fix(api): avoid circular deps in diffusion pipeline cache

api/onnx_web/__init__.py

@@ -5,9 +5,11 @@
   upscale_resrgan,
   upscale_stable_diffusion,
 )
-from .diffusion import (
+from .diffusion.load import (
   get_latents_from_seed,
   load_pipeline,
+)
+from .diffusion.run import (
   run_img2img_pipeline,
   run_inpaint_pipeline,
   run_txt2img_pipeline,

api/onnx_web/chain/blend_img2img.py

@@ -4,7 +4,7 @@
 from logging import getLogger
 from PIL import Image
 
-from ..diffusion import (
+from ..diffusion.load import (
     load_pipeline,
 )
 from ..params import (

api/onnx_web/chain/blend_inpaint.py

@@ -5,7 +5,7 @@
 from PIL import Image
 from typing import Callable, Tuple
 
-from ..diffusion import (
+from ..diffusion.load import (
     get_latents_from_seed,
     load_pipeline,
 )

api/onnx_web/chain/correct_codeformer.py

@@ -1,11 +1,14 @@
 from basicsr.utils import img2tensor, tensor2img
 from basicsr.utils.download_util import load_file_from_url
 from facexlib.utils.face_restoration_helper import FaceRestoreHelper
+from logging import getLogger
 from PIL import Image
 from torchvision.transforms.functional import normalize
 
 import torch
 
+logger = getLogger(__name__)
+
 pretrain_model_url = {
     'restoration': 'https://github.com/sczhou/CodeFormer/releases/download/v0.1.0/codeformer.pth',
 }
@@ -41,7 +44,7 @@ def correct_codeformer(image: Image.Image) -> Image.Image:
     # get face landmarks for each face
     num_det_faces = face_helper.get_face_landmarks_5(
         only_center_face=args.only_center_face, resize=640, eye_dist_threshold=5)
-    print(f'\tdetect {num_det_faces} faces')
+    logger.info('detect %s faces', num_det_faces)
     # align and warp each face
     face_helper.align_warp_face()
 
@@ -59,7 +62,7 @@ def correct_codeformer(image: Image.Image) -> Image.Image:
             del output
             torch.cuda.empty_cache()
         except Exception as error:
-            print(f'\tFailed inference for CodeFormer: {error}')
+            logger.error('Failed inference for CodeFormer: %s', error)
             restored_face = tensor2img(cropped_face_t, rgb2bgr=True, min_max=(-1, 1))
 
         restored_face = restored_face.astype('uint8')

api/onnx_web/chain/source_txt2img.py

@@ -4,7 +4,7 @@
 from logging import getLogger
 from PIL import Image
 
-from ..diffusion import (
+from ..diffusion.load import (
     get_latents_from_seed,
     load_pipeline,
 )

api/onnx_web/chain/upscale_outpaint.py

@@ -5,7 +5,7 @@
 from PIL import Image
 from typing import Callable, Tuple
 
-from ..diffusion import (
+from ..diffusion.load import (
     get_latents_from_seed,
     load_pipeline,
 )

api/onnx_web/diffusion/load.py

@@ -0,0 +1,81 @@
+from diffusers import (
+    DiffusionPipeline,
+)
+from logging import getLogger
+from typing import Any, Optional
+
+from ..params import (
+    Size,
+)
+
+import gc
+import numpy as np
+import torch
+
+logger = getLogger(__name__)
+
+last_pipeline_instance = None
+last_pipeline_options = (None, None, None)
+last_pipeline_scheduler = None
+
+
+def get_latents_from_seed(seed: int, size: Size) -> np.ndarray:
+    '''
+    From https://www.travelneil.com/stable-diffusion-updates.html
+    '''
+    # 1 is batch size
+    latents_shape = (1, 4, size.height // 8, size.width // 8)
+    # Gotta use numpy instead of torch, because torch's randn() doesn't support DML
+    rng = np.random.default_rng(seed)
+    image_latents = rng.standard_normal(latents_shape).astype(np.float32)
+    return image_latents
+
+
+def load_pipeline(pipeline: DiffusionPipeline, model: str, provider: str, scheduler: Any, device: Optional[str] = None):
+    global last_pipeline_instance
+    global last_pipeline_scheduler
+    global last_pipeline_options
+
+    options = (pipeline, model, provider)
+    if last_pipeline_instance != None and last_pipeline_options == options:
+        logger.info('reusing existing diffusion pipeline')
+        pipe = last_pipeline_instance
+    else:
+        logger.info('unloading previous diffusion pipeline')
+        last_pipeline_instance = None
+        last_pipeline_scheduler = None
+        gc.collect()
+        torch.cuda.empty_cache()
+
+        logger.info('loading new diffusion pipeline')
+        pipe = pipeline.from_pretrained(
+            model,
+            provider=provider,
+            safety_checker=None,
+            scheduler=scheduler.from_pretrained(model, subfolder='scheduler')
+        )
+
+        if device is not None:
+            pipe = pipe.to(device)
+
+        last_pipeline_instance = pipe
+        last_pipeline_options = options
+        last_pipeline_scheduler = scheduler
+
+    if last_pipeline_scheduler != scheduler:
+        logger.info('loading new diffusion scheduler')
+        scheduler = scheduler.from_pretrained(
+            model, subfolder='scheduler')
+
+        if device is not None:
+            scheduler = scheduler.to(device)
+
+        pipe.scheduler = scheduler
+        last_pipeline_scheduler = scheduler
+
+    logger.info('running garbage collection during pipeline change')
+    gc.collect()
+
+    return pipe
+
+

api/onnx_web/diffusion.py → api/onnx_web/diffusion/run.py

@@ -1,106 +1,39 @@
 from diffusers import (
-    DiffusionPipeline,
-    # onnx
     OnnxStableDiffusionPipeline,
     OnnxStableDiffusionImg2ImgPipeline,
     OnnxStableDiffusionInpaintPipeline,
 )
 from logging import getLogger
 from PIL import Image, ImageChops
-from typing import Any, Optional
+from typing import Any
 
-from .chain import (
-    StageParams,
-)
-from .image import (
+from ..image import (
     expand_image,
 )
-from .params import (
+from ..params import (
     ImageParams,
     Border,
     Size,
+    StageParams,
 )
-from .upscale import (
+from ..upscale import (
     run_upscale_correction,
     UpscaleParams,
 )
-from .utils import (
+from ..utils import (
     is_debug,
     base_join,
     ServerContext,
 )
+from .load import (
+  get_latents_from_seed,
+  load_pipeline,
+)
 
-import gc
 import numpy as np
-import torch
 
 logger = getLogger(__name__)
 
-last_pipeline_instance = None
-last_pipeline_options = (None, None, None)
-last_pipeline_scheduler = None
-
-
-def get_latents_from_seed(seed: int, size: Size) -> np.ndarray:
-    '''
-    From https://www.travelneil.com/stable-diffusion-updates.html
-    '''
-    # 1 is batch size
-    latents_shape = (1, 4, size.height // 8, size.width // 8)
-    # Gotta use numpy instead of torch, because torch's randn() doesn't support DML
-    rng = np.random.default_rng(seed)
-    image_latents = rng.standard_normal(latents_shape).astype(np.float32)
-    return image_latents
-
-
-def load_pipeline(pipeline: DiffusionPipeline, model: str, provider: str, scheduler: Any, device: Optional[str] = None):
-    global last_pipeline_instance
-    global last_pipeline_scheduler
-    global last_pipeline_options
-
-    options = (pipeline, model, provider)
-    if last_pipeline_instance != None and last_pipeline_options == options:
-        logger.info('reusing existing diffusion pipeline')
-        pipe = last_pipeline_instance
-    else:
-        logger.info('unloading previous diffusion pipeline')
-        last_pipeline_instance = None
-        last_pipeline_scheduler = None
-        gc.collect()
-        torch.cuda.empty_cache()
-
-        logger.info('loading new diffusion pipeline')
-        pipe = pipeline.from_pretrained(
-            model,
-            provider=provider,
-            safety_checker=None,
-            scheduler=scheduler.from_pretrained(model, subfolder='scheduler')
-        )
-
-        if device is not None:
-            pipe = pipe.to(device)
-
-        last_pipeline_instance = pipe
-        last_pipeline_options = options
-        last_pipeline_scheduler = scheduler
-
-    if last_pipeline_scheduler != scheduler:
-        logger.info('loading new diffusion scheduler')
-        scheduler = scheduler.from_pretrained(
-            model, subfolder='scheduler')
-
-        if device is not None:
-            scheduler = scheduler.to(device)
-
-        pipe.scheduler = scheduler
-        last_pipeline_scheduler = scheduler
-
-    logger.info('running garbage collection during pipeline change')
-    gc.collect()
-
-    return pipe
-
-
 def run_txt2img_pipeline(
     ctx: ServerContext,
     params: ImageParams,

api/onnx_web/serve.py

@@ -35,7 +35,7 @@
     upscale_stable_diffusion,
     ChainPipeline,
 )
-from .diffusion import (
+from .diffusion.run import (
     run_img2img_pipeline,
     run_inpaint_pipeline,
     run_txt2img_pipeline,