RuntimeError: result type Float can't be cast to the desired output type Long---

data
1

This is my code, I get this error. Tried to converrt df[labels] to torch.float but the issue still exists.
Getting —> trainer.train() error.

# Convert to Hugging Face dataset
dataset = Dataset.from_pandas(df)

# Load tokenizer and model
tokenizer = BertTokenizer.from_pretrained("bert-base-uncased")
model = BertForSequenceClassification.from_pretrained("bert-base-uncased", num_labels=len(mlb.classes_), problem_type="multi_label_classification")

# Tokenize data
def preprocess_function(examples):
    return tokenizer(examples['job_title'], truncation=True, padding=True)

tokenized_dataset = dataset.map(preprocess_function, batched=True)

# Ensure labels are of type torch.float (this is required for multi-label classification)
def cast_to_float(example):
    example['labels'] = torch.tensor(example['labels'], dtype=torch.float)  # Convert labels to torch.float
    return example

tokenized_dataset = tokenized_dataset.map(cast_to_float)

# Training arguments
training_args = TrainingArguments(
    output_dir="./results",
    evaluation_strategy="epoch",
    save_strategy="epoch",
    num_train_epochs=3,
    per_device_train_batch_size=8,
    logging_dir="./logs",
)

# Trainer
trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=tokenized_dataset,
    eval_dataset=tokenized_dataset,  # Ideally, you should split this into train/test datasets.
)

# Train model
trainer.train()
2

It’s unclear which operation fails in your model, but I would assume you are trying to perform an inplace operation with an integral and floating type which is disallowed as explained in the docs:

An integral output tensor cannot accept a floating point tensor.

Example:

a = torch.tensor(1)
print(a.type())
# torch.LongTensor

b = torch.tensor(2.)
print(b.type())
# torch.FloatTensor

a * b
# tensor(2.)
a *= b
# RuntimeError: result type Float can't be cast to the desired output type Long
4

This is tokenized_dataset which I am going to train:
Dataset({
features: [‘job_title’, ‘labels’, ‘input_ids’, ‘token_type_ids’, ‘attention_mask’],
num_rows: 3
})

job_title and token_type_ids are list type objects other columns are torch.LongTensor type. Should I exclude them from training dataset?