fix img2img/inpaint paste params

CHANGELOG.md

@@ -19,6 +19,7 @@ Note: Release pending `diffusers==0.24`
   - better prompt display in process tab
   - increase maximum lora cache values
   - fix for python 3.9 compatibility
+  - fix img2img/inpaint paste params
 
 ## Update for 2023-11-23
 

modules/img2img.py

@@ -208,7 +208,6 @@ def img2img(id_task: str, mode: int, prompt: str, negative_prompt: str, prompt_s
         p.scale_by = scale_by
     p.scripts = modules.scripts.scripts_img2img
     p.script_args = args
-    p.extra_generation_params['Resize mode'] = resize_mode
     if mask:
         p.extra_generation_params["Mask blur"] = mask_blur
         p.extra_generation_params["Mask alpha"] = mask_alpha

modules/processing.py

@@ -606,10 +606,14 @@ def create_infotext(p: StableDiffusionProcessing, all_prompts=None, all_seeds=No
         args["Init image size"] = f"{getattr(p, 'init_img_width', 0)}x{getattr(p, 'init_img_height', 0)}"
         args["Init image hash"] = getattr(p, 'init_img_hash', None)
         args["Mask weight"] = getattr(p, "inpainting_mask_weight", shared.opts.inpainting_mask_weight) if p.is_using_inpainting_conditioning else None
-        args['Resize mode'] = getattr(p, 'resize_mode', None)
         args['Resize scale'] = getattr(p, 'scale_by', None)
         args["Mask blur"] = p.mask_blur if getattr(p, 'mask', None) is not None and getattr(p, 'mask_blur', 0) > 0 else None
         args["Denoising strength"] = getattr(p, 'denoising_strength', None)
+        # lookup by index
+        if getattr(p, 'resize_mode', None) is not None:
+            RESIZE_MODES = ["None", "Resize fixed", "Crop and resize", "Resize and fill", "Latent upscale"]
+            args['Resize mode'] = RESIZE_MODES[p.resize_mode]
+        # TODO missing-by-index: inpainting_fill, inpaint_full_res, inpainting_mask_invert
     if 'face' in p.ops:
         args["Face restoration"] = shared.opts.face_restoration_model
     if 'color' in p.ops:

modules/sd_models.py

@@ -473,6 +473,9 @@ def load_model_weights(model: torch.nn.Module, checkpoint_info: CheckpointInfo,
     model.sd_model_hash = checkpoint_info.calculate_shorthash()
     model.sd_model_checkpoint = checkpoint_info.filename
     model.sd_checkpoint_info = checkpoint_info
+    model.is_sdxl = False # a1111 compatibility item
+    model.is_sd2 = False # a1111 compatibility item
+    model.is_sd1 = True # a1111 compatibility item
     shared.opts.data["sd_checkpoint_hash"] = checkpoint_info.sha256
     model.logvar = model.logvar.to(devices.device)  # fix for training
     sd_vae.delete_base_vae()
@@ -549,10 +552,6 @@ class ModelData:
             with self.lock:
                 try:
                     self.sd_model = reload_model_weights(op='model')
-                    if self.sd_model is not None:
-                        self.sd_model.is_sdxl = False # a1111 compatibility item
-                        self.sd_model.is_sd2 = False # a1111 compatibility item
-                        self.sd_model.is_sd1 = True # a1111 compatibility item
                     self.initial = False
                 except Exception as e:
                     shared.log.error("Failed to load stable diffusion model")
@@ -1107,12 +1106,16 @@ def load_model(checkpoint_info=None, already_loaded_state_dict=None, timer=None,
     sd_model = None
     stdout = io.StringIO()
     with contextlib.redirect_stdout(stdout):
+        """
         try:
             clip_is_included_into_sd = sd1_clip_weight in state_dict or sd2_clip_weight in state_dict
             with sd_disable_initialization.DisableInitialization(disable_clip=clip_is_included_into_sd):
                 sd_model = instantiate_from_config(sd_config.model)
-        except Exception:
+        except Exception as e:
+            shared.log.error(f'LDM: instantiate from config: {e}')
             sd_model = instantiate_from_config(sd_config.model)
+        """
+        sd_model = instantiate_from_config(sd_config.model)
     for line in stdout.getvalue().splitlines():
         if len(line) > 0:
             shared.log.info(f'LDM: {line.strip()}')

repositories/ldm/models/diffusion/ddpm.py

@@ -81,7 +81,7 @@ class DDPM(pl.LightningModule):
         super().__init__()
         assert parameterization in ["eps", "x0", "v"], 'currently only supporting "eps" and "x0" and "v"'
         self.parameterization = parameterization
-        print(f"{self.__class__.__name__}: Running in {self.parameterization}-prediction mode")
+        print(f"{self.__class__.__name__}: mode={self.parameterization}")
         self.cond_stage_model = None
         self.clip_denoised = clip_denoised
         self.log_every_t = log_every_t

repositories/ldm/modules/encoders/modules.py

@@ -301,8 +301,8 @@ class FrozenCLIPT5Encoder(AbstractEncoder):
         super().__init__()
         self.clip_encoder = FrozenCLIPEmbedder(clip_version, device, max_length=clip_max_length)
         self.t5_encoder = FrozenT5Embedder(t5_version, device, max_length=t5_max_length)
-        print(f"{self.clip_encoder.__class__.__name__} has {count_params(self.clip_encoder) * 1.e-6:.2f} M parameters, "
-              f"{self.t5_encoder.__class__.__name__} comes with {count_params(self.t5_encoder) * 1.e-6:.2f} M params.")
+        print(f"{self.clip_encoder.__class__.__name__} params={count_params(self.clip_encoder) * 1.e-6:.2f} M "
+              f"{self.t5_encoder.__class__.__name__} params={count_params(self.t5_encoder) * 1.e-6:.2f} M")
 
     def encode(self, text):
         return self(text)

repositories/ldm/util.py

@@ -75,7 +75,7 @@ def mean_flat(tensor):
 def count_params(model, verbose=False):
     total_params = sum(p.numel() for p in model.parameters())
     if verbose:
-        print(f"{model.__class__.__name__} has {total_params*1.e-6:.2f} M params.")
+        print(f"{model.__class__.__name__} params={total_params*1.e-6:.2f}M")
     return total_params