Error: Giliner Training data
Opened this issue · 2 comments
I am facing the error in the Training data arguments in the dataset in this line {
training_args = TrainingArguments(
output_dir="models",
learning_rate=5e-6,
weight_decay=0.01,
others_lr=1e-5,
others_weight_decay=0.01,
lr_scheduler_type="linear", #cosine
warmup_ratio=0.1,
per_device_train_batch_size=8,
per_device_eval_batch_size=8,
num_train_epochs=3,
evaluation_strategy="epoch",
save_steps = 1000,
save_total_limit=10,
dataloader_num_workers = 8,
use_cpu = True,
report_to="none",
)
trainer = Trainer(
model=model,
args=training_args,
train_dataset=train_dataset,
eval_dataset=test_dataset,
tokenizer=model.data_processor.transformer_tokenizer,
data_collator=data_collator,
)
trainer.train()
}
Iam getting an error: BackendCompilerFailed: backend='inductor' raised:
CppCompileError: C++ compile error
Output:
/tmp/torchinductor_ubuntu/tw/ctwif4gnjxuyqq47csmx5l4wpoyqgb6kbcqb556l2sakhglmvmzd.cpp: In function ‘void kernel(float*, const long int*, const float*, const float*, float*, long int, long int)’:
/tmp/torchinductor_ubuntu/tw/ctwif4gnjxuyqq47csmx5l4wpoyqgb6kbcqb556l2sakhglmvmzd.cpp:32:46: error: no match for ‘operator==’ (operand types are ‘at::vec::CPU_CAPABILITY::VectorizedN<long int, 2>’ and ‘int’)
32 | auto tmp5 = tmp4 == 0;
| ~~~~ ^~ ~
| | |
| | int
| at::vec::CPU_CAPABILITY::VectorizedN<long int, 2>
Hi, can retry without compiling the model ?
I have tried but the same result