fix(api): pass job context and device to upscaling

api/onnx_web/diffusion/run.py

@@ -69,7 +69,7 @@ def run_txt2img_pipeline(
     )
     image = result.images[0]
     image = run_upscale_correction(
-        server, StageParams(), params, image, upscale=upscale)
+        job, server, StageParams(), params, image, upscale=upscale)
 
     dest = save_image(server, output, image)
     save_params(server, output, params, size, upscale=upscale)
@@ -109,7 +109,7 @@ def run_img2img_pipeline(
     )
     image = result.images[0]
     image = run_upscale_correction(
-        server, StageParams(), params, image, upscale=upscale)
+        job, server, StageParams(), params, image, upscale=upscale)
 
     dest = save_image(server, output, image)
     size = Size(*source_image.size)
@@ -141,7 +141,6 @@ def run_inpaint_pipeline(
     # progress = job.get_progress_callback()
     stage = StageParams()
 
-    # TODO: pass device, progress
     image = upscale_outpaint(
         server,
         stage,
@@ -162,7 +161,7 @@ def run_inpaint_pipeline(
             'output image size does not match source, skipping post-blend')
 
     image = run_upscale_correction(
-        server, stage, params, image, upscale=upscale)
+        job, server, stage, params, image, upscale=upscale)
 
     dest = save_image(server, output, image)
     save_params(server, output, params, size, upscale=upscale, border=border)
@@ -186,9 +185,8 @@ def run_upscale_pipeline(
     # progress = job.get_progress_callback()
     stage = StageParams()
 
-    # TODO: pass device, progress
     image = run_upscale_correction(
-        server, stage, params, source_image, upscale=upscale)
+        job, server, stage, params, source_image, upscale=upscale)
 
     dest = save_image(server, output, image)
     save_params(server, output, params, size, upscale=upscale)

api/onnx_web/serve.py

@@ -41,6 +41,7 @@
     ChainPipeline,
 )
 from .device_pool import (
+    DeviceParams,
     DevicePoolExecutor,
 )
 from .diffusion.run import (
@@ -168,14 +169,23 @@ def url_from_rule(rule) -> str:
     return url_for(rule.endpoint, **options)
 
 
-def pipeline_from_request() -> Tuple[ImageParams, Size]:
+def pipeline_from_request() -> Tuple[DeviceParams, ImageParams, Size]:
     user = request.remote_addr
 
+    # platform stuff
+    device_name = request.args.get('platform', available_platforms[0].device)
+    device = None
+
+    for platform in available_platforms:
+        if platform.device == device_name:
+            device = available_platforms[0]
+
+    if device is None:
+        raise Exception('unknown device')
+
     # pipeline stuff
     model = get_not_empty(request.args, 'model', get_config_value('model'))
     model_path = get_model_path(model)
-    provider = get_from_map(request.args, 'platform',
-                            platform_providers, get_config_value('platform'))
     scheduler = get_from_map(request.args, 'scheduler',
                              pipeline_schedulers, get_config_value('scheduler'))
 
@@ -213,12 +223,12 @@ def pipeline_from_request() -> Tuple[ImageParams, Size]:
         seed = np.random.randint(np.iinfo(np.int32).max)
 
     logger.info("request from %s: %s rounds of %s using %s on %s, %sx%s, %s, %s - %s",
-                user, steps, scheduler.__name__, model_path, provider, width, height, cfg, seed, prompt)
+                user, steps, scheduler.__name__, model_path, device.provider, width, height, cfg, seed, prompt)
 
-    params = ImageParams(model_path, provider, scheduler, prompt,
+    params = ImageParams(model_path, scheduler, prompt,
                          negative_prompt, cfg, steps, seed)
     size = Size(width, height)
-    return (params, size)
+    return (device, params, size)
 
 
 def border_from_request() -> Border:

api/onnx_web/upscale.py

@@ -7,6 +7,9 @@
     upscale_resrgan,
     ChainPipeline,
 )
+from .device_pool import (
+    JobContext,
+)
 from .params import (
     ImageParams,
     SizeChart,
@@ -21,7 +24,8 @@
 
 
 def run_upscale_correction(
-    ctx: ServerContext,
+    job: JobContext,
+    server: ServerContext,
     stage: StageParams,
     params: ImageParams,
     image: Image.Image,
@@ -51,4 +55,4 @@ def run_upscale_correction(
                             outscale=1)
         chain.append((correct_gfpgan, stage, None))
 
-    return chain(ctx, params, image, prompt=params.prompt, upscale=upscale)
+    return chain(job, server, params, image, prompt=params.prompt, upscale=upscale)

common/pipelines/outpaint.json

@@ -12,6 +12,7 @@
       "type": "upscale-outpaint",
       "params": {
         "border": 256,
+        "model": "stable-diffusion-onnx-v1-inpainting",
         "prompt": "a magical wizard in a robe fighting a dragon"
       }
     },