenizer=tokenizer, mlm=False)

from transformers import DataCollatorForLanguageModeling
data_collator = DataCollatorForLanguageModeling(tokenizer=tokenizer, mlm=False, return_tensors="tf")

from transformers import AutoModelForCausalLM, TrainingArguments, Trainer
model = AutoModelForCausalLM.from_pretrained("distilbert/distilgpt2")

training_args = TrainingArguments(
    output_dir="my_awesome_eli5_clm-model",
    eval_strategy="epoch",
    learning_rate=2e-5,
    weight_decay=0.01,
    push_to_hub=True,
)

trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=lm_dataset["train"],
    eval_dataset=lm_dataset["test"],
    data_collator=data_collator,
    tokenizer=tokenizer,
)

trainer.train()

import math
eval_results = trainer.evaluate()
print(f"Perplexity: {math.exp(eval_results['eval_loss']):.2f}")

trainer.push_to_hub()

from transformers import create_optimizer, AdamWeightDecay
optimizer = AdamWeightDecay(learning_rate=2e-5, weight_decay_rate=0.01)

from transformers import TFAutoModelForCausalLM
model = TFAutoModelForCausalLM.from_pretrained("distilbert/distilgpt2")

tf_train_set = model.prepare_tf_dataset(
    lm_dataset["train"],
    shuffle=True,
    batch_size=16,
    collate_fn=data_collator,
)

tf_test_set = model.prepare_tf_dataset(
    lm_dataset["test"],
    shuffle=False,
    batch_size=16,
    collate_fn=data_collator,
)

import tensorflow as tf
model.compile(optimizer=optimizer)  # 无需指定损失函数

from transformers.keras_callbacks import PushToHubCallback
callback = PushToHubCallback(
    output_dir="my_awesome_eli5_clm-model",
    tokenizer=tokenizer,
)

model.fit(x=tf_train_set, validation_data=tf_test_set, epochs=3, callbacks=[callback])

prompt = "Somatic hypermutation allows the immune system to"

from transformers import pipeline
generator = pipeline("text-generation", model="username/my_awesome_eli5_clm-model")
generator(prompt)

from transformers import AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("username/my_awesome_eli5_clm-model")
inputs = tokenizer(prompt, return_tensors="pt").input_ids

from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("username/my_awesome_eli5_clm-model")
outputs = model.generate(inputs, max_new_tokens=100, do_sample=True, top_k=50, top_p=0.95)

tokenizer.batch_decode(outputs, skip_special_tokens=True)

from transformers import AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("username/my_awesome_eli5_clm-model")
inputs = tokenizer(prompt, return_tensors="tf").input_ids

from transformers import TFAutoModelForCausalLM
model = TFAutoModelForCausalLM.from_pretrained("username/my_awesome_eli5_clm-model")
outputs = model.generate(input_ids=inputs, max_new_tokens=100, do_sample=True, top_k=50, top_p=0.95)

tokenizer.batch_decode(outputs, skip_special_tokens=True)