erge branch 'master' of http://git.nakadashi.lol/akko/diffvg

apps/imgs/Triangle.obj

@@ -1,5 +0,0 @@
-v 100 150
-v 42.3 50
-v 157.7 50
-
-f 1 2 3

apps/test_triangle.py

@@ -3,140 +3,146 @@ import torch
 import skimage
 import numpy as np
 
-# Use GPU if available
-pydiffvg.set_use_gpu(torch.cuda.is_available())
+def run():
 
-canvas_width, canvas_height = 256, 256
-num_control_points = torch.tensor([2, 2, 2])
-points = torch.tensor([[20.0,  30.0], # base
-                       [50.0,  60.0], # control point
-                       [ 90.0, 198.0], # control point
-                       [ 60.0, 218.0], # base
-                       [ 90.0, 180.0], # control point
-                       [200.0,  85.0], # control point
-                       [230.0,  90.0], # base
-                       [220.0,  70.0], # control point
-                       [130.0,  55.0]]) # control point
-path = pydiffvg.Path(num_control_points = num_control_points,
-                     points = points,
-                     is_closed = True)
-shapes = [path]
-path_group = pydiffvg.ShapeGroup(shape_ids = torch.tensor([0]),
-                                 fill_color = torch.tensor([0.3, 0.6, 0.3, 1.0]))
-shape_groups = [path_group]
-scene_args = pydiffvg.RenderFunction.serialize_scene(\
-    canvas_width, canvas_height, shapes, shape_groups)
+    # Use GPU if available
+    pydiffvg.set_use_gpu(torch.cuda.is_available())
 
-render = pydiffvg.RenderFunction.apply
-img = render(256, # width
-             256, # height
-             2,   # num_samples_x
-             2,   # num_samples_y
-             0,   # seed
-             None,
-             *scene_args)
-# The output image is in linear RGB space. Do Gamma correction before saving the image.
-pydiffvg.imwrite(img.cpu(), 'results/test_curve/target.png', gamma=2.2)
-target = img.clone()
+    canvas_width, canvas_height = 256, 256
+    num_control_points = torch.tensor([2, 2, 2])
+    points = torch.tensor([[20.0,  30.0], # base
+                        [50.0,  60.0], # control point
+                        [ 90.0, 198.0], # control point
+                        [ 60.0, 218.0], # base
+                        [ 90.0, 180.0], # control point
+                        [200.0,  85.0], # control point
+                        [230.0,  90.0], # base
+                        [220.0,  70.0], # control point
+                        [130.0,  55.0]]) # control point
+    path = pydiffvg.Path(num_control_points = num_control_points,
+                        points = points,
+                        is_closed = True)
+    shapes = [path]
+    path_group = pydiffvg.ShapeGroup(shape_ids = torch.tensor([0]),
+                                    fill_color = torch.tensor([0.3, 0.6, 0.3, 1.0]))
+    shape_groups = [path_group]
+    scene_args = pydiffvg.RenderFunction.serialize_scene(\
+        canvas_width, canvas_height, shapes, shape_groups)
 
-# Load the obj file, get the vertices/control points
-obj = "imgs/Triangle.obj"
-vertices_tmp, faces_tmp = pydiffvg.obj_to_scene(obj)
-print(float(vertices_tmp[1][1]))
-print(faces_tmp)
+    render = pydiffvg.RenderFunction.apply
+    img = render(256, # width
+                256, # height
+                2,   # num_samples_x
+                2,   # num_samples_y
+                0,   # seed
+                None,
+                *scene_args)
+    # The output image is in linear RGB space. Do Gamma correction before saving the image.
+    pydiffvg.imwrite(img.cpu(), 'results/test_curve/target.png', gamma=2.2)
+    target = img.clone()
 
-vertices = []
-faces = []
-for v in vertices_tmp:
-    vertices.append([float(v[1]), float(v[2])])
-for f in faces_tmp:
-    vs_count = len(f)
-    tmp = []
-    for v in f[1:]:
-        tmp.append(int(v))
-    faces.append(tmp)
-print(vertices)
-print(faces)
+    # Load the obj file, get the vertices/control points
+    obj = "imgs/Triangle.obj"
+    vertices_tmp, faces_tmp = pydiffvg.obj_to_scene(obj)
+    print(float(vertices_tmp[1][1]))
+    print(faces_tmp)
 
-# Ternary subdivision
-# Move the path to produce initial guess
-# normalize points for easier learning rate
-points_n = torch.tensor([[vertices[0][0]/256.0, vertices[0][1]/256.0], # base
-                         [70.0/256.0,  140.0/256.0], # control point
-                         [50.0/256.0, 100.0/256.0], # control point
-                         [vertices[1][0]/256.0, vertices[1][1]/256.0], # base
-                         [80.0/256.0, 40.0/256.0], # control point
-                         [120.0/256.0,  40.0/256.0], # control point
-                         [vertices[2][0]/256.0, vertices[2][1]/256.0], # base
-                         [150.0/256.0,  100.0/256.0], # control point
-                         [130.0/256.0,  140.0/256.0]], # control point
-                        requires_grad = True) 
-color = torch.tensor([0.3, 0.2, 0.5, 1.0], requires_grad=True)
-path.points = points_n * 256
-path_group.fill_color = color
-scene_args = pydiffvg.RenderFunction.serialize_scene(\
-    canvas_width, canvas_height, shapes, shape_groups)
-img = render(256, # width
-             256, # height
-             2,   # num_samples_x
-             2,   # num_samples_y
-             1,   # seed
-             None,
-             *scene_args)
-pydiffvg.imwrite(img.cpu(), 'results/test_curve/init.png', gamma=2.2)
+    vertices = []
+    faces = []
+    for v in vertices_tmp:
+        vertices.append([float(v[1]), float(v[2])])
+    for f in faces_tmp:
+        vs_count = len(f)
+        tmp = []
+        for v in f[1:]:
+            tmp.append(int(v))
+        faces.append(tmp)
+    print(vertices)
+    print(faces)
 
-# Optimize
-optimizer = torch.optim.Adam([points_n, color], lr=1e-2)
-# Run 100 Adam iterations.
-for t in range(100):
-    print('iteration:', t)
-    optimizer.zero_grad()
-    # Forward pass: render the image.
+    # Ternary subdivision
+    # Move the path to produce initial guess
+    # normalize points for easier learning rate
+    points_n = torch.tensor([[vertices[0][0]/256.0, vertices[0][1]/256.0], # base
+                            [70.0/256.0,  140.0/256.0], # control point
+                            [50.0/256.0, 100.0/256.0], # control point
+                            [vertices[1][0]/256.0, vertices[1][1]/256.0], # base
+                            [80.0/256.0, 40.0/256.0], # control point
+                            [120.0/256.0,  40.0/256.0], # control point
+                            [vertices[2][0]/256.0, vertices[2][1]/256.0], # base
+                            [150.0/256.0,  100.0/256.0], # control point
+                            [130.0/256.0,  140.0/256.0]], # control point
+                            requires_grad = True)
+    color = torch.tensor([0.3, 0.2, 0.5, 1.0], requires_grad=True)
     path.points = points_n * 256
     path_group.fill_color = color
+    scene_args = pydiffvg.RenderFunction.serialize_scene(\
+        canvas_width, canvas_height, shapes, shape_groups)
+    return scene_args
+    img = render(256, # width
+                256, # height
+                2,   # num_samples_x
+                2,   # num_samples_y
+                1,   # seed
+                None,
+                *scene_args)
+    pydiffvg.imwrite(img.cpu(), 'results/test_curve/init.png', gamma=2.2)
+
+    # Optimize
+    optimizer = torch.optim.Adam([points_n, color], lr=1e-2)
+    # Run 100 Adam iterations.
+    for t in range(100):
+        print('iteration:', t)
+        optimizer.zero_grad()
+        # Forward pass: render the image.
+        path.points = points_n * 256
+        path_group.fill_color = color
+        scene_args = pydiffvg.RenderFunction.serialize_scene(\
+            canvas_width, canvas_height, shapes, shape_groups)
+        img = render(256,   # width
+                    256,   # height
+                    2,     # num_samples_x
+                    2,     # num_samples_y
+                    t+1,   # seed
+                    None,
+                    *scene_args)
+        # Save the intermediate render.
+        pydiffvg.imwrite(img.cpu(),
+                         'results/test_curve/iter_{}.png'.format(t), gamma=2.2)
+        # Compute the loss function. Here it is L2.
+        loss = (img - target).pow(2).sum()
+        print('loss:', loss.item())
+
+        # Backpropagate the gradients.
+        loss.backward()
+        # Print the gradients
+        print('points_n.grad:', points_n.grad)
+        print('color.grad:', color.grad)
+
+        # Take a gradient descent step.
+        optimizer.step()
+        # Print the current params.
+        print('points:', path.points)
+        print('color:', path_group.fill_color)
+
+    # Render the final result.
     scene_args = pydiffvg.RenderFunction.serialize_scene(\
         canvas_width, canvas_height, shapes, shape_groups)
     img = render(256,   # width
-                 256,   # height
-                 2,     # num_samples_x
-                 2,     # num_samples_y
-                 t+1,   # seed
-                 None,
-                 *scene_args)
-    # Save the intermediate render.
-    pydiffvg.imwrite(img.cpu(), 'results/test_curve/iter_{}.png'.format(t), gamma=2.2)
-    # Compute the loss function. Here it is L2.
-    loss = (img - target).pow(2).sum()
-    print('loss:', loss.item())
+                256,   # height
+                2,     # num_samples_x
+                2,     # num_samples_y
+                102,    # seed
+                None,
+                *scene_args)
+    # Save the images and differences.
+    pydiffvg.imwrite(img.cpu(), 'results/test_curve/final.png')
 
-    # Backpropagate the gradients.
-    loss.backward()
-    # Print the gradients
-    print('points_n.grad:', points_n.grad)
-    print('color.grad:', color.grad)
-
-    # Take a gradient descent step.
-    optimizer.step()
-    # Print the current params.
-    print('points:', path.points)
-    print('color:', path_group.fill_color)
-
-# Render the final result.
-scene_args = pydiffvg.RenderFunction.serialize_scene(\
-    canvas_width, canvas_height, shapes, shape_groups)
-img = render(256,   # width
-             256,   # height
-             2,     # num_samples_x
-             2,     # num_samples_y
-             102,    # seed
-             None,
-             *scene_args)
-# Save the images and differences.
-pydiffvg.imwrite(img.cpu(), 'results/test_curve/final.png')
-
-# Convert the intermediate renderings to a video.
-from subprocess import call
-call(["ffmpeg", "-framerate", "24", "-i",
-    "results/test_curve/iter_%d.png", "-vb", "20M",
-    "results/test_curve/out.mp4"])
+    # Convert the intermediate renderings to a video.
+    from subprocess import call
+    call(["ffmpeg", "-framerate", "24", "-i",
+        "results/test_curve/iter_%d.png", "-vb", "20M",
+        "results/test_curve/out.mp4"])
 
+if __name__ == "__main__":
+    run()

color.h

@@ -7,7 +7,8 @@
 enum class ColorType {
     Constant,
     LinearGradient,
-    RadialGradient
+    RadialGradient,
+    PatchGradient
 };
 
 struct Constant {
@@ -61,3 +62,13 @@ struct RadialGradient {
     float *stop_offsets;
     float *stop_colors; // rgba
 };
+
+/*
+struct PatchGradient {
+    PatchGradient(const Vector2f &topLeft,
+                  const Vector2f &topRight,
+                  const Vector2f &bottomLeft,
+                  const Vector2f &bottomRight,
+                  const )
+}
+*/

diffvg.cpp

@@ -1490,6 +1490,11 @@ void render(std::shared_ptr<Scene> scene,
             bool use_prefiltering,
             ptr<float> eval_positions,
             int num_eval_positions) {
+
+    /////////////////
+    // Setup stuff //
+    /////////////////
+
 #ifdef __NVCC__
     int old_device_id = -1;
     if (scene->use_gpu) {
@@ -1519,6 +1524,11 @@ void render(std::shared_ptr<Scene> scene,
         }
     }
 
+    ///////////////////////////////////////////////////////////////////
+    // IDK what this does but I think it relates to the prefiltering //
+    // TODO                                                          //
+    ///////////////////////////////////////////////////////////////////
+
     if (render_image.get() != nullptr || d_render_image.get() != nullptr ||
         render_sdf.get() != nullptr || d_render_sdf.get() != nullptr) {
         if (weight_image != nullptr) {
@@ -1533,6 +1543,11 @@ void render(std::shared_ptr<Scene> scene,
             }, width * height * num_samples_x * num_samples_y, scene->use_gpu);
         }
 
+
+        ////////////////////////////////////////////////
+        // Think this relates to the actual rendering //
+        ////////////////////////////////////////////////
+
         auto num_samples = eval_positions.get() == nullptr ?
             width * height * num_samples_x * num_samples_y : num_eval_positions;
         parallel_for(render_kernel{
@@ -1555,6 +1570,11 @@ void render(std::shared_ptr<Scene> scene,
         }, num_samples, scene->use_gpu);
     }
 
+    /////////////////////////////////////////////////////////////////////
+    // The reason for this is described in the diffvg paper            //
+    // I don't think it's especially important for the gradient meshes //
+    /////////////////////////////////////////////////////////////////////
+
     // Boundary sampling
     if (!use_prefiltering && d_render_image.get() != nullptr) {
         auto num_samples = width * height * num_samples_x * num_samples_y;

pydiffvg/parse_svg.py

@@ -586,13 +586,17 @@ def obj_to_scene(filename):
     """
         Load from a obj file and convert to PyTorch tensors.
     """
-    f = open(filename)
-    lines = f.readlines()
+    with open(filename, 'r') as f:
+        lines = f.readlines()
+
     data_lines = []
     vertices = []
     faces = []
     for line in lines:
         l=line.split()
+        # Ignore empty lines
+        if not l:
+            continue
         #vertex
         if(l[0] == "v"):
             vertices.append(l)

pydiffvg/render_pytorch.py

@@ -7,6 +7,9 @@ import warnings
 
 print_timing = False
 
+def popmult(lst, n):
+    return (lst[:n], lst[n:])
+
 def set_print_timing(val):
     global print_timing
     print_timing=val
@@ -35,98 +38,128 @@ class RenderFunction(torch.autograd.Function):
         """
         num_shapes = len(shapes)
         num_shape_groups = len(shape_groups)
-        args = []
-        args.append(canvas_width)
-        args.append(canvas_height)
-        args.append(num_shapes)
-        args.append(num_shape_groups)
-        args.append(output_type)
-        args.append(use_prefiltering)
-        args.append(eval_positions.to(pydiffvg.get_device()))
+        args = [canvas_width,
+                canvas_height,
+                num_shapes,
+                num_shape_groups,
+                output_type,
+                use_prefiltering,
+                eval_positions.to(pydiffvg.get_device())]
+
         for shape in shapes:
             use_thickness = False
             if isinstance(shape, pydiffvg.Circle):
-                assert(shape.center.is_contiguous())
-                args.append(diffvg.ShapeType.circle)
-                args.append(shape.radius.cpu())
-                args.append(shape.center.cpu())
+                assert shape.center.is_contiguous()
+                args += [
+                    diffvg.ShapeType.circle,
+                    shape.radius.cpu(),
+                    shape.center.cpu()
+                ]
             elif isinstance(shape, pydiffvg.Ellipse):
-                assert(shape.radius.is_contiguous())
-                assert(shape.center.is_contiguous())
-                args.append(diffvg.ShapeType.ellipse)
-                args.append(shape.radius.cpu())
-                args.append(shape.center.cpu())
+                assert shape.radius.is_contiguous()
+                assert shape.center.is_contiguous()
+                args += [
+                    diffvg.ShapeType.ellipse,
+                    shape.radius.cpu(),
+                    shape.center.cpu()
+                ]
             elif isinstance(shape, pydiffvg.Path):
-                assert(shape.num_control_points.is_contiguous())
-                assert(shape.points.is_contiguous())
-                assert(shape.points.shape[1] == 2)
-                assert(torch.isfinite(shape.points).all())
-                args.append(diffvg.ShapeType.path)
-                args.append(shape.num_control_points.to(torch.int32).cpu())
-                args.append(shape.points.cpu())
+                assert shape.num_control_points.is_contiguous()
+                assert shape.points.is_contiguous()
+                assert shape.points.shape[1] == 2
+                assert torch.isfinite(shape.points).all()
+
+                args += [
+                    diffvg.ShapeType.path,
+                    shape.num_control_points.to(torch.int32).cpu(),
+                    shape.points.cpu()
+                ]
+
                 if len(shape.stroke_width.shape) > 0 and shape.stroke_width.shape[0] > 1:
-                    assert(torch.isfinite(shape.stroke_width).all())
+                    assert torch.isfinite(shape.stroke_width).all()
                     use_thickness = True
                     args.append(shape.stroke_width.cpu())
                 else:
                     args.append(None)
-                args.append(shape.is_closed)
-                args.append(shape.use_distance_approx)
+
+                args += [shape.is_closed, shape.use_distance_approx]
             elif isinstance(shape, pydiffvg.Polygon):
-                assert(shape.points.is_contiguous())
-                assert(shape.points.shape[1] == 2)
+                assert shape.points.is_contiguous()
+                assert shape.points.shape[1] == 2
+
                 args.append(diffvg.ShapeType.path)
+
                 if shape.is_closed:
                     args.append(torch.zeros(shape.points.shape[0], dtype = torch.int32))
                 else:
                     args.append(torch.zeros(shape.points.shape[0] - 1, dtype = torch.int32))
-                args.append(shape.points.cpu())
-                args.append(None)  
-                args.append(shape.is_closed)
-                args.append(False) # use_distance_approx
+
+                args += [
+                    shape.points.cpu(),
+                    None,
+                    shape.is_closed(),
+                    False # use_distance_approx
+                ]
+
             elif isinstance(shape, pydiffvg.Rect):
-                assert(shape.p_min.is_contiguous())
-                assert(shape.p_max.is_contiguous())
-                args.append(diffvg.ShapeType.rect)
-                args.append(shape.p_min.cpu())
-                args.append(shape.p_max.cpu())
+                assert shape.p_min.is_contiguous()
+                assert shape.p_max.is_contiguous()
+
+                args += [
+                    diffvg.ShapeType.rect,
+                    shape.p_min.cpu(),
+                    shape.p_max.cpu()
+                ]
             else:
-                assert(False)
+                assert False
+
             if use_thickness:
                 args.append(torch.tensor(0.0))
             else:
                 args.append(shape.stroke_width.cpu())
 
         for shape_group in shape_groups:
-            assert(shape_group.shape_ids.is_contiguous())
+            assert shape_group.shape_ids.is_contiguous()
             args.append(shape_group.shape_ids.to(torch.int32).cpu())
             # Fill color
             if shape_group.fill_color is None:
                 args.append(None)
+
             elif isinstance(shape_group.fill_color, torch.Tensor):
-                assert(shape_group.fill_color.is_contiguous())
-                args.append(diffvg.ColorType.constant)
-                args.append(shape_group.fill_color.cpu())
+                assert shape_group.fill_color.is_contiguous()
+
+                args += [
+                    diffvg.ColorType.constant,
+                    shape_group.fill_color.cpu()
+                ]
+
             elif isinstance(shape_group.fill_color, pydiffvg.LinearGradient):
-                assert(shape_group.fill_color.begin.is_contiguous())
-                assert(shape_group.fill_color.end.is_contiguous())
-                assert(shape_group.fill_color.offsets.is_contiguous())
-                assert(shape_group.fill_color.stop_colors.is_contiguous())
-                args.append(diffvg.ColorType.linear_gradient)
-                args.append(shape_group.fill_color.begin.cpu())
-                args.append(shape_group.fill_color.end.cpu())
-                args.append(shape_group.fill_color.offsets.cpu())
-                args.append(shape_group.fill_color.stop_colors.cpu())
+                assert shape_group.fill_color.begin.is_contiguous()
+                assert shape_group.fill_color.end.is_contiguous()
+                assert shape_group.fill_color.offsets.is_contiguous()
+                assert shape_group.fill_color.stop_colors.is_contiguous()
+
+                args += [
+                    diffvg.ColorType.linear_gradient,
+                    shape_group.fill_color.begin.cpu(),
+                    shape_group.fill_color.end.cpu(),
+                    shape_group.fill_color.offsets.cpu(),
+                    shape_group.fill_color.stop_colors.cpu(),
+                ]
+
             elif isinstance(shape_group.fill_color, pydiffvg.RadialGradient):
-                assert(shape_group.fill_color.center.is_contiguous())
-                assert(shape_group.fill_color.radius.is_contiguous())
-                assert(shape_group.fill_color.offsets.is_contiguous())
-                assert(shape_group.fill_color.stop_colors.is_contiguous())
-                args.append(diffvg.ColorType.radial_gradient)
-                args.append(shape_group.fill_color.center.cpu())
-                args.append(shape_group.fill_color.radius.cpu())
-                args.append(shape_group.fill_color.offsets.cpu())
-                args.append(shape_group.fill_color.stop_colors.cpu())
+                assert shape_group.fill_color.center.is_contiguous()
+                assert shape_group.fill_color.radius.is_contiguous()
+                assert shape_group.fill_color.offsets.is_contiguous()
+                assert shape_group.fill_color.stop_colors.is_contiguous()
+
+                args += [
+                    diffvg.ColorType.radial_gradient,
+                    shape_group.fill_color.center.cpu(),
+                    shape_group.fill_color.radius.cpu(),
+                    shape_group.fill_color.offsets.cpu(),
+                    shape_group.fill_color.stop_colors.cpu()
+                ]
 
             if shape_group.fill_color is not None:
                 # go through the underlying shapes and check if they are all closed
@@ -138,35 +171,44 @@ class RenderFunction(torch.autograd.Function):
             # Stroke color
             if shape_group.stroke_color is None:
                 args.append(None)
+
             elif isinstance(shape_group.stroke_color, torch.Tensor):
-                assert(shape_group.stroke_color.is_contiguous())
-                args.append(diffvg.ColorType.constant)
-                args.append(shape_group.stroke_color.cpu())
+                assert shape_group.stroke_color.is_contiguous()
+                args += [ diffvg.ColorType.constant, shape_group.stroke_color.cpu() ]
+
             elif isinstance(shape_group.stroke_color, pydiffvg.LinearGradient):
-                assert(shape_group.stroke_color.begin.is_contiguous())
-                assert(shape_group.stroke_color.end.is_contiguous())
-                assert(shape_group.stroke_color.offsets.is_contiguous())
-                assert(shape_group.stroke_color.stop_colors.is_contiguous())
-                assert(torch.isfinite(shape_group.stroke_color.stop_colors).all())
-                args.append(diffvg.ColorType.linear_gradient)
-                args.append(shape_group.stroke_color.begin.cpu())
-                args.append(shape_group.stroke_color.end.cpu())
-                args.append(shape_group.stroke_color.offsets.cpu())
-                args.append(shape_group.stroke_color.stop_colors.cpu())
+                assert shape_group.stroke_color.begin.is_contiguous()
+                assert shape_group.stroke_color.end.is_contiguous()
+                assert shape_group.stroke_color.offsets.is_contiguous()
+                assert shape_group.stroke_color.stop_colors.is_contiguous()
+                assert torch.isfinite(shape_group.stroke_color.stop_colors).all()
+
+                args += [
+                    diffvg.ColorType.linear_gradient,
+                    shape_group.stroke_color.begin.cpu(),
+                    shape_group.stroke_color.end.cpu(),
+                    shape_group.stroke_color.offsets.cpu(),
+                    shape_group.stroke_color.stop_colors.cpu()
+                ]
+
             elif isinstance(shape_group.stroke_color, pydiffvg.RadialGradient):
-                assert(shape_group.stroke_color.center.is_contiguous())
-                assert(shape_group.stroke_color.radius.is_contiguous())
-                assert(shape_group.stroke_color.offsets.is_contiguous())
-                assert(shape_group.stroke_color.stop_colors.is_contiguous())
-                assert(torch.isfinite(shape_group.stroke_color.stop_colors).all())
-                args.append(diffvg.ColorType.radial_gradient)
-                args.append(shape_group.stroke_color.center.cpu())
-                args.append(shape_group.stroke_color.radius.cpu())
-                args.append(shape_group.stroke_color.offsets.cpu())
-                args.append(shape_group.stroke_color.stop_colors.cpu())
-            args.append(shape_group.use_even_odd_rule)
-            # Transformation
-            args.append(shape_group.shape_to_canvas.contiguous().cpu())
+                assert shape_group.stroke_color.center.is_contiguous()
+                assert shape_group.stroke_color.radius.is_contiguous()
+                assert shape_group.stroke_color.offsets.is_contiguous()
+                assert shape_group.stroke_color.stop_colors.is_contiguous()
+                assert torch.isfinite(shape_group.stroke_color.stop_colors).all()
+
+                args += [
+                    diffvg.ColorType.radial_gradient,
+                    shape_group.stroke_color.center.cpu(),
+                    shape_group.stroke_color.radius.cpu(),
+                    shape_group.stroke_color.offsets.cpu(),
+                    shape_group.stroke_color.stop_colors.cpu()
+                ]
+
+            args += [ shape_group.use_even_odd_rule,
+                      shape_group.shape_to_canvas.contiguous().cpu() ]
+
         args.append(filter.type)
         args.append(filter.radius.cpu())
         return args
@@ -184,52 +226,31 @@ class RenderFunction(torch.autograd.Function):
             Forward rendering pass.
         """
         # Unpack arguments
-        current_index = 0
-        canvas_width = args[current_index]
-        current_index += 1
-        canvas_height = args[current_index]
-        current_index += 1
-        num_shapes = args[current_index]
-        current_index += 1
-        num_shape_groups = args[current_index]
-        current_index += 1
-        output_type = args[current_index]
-        current_index += 1
-        use_prefiltering = args[current_index]
-        current_index += 1
-        eval_positions = args[current_index]
-        current_index += 1
+        args = list(args)
+
+        (canvas_width, canvas_height, num_shapes,
+         num_shape_groups, output_type, use_prefiltering, eval_positions), args = popmult(args, 7)
+
         shapes = []
         shape_groups = []
         shape_contents = [] # Important to avoid GC deleting the shapes
         color_contents = [] # Same as above
         for shape_id in range(num_shapes):
-            shape_type = args[current_index]
-            current_index += 1
+            shape_type = args.pop(0)
             if shape_type == diffvg.ShapeType.circle:
-                radius = args[current_index]
-                current_index += 1
-                center = args[current_index]
-                current_index += 1
+                (radius, center), args = popmult(args, 2)
+
                 shape = diffvg.Circle(radius, diffvg.Vector2f(center[0], center[1]))
             elif shape_type == diffvg.ShapeType.ellipse:
-                radius = args[current_index]
-                current_index += 1
-                center = args[current_index]
-                current_index += 1
+                (radius, center), args = popmult(args, 2)
+
                 shape = diffvg.Ellipse(diffvg.Vector2f(radius[0], radius[1]),
                                        diffvg.Vector2f(center[0], center[1]))
             elif shape_type == diffvg.ShapeType.path:
-                num_control_points = args[current_index]
-                current_index += 1
-                points = args[current_index]
-                current_index += 1
-                thickness = args[current_index]
-                current_index += 1
-                is_closed = args[current_index]
-                current_index += 1
-                use_distance_approx = args[current_index]
-                current_index += 1
+
+                (num_control_points, points, thickness,
+                 is_closed, use_distance_approx), args = popmult(args, 5)
+
                 shape = diffvg.Path(diffvg.int_ptr(num_control_points.data_ptr()),
                                     diffvg.float_ptr(points.data_ptr()),
                                     diffvg.float_ptr(thickness.data_ptr() if thickness is not None else 0),
@@ -238,55 +259,36 @@ class RenderFunction(torch.autograd.Function):
                                     is_closed,
                                     use_distance_approx)
             elif shape_type == diffvg.ShapeType.rect:
-                p_min = args[current_index]
-                current_index += 1
-                p_max = args[current_index]
-                current_index += 1
+                (p_min, p_max), args = popmult(args, 2)
                 shape = diffvg.Rect(diffvg.Vector2f(p_min[0], p_min[1]),
                                     diffvg.Vector2f(p_max[0], p_max[1]))
             else:
-                assert(False)
-            stroke_width = args[current_index]
-            current_index += 1
+                assert False
+
+            stroke_width = args.pop(0)
             shapes.append(diffvg.Shape(\
                 shape_type, shape.get_ptr(), stroke_width.item()))
             shape_contents.append(shape)
 
         for shape_group_id in range(num_shape_groups):
-            shape_ids = args[current_index]
-            current_index += 1
-            fill_color_type = args[current_index]
-            current_index += 1
+            (shape_ids, fill_color_type), args = popmult(args, 2)
             if fill_color_type == diffvg.ColorType.constant:
-                color = args[current_index]
-                current_index += 1
+                color = args.pop(0)
                 fill_color = diffvg.Constant(\
                     diffvg.Vector4f(color[0], color[1], color[2], color[3]))
             elif fill_color_type == diffvg.ColorType.linear_gradient:
-                beg = args[current_index]
-                current_index += 1
-                end = args[current_index]
-                current_index += 1
-                offsets = args[current_index]
-                current_index += 1
-                stop_colors = args[current_index]
-                current_index += 1
-                assert(offsets.shape[0] == stop_colors.shape[0])
+                (beg, end, offsets, stop_colors), args = popmult(args, 4)
+
+                assert offsets.shape[0] == stop_colors.shape[0]
                 fill_color = diffvg.LinearGradient(diffvg.Vector2f(beg[0], beg[1]),
                                                    diffvg.Vector2f(end[0], end[1]),
                                                    offsets.shape[0],
                                                    diffvg.float_ptr(offsets.data_ptr()),
                                                    diffvg.float_ptr(stop_colors.data_ptr()))
             elif fill_color_type == diffvg.ColorType.radial_gradient:
-                center = args[current_index]
-                current_index += 1
-                radius = args[current_index]
-                current_index += 1
-                offsets = args[current_index]
-                current_index += 1
-                stop_colors = args[current_index]
-                current_index += 1
-                assert(offsets.shape[0] == stop_colors.shape[0])
+                (center, radius, offsets, stop_colors), args = popmult(args, 4)
+
+                assert offsets.shape[0] == stop_colors.shape[0]
                 fill_color = diffvg.RadialGradient(diffvg.Vector2f(center[0], center[1]),
                                                    diffvg.Vector2f(radius[0], radius[1]),
                                                    offsets.shape[0],
@@ -295,39 +297,28 @@ class RenderFunction(torch.autograd.Function):
             elif fill_color_type is None:
                 fill_color = None
             else:
-                assert(False)
-            stroke_color_type = args[current_index]
-            current_index += 1
+                assert False
+
+            stroke_color_type = args.pop(0)
             if stroke_color_type == diffvg.ColorType.constant:
-                color = args[current_index]
-                current_index += 1
+                color = args.pop(0)
                 stroke_color = diffvg.Constant(\
                     diffvg.Vector4f(color[0], color[1], color[2], color[3]))
+
             elif stroke_color_type == diffvg.ColorType.linear_gradient:
-                beg = args[current_index]
-                current_index += 1
-                end = args[current_index]
-                current_index += 1
-                offsets = args[current_index]
-                current_index += 1
-                stop_colors = args[current_index]
-                current_index += 1
-                assert(offsets.shape[0] == stop_colors.shape[0])
+                (beg, end, offsets, stop_colors), args = popmult(args, 4)
+
+                assert offsets.shape[0] == stop_colors.shape[0]
                 stroke_color = diffvg.LinearGradient(diffvg.Vector2f(beg[0], beg[1]),
                                                      diffvg.Vector2f(end[0], end[1]),
                                                      offsets.shape[0],
                                                      diffvg.float_ptr(offsets.data_ptr()),
                                                      diffvg.float_ptr(stop_colors.data_ptr()))
+
             elif stroke_color_type == diffvg.ColorType.radial_gradient:
-                center = args[current_index]
-                current_index += 1
-                radius = args[current_index]
-                current_index += 1
-                offsets = args[current_index]
-                current_index += 1
-                stop_colors = args[current_index]
-                current_index += 1
-                assert(offsets.shape[0] == stop_colors.shape[0])
+                (center, radius, offsets, stop_colors), args = popmult(args, 4)
+
+                assert offsets.shape[0] == stop_colors.shape[0]
                 stroke_color = diffvg.RadialGradient(diffvg.Vector2f(center[0], center[1]),
                                                      diffvg.Vector2f(radius[0], radius[1]),
                                                      offsets.shape[0],
@@ -335,17 +326,18 @@ class RenderFunction(torch.autograd.Function):
                                                      diffvg.float_ptr(stop_colors.data_ptr()))
             elif stroke_color_type is None:
                 stroke_color = None
+
             else:
-                assert(False)
-            use_even_odd_rule = args[current_index]
-            current_index += 1
-            shape_to_canvas = args[current_index]
-            current_index += 1
+                assert False
+
+            (use_even_odd_rule, shape_to_canvas), args = popmult(args, 2)
 
             if fill_color is not None:
                 color_contents.append(fill_color)
+
             if stroke_color is not None:
                 color_contents.append(stroke_color)
+
             shape_groups.append(diffvg.ShapeGroup(\
                 diffvg.int_ptr(shape_ids.data_ptr()),
                 shape_ids.shape[0],
@@ -356,10 +348,7 @@ class RenderFunction(torch.autograd.Function):
                 use_even_odd_rule,
                 diffvg.float_ptr(shape_to_canvas.data_ptr())))
 
-        filter_type = args[current_index]
-        current_index += 1
-        filter_radius = args[current_index]
-        current_index += 1
+        (filter_type, filter_radius), args = popmult(args, 2)
         filt = diffvg.Filter(filter_type, filter_radius)
 
         start = time.time()
@@ -367,15 +356,16 @@ class RenderFunction(torch.autograd.Function):
             shapes, shape_groups, filt, pydiffvg.get_use_gpu(),
             pydiffvg.get_device().index if pydiffvg.get_device().index is not None else -1)
         time_elapsed = time.time() - start
+
         global print_timing
         if print_timing:
             print('Scene construction, time: %.5f s' % time_elapsed)
 
         if output_type == OutputType.color:
-            assert(eval_positions.shape[0] == 0)
+            assert eval_positions.shape[0] == 0
             rendered_image = torch.zeros(height, width, 4, device = pydiffvg.get_device())
         else:
-            assert(output_type == OutputType.sdf)          
+            assert output_type == OutputType.sdf
             if eval_positions.shape[0] == 0:
                 rendered_image = torch.zeros(height, width, 1, device = pydiffvg.get_device())
             else:
@@ -386,9 +376,10 @@ class RenderFunction(torch.autograd.Function):
             if background_image.shape[2] == 3:
                 raise NotImplementedError('Background image must have 4 channels, not 3. Add a fourth channel with all ones via torch.ones().')
             background_image = background_image.contiguous()
-            assert(background_image.shape[0] == rendered_image.shape[0])
-            assert(background_image.shape[1] == rendered_image.shape[1])
-            assert(background_image.shape[2] == 4)
+
+            assert background_image.shape[0] == rendered_image.shape[0]
+            assert background_image.shape[1] == rendered_image.shape[1]
+            assert background_image.shape[2] == 4
 
         start = time.time()
         diffvg.render(scene,
@@ -407,7 +398,7 @@ class RenderFunction(torch.autograd.Function):
                       use_prefiltering,
                       diffvg.float_ptr(eval_positions.data_ptr()),
                       eval_positions.shape[0])
-        assert(torch.isfinite(rendered_image).all())
+        assert torch.isfinite(rendered_image).all()
         time_elapsed = time.time() - start
         if print_timing:
             print('Forward pass, time: %.5f s' % time_elapsed)
@@ -438,55 +429,33 @@ class RenderFunction(torch.autograd.Function):
                     *args):
         if not grad_img.is_contiguous():
             grad_img = grad_img.contiguous()
-        assert(torch.isfinite(grad_img).all())
+        assert torch.isfinite(grad_img).all()
 
+        args = list(args)
         # Unpack arguments
-        current_index = 0
-        canvas_width = args[current_index]
-        current_index += 1
-        canvas_height = args[current_index]
-        current_index += 1
-        num_shapes = args[current_index]
-        current_index += 1
-        num_shape_groups = args[current_index]
-        current_index += 1
-        output_type = args[current_index]
-        current_index += 1
-        use_prefiltering = args[current_index]
-        current_index += 1
-        eval_positions = args[current_index]
-        current_index += 1        
+        (canvas_width, canvas_height, num_shapes,
+         num_shape_groups, output_type, use_prefiltering,
+         eval_positions), args = popmult(args, 7)
+
         shapes = []
         shape_groups = []
         shape_contents = [] # Important to avoid GC deleting the shapes
         color_contents = [] # Same as above
         for shape_id in range(num_shapes):
-            shape_type = args[current_index]
-            current_index += 1
+            shape_type = args.pop(0)
             if shape_type == diffvg.ShapeType.circle:
-                radius = args[current_index]
-                current_index += 1
-                center = args[current_index]
-                current_index += 1
+                (radius, center), args = popmult(args, 2)
+
                 shape = diffvg.Circle(radius, diffvg.Vector2f(center[0], center[1]))
             elif shape_type == diffvg.ShapeType.ellipse:
-                radius = args[current_index]
-                current_index += 1
-                center = args[current_index]
-                current_index += 1
+                (radius, center), args = popmult(args, 2)
+
                 shape = diffvg.Ellipse(diffvg.Vector2f(radius[0], radius[1]),
                                        diffvg.Vector2f(center[0], center[1]))
             elif shape_type == diffvg.ShapeType.path:
-                num_control_points = args[current_index]
-                current_index += 1
-                points = args[current_index]
-                current_index += 1
-                thickness = args[current_index]
-                current_index += 1
-                is_closed = args[current_index]
-                current_index += 1
-                use_distance_approx = args[current_index]
-                current_index += 1
+                (num_control_points, points, thickness,
+                 is_closed, use_distance_approx), args = popmult(args, 5)
+
                 shape = diffvg.Path(diffvg.int_ptr(num_control_points.data_ptr()),
                                     diffvg.float_ptr(points.data_ptr()),
                                     diffvg.float_ptr(thickness.data_ptr() if thickness is not None else 0),
@@ -495,114 +464,93 @@ class RenderFunction(torch.autograd.Function):
                                     is_closed,
                                     use_distance_approx)
             elif shape_type == diffvg.ShapeType.rect:
-                p_min = args[current_index]
-                current_index += 1
-                p_max = args[current_index]
-                current_index += 1
+                (p_min, p_max), args = popmult(args, 2)
+
                 shape = diffvg.Rect(diffvg.Vector2f(p_min[0], p_min[1]),
                                     diffvg.Vector2f(p_max[0], p_max[1]))
             else:
-                assert(False)
-            stroke_width = args[current_index]
-            current_index += 1
+                assert False
+
+            stroke_width = args.pop(0)
+
             shapes.append(diffvg.Shape(\
                 shape_type, shape.get_ptr(), stroke_width.item()))
             shape_contents.append(shape)
 
         for shape_group_id in range(num_shape_groups):
-            shape_ids = args[current_index]
-            current_index += 1
-            fill_color_type = args[current_index]
-            current_index += 1
+            (shape_ids, fill_color_type), args = popmult(args, 2)
+
             if fill_color_type == diffvg.ColorType.constant:
-                color = args[current_index]
-                current_index += 1
+                color = args.pop(0)
+
                 fill_color = diffvg.Constant(\
                     diffvg.Vector4f(color[0], color[1], color[2], color[3]))
+
             elif fill_color_type == diffvg.ColorType.linear_gradient:
-                beg = args[current_index]
-                current_index += 1
-                end = args[current_index]
-                current_index += 1
-                offsets = args[current_index]
-                current_index += 1
-                stop_colors = args[current_index]
-                current_index += 1
-                assert(offsets.shape[0] == stop_colors.shape[0])
+                (beg, end, offsets, stop_colors), args = popmult(args, 4)
+
+                assert offsets.shape[0] == stop_colors.shape[0]
                 fill_color = diffvg.LinearGradient(diffvg.Vector2f(beg[0], beg[1]),
                                                    diffvg.Vector2f(end[0], end[1]),
                                                    offsets.shape[0],
                                                    diffvg.float_ptr(offsets.data_ptr()),
                                                    diffvg.float_ptr(stop_colors.data_ptr()))
+
             elif fill_color_type == diffvg.ColorType.radial_gradient:
-                center = args[current_index]
-                current_index += 1
-                radius = args[current_index]
-                current_index += 1
-                offsets = args[current_index]
-                current_index += 1
-                stop_colors = args[current_index]
-                current_index += 1
-                assert(offsets.shape[0] == stop_colors.shape[0])
+                (center, radius, offsets, stop_colors), args = popmult(args, 4)
+
+                assert offsets.shape[0] == stop_colors.shape[0]
                 fill_color = diffvg.RadialGradient(diffvg.Vector2f(center[0], center[1]),
                                                    diffvg.Vector2f(radius[0], radius[1]),
                                                    offsets.shape[0],
                                                    diffvg.float_ptr(offsets.data_ptr()),
                                                    diffvg.float_ptr(stop_colors.data_ptr()))
+
             elif fill_color_type is None:
                 fill_color = None
+
             else:
-                assert(False)
-            stroke_color_type = args[current_index]
-            current_index += 1
+                assert False
+
+            stroke_color_type = args.pop(0)
+
             if stroke_color_type == diffvg.ColorType.constant:
-                color = args[current_index]
-                current_index += 1
+                color = args.pop(0)
                 stroke_color = diffvg.Constant(\
                     diffvg.Vector4f(color[0], color[1], color[2], color[3]))
             elif stroke_color_type == diffvg.ColorType.linear_gradient:
-                beg = args[current_index]
-                current_index += 1
-                end = args[current_index]
-                current_index += 1
-                offsets = args[current_index]
-                current_index += 1
-                stop_colors = args[current_index]
-                current_index += 1
-                assert(offsets.shape[0] == stop_colors.shape[0])
+                (beg, end, offsets, stop_colors) = popmult(args, 4)
+
+                assert offsets.shape[0] == stop_colors.shape[0]
                 stroke_color = diffvg.LinearGradient(diffvg.Vector2f(beg[0], beg[1]),
                                                      diffvg.Vector2f(end[0], end[1]),
                                                      offsets.shape[0],
                                                      diffvg.float_ptr(offsets.data_ptr()),
                                                      diffvg.float_ptr(stop_colors.data_ptr()))
             elif stroke_color_type == diffvg.ColorType.radial_gradient:
-                center = args[current_index]
-                current_index += 1
-                radius = args[current_index]
-                current_index += 1
-                offsets = args[current_index]
-                current_index += 1
-                stop_colors = args[current_index]
-                current_index += 1
-                assert(offsets.shape[0] == stop_colors.shape[0])
+                (center, radius, offsets, stop_colors), args = popmult(args, 4)
+
+                assert offsets.shape[0] == stop_colors.shape[0]
                 stroke_color = diffvg.RadialGradient(diffvg.Vector2f(center[0], center[1]),
                                                      diffvg.Vector2f(radius[0], radius[1]),
                                                      offsets.shape[0],
                                                      diffvg.float_ptr(offsets.data_ptr()),
                                                      diffvg.float_ptr(stop_colors.data_ptr()))
+
             elif stroke_color_type is None:
                 stroke_color = None
+
             else:
-                assert(False)
-            use_even_odd_rule = args[current_index]
-            current_index += 1
-            shape_to_canvas = args[current_index]
-            current_index += 1
+                assert False
+
+            (use_even_odd_rule, shape_to_canvas), args = popmult(args, 2)
 
             if fill_color is not None:
                 color_contents.append(fill_color)
+
             if stroke_color is not None:
                 color_contents.append(stroke_color)
+
             shape_groups.append(diffvg.ShapeGroup(\
                 diffvg.int_ptr(shape_ids.data_ptr()),
                 shape_ids.shape[0],
@@ -613,31 +561,32 @@ class RenderFunction(torch.autograd.Function):
                 use_even_odd_rule,
                 diffvg.float_ptr(shape_to_canvas.data_ptr())))
 
-        filter_type = args[current_index]
-        current_index += 1
-        filter_radius = args[current_index]
-        current_index += 1
+        (filter_type, filter_radius), args = popmult(args, 2)
+
         filt = diffvg.Filter(filter_type, filter_radius)
 
         scene = diffvg.Scene(canvas_width, canvas_height,
-            shapes, shape_groups, filt, pydiffvg.get_use_gpu(),
-            pydiffvg.get_device().index if pydiffvg.get_device().index is not None else -1)
+                             shapes, shape_groups, filt, pydiffvg.get_use_gpu(),
+                             pydiffvg.get_device().index if pydiffvg.get_device().index is not None else -1)
 
         if output_type == OutputType.color:
-            assert(grad_img.shape[2] == 4)
+            assert grad_img.shape[2] == 4
         else:
-            assert(grad_img.shape[2] == 1)
+            assert grad_img.shape[2] == 1
 
         if background_image is not None:
             background_image = background_image.to(pydiffvg.get_device())
             if background_image.shape[2] == 3:
-                background_image = torch.cat((\
-                    background_image, torch.ones(background_image.shape[0], background_image.shape[1], 1,
-                        device = background_image.device)), dim = 2)
+                background_image = torch.cat((
+                    background_image, torch.ones(background_image.shape[0],
+                                                 background_image.shape[1],
+                                                 1,
+                                                 device=background_image.device)),
+                                             dim=2)
             background_image = background_image.contiguous()
-            assert(background_image.shape[0] == rendered_image.shape[0])
-            assert(background_image.shape[1] == rendered_image.shape[1])
-            assert(background_image.shape[2] == 4)
+            assert background_image.shape[0] == rendered_image.shape[0]
+            assert background_image.shape[1] == rendered_image.shape[1]
+            assert background_image.shape[2] == 4
 
         translation_grad_image = \
             torch.zeros(height, width, 2, device = pydiffvg.get_device())
@@ -661,7 +610,7 @@ class RenderFunction(torch.autograd.Function):
         time_elapsed = time.time() - start
         if print_timing:
             print('Gradient pass, time: %.5f s' % time_elapsed)
-        assert(torch.isfinite(translation_grad_image).all())
+        assert torch.isfinite(translation_grad_image).all()
 
         return translation_grad_image
 
@@ -670,7 +619,7 @@ class RenderFunction(torch.autograd.Function):
                  grad_img):
         if not grad_img.is_contiguous():
             grad_img = grad_img.contiguous()
-        assert(torch.isfinite(grad_img).all())
+        assert torch.isfinite(grad_img).all()
 
         scene = ctx.scene
         width = ctx.width
@@ -710,20 +659,11 @@ class RenderFunction(torch.autograd.Function):
         if print_timing:
             print('Backward pass, time: %.5f s' % time_elapsed)
 
-        d_args = []
-        d_args.append(None) # width
-        d_args.append(None) # height
-        d_args.append(None) # num_samples_x
-        d_args.append(None) # num_samples_y
-        d_args.append(None) # seed
-        d_args.append(d_background_image)
-        d_args.append(None) # canvas_width
-        d_args.append(None) # canvas_height
-        d_args.append(None) # num_shapes
-        d_args.append(None) # num_shape_groups
-        d_args.append(None) # output_type
-        d_args.append(None) # use_prefiltering
-        d_args.append(None) # eval_positions
+        # [width, height, num_samples_x, num_samples_y, seed,
+        #  d_background_image, canvas_width, canvas_height, num_shapes,
+        #  num_shape_groups, output_type, use_prefiltering, _eval_positions]
+        d_args = [None] * 5 + [d_background_image] + [None] * 7
+
         for shape_id in range(scene.num_shapes):
             d_args.append(None) # type
             d_shape = scene.get_d_shape(shape_id)
@@ -731,21 +671,21 @@ class RenderFunction(torch.autograd.Function):
             if d_shape.type == diffvg.ShapeType.circle:
                 d_circle = d_shape.as_circle()
                 radius = torch.tensor(d_circle.radius)
-                assert(torch.isfinite(radius).all())
+                assert torch.isfinite(radius).all()
                 d_args.append(radius)
                 c = d_circle.center
                 c = torch.tensor((c.x, c.y))
-                assert(torch.isfinite(c).all())
+                assert torch.isfinite(c).all()
                 d_args.append(c)
             elif d_shape.type == diffvg.ShapeType.ellipse:
                 d_ellipse = d_shape.as_ellipse()
                 r = d_ellipse.radius
                 r = torch.tensor((d_ellipse.radius.x, d_ellipse.radius.y))
-                assert(torch.isfinite(r).all())
+                assert torch.isfinite(r).all()
                 d_args.append(r)
                 c = d_ellipse.center
                 c = torch.tensor((c.x, c.y))
-                assert(torch.isfinite(c).all())
+                assert torch.isfinite(c).all()
                 d_args.append(c)
             elif d_shape.type == diffvg.ShapeType.path:
                 d_path = d_shape.as_path()
@@ -757,9 +697,9 @@ class RenderFunction(torch.autograd.Function):
                     d_path.copy_to(diffvg.float_ptr(points.data_ptr()), diffvg.float_ptr(thickness.data_ptr()))
                 else:
                     d_path.copy_to(diffvg.float_ptr(points.data_ptr()), diffvg.float_ptr(0))
-                assert(torch.isfinite(points).all())
+                assert torch.isfinite(points).all()
                 if thickness is not None:
-                    assert(torch.isfinite(thickness).all())
+                    assert torch.isfinite(thickness).all()
                 d_args.append(None) # num_control_points
                 d_args.append(points)
                 d_args.append(thickness)
@@ -769,17 +709,17 @@ class RenderFunction(torch.autograd.Function):
                 d_rect = d_shape.as_rect()
                 p_min = torch.tensor((d_rect.p_min.x, d_rect.p_min.y))
                 p_max = torch.tensor((d_rect.p_max.x, d_rect.p_max.y))
-                assert(torch.isfinite(p_min).all())
-                assert(torch.isfinite(p_max).all())
+                assert torch.isfinite(p_min).all()
+                assert torch.isfinite(p_max).all()
                 d_args.append(p_min)
                 d_args.append(p_max)
             else:
-                assert(False)
+                assert False
             if use_thickness:
                 d_args.append(None)
             else:
                 w = torch.tensor((d_shape.stroke_width))
-                assert(torch.isfinite(w).all())
+                assert torch.isfinite(w).all()
                 d_args.append(w)
 
         for group_id in range(scene.num_shape_groups):
@@ -802,7 +742,7 @@ class RenderFunction(torch.autograd.Function):
                     d_linear_gradient.copy_to(\
                         diffvg.float_ptr(offsets.data_ptr()),
                         diffvg.float_ptr(stop_colors.data_ptr()))
-                    assert(torch.isfinite(stop_colors).all())
+                    assert torch.isfinite(stop_colors).all()
                     d_args.append(offsets)
                     d_args.append(stop_colors)
                 elif d_shape_group.fill_color_type == diffvg.ColorType.radial_gradient:
@@ -816,11 +756,11 @@ class RenderFunction(torch.autograd.Function):
                     d_radial_gradient.copy_to(\
                         diffvg.float_ptr(offsets.data_ptr()),
                         diffvg.float_ptr(stop_colors.data_ptr()))
-                    assert(torch.isfinite(stop_colors).all())
+                    assert torch.isfinite(stop_colors).all()
                     d_args.append(offsets)
                     d_args.append(stop_colors)
                 else:
-                    assert(False)
+                    assert False
             d_args.append(None) # stroke_color_type
             if d_shape_group.has_stroke_color():
                 if d_shape_group.stroke_color_type == diffvg.ColorType.constant:
@@ -838,7 +778,7 @@ class RenderFunction(torch.autograd.Function):
                     d_linear_gradient.copy_to(\
                         diffvg.float_ptr(offsets.data_ptr()),
                         diffvg.float_ptr(stop_colors.data_ptr()))
-                    assert(torch.isfinite(stop_colors).all())
+                    assert torch.isfinite(stop_colors).all()
                     d_args.append(offsets)
                     d_args.append(stop_colors)
                 elif d_shape_group.fill_color_type == diffvg.ColorType.radial_gradient:
@@ -852,15 +792,15 @@ class RenderFunction(torch.autograd.Function):
                     d_radial_gradient.copy_to(\
                         diffvg.float_ptr(offsets.data_ptr()),
                         diffvg.float_ptr(stop_colors.data_ptr()))
-                    assert(torch.isfinite(stop_colors).all())
+                    assert torch.isfinite(stop_colors).all()
                     d_args.append(offsets)
                     d_args.append(stop_colors)
                 else:
-                    assert(False)
+                    assert False
             d_args.append(None) # use_even_odd_rule
             d_shape_to_canvas = torch.zeros((3, 3))
             d_shape_group.copy_to(diffvg.float_ptr(d_shape_to_canvas.data_ptr()))
-            assert(torch.isfinite(d_shape_to_canvas).all())
+            assert torch.isfinite(d_shape_to_canvas).all()
             d_args.append(d_shape_to_canvas)
         d_args.append(None) # filter_type
         d_args.append(torch.tensor(scene.get_d_filter_radius()))

pyproject.toml

@@ -5,8 +5,11 @@ description = ""
 authors = ["Marco Lee <marco@goodnotesapp.com>"]
 
 [tool.poetry.dependencies]
-python = ">=3.8,<3.11"
+python = ">=3.9,<3.11"
 pygame = "^2.0.1"
+torch = "^2.0.1"
+ipython = "^8.13.2"
+pytest-metadata = "^2.0.4"
 
 [tool.poetry.dev-dependencies]
 torch = "^2.0.0"