Root directory not created when calling .save_all_adapter

[eugene-yang]

Environment info

• adapter-transformers version: 3.0.1
• Platform: Arch Linux
• Python version: 3.10
• PyTorch version (GPU?):
• Tensorflow version (GPU?):
• Using GPU in script?:
• Using distributed or parallel set-up in script?:

Information

Model I am using (Bert, XLNet ...): XLMR

Language I am using the model on (English, Chinese ...): English

Adapter setup I am using (if any):

The problem arises when using:

• the official example scripts: (give details below)
• my own modified scripts: (give details below)

The tasks I am working on is:

• an official GLUE/SQUaD task: (give the name)
• my own task or dataset: (give details below)

To reproduce

Steps to reproduce the behavior:

from transformers import AutoConfig, AutoAdapterModel, AdapterConfig
model = AutoAdapterModel.from_pretrained('xlm-roberta-base')
config = AdapterConfig.load("pfeiffer", non_linearity="relu", reduction_factor=2)
model.load_adapter('en/wiki@ukp', config=config)
model.save_all_adapters('./test_save_adapters/') 

Trace

---------------------------------------------------------------------------
FileNotFoundError                         Traceback (most recent call last)
Input In [1], in <cell line: 5>()
      3 config = AdapterConfig.load("pfeiffer", non_linearity="relu", reduction_factor=2)
      4 model.load_adapter('en/wiki@ukp', config=config)
----> 5 model.save_all_adapters('./test_save_adapters/')

File ~/.conda/envs/adapter/lib/python3.10/site-packages/transformers/adapters/model_mixin.py:934, in ModelWithHeadsAdaptersMixin.save_all_adapters(self, save_directory, with_head, meta_dict, custom_weights_loaders)
    932 else:
    933     meta_dict = {"config_id": h}
--> 934 self.save_adapter(
    935     save_path,
    936     name,
    937     meta_dict=meta_dict,
    938     with_head=with_head,
    939     custom_weights_loaders=custom_weights_loaders,
    940 )

File ~/.conda/envs/adapter/lib/python3.10/site-packages/transformers/adapters/model_mixin.py:869, in ModelWithHeadsAdaptersMixin.save_adapter(self, save_directory, adapter_name, with_head, meta_dict, custom_weights_loaders)
    867         custom_weights_loaders = []
    868     custom_weights_loaders.append(PredictionHeadLoader(self, error_on_missing=False))
--> 869 super().save_adapter(
    870     save_directory,
    871     adapter_name,
    872     meta_dict=meta_dict,
    873     custom_weights_loaders=custom_weights_loaders,
    874 )

File ~/.conda/envs/adapter/lib/python3.10/site-packages/transformers/adapters/model_mixin.py:395, in ModelAdaptersMixin.save_adapter(self, save_directory, adapter_name, meta_dict, custom_weights_loaders)
    383 """
    384 Saves an adapter and its configuration file to a directory so that it can be shared or reloaded using
    385 `load_adapter()`.
   (...)
    392     ValueError: If the given adapter name is invalid.
    393 """
    394 loader = AdapterLoader(self)
--> 395 loader.save(save_directory, adapter_name, meta_dict)
    396 # save additional custom weights
    397 if custom_weights_loaders:

File ~/.conda/envs/adapter/lib/python3.10/site-packages/transformers/adapters/loading.py:360, in AdapterLoader.save(self, save_directory, name, meta_dict)
    351 """
    352 Saves an adapter and its configuration file to a directory, so that it can be reloaded using the `load()`
    353 method.
   (...)
    357     task_name (str): the name of the adapter to be saved
    358 """
    359 if not exists(save_directory):
--> 360     mkdir(save_directory)
    361 else:
    362     assert isdir(
    363         save_directory
    364     ), "Saving path should be a directory where adapter and configuration can be saved."

FileNotFoundError: [Errno 2] No such file or directory: './test_save_adapters/en'

This applys to .save_adapter_fusions as well.

Expected behavior

Should create the root directory first before saving each adapter.
https://github.com/adapter-hub/adapter-transformers/blob/master/src/transformers/adapters/model_mixin.py#L952

[calpt]

Thanks for bringing this up, this issue should be fixed with #375.