使用finetune_task_lora.sh微调llava_1.5模型无法运行

[zhangye0402]

报错信息如图，问题：第一轮训练显示loss为0后直接结束训练 我的环境如下：

训练脚本如下：

我曾经用这个脚本正确运行过训练程序，但是在重新配置环境后出现了这样的错误，请问是哪里出错了呢，谢谢大家！

[MarkDeng1]

我的微调 从HF下载的模型总是报错，说结构不匹配。。
Traceback (most recent call last):
File "/fred/oz337/zdeng/prompt_stealing_ours/LLaVA/llava/train/train.py", line 1006, in
train()
File "/fred/oz337/zdeng/prompt_stealing_ours/LLaVA/llava/train/train.py", line 834, in train
model = LlavaMptForCausalLM.from_pretrained(
File "/fred/oz337/zdeng/.conda/envs/llava/lib/python3.10/site-packages/transformers/modeling_utils.py", line 3850, in from_pretrained
) = cls._load_pretrained_model(
File "/fred/oz337/zdeng/.conda/envs/llava/lib/python3.10/site-packages/transformers/modeling_utils.py", line 4146, in _load_pretrained_model
raise ValueError(
ValueError: The state dictionary of the model you are trying to load is corrupted. Are you sure it was properly saved?