ValueError: You should supply an encoding or a list of encodings to this method that includes input_ids, but you provided ['label']

[atrompeterog]

Describe the bug

I am trying to create a PEFT model from DISTILBERT model, and run a training loop. However, the trainer.train() is giving me this error: ValueError: You should supply an encoding or a list of encodings to this method that includes input_ids, but you provided ['label']
Here is my code:

Steps to reproduce the bug

#Creating a PEFT Config
from peft import LoraConfig
from transformers import AutoTokenizer, AutoModelForSequenceClassification
from peft import get_peft_model

lora_config = LoraConfig(
task_type="SEQ_CLASS",
r=8,
lora_alpha=32,
target_modules=["q_lin", "k_lin", "v_lin"],
lora_dropout=0.01,
)

#Converting a Transformers Model into a PEFT Model
model = AutoModelForSequenceClassification.from_pretrained(
"distilbert-base-uncased",
num_labels=2, #Binary classification, 1 = positive, 0 = negative

)
lora_model = get_peft_model(model, lora_config)

print(lora_model)

Tokenize data set
from datasets import load_dataset
from transformers import AutoTokenizer, AutoModelForSequenceClassification

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

Load the train and test splits dataset

dataset = load_dataset("fancyzhx/amazon_polarity")

#create a smaller subset for train and test
subset_size = 5000
small_train_dataset = dataset["train"].shuffle(seed=42).select(range(subset_size))
small_test_dataset = dataset["test"].shuffle(seed=42).select(range(subset_size))

#Tokenize data
def tokenize_function(example):
return tokenizer(example["content"], padding="max_length", truncation=True)

tokenized_train_dataset = small_train_dataset.map(tokenize_function, batched=True)
tokenized_test_dataset = small_test_dataset.map(tokenize_function, batched=True)

train_lora = tokenized_train_dataset.rename_column('label', 'labels')
test_lora = tokenized_test_dataset.rename_column('label', 'labels')

print(tokenized_train_dataset.column_names)
print(tokenized_test_dataset.column_names)

#Train the PEFT model

import numpy as np
from transformers import Trainer, TrainingArguments, default_data_collator, DataCollatorWithPadding
from datasets import load_dataset
from transformers import AutoTokenizer, AutoModelForSequenceClassification

def compute_metrics(eval_pred):
predictions, labels = eval_pred
predictions = np.argmax(predictions, axis=1)
return {"accuracy": (predictions == labels).mean()}

trainer = Trainer(
model=lora_model,
args=TrainingArguments(
output_dir=".",
learning_rate=2e-3,
# Reduce the batch size if you don't have enough memory
per_device_train_batch_size=1,
per_device_eval_batch_size=1,
num_train_epochs=3,
weight_decay=0.01,
evaluation_strategy="epoch",
save_strategy="epoch",
load_best_model_at_end=True,
),
train_dataset=tokenized_train_dataset,
eval_dataset=tokenized_test_dataset,
tokenizer=tokenizer,
data_collator=DataCollatorWithPadding(tokenizer=tokenizer, return_tensors="pt"),
compute_metrics=compute_metrics,
)

trainer.train()

Expected behavior

Example of output:
[558/558 01:04, Epoch XX]

Epoch	Training Loss	Validation Loss	Accuracy
1	No log	0.046478	0.988341
2	0.052800	0.048840	0.988341

Environment info

Using python and jupyter notbook

[atrompeterog]

Now i have other error