Move cache folder to huggingface/hub for consistency with hf_hub (#18492

)

* Move cache folder to just huggingface

* Thank you VsCode for this needless import

* Move to hub

* Forgot one

docs/source/en/installation.mdx

@@ -139,11 +139,11 @@ conda install -c huggingface transformers
 
 ## Cache setup
 
-Pretrained models are downloaded and locally cached at: `~/.cache/huggingface/transformers/`. This is the default directory given by the shell environment variable `TRANSFORMERS_CACHE`. On Windows, the default directory is given by `C:\Users\username\.cache\huggingface\transformers`. You can change the shell environment variables shown below - in order of priority - to specify a different cache directory:
+Pretrained models are downloaded and locally cached at: `~/.cache/huggingface/hub`. This is the default directory given by the shell environment variable `TRANSFORMERS_CACHE`. On Windows, the default directory is given by `C:\Users\username\.cache\huggingface\hub`. You can change the shell environment variables shown below - in order of priority - to specify a different cache directory:
 
-1. Shell environment variable (default): `TRANSFORMERS_CACHE`.
-2. Shell environment variable: `HF_HOME` + `transformers/`.
-3. Shell environment variable: `XDG_CACHE_HOME` + `/huggingface/transformers`.
+1. Shell environment variable (default): `HUGGINGFACE_HUB_CACHE` or `TRANSFORMERS_CACHE`.
+2. Shell environment variable: `HF_HOME`.
+3. Shell environment variable: `XDG_CACHE_HOME` + `/huggingface`.
 
 <Tip>
 

src/transformers/utils/hub.py

@@ -81,7 +81,7 @@ def is_offline_mode():
 hf_cache_home = os.path.expanduser(
     os.getenv("HF_HOME", os.path.join(os.getenv("XDG_CACHE_HOME", "~/.cache"), "huggingface"))
 )
-default_cache_path = os.path.join(hf_cache_home, "transformers")
+default_cache_path = os.path.join(hf_cache_home, "hub")
 
 # Onetime move from the old location to the new one if no ENV variable has been set.
 if (
@@ -102,7 +102,8 @@ def is_offline_mode():
 
 PYTORCH_PRETRAINED_BERT_CACHE = os.getenv("PYTORCH_PRETRAINED_BERT_CACHE", default_cache_path)
 PYTORCH_TRANSFORMERS_CACHE = os.getenv("PYTORCH_TRANSFORMERS_CACHE", PYTORCH_PRETRAINED_BERT_CACHE)
-TRANSFORMERS_CACHE = os.getenv("TRANSFORMERS_CACHE", PYTORCH_TRANSFORMERS_CACHE)
+HUGGINGFACE_HUB_CACHE = os.getenv("HUGGINGFACE_HUB_CACHE", PYTORCH_TRANSFORMERS_CACHE)
+TRANSFORMERS_CACHE = os.getenv("TRANSFORMERS_CACHE", HUGGINGFACE_HUB_CACHE)
 HF_MODULES_CACHE = os.getenv("HF_MODULES_CACHE", os.path.join(hf_cache_home, "modules"))
 TRANSFORMERS_DYNAMIC_MODULE_NAME = "transformers_modules"
 SESSION_ID = uuid4().hex
@@ -1475,9 +1476,16 @@ def move_to_new_cache(file, repo, filename, revision, etag, commit_hash):
     clean_files_for(file)
 
 
-def move_cache(cache_dir=None, token=None):
+def move_cache(cache_dir=None, new_cache_dir=None, token=None):
+    if new_cache_dir is None:
+        new_cache_dir = TRANSFORMERS_CACHE
     if cache_dir is None:
-        cache_dir = TRANSFORMERS_CACHE
+        # Migrate from old cache in .cache/huggingface/hub
+        old_cache = Path(TRANSFORMERS_CACHE).parent / "transformers"
+        if os.path.isdir(str(old_cache)):
+            cache_dir = str(old_cache)
+        else:
+            cache_dir = new_cache_dir
     if token is None:
         token = HfFolder.get_token()
     cached_files = get_all_cached_files(cache_dir=cache_dir)