微调Llama2模型时遇TypeError: bool类型不可迭代问题求助

我使用包含22条训练数据、8条验证数据的小型数据集微调Llama2模型，使用SFT Trainer时持续报错，尝试调整训练参数、PEFT配置均未解决问题，目标是完成小样本训练后进行推理测试。在TrainingArguments中添加remove_unused_columns=False后，出现如下错误：

TypeError: argument of type 'bool' is not iterable

我的TrainingArguments配置如下：

TrainingArguments(
_n_gpu=2,
adafactor=False,
adam_beta1=0.9,
adam_beta2=0.999,
adam_epsilon=1e-08,
auto_find_batch_size=False,
bf16=False,
bf16_full_eval=False,
data_seed=None,
dataloader_drop_last=False,
dataloader_num_workers=0,
dataloader_pin_memory=True,
ddp_backend=None,
ddp_broadcast_buffers=None,
ddp_bucket_cap_mb=None,
ddp_find_unused_parameters=None,
ddp_timeout=1800,
debug=[],
deepspeed=None,
disable_tqdm=False,
dispatch_batches=None,
do_eval=True,
do_predict=False,
do_train=False,
eval_accumulation_steps=None,
eval_delay=0,
eval_steps=100,
evaluation_strategy=IntervalStrategy.EPOCH,
fp16=False,
fp16_backend=auto,
fp16_full_eval=False,
fp16_opt_level=O1,
fsdp=[],
fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_grad_ckpt': False},
fsdp_min_num_params=0,
fsdp_transformer_layer_cls_to_wrap=None,
full_determinism=False,
gradient_accumulation_steps=4,
gradient_checkpointing=False,
greater_is_better=None,
group_by_length=False,
half_precision_backend=auto,
hub_always_push=False,
hub_model_id=None,
hub_private_repo=False,
hub_strategy=HubStrategy.EVERY_SAVE,
hub_token=<HUB_TOKEN>,
ignore_data_skip=False,
include_inputs_for_metrics=False,
include_tokens_per_second=False,
jit_mode_eval=False,
label_names=None,
label_smoothing_factor=0.0,
learning_rate=5e-05,
length_column_name=length,
load_best_model_at_end=False,
local_rank=0,
log_level=passive,
log_level_replica=warning,
log_on_each_node=True,
logging_dir=test_trainer/runs/Nov23_10-56-58_frontierrhel8,
logging_first_step=False,
logging_nan_inf_filter=True,
logging_steps=500,
logging_strategy=IntervalStrategy.STEPS,
lr_scheduler_type=SchedulerType.LINEAR,
max_grad_norm=1.0,
max_steps=20,
metric_for_best_model=None,
mp_parameters=,
no_cuda=False,
num_train_epochs=2,
optim=OptimizerNames.ADAMW_TORCH,
optim_args=None,
output_dir=test_trainer,
overwrite_output_dir=False,
past_index=-1,
per_device_eval_batch_size=4,
per_device_train_batch_size=4,
prediction_loss_only=False,
push_to_hub=False,
push_to_hub_model_id=None,
push_to_hub_organization=None,
push_to_hub_token=<PUSH_TO_HUB_TOKEN>,
ray_scope=last,
remove_unused_columns=False,
report_to=[],
resume_from_checkpoint=None,
run_name=test_trainer,
save_on_each_node=False,
save_safetensors=False,
save_steps=500,
save_strategy=IntervalStrategy.STEPS,
save_total_limit=None,
seed=42,
sharded_ddp=[],
skip_memory_metrics=True,
tf32=None,
torch_compile=False,
torch_compile_backend=None,
torch_compile_mode=None,
torchdynamo=None,
tpu_metrics_debug=False,
tpu_num_cores=None,
use_cpu=False,
use_ipex=False,
use_legacy_prediction_loop=False,
use_mps_device=False,
warmup_ratio=0.0,
warmup_steps=10,
weight_decay=0.0,
)

• 修复mp_parameters参数：将mp_parameters=改为mp_parameters=""，该参数需为字符串类型，空值不能直接留空，否则会触发类型迭代错误。
• 开启训练开关：把do_train=False改为do_train=True，否则训练流程无法启动，还会引发后续逻辑异常。
• 指定label_names参数：结合SFT Trainer的特性，添加label_names=["labels"]，配合remove_unused_columns=False一起使用，确保数据列被正确识别处理。
• 适配小数据集调整batch size：训练集仅22条数据，当前per_device_train_batch_size=4+gradient_accumulation_steps=4的有效batch size为16，建议将per_device_train_batch_size改为2，避免数据加载时出现维度不匹配问题。

内容的提问来源于stack exchange，提问作者Akshat Jain