hires fix checkpoint selection

modules/generation_parameters_copypaste.py

@@ -280,6 +280,9 @@ Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 965400086, Size: 512x512, Model
     if "Hires sampler" not in res:
         res["Hires sampler"] = "Use same sampler"
 
+    if "Hires checkpoint" not in res:
+        res["Hires checkpoint"] = "Use same checkpoint"
+
     if "Hires prompt" not in res:
         res["Hires prompt"] = ""
 

modules/processing.py

@@ -935,7 +935,7 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
     cached_hr_uc = [None, None]
     cached_hr_c = [None, None]
 
-    def __init__(self, enable_hr: bool = False, denoising_strength: float = 0.75, firstphase_width: int = 0, firstphase_height: int = 0, hr_scale: float = 2.0, hr_upscaler: str = None, hr_second_pass_steps: int = 0, hr_resize_x: int = 0, hr_resize_y: int = 0, hr_sampler_name: str = None, hr_prompt: str = '', hr_negative_prompt: str = '', **kwargs):
+    def __init__(self, enable_hr: bool = False, denoising_strength: float = 0.75, firstphase_width: int = 0, firstphase_height: int = 0, hr_scale: float = 2.0, hr_upscaler: str = None, hr_second_pass_steps: int = 0, hr_resize_x: int = 0, hr_resize_y: int = 0, hr_checkpoint_name: str = None, hr_sampler_name: str = None, hr_prompt: str = '', hr_negative_prompt: str = '', **kwargs):
         super().__init__(**kwargs)
         self.enable_hr = enable_hr
         self.denoising_strength = denoising_strength
@@ -946,11 +946,14 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
         self.hr_resize_y = hr_resize_y
         self.hr_upscale_to_x = hr_resize_x
         self.hr_upscale_to_y = hr_resize_y
+        self.hr_checkpoint_name = hr_checkpoint_name
+        self.hr_checkpoint_info = None
         self.hr_sampler_name = hr_sampler_name
         self.hr_prompt = hr_prompt
         self.hr_negative_prompt = hr_negative_prompt
         self.all_hr_prompts = None
         self.all_hr_negative_prompts = None
+        self.latent_scale_mode = None
 
         if firstphase_width != 0 or firstphase_height != 0:
             self.hr_upscale_to_x = self.width
@@ -973,6 +976,14 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
 
     def init(self, all_prompts, all_seeds, all_subseeds):
         if self.enable_hr:
+            if self.hr_checkpoint_name:
+                self.hr_checkpoint_info = sd_models.get_closet_checkpoint_match(self.hr_checkpoint_name)
+
+                if self.hr_checkpoint_info is None:
+                    raise Exception(f'Could not find checkpoint with name {self.hr_checkpoint_name}')
+
+                self.extra_generation_params["Hires checkpoint"] = self.hr_checkpoint_info.short_title
+
             if self.hr_sampler_name is not None and self.hr_sampler_name != self.sampler_name:
                 self.extra_generation_params["Hires sampler"] = self.hr_sampler_name
 
@@ -982,6 +993,11 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
             if tuple(self.hr_negative_prompt) != tuple(self.negative_prompt):
                 self.extra_generation_params["Hires negative prompt"] = self.hr_negative_prompt
 
+            self.latent_scale_mode = shared.latent_upscale_modes.get(self.hr_upscaler, None) if self.hr_upscaler is not None else shared.latent_upscale_modes.get(shared.latent_upscale_default_mode, "nearest")
+            if self.enable_hr and self.latent_scale_mode is None:
+                if not any(x.name == self.hr_upscaler for x in shared.sd_upscalers):
+                    raise Exception(f"could not find upscaler named {self.hr_upscaler}")
+
             if opts.use_old_hires_fix_width_height and self.applied_old_hires_behavior_to != (self.width, self.height):
                 self.hr_resize_x = self.width
                 self.hr_resize_y = self.height
@@ -1020,14 +1036,6 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
                     self.truncate_x = (self.hr_upscale_to_x - target_w) // opt_f
                     self.truncate_y = (self.hr_upscale_to_y - target_h) // opt_f
 
-            # special case: the user has chosen to do nothing
-            if self.hr_upscale_to_x == self.width and self.hr_upscale_to_y == self.height:
-                self.enable_hr = False
-                self.denoising_strength = None
-                self.extra_generation_params.pop("Hires upscale", None)
-                self.extra_generation_params.pop("Hires resize", None)
-                return
-
             if not state.processing_has_refined_job_count:
                 if state.job_count == -1:
                     state.job_count = self.n_iter
@@ -1045,17 +1053,22 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
     def sample(self, conditioning, unconditional_conditioning, seeds, subseeds, subseed_strength, prompts):
         self.sampler = sd_samplers.create_sampler(self.sampler_name, self.sd_model)
 
-        latent_scale_mode = shared.latent_upscale_modes.get(self.hr_upscaler, None) if self.hr_upscaler is not None else shared.latent_upscale_modes.get(shared.latent_upscale_default_mode, "nearest")
-        if self.enable_hr and latent_scale_mode is None:
-            if not any(x.name == self.hr_upscaler for x in shared.sd_upscalers):
-                raise Exception(f"could not find upscaler named {self.hr_upscaler}")
-
         x = create_random_tensors([opt_C, self.height // opt_f, self.width // opt_f], seeds=seeds, subseeds=subseeds, subseed_strength=self.subseed_strength, seed_resize_from_h=self.seed_resize_from_h, seed_resize_from_w=self.seed_resize_from_w, p=self)
         samples = self.sampler.sample(self, x, conditioning, unconditional_conditioning, image_conditioning=self.txt2img_image_conditioning(x))
 
         if not self.enable_hr:
             return samples
 
+        current = shared.sd_model.sd_checkpoint_info
+        try:
+            if self.hr_checkpoint_info is not None:
+                sd_models.reload_model_weights(info=self.hr_checkpoint_info)
+
+            return self.sample_hr_pass(samples, seeds, subseeds, subseed_strength, prompts)
+        finally:
+            sd_models.reload_model_weights(info=current)
+
+    def sample_hr_pass(self, samples, seeds, subseeds, subseed_strength, prompts):
         self.is_hr_pass = True
 
         target_width = self.hr_upscale_to_x
@@ -1073,11 +1086,11 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
             info = create_infotext(self, self.all_prompts, self.all_seeds, self.all_subseeds, [], iteration=self.iteration, position_in_batch=index)
             images.save_image(image, self.outpath_samples, "", seeds[index], prompts[index], opts.samples_format, info=info, p=self, suffix="-before-highres-fix")
 
-        if latent_scale_mode is not None:
+        if self.latent_scale_mode is not None:
             for i in range(samples.shape[0]):
                 save_intermediate(samples, i)
 
-            samples = torch.nn.functional.interpolate(samples, size=(target_height // opt_f, target_width // opt_f), mode=latent_scale_mode["mode"], antialias=latent_scale_mode["antialias"])
+            samples = torch.nn.functional.interpolate(samples, size=(target_height // opt_f, target_width // opt_f), mode=self.latent_scale_mode["mode"], antialias=self.latent_scale_mode["antialias"])
 
             # Avoid making the inpainting conditioning unless necessary as
             # this does need some extra compute to decode / encode the image again.
@@ -1193,7 +1206,7 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
         self.hr_uc = None
         self.hr_c = None
 
-        if self.enable_hr:
+        if self.enable_hr and self.hr_checkpoint_info is None:
             if shared.opts.hires_fix_use_firstpass_conds:
                 self.calculate_hr_conds()
 

modules/sd_models.py

@@ -52,6 +52,7 @@ class CheckpointInfo:
         self.shorthash = self.sha256[0:10] if self.sha256 else None
 
         self.title = name if self.shorthash is None else f'{name} [{self.shorthash}]'
+        self.short_title = self.name_for_extra if self.shorthash is None else f'{self.name_for_extra} [{self.shorthash}]'
 
         self.ids = [self.hash, self.model_name, self.title, name, f'{name} [{self.hash}]'] + ([self.shorthash, self.sha256, f'{self.name} [{self.shorthash}]'] if self.shorthash else [])
 
@@ -81,6 +82,7 @@ class CheckpointInfo:
 
         checkpoints_list.pop(self.title)
         self.title = f'{self.name} [{self.shorthash}]'
+        self.short_title = f'{self.name_for_extra} [{self.shorthash}]'
         self.register()
 
         return self.shorthash
@@ -101,14 +103,8 @@ def setup_model():
     enable_midas_autodownload()
 
 
-def checkpoint_tiles():
+def checkpoint_tiles(use_short=False):
-    def convert(name):
+    return [x.short_title if use_short else x.title for x in checkpoints_list.values()]
-        return int(name) if name.isdigit() else name.lower()
-
-    def alphanumeric_key(key):
-        return [convert(c) for c in re.split('([0-9]+)', key)]
-
-    return sorted([x.title for x in checkpoints_list.values()], key=alphanumeric_key)
 
 
 def list_models():
@@ -131,11 +127,14 @@ def list_models():
     elif cmd_ckpt is not None and cmd_ckpt != shared.default_sd_model_file:
         print(f"Checkpoint in --ckpt argument not found (Possible it was moved to {model_path}: {cmd_ckpt}", file=sys.stderr)
 
-    for filename in sorted(model_list, key=str.lower):
+    for filename in model_list:
         checkpoint_info = CheckpointInfo(filename)
         checkpoint_info.register()
 
 
+re_strip_checksum = re.compile(r"\s*\[[^]]+]\s*$")
+
+
 def get_closet_checkpoint_match(search_string):
     checkpoint_info = checkpoint_aliases.get(search_string, None)
     if checkpoint_info is not None:
@@ -145,6 +144,11 @@ def get_closet_checkpoint_match(search_string):
     if found:
         return found[0]
 
+    search_string_without_checksum = re.sub(re_strip_checksum, '', search_string)
+    found = sorted([info for info in checkpoints_list.values() if search_string_without_checksum in info.title], key=lambda x: len(x.title))
+    if found:
+        return found[0]
+
     return None
 
 

modules/shared.py

@@ -220,12 +220,19 @@ class State:
             return
 
         import modules.sd_samplers
-        if opts.show_progress_grid:
-            self.assign_current_image(modules.sd_samplers.samples_to_image_grid(self.current_latent))
-        else:
-            self.assign_current_image(modules.sd_samplers.sample_to_image(self.current_latent))
 
-        self.current_image_sampling_step = self.sampling_step
+        try:
+            if opts.show_progress_grid:
+                self.assign_current_image(modules.sd_samplers.samples_to_image_grid(self.current_latent))
+            else:
+                self.assign_current_image(modules.sd_samplers.sample_to_image(self.current_latent))
+
+            self.current_image_sampling_step = self.sampling_step
+
+        except Exception:
+            # when switching models during genration, VAE would be on CPU, so creating an image will fail.
+            # we silently ignore this error
+            errors.record_exception()
 
     def assign_current_image(self, image):
         self.current_image = image
@@ -512,7 +519,7 @@ options_templates.update(options_section(('ui', "User interface"), {
     "ui_tab_order": OptionInfo([], "UI tab order", ui_components.DropdownMulti, lambda: {"choices": list(tab_names)}).needs_restart(),
     "hidden_tabs": OptionInfo([], "Hidden UI tabs", ui_components.DropdownMulti, lambda: {"choices": list(tab_names)}).needs_restart(),
     "ui_reorder_list": OptionInfo([], "txt2img/img2img UI item order", ui_components.DropdownMulti, lambda: {"choices": list(shared_items.ui_reorder_categories())}).info("selected items appear first").needs_restart(),
-    "hires_fix_show_sampler": OptionInfo(False, "Hires fix: show hires sampler selection").needs_restart(),
+    "hires_fix_show_sampler": OptionInfo(False, "Hires fix: show hires checkpoint and sampler selection").needs_restart(),
     "hires_fix_show_prompts": OptionInfo(False, "Hires fix: show hires prompt and negative prompt").needs_restart(),
     "disable_token_counters": OptionInfo(False, "Disable prompt token counters").needs_restart(),
 }))

modules/txt2img.py

@@ -9,7 +9,7 @@ from modules.ui import plaintext_to_html
 import gradio as gr
 
 
-def txt2img(id_task: str, prompt: str, negative_prompt: str, prompt_styles, steps: int, sampler_index: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, enable_hr: bool, denoising_strength: float, hr_scale: float, hr_upscaler: str, hr_second_pass_steps: int, hr_resize_x: int, hr_resize_y: int, hr_sampler_index: int, hr_prompt: str, hr_negative_prompt, override_settings_texts, request: gr.Request, *args):
+def txt2img(id_task: str, prompt: str, negative_prompt: str, prompt_styles, steps: int, sampler_index: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, enable_hr: bool, denoising_strength: float, hr_scale: float, hr_upscaler: str, hr_second_pass_steps: int, hr_resize_x: int, hr_resize_y: int, hr_checkpoint_name: str, hr_sampler_index: int, hr_prompt: str, hr_negative_prompt, override_settings_texts, request: gr.Request, *args):
     override_settings = create_override_settings_dict(override_settings_texts)
 
     p = processing.StableDiffusionProcessingTxt2Img(
@@ -41,6 +41,7 @@ def txt2img(id_task: str, prompt: str, negative_prompt: str, prompt_styles, step
         hr_second_pass_steps=hr_second_pass_steps,
         hr_resize_x=hr_resize_x,
         hr_resize_y=hr_resize_y,
+        hr_checkpoint_name=None if hr_checkpoint_name == 'Use same checkpoint' else hr_checkpoint_name,
         hr_sampler_name=sd_samplers.samplers_for_img2img[hr_sampler_index - 1].name if hr_sampler_index != 0 else None,
         hr_prompt=hr_prompt,
         hr_negative_prompt=hr_negative_prompt,

modules/ui.py

@@ -476,6 +476,10 @@ def create_ui():
                                 hr_resize_y = gr.Slider(minimum=0, maximum=2048, step=8, label="Resize height to", value=0, elem_id="txt2img_hr_resize_y")
 
                             with FormRow(elem_id="txt2img_hires_fix_row3", variant="compact", visible=opts.hires_fix_show_sampler) as hr_sampler_container:
+                                checkpoint_choices = lambda: ["Use same checkpoint"] + modules.sd_models.checkpoint_tiles(use_short=True)
+                                hr_checkpoint_name = gr.Dropdown(label='Hires checkpoint', elem_id="hr_checkpoint", choices=checkpoint_choices(), value="Use same checkpoint")
+                                create_refresh_button(hr_checkpoint_name, modules.sd_models.list_models, lambda: {"choices": checkpoint_choices()}, "hr_checkpoint_refresh")
+
                                 hr_sampler_index = gr.Dropdown(label='Hires sampling method', elem_id="hr_sampler", choices=["Use same sampler"] + [x.name for x in samplers_for_img2img], value="Use same sampler", type="index")
 
                             with FormRow(elem_id="txt2img_hires_fix_row4", variant="compact", visible=opts.hires_fix_show_prompts) as hr_prompts_container:
@@ -553,6 +557,7 @@ def create_ui():
                     hr_second_pass_steps,
                     hr_resize_x,
                     hr_resize_y,
+                    hr_checkpoint_name,
                     hr_sampler_index,
                     hr_prompt,
                     hr_negative_prompt,
@@ -630,8 +635,9 @@ def create_ui():
                 (hr_second_pass_steps, "Hires steps"),
                 (hr_resize_x, "Hires resize-1"),
                 (hr_resize_y, "Hires resize-2"),
+                (hr_checkpoint_name, "Hires checkpoint"),
                 (hr_sampler_index, "Hires sampler"),
-                (hr_sampler_container, lambda d: gr.update(visible=True) if d.get("Hires sampler", "Use same sampler") != "Use same sampler" else gr.update()),
+                (hr_sampler_container, lambda d: gr.update(visible=True) if d.get("Hires sampler", "Use same sampler") != "Use same sampler" or d.get("Hires checkpoint", "Use same checkpoint") != "Use same checkpoint" else gr.update()),
                 (hr_prompt, "Hires prompt"),
                 (hr_negative_prompt, "Hires negative prompt"),
                 (hr_prompts_container, lambda d: gr.update(visible=True) if d.get("Hires prompt", "") != "" or d.get("Hires negative prompt", "") != "" else gr.update()),