XLNet Blurr Error

gistfile1.txt

import pandas as pd
import torch
from transformers import *
from fastai.text.all import *
from blurr.data.all import *
from blurr.modeling.all import *

pretrained_model_name = "xlnet-base-cased"
model_name = pretrained_model_name.split("-")[0]

root_dir = Path(os.getcwd())
resource_dir = root_dir / "resources"
data_dir = resource_dir / "data"
model_dir = resource_dir / "models"

train_path = data_dir/"train.tsv"
classifier_path = "{}_Classifier".format(model_name.upper())

train_data = pd.read_csv(train_path,header=0,index_col=0,sep="\t",skip_blank_lines=True)
task = HF_TASKS_AUTO.SequenceClassification

config = AutoConfig.from_pretrained(pretrained_model_name)
config.num_labels = len(train_data["target"].unique())

hf_arch, hf_config, hf_tokenizer, hf_model = BLURR_MODEL_HELPER.get_hf_objects(pretrained_model_name,  
                                                                               task=task,
                                                                               config=config)

blocks = (HF_TextBlock(hf_arch=hf_arch, hf_tokenizer=hf_tokenizer), CategoryBlock)

dblock = DataBlock(blocks=blocks, 
                   get_x=ColReader('text'), get_y=ColReader('target'), 
                   splitter=RandomSplitter(0.2, seed=321))

dls = dblock.dataloaders(train_data, bs=32)

model = HF_BaseModelWrapper(hf_model)

learn = Learner(dls, 
                model,
                path=resource_dir,
                opt_func=partial(Adam, decouple_wd=True),
                loss_func=CrossEntropyLossFlat(),
                metrics=[accuracy],
                cbs=[HF_BaseModelCallback],
                splitter=hf_splitter).to_fp16()

learn.create_opt() 
learn.freeze()

learn.fit_one_cycle(10, 1e-3)

learn.export(classifier_path)