Dev/offline (#356)

* update diffusers requirements.txt
* diffusers image: possibility to prefer offline mode

Useful to load private models successfully without providing any token if
the said model is already cached

docker_images/diffusers/Dockerfile

@@ -42,7 +42,6 @@ ARG max_workers=1
 ENV MAX_WORKERS=$max_workers
 ENV HUGGINGFACE_HUB_CACHE=/data
 ENV DIFFUSERS_CACHE=/data
-ENV HF_HOME=/data
 
 # Necessary on GPU environment docker.
 # TIMEOUT env variable is used by nvcr.io/nvidia/pytorch:xx for another purpose

docker_images/diffusers/app/pipelines/image_to_image.py

@@ -3,7 +3,7 @@
 import os
 
 import torch
-from app import idle, timing
+from app import idle, timing, validation
 from app.pipelines import Pipeline
 from diffusers import (
     AltDiffusionImg2ImgPipeline,
@@ -26,7 +26,7 @@
     StableUnCLIPImg2ImgPipeline,
     StableUnCLIPPipeline,
 )
-from huggingface_hub import hf_hub_download, model_info
+from huggingface_hub import file_download, hf_api, hf_hub_download, model_info, utils
 from PIL import Image
 
 
@@ -36,7 +36,36 @@
 class ImageToImagePipeline(Pipeline):
     def __init__(self, model_id: str):
         use_auth_token = os.getenv("HF_API_TOKEN")
-        model_data = model_info(model_id, token=use_auth_token)
+        self.use_auth_token = use_auth_token
+        # This should allow us to make the image work with private models when no token is provided, if the said model
+        # is already in local cache
+        self.offline_preferred = validation.str_to_bool(os.getenv("OFFLINE_PREFERRED"))
+        fetched = False
+        if self.offline_preferred:
+            cache_root = os.getenv(
+                "DIFFUSERS_CACHE", os.getenv("HUGGINGFACE_HUB_CACHE", "")
+            )
+            folder_name = file_download.repo_folder_name(
+                repo_id=model_id, repo_type="model"
+            )
+            folder_path = os.path.join(cache_root, folder_name)
+            logger.debug("Cache folder path %s", folder_path)
+            filename = os.path.join(folder_path, "hub_model_info.json")
+            try:
+                with open(filename, "r") as f:
+                    model_data = json.load(f)
+            except OSError:
+                logger.info(
+                    "No cached model info found in file %s found for model %s. Fetching on the hub",
+                    filename,
+                    model_id,
+                )
+            else:
+                model_data = hf_api.ModelInfo(**model_data)
+                fetched = True
+
+        if not fetched:
+            model_data = model_info(model_id, token=self.use_auth_token)
 
         kwargs = (
             {"safety_checker": None}
@@ -55,9 +84,26 @@ def __init__(self, model_id: str):
                 config_file_name = file_name
                 break
         if config_file_name:
-            config_file = hf_hub_download(
-                model_id, config_file_name, token=use_auth_token
-            )
+            fetched = False
+            if self.offline_preferred:
+                try:
+                    config_file = hf_hub_download(
+                        model_id,
+                        config_file_name,
+                        token=self.use_auth_token,
+                        local_files_only=True,
+                    )
+                except utils.LocalEntryNotFoundError:
+                    logger.info("Unable to fetch model index in local cache")
+                else:
+                    fetched = True
+            if not fetched:
+                config_file = hf_hub_download(
+                    model_id,
+                    config_file_name,
+                    token=self.use_auth_token,
+                )
+
             with open(config_file, "r") as f:
                 config_dict = json.load(f)
 

docker_images/diffusers/app/pipelines/text_to_image.py

@@ -4,7 +4,7 @@
 from typing import TYPE_CHECKING
 
 import torch
-from app import idle, lora, timing
+from app import idle, lora, timing, validation
 from app.pipelines import Pipeline
 from diffusers import (
     AutoencoderKL,
@@ -13,7 +13,7 @@
     EulerAncestralDiscreteScheduler,
 )
 from diffusers.schedulers.scheduling_utils import KarrasDiffusionSchedulers
-from huggingface_hub import hf_hub_download, model_info
+from huggingface_hub import file_download, hf_api, hf_hub_download, model_info, utils
 
 
 logger = logging.getLogger(__name__)
@@ -27,7 +27,35 @@ def __init__(self, model_id: str):
         self.current_lora_adapter = None
         self.model_id = None
         self.use_auth_token = os.getenv("HF_API_TOKEN")
-        model_data = model_info(model_id, token=self.use_auth_token)
+        # This should allow us to make the image work with private models when no token is provided, if the said model
+        # is already in local cache
+        self.offline_preferred = validation.str_to_bool(os.getenv("OFFLINE_PREFERRED"))
+        fetched = False
+        if self.offline_preferred:
+            cache_root = os.getenv(
+                "DIFFUSERS_CACHE", os.getenv("HUGGINGFACE_HUB_CACHE", "")
+            )
+            folder_name = file_download.repo_folder_name(
+                repo_id=model_id, repo_type="model"
+            )
+            folder_path = os.path.join(cache_root, folder_name)
+            logger.debug("Cache folder path %s", folder_path)
+            filename = os.path.join(folder_path, "hub_model_info.json")
+            try:
+                with open(filename, "r") as f:
+                    model_data = json.load(f)
+            except OSError:
+                logger.info(
+                    "No cached model info found in file %s found for model %s. Fetching on the hub",
+                    filename,
+                    model_id,
+                )
+            else:
+                model_data = hf_api.ModelInfo(**model_data)
+                fetched = True
+
+        if not fetched:
+            model_data = model_info(model_id, token=self.use_auth_token)
 
         kwargs = (
             {"safety_checker": None}
@@ -44,9 +72,26 @@ def __init__(self, model_id: str):
         if self._is_lora(model_data):
             model_type = "LoraModel"
         elif has_model_index:
-            config_file = hf_hub_download(
-                model_id, "model_index.json", token=self.use_auth_token
-            )
+            fetched = False
+            if self.offline_preferred:
+                try:
+                    config_file = hf_hub_download(
+                        model_id,
+                        "model_index.json",
+                        token=self.use_auth_token,
+                        local_files_only=True,
+                    )
+                except utils.LocalEntryNotFoundError:
+                    logger.info("Unable to fetch model index in local cache")
+                else:
+                    fetched = True
+
+            if not fetched:
+                config_file = hf_hub_download(
+                    model_id,
+                    "model_index.json",
+                    token=self.use_auth_token,
+                )
             with open(config_file, "r") as f:
                 config_dict = json.load(f)
             model_type = config_dict.get("_class_name", None)

docker_images/diffusers/app/validation.py

@@ -0,0 +1,8 @@
+import re
+
+
+STR_TO_BOOL = re.compile(r"^\s*true|yes|1\s*$", re.IGNORECASE)
+
+
+def str_to_bool(s):
+    return STR_TO_BOOL.match(str(s))

docker_images/diffusers/requirements.txt

@@ -1,8 +1,7 @@
 starlette==0.27.0
 # to be replaced with api-inference-community==0.0.33 as soon as released
 git+https://github.com/huggingface/api-inference-community.git@b3ef3f3a6015ed988ce77f71935e45006be5b054
-# to be replaced with diffusers==0.21.5 as soon as released
-git+https://github.com/huggingface/diffusers.git@ed2f956072a3b446d984f359ba6c427c259ab4ee
+diffusers==0.23.1
 transformers==4.31.0
 accelerate==0.21.0
 hf_transfer==0.1.3