fix: default to empty reference on pix2pix models to avoid segfault

Author: wbruna

stable-diffusion.cpp

@@ -448,6 +448,10 @@ class StableDiffusionGGML {
             diffusion_model->alloc_params_buffer();
             diffusion_model->get_param_tensors(tensors);
 
+            if (sd_version_is_unet_edit(version)) {
+                vae_decode_only = false;
+            }
+
             if (high_noise_diffusion_model) {
                 high_noise_diffusion_model->alloc_params_buffer();
                 high_noise_diffusion_model->get_param_tensors(tensors);
@@ -2319,23 +2323,36 @@ sd_image_t* generate_image(sd_ctx_t* sd_ctx, const sd_img_gen_params_t* sd_img_g
         init_latent = generate_init_latent(sd_ctx, work_ctx, width, height);
     }
 
+    sd_guidance_params_t guidance = sd_img_gen_params->sample_params.guidance;
+    std::vector<sd_image_t*> ref_images;
+    for (int i = 0; i < sd_img_gen_params->ref_images_count; i++) {
+        ref_images.push_back(&sd_img_gen_params->ref_images[i]);
+    }
+
+    std::vector<uint8_t> empty_image_data;
+    sd_image_t empty_image = {(uint32_t)width, (uint32_t)height, 3, nullptr};
+    if (ref_images.empty() && sd_version_is_unet_edit(sd_ctx->sd->version))
+    {
+        LOG_WARN("This model needs at least one reference image; using an empty reference");
+        empty_image_data.reserve(width * height * 3);
+        ref_images.push_back(&empty_image);
+        empty_image.data = empty_image_data.data();
+        guidance.img_cfg = 0.f;
+    }
+
     if (sd_img_gen_params->ref_images_count > 0) {
         LOG_INFO("EDIT mode");
     }
-    else if (sd_ctx->sd->version == VERSION_SD1_PIX2PIX || sd_ctx->sd->version == VERSION_SDXL_PIX2PIX) {
-        LOG_ERROR("This model needs at least one reference image");
-        return NULL;
-    }
 
     std::vector<ggml_tensor*> ref_latents;
-    for (int i = 0; i < sd_img_gen_params->ref_images_count; i++) {
+    for (int i = 0; i < ref_images.size(); i++) {
         ggml_tensor* img = ggml_new_tensor_4d(work_ctx,
                                               GGML_TYPE_F32,
-                                              sd_img_gen_params->ref_images[i].width,
-                                              sd_img_gen_params->ref_images[i].height,
+                                              ref_images[i]->width,
+                                              ref_images[i]->height,
                                               3,
                                               1);
-        sd_image_to_tensor(sd_img_gen_params->ref_images[i].data, img);
+        sd_image_to_tensor(ref_images[i]->data, img);
 
         ggml_tensor* latent = NULL;
         if (sd_ctx->sd->use_tiny_autoencoder) {
@@ -2368,7 +2385,7 @@ sd_image_t* generate_image(sd_ctx_t* sd_ctx, const sd_img_gen_params_t* sd_img_g
                                                         SAFE_STR(sd_img_gen_params->prompt),
                                                         SAFE_STR(sd_img_gen_params->negative_prompt),
                                                         sd_img_gen_params->clip_skip,
-                                                        sd_img_gen_params->sample_params.guidance,
+                                                        guidance,
                                                         sd_img_gen_params->sample_params.eta,
                                                         width,
                                                         height,