fix: improve VAE tiling (#372)

* fix and improve: VAE tiling
- properly handle the upper left corner interpolating both x and y
- refactor out lerp
- use smootherstep to preserve more detail and spend less area blending

* actually fix vae tile merging

Co-authored-by: stduhpf <stephduh@live.fr>

* remove the now unused lerp function

---------

Co-authored-by: stduhpf <stephduh@live.fr>

README.md

@@ -192,7 +192,10 @@ arguments:
   -M, --mode [MODEL]                 run mode (txt2img or img2img or convert, default: txt2img)
   -t, --threads N                    number of threads to use during computation (default: -1).
                                      If threads <= 0, then threads will be set to the number of CPU physical cores
-  -m, --model [MODEL]                path to model
+  -m, --model [MODEL]                path to full model
+  --diffusion-model                  path to the standalone diffusion model
+  --clip_l                           path to the clip-l text encoder
+  --t5xxl                            path to the the t5xxl text encoder.
   --vae [VAE]                        path to vae
   --taesd [TAESD_PATH]               path to taesd. Using Tiny AutoEncoder for fast decoding (low quality)
   --control-net [CONTROL_PATH]       path to control net model
@@ -227,6 +230,8 @@ arguments:
   --clip-skip N                      ignore last layers of CLIP network; 1 ignores none, 2 ignores one layer (default: -1)
                                      <= 0 represents unspecified, will be 1 for SD1.x, 2 for SD2.x
   --vae-tiling                       process vae in tiles to reduce memory usage
+  --vae-on-cpu                       keep vae in cpu (for low vram)
+  --clip-on-cpu                      keep clip in cpu (for low vram).
   --control-net-cpu                  keep controlnet in cpu (for low vram)
   --canny                            apply canny preprocessor (edge detection)
   --color                            Colors the logging tags according to level

examples/cli/main.cpp

@@ -173,7 +173,10 @@ void print_usage(int argc, const char* argv[]) {
     printf("  -M, --mode [MODEL]                 run mode (txt2img or img2img or convert, default: txt2img)\n");
     printf("  -t, --threads N                    number of threads to use during computation (default: -1).\n");
     printf("                                     If threads <= 0, then threads will be set to the number of CPU physical cores\n");
-    printf("  -m, --model [MODEL]                path to model\n");
+    printf("  -m, --model [MODEL]                path to full model\n");
+    printf("  --diffusion-model                  path to the standalone diffusion model\n");
+    printf("  --clip_l                           path to the clip-l text encoder\n");
+    printf("  --t5xxl                            path to the the t5xxl text encoder.\n");
     printf("  --vae [VAE]                        path to vae\n");
     printf("  --taesd [TAESD_PATH]               path to taesd. Using Tiny AutoEncoder for fast decoding (low quality)\n");
     printf("  --control-net [CONTROL_PATH]       path to control net model\n");
@@ -208,6 +211,8 @@ void print_usage(int argc, const char* argv[]) {
     printf("  --clip-skip N                      ignore last layers of CLIP network; 1 ignores none, 2 ignores one layer (default: -1)\n");
     printf("                                     <= 0 represents unspecified, will be 1 for SD1.x, 2 for SD2.x\n");
     printf("  --vae-tiling                       process vae in tiles to reduce memory usage\n");
+    printf("  --vae-on-cpu                       keep vae in cpu (for low vram)\n");
+    printf("  --clip-on-cpu                      keep clip in cpu (for low vram).\n");
     printf("  --control-net-cpu                  keep controlnet in cpu (for low vram)\n");
     printf("  --canny                            apply canny preprocessor (edge detection)\n");
     printf("  --color                            Colors the logging tags according to level\n");

ggml_extend.hpp

@@ -353,6 +353,12 @@ __STATIC_INLINE__ void ggml_split_tensor_2d(struct ggml_tensor* input,
     }
 }
 
+// unclamped -> expects x in the range [0-1]
+__STATIC_INLINE__ float ggml_smootherstep_f32(const float x) {
+    GGML_ASSERT(x >= 0.f && x <= 1.f);
+    return x * x * x * (x * (6.0f * x - 15.0f) + 10.0f);
+}
+
 __STATIC_INLINE__ void ggml_merge_tensor_2d(struct ggml_tensor* input,
                                             struct ggml_tensor* output,
                                             int x,
@@ -361,6 +367,10 @@ __STATIC_INLINE__ void ggml_merge_tensor_2d(struct ggml_tensor* input,
     int64_t width    = input->ne[0];
     int64_t height   = input->ne[1];
     int64_t channels = input->ne[2];
+
+    int64_t img_width    = output->ne[0];
+    int64_t img_height   = output->ne[1];
+
     GGML_ASSERT(input->type == GGML_TYPE_F32 && output->type == GGML_TYPE_F32);
     for (int iy = 0; iy < height; iy++) {
         for (int ix = 0; ix < width; ix++) {
@@ -368,16 +378,23 @@ __STATIC_INLINE__ void ggml_merge_tensor_2d(struct ggml_tensor* input,
                 float new_value = ggml_tensor_get_f32(input, ix, iy, k);
                 if (overlap > 0) {  // blend colors in overlapped area
                     float old_value = ggml_tensor_get_f32(output, x + ix, y + iy, k);
-                    if (x > 0 && ix < overlap) {  // in overlapped horizontal
-                        ggml_tensor_set_f32(output, old_value + (new_value - old_value) * (ix / (1.0f * overlap)), x + ix, y + iy, k);
-                        continue;
-                    }
-                    if (y > 0 && iy < overlap) {  // in overlapped vertical
-                        ggml_tensor_set_f32(output, old_value + (new_value - old_value) * (iy / (1.0f * overlap)), x + ix, y + iy, k);
-                        continue;
-                    }
+
+                    const float x_f_0 = (x > 0) ? ix / float(overlap) : 1;
+                    const float x_f_1 = (x < (img_width - width)) ? (width - ix) / float(overlap) : 1 ;
+                    const float y_f_0 = (y > 0) ? iy / float(overlap) : 1;
+                    const float y_f_1 = (y < (img_height - height)) ? (height - iy) / float(overlap) : 1;
+
+                    const float x_f = std::min(std::min(x_f_0, x_f_1), 1.f);
+                    const float y_f = std::min(std::min(y_f_0, y_f_1), 1.f);
+
+                    ggml_tensor_set_f32(
+                        output,
+                        old_value + new_value * ggml_smootherstep_f32(y_f) * ggml_smootherstep_f32(x_f),
+                        x + ix, y + iy, k
+                    );
+                } else {
+                    ggml_tensor_set_f32(output, new_value, x + ix, y + iy, k);
                 }
-                ggml_tensor_set_f32(output, new_value, x + ix, y + iy, k);
             }
         }
     }