StableDiffusionInpaintPipeline的Example报错:OSError: Unable to load weights from pytorch checkpoint..._loading pipeline components.

项目地址： 

https://huggingface.co/docs/diffusers/api/pipelines/stable_diffusion/inpainthttps://huggingface.co/docs/diffusers/api/pipelines/stable_diffusion/inpaint

背景：

运行的是以下这个Example，在上一次报错后，已经把模型下载到了本地并对应修改了执行脚本中的内容

上一次是网络连接错误，解决方法是：https://blog.csdn.net/chenzhuochen7788/article/details/135825187?spm=1001.2014.3001.5501完整报错内容如下： 

(myconda) root@P4Qmo1:/mnt/Diffusers/inpainting# ls
SD_inpaint_eg.py  runwayml--stable-diffusion-inpainting
(myconda) root@P4Qmo1:/mnt/Diffusers/inpainting# python3 SD_inpaint_eg.py
Loading pipeline components...:  43%|█████▌       | 3/7 [00:29<00:39,  9.92s/it]
Traceback (most recent call last):
  File "/root/miniconda3/envs/myconda/lib/python3.8/site-packages/transformers/modeling_utils.py", line 533, in load_state_dict
    return torch.load(
  File "/root/miniconda3/envs/myconda/lib/python3.8/site-packages/torch/serialization.py", line 712, in load
    return _load(opened_zipfile, map_location, pickle_module, **pickle_load_args)
  File "/root/miniconda3/envs/myconda/lib/python3.8/site-packages/torch/serialization.py", line 1044, in _load
    unpickler = UnpicklerWrapper(data_file, **pickle_load_args)
TypeError: 'weights_only' is an invalid keyword argument for Unpickler()
 
During handling of the above exception, another exception occurred:
 
Traceback (most recent call last):
  File "/root/miniconda3/envs/myconda/lib/python3.8/site-packages/transformers/modeling_utils.py", line 542, in load_state_dict
    if f.read(7) == "version":
  File "/root/miniconda3/envs/myconda/lib/python3.8/codecs.py", line 322, in decode
    (result, consumed) = self._buffer_decode(data, self.errors, final)
UnicodeDecodeError: 'utf-8' codec can't decode byte 0x80 in position 64: invalid start byte
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
  File "SD_inpaint_eg.py", line 20, in <module>
    pipe = StableDiffusionInpaintPipeline.from_pretrained(
  File "/root/miniconda3/envs/myconda/lib/python3.8/site-packages/huggingface_hub/utils/_validators.py", line 118, in _inner_fn
    return fn(*args, **kwargs)
  File "/root/miniconda3/envs/myconda/lib/python3.8/site-packages/diffusers/pipelines/pipeline_utils.py", line 1271, in from_pretrained
    loaded_sub_model = load_sub_model(
  File "/root/miniconda3/envs/myconda/lib/python3.8/site-packages/diffusers/pipelines/pipeline_utils.py", line 525, in load_sub_model
    loaded_sub_model = load_method(os.path.join(cached_folder, name), **loading_kwargs)
  File "/root/miniconda3/envs/myconda/lib/python3.8/site-packages/transformers/modeling_utils.py", line 3527, in from_pretrained
    state_dict = load_state_dict(resolved_archive_file)
  File "/root/miniconda3/envs/myconda/lib/python3.8/site-packages/transformers/modeling_utils.py", line 554, in load_state_dict
    raise OSError(
OSError: Unable to load weights from pytorch checkpoint file for '/mnt/Diffusers/inpainting/runwayml--stable-diffusion-inpainting/text_encoder/pytorch_model.bin' at '/mnt/Diffusers/inpainting/runwayml--stable-diffusion-inpainting/text_encoder/pytorch_model.bin'. If you tried to load a PyTorch model from a TF 2.0 checkpoint, please set from_tf=True.

问题：

查看第一个报错信息：

Traceback (most recent call last):
  File "/root/miniconda3/envs/myconda/lib/python3.8/site-packages/transformers/modeling_utils.py", line 533, in load_state_dict
    return torch.load(
  File "/root/miniconda3/envs/myconda/lib/python3.8/site-packages/torch/serialization.py", line 712, in load
    return _load(opened_zipfile, map_location, pickle_module, **pickle_load_args)
  File "/root/miniconda3/envs/myconda/lib/python3.8/site-packages/torch/serialization.py", line 1044, in _load
    unpickler = UnpicklerWrapper(data_file, **pickle_load_args)
TypeError: 'weights_only' is an invalid keyword argument for Unpickler()

说的是‘weights_only'参数有问题，因为模型和代码都是官网上搞到的，如果有问题早就被别人发现了，所以怀疑是不是版本问题。

第一行：问题出在transformers这个库，感觉上是版本不兼容。

解决：

分别查看了torch和tansformers的版本：

>>> import transformers
>>> print(transformers.__version__)
4.37.0
>>> import torch
>>> print(torch.__version__)
1.11.0+cu113

torch是租用的云服务器上预装的，版本不算新。

然后觉得可能是自己装的transformers这个库版本太新了。

尝试降版本到4.36，报错解决了。（装好后pip check一下检查有没有冲突）

(myconda) root@wnyg7e:/mnt/Diffusers/sd-inpainting# pip install transformers==4.36
(myconda) root@wnyg7e:/mnt/Diffusers/sd-inpainting# pip check

之后可以运行代码：

(myconda) root@wnyg7e:/mnt/Diffusers/sd-inpainting# python3 SD_inpaint_eg.py
Loading pipeline components...:  43%|█████████████████▏                      | 3/7 [00:08<00:12,  3.10s/it]`text_config_dict` is provided which will be used to initialize `CLIPTextConfig`. The value `text_config["id2label"]` will be overriden.
`text_config_dict` is provided which will be used to initialize `CLIPTextConfig`. The value `text_config["bos_token_id"]` will be overriden.
`text_config_dict` is provided which will be used to initialize `CLIPTextConfig`. The value `text_config["eos_token_id"]` will be overriden.
Loading pipeline components...: 100%|████████████████████████████████████████| 7/7 [00:30<00:00,  4.37s/it]
100%|██████████████████████████████████████████████████████████████████████| 50/50 [00:09<00:00,  5.54it/s]
(myconda) root@wnyg7e:/mnt/Diffusers/sd-inpainting# vim SD_inpaint_eg.py
(myconda) root@wnyg7e:/mnt/Diffusers/sd-inpainting# ls
SD_inpaint.png  SD_inpaint_eg.py  stable-diffusion-inpainting