add support for switching model checkpoints at runtime

modules/images.py

@@ -274,7 +274,7 @@ def apply_filename_pattern(x, p, seed, prompt):
         x = x.replace("[height]", str(p.height))
         x = x.replace("[sampler]", sd_samplers.samplers[p.sampler_index].name)
 
-    x = x.replace("[model_hash]", shared.sd_model_hash)
+    x = x.replace("[model_hash]", shared.sd_model.sd_model_hash)
     x = x.replace("[date]", datetime.date.today().isoformat())
 
     if cmd_opts.hide_ui_dir_config:

modules/processing.py

@@ -227,7 +227,7 @@ def process_images(p: StableDiffusionProcessing) -> Processed:
             "Seed": all_seeds[index],
             "Face restoration": (opts.face_restoration_model if p.restore_faces else None),
             "Size": f"{p.width}x{p.height}",
-            "Model hash": (None if not opts.add_model_hash_to_info or not shared.sd_model_hash else shared.sd_model_hash),
+            "Model hash": (None if not opts.add_model_hash_to_info or not shared.sd_model.sd_model_hash else shared.sd_model.sd_model_hash),
             "Batch size": (None if p.batch_size < 2 else p.batch_size),
             "Batch pos": (None if p.batch_size < 2 else position_in_batch),
             "Variation seed": (None if p.subseed_strength == 0 else all_subseeds[index]),

modules/sd_models.py

@@ -0,0 +1,148 @@
+import glob
+import os.path
+import sys
+from collections import namedtuple
+import torch
+from omegaconf import OmegaConf
+
+
+from ldm.util import instantiate_from_config
+
+from modules import shared
+
+CheckpointInfo = namedtuple("CheckpointInfo", ['filename', 'title', 'hash'])
+checkpoints_list = {}
+
+try:
+    # this silences the annoying "Some weights of the model checkpoint were not used when initializing..." message at start.
+
+    from transformers import logging
+
+    logging.set_verbosity_error()
+except Exception:
+    pass
+
+
+def list_models():
+    checkpoints_list.clear()
+
+    model_dir = os.path.abspath(shared.cmd_opts.ckpt_dir)
+
+    def modeltitle(path, h):
+        abspath = os.path.abspath(path)
+
+        if abspath.startswith(model_dir):
+            name = abspath.replace(model_dir, '')
+        else:
+            name = os.path.basename(path)
+
+        if name.startswith("\\") or name.startswith("/"):
+            name = name[1:]
+
+        return f'{name} [{h}]'
+
+    cmd_ckpt = shared.cmd_opts.ckpt
+    if os.path.exists(cmd_ckpt):
+        h = model_hash(cmd_ckpt)
+        title = modeltitle(cmd_ckpt, h)
+        checkpoints_list[title] = CheckpointInfo(cmd_ckpt, title, h)
+    elif cmd_ckpt is not None and cmd_ckpt != shared.default_sd_model_file:
+        print(f"Checkpoint in --ckpt argument not found: {cmd_ckpt}", file=sys.stderr)
+
+    if os.path.exists(model_dir):
+        for filename in glob.glob(model_dir + '/**/*.ckpt', recursive=True):
+            h = model_hash(filename)
+            title = modeltitle(filename, h)
+            checkpoints_list[title] = CheckpointInfo(filename, title, h)
+
+
+def model_hash(filename):
+    try:
+        with open(filename, "rb") as file:
+            import hashlib
+            m = hashlib.sha256()
+
+            file.seek(0x100000)
+            m.update(file.read(0x10000))
+            return m.hexdigest()[0:8]
+    except FileNotFoundError:
+        return 'NOFILE'
+
+
+def select_checkpoint():
+    model_checkpoint = shared.opts.sd_model_checkpoint
+    checkpoint_info = checkpoints_list.get(model_checkpoint, None)
+    if checkpoint_info is not None:
+        return checkpoint_info
+
+    if len(checkpoints_list) == 0:
+        print(f"Checkpoint {model_checkpoint} not found and no other checkpoints found", file=sys.stderr)
+        return None
+
+    checkpoint_info = next(iter(checkpoints_list.values()))
+    if model_checkpoint is not None:
+        print(f"Checkpoint {model_checkpoint} not found; loading fallback {checkpoint_info.title}", file=sys.stderr)
+
+    return checkpoint_info
+
+
+def load_model_weights(model, checkpoint_file, sd_model_hash):
+    print(f"Loading weights [{sd_model_hash}] from {checkpoint_file}")
+
+    pl_sd = torch.load(checkpoint_file, map_location="cpu")
+    if "global_step" in pl_sd:
+        print(f"Global Step: {pl_sd['global_step']}")
+    sd = pl_sd["state_dict"]
+
+    model.load_state_dict(sd, strict=False)
+
+    if shared.cmd_opts.opt_channelslast:
+        model.to(memory_format=torch.channels_last)
+
+    if not shared.cmd_opts.no_half:
+        model.half()
+
+    model.sd_model_hash = sd_model_hash
+    model.sd_model_checkpint = checkpoint_file
+
+
+def load_model():
+    from modules import lowvram, sd_hijack
+    checkpoint_info = select_checkpoint()
+
+    sd_config = OmegaConf.load(shared.cmd_opts.config)
+    sd_model = instantiate_from_config(sd_config.model)
+    load_model_weights(sd_model, checkpoint_info.filename, checkpoint_info.hash)
+
+    if shared.cmd_opts.lowvram or shared.cmd_opts.medvram:
+        lowvram.setup_for_low_vram(sd_model, shared.cmd_opts.medvram)
+    else:
+        sd_model.to(shared.device)
+
+    sd_hijack.model_hijack.hijack(sd_model)
+
+    sd_model.eval()
+
+    print(f"Model loaded.")
+    return sd_model
+
+
+def reload_model_weights(sd_model):
+    from modules import lowvram, devices
+    checkpoint_info = select_checkpoint()
+
+    if sd_model.sd_model_checkpint == checkpoint_info.filename:
+        return
+
+    if shared.cmd_opts.lowvram or shared.cmd_opts.medvram:
+        lowvram.send_everything_to_cpu()
+    else:
+        sd_model.to(devices.cpu)
+
+    load_model_weights(sd_model, checkpoint_info.filename, checkpoint_info.hash)
+
+    if not shared.cmd_opts.lowvram and not shared.cmd_opts.medvram:
+        sd_model.to(devices.device)
+
+    print(f"Weights loaded.")
+    return sd_model

modules/shared.py

@@ -13,14 +13,15 @@ from modules.devices import get_optimal_device
 import modules.styles
 import modules.interrogate
 import modules.memmon
+import modules.sd_models
 
 sd_model_file = os.path.join(script_path, 'model.ckpt')
-if not os.path.exists(sd_model_file):
-    sd_model_file = "models/ldm/stable-diffusion-v1/model.ckpt"
+default_sd_model_file = sd_model_file
 
 parser = argparse.ArgumentParser()
 parser.add_argument("--config", type=str, default=os.path.join(sd_path, "configs/stable-diffusion/v1-inference.yaml"), help="path to config which constructs model",)
-parser.add_argument("--ckpt", type=str, default=os.path.join(sd_path, sd_model_file), help="path to checkpoint of model",)
+parser.add_argument("--ckpt", type=str, default=sd_model_file, help="path to checkpoint of stable diffusion model; this checkpoint will be added to the list of checkpoints and loaded by default if you don't have a checkpoint selected in settings",)
+parser.add_argument("--ckpt-dir", type=str, default=os.path.join(script_path, 'models'), help="path to directory with stable diffusion checkpoints",)
 parser.add_argument("--gfpgan-dir", type=str, help="GFPGAN directory", default=('./src/gfpgan' if os.path.exists('./src/gfpgan') else './GFPGAN'))
 parser.add_argument("--gfpgan-model", type=str, help="GFPGAN model file name", default='GFPGANv1.3.pth')
 parser.add_argument("--no-half", action='store_true', help="do not switch the model to 16-bit floats")
@@ -88,13 +89,17 @@ interrogator = modules.interrogate.InterrogateModels("interrogate")
 
 face_restorers = []
 
+modules.sd_models.list_models()
+
+
 class Options:
     class OptionInfo:
-        def __init__(self, default=None, label="", component=None, component_args=None):
+        def __init__(self, default=None, label="", component=None, component_args=None, onchange=None):
             self.default = default
             self.label = label
             self.component = component
             self.component_args = component_args
+            self.onchange = onchange
 
     data = None
     hide_dirs = {"visible": False} if cmd_opts.hide_ui_dir_config else None
@@ -150,6 +155,7 @@ class Options:
         "interrogate_clip_min_length": OptionInfo(24, "Interrogate: minimum description length (excluding artists, etc..)", gr.Slider, {"minimum": 1, "maximum": 128, "step": 1}),
         "interrogate_clip_max_length": OptionInfo(48, "Interrogate: maximum description length", gr.Slider, {"minimum": 1, "maximum": 256, "step": 1}),
         "interrogate_clip_dict_limit": OptionInfo(1500, "Interrogate: maximum number of lines in text file (0 = No limit)"),
+        "sd_model_checkpoint": OptionInfo(None, "Stable Diffusion checkpoint", gr.Radio, lambda: {"choices": [x.title for x in modules.sd_models.checkpoints_list.values()]}),
     }
 
     def __init__(self):
@@ -180,6 +186,10 @@ class Options:
         with open(filename, "r", encoding="utf8") as file:
             self.data = json.load(file)
 
+    def onchange(self, key, func):
+        item = self.data_labels.get(key)
+        item.onchange = func
+
 
 opts = Options()
 if os.path.exists(config_filename):
@@ -188,7 +198,6 @@ if os.path.exists(config_filename):
 sd_upscalers = []
 
 sd_model = None
-sd_model_hash = ''
 
 progress_print_out = sys.stdout
 

modules/ui.py

@@ -758,7 +758,12 @@ def create_ui(txt2img, img2img, run_extras, run_pnginfo):
             if comp_args and isinstance(comp_args, dict) and comp_args.get('visible') is False:
                 continue
 
+            oldval = opts.data.get(key, None)
             opts.data[key] = value
+
+            if oldval != value and opts.data_labels[key].onchange is not None:
+                opts.data_labels[key].onchange()
+
             up.append(comp.update(value=value))
 
         opts.save(shared.config_filename)

webui.py

@@ -3,13 +3,8 @@ import threading
 
 from modules.paths import script_path
 
-import torch
-from omegaconf import OmegaConf
-
 import signal
 
-from ldm.util import instantiate_from_config
-
 from modules.shared import opts, cmd_opts, state
 import modules.shared as shared
 import modules.ui
@@ -24,6 +19,7 @@ import modules.extras
 import modules.lowvram
 import modules.txt2img
 import modules.img2img
+import modules.sd_models
 
 
 modules.codeformer_model.setup_codeformer()
@@ -33,31 +29,19 @@ shared.face_restorers.append(modules.face_restoration.FaceRestoration())
 esrgan.load_models(cmd_opts.esrgan_models_path)
 realesrgan.setup_realesrgan()
 
-
-def load_model_from_config(config, ckpt, verbose=False):
-    print(f"Loading model [{shared.sd_model_hash}] from {ckpt}")
-    pl_sd = torch.load(ckpt, map_location="cpu")
-    if "global_step" in pl_sd:
-        print(f"Global Step: {pl_sd['global_step']}")
-    sd = pl_sd["state_dict"]
-
-    model = instantiate_from_config(config.model)
-    m, u = model.load_state_dict(sd, strict=False)
-    if len(m) > 0 and verbose:
-        print("missing keys:")
-        print(m)
-    if len(u) > 0 and verbose:
-        print("unexpected keys:")
-        print(u)
-    if cmd_opts.opt_channelslast:
-        model = model.to(memory_format=torch.channels_last)
-    model.eval()
-    return model
-
-
 queue_lock = threading.Lock()
 
 
+def wrap_queued_call(func):
+    def f(*args, **kwargs):
+        with queue_lock:
+            res = func(*args, **kwargs)
+
+        return res
+
+    return f
+
+
 def wrap_gradio_gpu_call(func):
     def f(*args, **kwargs):
         shared.state.sampling_step = 0
@@ -80,33 +64,8 @@ def wrap_gradio_gpu_call(func):
 
 modules.scripts.load_scripts(os.path.join(script_path, "scripts"))
 
-try:
-    # this silences the annoying "Some weights of the model checkpoint were not used when initializing..." message at start.
-
-    from transformers import logging
-
-    logging.set_verbosity_error()
-except Exception:
-    pass
-
-with open(cmd_opts.ckpt, "rb") as file:
-    import hashlib
-    m = hashlib.sha256()
-
-    file.seek(0x100000)
-    m.update(file.read(0x10000))
-    shared.sd_model_hash = m.hexdigest()[0:8]
-
-sd_config = OmegaConf.load(cmd_opts.config)
-shared.sd_model = load_model_from_config(sd_config, cmd_opts.ckpt)
-shared.sd_model = (shared.sd_model if cmd_opts.no_half else shared.sd_model.half())
-
-if cmd_opts.lowvram or cmd_opts.medvram:
-    modules.lowvram.setup_for_low_vram(shared.sd_model, cmd_opts.medvram)
-else:
-    shared.sd_model = shared.sd_model.to(shared.device)
-
-modules.sd_hijack.model_hijack.hijack(shared.sd_model)
+shared.sd_model = modules.sd_models.load_model()
+shared.opts.onchange("sd_model_checkpoint", wrap_queued_call(lambda: modules.sd_models.reload_model_weights(shared.sd_model)))
 
 
 def webui():