fix(api): replace some numpy RNGs with torch equivalent

api/onnx_web/chain/blend_img2img.py

@@ -1,6 +1,6 @@
 from logging import getLogger
 
-import numpy as np
+import torch
 from diffusers import OnnxStableDiffusionImg2ImgPipeline
 from PIL import Image
 
@@ -33,7 +33,7 @@ def blend_img2img(
         job.get_device(),
     )
 
-    rng = np.random.RandomState(params.seed)
+    rng = torch.manual_seed(params.seed)
 
     result = pipe(
         prompt,

api/onnx_web/chain/blend_inpaint.py

@@ -1,7 +1,7 @@
 from logging import getLogger
 from typing import Callable, Tuple
 
-import numpy as np
+import torch
 from diffusers import OnnxStableDiffusionInpaintPipeline
 from PIL import Image
 
@@ -67,7 +67,7 @@ def blend_inpaint(
         )
 
         latents = get_latents_from_seed(params.seed, size)
-        rng = np.random.RandomState(params.seed)
+        rng = torch.manual_seed(params.seed)
 
         result = pipe(
             params.prompt,

api/onnx_web/chain/source_txt2img.py

@@ -1,6 +1,7 @@
 from logging import getLogger
 
 import numpy as np
+import torch
 from diffusers import OnnxStableDiffusionPipeline
 from PIL import Image
 
@@ -36,7 +37,7 @@ def source_txt2img(
     )
 
     latents = get_latents_from_seed(params.seed, size)
-    rng = np.random.RandomState(params.seed)
+    rng = torch.manual_seed(params.seed)
 
     result = pipe(
         prompt,

api/onnx_web/chain/upscale_outpaint.py

@@ -1,7 +1,7 @@
 from logging import getLogger
 from typing import Callable, Tuple
 
-import numpy as np
+import torch
 from diffusers import OnnxStableDiffusionInpaintPipeline
 from PIL import Image, ImageDraw
 
@@ -73,7 +73,7 @@ def upscale_outpaint(
         )
 
         latents = get_tile_latents(full_latents, dims)
-        rng = np.random.RandomState(params.seed)
+        rng = torch.manual_seed(params.seed)
 
         result = pipe.inpaint(
             image,

api/onnx_web/diffusion/load.py

@@ -19,7 +19,8 @@ latent_factor = 8
 
 def get_latents_from_seed(seed: int, size: Size, batch: int = 1) -> np.ndarray:
     """
-    From https://www.travelneil.com/stable-diffusion-updates.html
+    From https://www.travelneil.com/stable-diffusion-updates.html.
+    This one needs to use np.random because of the return type.
     """
     latents_shape = (
         batch,

api/onnx_web/diffusion/run.py

@@ -2,6 +2,7 @@ from logging import getLogger
 from typing import Any
 
 import numpy as np
+import torch
 from diffusers import OnnxStableDiffusionImg2ImgPipeline, OnnxStableDiffusionPipeline
 from PIL import Image, ImageChops
 
@@ -29,7 +30,7 @@ def run_txt2img_pipeline(
     )
 
     latents = get_latents_from_seed(params.seed, size)
-    rng = np.random.RandomState(params.seed)
+    rng = torch.manual_seed(params.seed)
 
     progress = job.get_progress_callback()
     result = pipe.text2img(
@@ -74,7 +75,7 @@ def run_img2img_pipeline(
         job.get_device(),
     )
 
-    rng = np.random.RandomState(params.seed)
+    rng = torch.manual_seed(params.seed)
 
     progress = job.get_progress_callback()
     result = pipe.img2img(

api/onnx_web/serve.py

@@ -215,6 +215,7 @@ def pipeline_from_request() -> Tuple[DeviceParams, ImageParams, Size]:
 
     seed = int(request.args.get("seed", -1))
     if seed == -1:
+        # this one can safely use np.random because it produces a single value
         seed = np.random.randint(np.iinfo(np.int32).max)
 
     logger.info(