Merge pull request #11757 from AUTOMATIC1111/sdxl

SD XL support

extensions-builtin/Lora/lora.py

@@ -68,6 +68,14 @@ def convert_diffusers_name_to_compvis(key, is_sd2):
 
         return f"transformer_text_model_encoder_layers_{m[0]}_{m[1]}"
 
+    if match(m, r"lora_te2_text_model_encoder_layers_(\d+)_(.+)"):
+        if 'mlp_fc1' in m[1]:
+            return f"1_model_transformer_resblocks_{m[0]}_{m[1].replace('mlp_fc1', 'mlp_c_fc')}"
+        elif 'mlp_fc2' in m[1]:
+            return f"1_model_transformer_resblocks_{m[0]}_{m[1].replace('mlp_fc2', 'mlp_c_proj')}"
+        else:
+            return f"1_model_transformer_resblocks_{m[0]}_{m[1].replace('self_attn', 'attn')}"
+
     return key
 
 
@@ -147,10 +155,20 @@ class LoraUpDownModule:
 def assign_lora_names_to_compvis_modules(sd_model):
     lora_layer_mapping = {}
 
-    for name, module in shared.sd_model.cond_stage_model.wrapped.named_modules():
-        lora_name = name.replace(".", "_")
-        lora_layer_mapping[lora_name] = module
-        module.lora_layer_name = lora_name
+    if shared.sd_model.is_sdxl:
+        for i, embedder in enumerate(shared.sd_model.conditioner.embedders):
+            if not hasattr(embedder, 'wrapped'):
+                continue
+
+            for name, module in embedder.wrapped.named_modules():
+                lora_name = f'{i}_{name.replace(".", "_")}'
+                lora_layer_mapping[lora_name] = module
+                module.lora_layer_name = lora_name
+    else:
+        for name, module in shared.sd_model.cond_stage_model.wrapped.named_modules():
+            lora_name = name.replace(".", "_")
+            lora_layer_mapping[lora_name] = module
+            module.lora_layer_name = lora_name
 
     for name, module in shared.sd_model.model.named_modules():
         lora_name = name.replace(".", "_")
@@ -173,10 +191,10 @@ def load_lora(name, lora_on_disk):
     keys_failed_to_match = {}
     is_sd2 = 'model_transformer_resblocks' in shared.sd_model.lora_layer_mapping
 
-    for key_diffusers, weight in sd.items():
-        key_diffusers_without_lora_parts, lora_key = key_diffusers.split(".", 1)
-        key = convert_diffusers_name_to_compvis(key_diffusers_without_lora_parts, is_sd2)
+    for key_lora, weight in sd.items():
+        key_lora_without_lora_parts, lora_key = key_lora.split(".", 1)
 
+        key = convert_diffusers_name_to_compvis(key_lora_without_lora_parts, is_sd2)
         sd_module = shared.sd_model.lora_layer_mapping.get(key, None)
 
         if sd_module is None:
@@ -184,8 +202,16 @@ def load_lora(name, lora_on_disk):
             if m:
                 sd_module = shared.sd_model.lora_layer_mapping.get(m.group(1), None)
 
+        # SDXL loras seem to already have correct compvis keys, so only need to replace "lora_unet" with "diffusion_model"
+        if sd_module is None and "lora_unet" in key_lora_without_lora_parts:
+            key = key_lora_without_lora_parts.replace("lora_unet", "diffusion_model")
+            sd_module = shared.sd_model.lora_layer_mapping.get(key, None)
+        elif sd_module is None and "lora_te1_text_model" in key_lora_without_lora_parts:
+            key = key_lora_without_lora_parts.replace("lora_te1_text_model", "0_transformer_text_model")
+            sd_module = shared.sd_model.lora_layer_mapping.get(key, None)
+
         if sd_module is None:
-            keys_failed_to_match[key_diffusers] = key
+            keys_failed_to_match[key_lora] = key
             continue
 
         lora_module = lora.modules.get(key, None)
@@ -208,9 +234,9 @@ def load_lora(name, lora_on_disk):
         elif type(sd_module) == torch.nn.Conv2d and weight.shape[2:] == (3, 3):
             module = torch.nn.Conv2d(weight.shape[1], weight.shape[0], (3, 3), bias=False)
         else:
-            print(f'Lora layer {key_diffusers} matched a layer with unsupported type: {type(sd_module).__name__}')
+            print(f'Lora layer {key_lora} matched a layer with unsupported type: {type(sd_module).__name__}')
             continue
-            raise AssertionError(f"Lora layer {key_diffusers} matched a layer with unsupported type: {type(sd_module).__name__}")
+            raise AssertionError(f"Lora layer {key_lora} matched a layer with unsupported type: {type(sd_module).__name__}")
 
         with torch.no_grad():
             module.weight.copy_(weight)
@@ -222,7 +248,7 @@ def load_lora(name, lora_on_disk):
         elif lora_key == "lora_down.weight":
             lora_module.down = module
         else:
-            raise AssertionError(f"Bad Lora layer name: {key_diffusers} - must end in lora_up.weight, lora_down.weight or alpha")
+            raise AssertionError(f"Bad Lora layer name: {key_lora} - must end in lora_up.weight, lora_down.weight or alpha")
 
     if keys_failed_to_match:
         print(f"Failed to match keys when loading Lora {lora_on_disk.filename}: {keys_failed_to_match}")

modules/hypernetworks/hypernetwork.py

@@ -378,7 +378,7 @@ def apply_hypernetworks(hypernetworks, context, layer=None):
     return context_k, context_v
 
 
-def attention_CrossAttention_forward(self, x, context=None, mask=None):
+def attention_CrossAttention_forward(self, x, context=None, mask=None, **kwargs):
     h = self.heads
 
     q = self.to_q(x)

modules/launch_utils.py

@@ -237,11 +237,13 @@ def prepare_environment():
     openclip_package = os.environ.get('OPENCLIP_PACKAGE', "https://github.com/mlfoundations/open_clip/archive/bb6e834e9c70d9c27d0dc3ecedeebeaeb1ffad6b.zip")
 
     stable_diffusion_repo = os.environ.get('STABLE_DIFFUSION_REPO', "https://github.com/Stability-AI/stablediffusion.git")
+    stable_diffusion_xl_repo = os.environ.get('STABLE_DIFFUSION_XL_REPO', "https://github.com/Stability-AI/generative-models.git")
     k_diffusion_repo = os.environ.get('K_DIFFUSION_REPO', 'https://github.com/crowsonkb/k-diffusion.git')
     codeformer_repo = os.environ.get('CODEFORMER_REPO', 'https://github.com/sczhou/CodeFormer.git')
     blip_repo = os.environ.get('BLIP_REPO', 'https://github.com/salesforce/BLIP.git')
 
     stable_diffusion_commit_hash = os.environ.get('STABLE_DIFFUSION_COMMIT_HASH', "cf1d67a6fd5ea1aa600c4df58e5b47da45f6bdbf")
+    stable_diffusion_xl_commit_hash = os.environ.get('STABLE_DIFFUSION_XL_COMMIT_HASH', "5c10deee76adad0032b412294130090932317a87")
     k_diffusion_commit_hash = os.environ.get('K_DIFFUSION_COMMIT_HASH', "c9fe758757e022f05ca5a53fa8fac28889e4f1cf")
     codeformer_commit_hash = os.environ.get('CODEFORMER_COMMIT_HASH', "c5b4593074ba6214284d6acd5f1719b6c5d739af")
     blip_commit_hash = os.environ.get('BLIP_COMMIT_HASH', "48211a1594f1321b00f14c9f7a5b4813144b2fb9")
@@ -299,6 +301,7 @@ def prepare_environment():
     os.makedirs(os.path.join(script_path, dir_repos), exist_ok=True)
 
     git_clone(stable_diffusion_repo, repo_dir('stable-diffusion-stability-ai'), "Stable Diffusion", stable_diffusion_commit_hash)
+    git_clone(stable_diffusion_xl_repo, repo_dir('generative-models'), "Stable Diffusion XL", stable_diffusion_xl_commit_hash)
     git_clone(k_diffusion_repo, repo_dir('k-diffusion'), "K-diffusion", k_diffusion_commit_hash)
     git_clone(codeformer_repo, repo_dir('CodeFormer'), "CodeFormer", codeformer_commit_hash)
     git_clone(blip_repo, repo_dir('BLIP'), "BLIP", blip_commit_hash)
@@ -323,6 +326,7 @@ def prepare_environment():
         exit(0)
 
 
+
 def configure_for_tests():
     if "--api" not in sys.argv:
         sys.argv.append("--api")

modules/lowvram.py

@@ -53,19 +53,46 @@ def setup_for_low_vram(sd_model, use_medvram):
         send_me_to_gpu(first_stage_model, None)
         return first_stage_model_decode(z)
 
-    # for SD1, cond_stage_model is CLIP and its NN is in the tranformer frield, but for SD2, it's open clip, and it's in model field
-    if hasattr(sd_model.cond_stage_model, 'model'):
-        sd_model.cond_stage_model.transformer = sd_model.cond_stage_model.model
-
-    # remove several big modules: cond, first_stage, depth/embedder (if applicable), and unet from the model and then
-    # send the model to GPU. Then put modules back. the modules will be in CPU.
-    stored = sd_model.cond_stage_model.transformer, sd_model.first_stage_model, getattr(sd_model, 'depth_model', None), getattr(sd_model, 'embedder', None), sd_model.model
-    sd_model.cond_stage_model.transformer, sd_model.first_stage_model, sd_model.depth_model, sd_model.embedder, sd_model.model = None, None, None, None, None
+    to_remain_in_cpu = [
+        (sd_model, 'first_stage_model'),
+        (sd_model, 'depth_model'),
+        (sd_model, 'embedder'),
+        (sd_model, 'model'),
+        (sd_model, 'embedder'),
+    ]
+
+    is_sdxl = hasattr(sd_model, 'conditioner')
+    is_sd2 = not is_sdxl and hasattr(sd_model.cond_stage_model, 'model')
+
+    if is_sdxl:
+        to_remain_in_cpu.append((sd_model, 'conditioner'))
+    elif is_sd2:
+        to_remain_in_cpu.append((sd_model.cond_stage_model, 'model'))
+    else:
+        to_remain_in_cpu.append((sd_model.cond_stage_model, 'transformer'))
+
+    # remove several big modules: cond, first_stage, depth/embedder (if applicable), and unet from the model
+    stored = []
+    for obj, field in to_remain_in_cpu:
+        module = getattr(obj, field, None)
+        stored.append(module)
+        setattr(obj, field, None)
+
+    # send the model to GPU.
     sd_model.to(devices.device)
-    sd_model.cond_stage_model.transformer, sd_model.first_stage_model, sd_model.depth_model, sd_model.embedder, sd_model.model = stored
+
+    # put modules back. the modules will be in CPU.
+    for (obj, field), module in zip(to_remain_in_cpu, stored):
+        setattr(obj, field, module)
 
     # register hooks for those the first three models
-    sd_model.cond_stage_model.transformer.register_forward_pre_hook(send_me_to_gpu)
+    if is_sdxl:
+        sd_model.conditioner.register_forward_pre_hook(send_me_to_gpu)
+    elif is_sd2:
+        sd_model.cond_stage_model.model.register_forward_pre_hook(send_me_to_gpu)
+    else:
+        sd_model.cond_stage_model.transformer.register_forward_pre_hook(send_me_to_gpu)
+
     sd_model.first_stage_model.register_forward_pre_hook(send_me_to_gpu)
     sd_model.first_stage_model.encode = first_stage_model_encode_wrap
     sd_model.first_stage_model.decode = first_stage_model_decode_wrap
@@ -73,11 +100,9 @@ def setup_for_low_vram(sd_model, use_medvram):
         sd_model.depth_model.register_forward_pre_hook(send_me_to_gpu)
     if sd_model.embedder:
         sd_model.embedder.register_forward_pre_hook(send_me_to_gpu)
-    parents[sd_model.cond_stage_model.transformer] = sd_model.cond_stage_model
 
-    if hasattr(sd_model.cond_stage_model, 'model'):
-        sd_model.cond_stage_model.model = sd_model.cond_stage_model.transformer
-        del sd_model.cond_stage_model.transformer
+    if hasattr(sd_model, 'cond_stage_model'):
+        parents[sd_model.cond_stage_model.transformer] = sd_model.cond_stage_model
 
     if use_medvram:
         sd_model.model.register_forward_pre_hook(send_me_to_gpu)

modules/paths.py

@@ -5,6 +5,21 @@ from modules.paths_internal import models_path, script_path, data_path, extensio
 import modules.safe  # noqa: F401
 
 
+def mute_sdxl_imports():
+    """create fake modules that SDXL wants to import but doesn't actually use for our purposes"""
+
+    class Dummy:
+        pass
+
+    module = Dummy()
+    module.LPIPS = None
+    sys.modules['taming.modules.losses.lpips'] = module
+
+    module = Dummy()
+    module.StableDataModuleFromConfig = None
+    sys.modules['sgm.data'] = module
+
+
 # data_path = cmd_opts_pre.data
 sys.path.insert(0, script_path)
 
@@ -18,8 +33,11 @@ for possible_sd_path in possible_sd_paths:
 
 assert sd_path is not None, f"Couldn't find Stable Diffusion in any of: {possible_sd_paths}"
 
+mute_sdxl_imports()
+
 path_dirs = [
     (sd_path, 'ldm', 'Stable Diffusion', []),
+    (os.path.join(sd_path, '../generative-models'), 'sgm', 'Stable Diffusion XL', ["sgm"]),
     (os.path.join(sd_path, '../CodeFormer'), 'inference_codeformer.py', 'CodeFormer', []),
     (os.path.join(sd_path, '../BLIP'), 'models/blip.py', 'BLIP', []),
     (os.path.join(sd_path, '../k-diffusion'), 'k_diffusion/sampling.py', 'k_diffusion', ["atstart"]),
@@ -35,6 +53,13 @@ for d, must_exist, what, options in path_dirs:
         d = os.path.abspath(d)
         if "atstart" in options:
             sys.path.insert(0, d)
+        elif "sgm" in options:
+            # Stable Diffusion XL repo has scripts dir with __init__.py in it which ruins every extension's scripts dir, so we
+            # import sgm and remove it from sys.path so that when a script imports scripts.something, it doesbn't use sgm's scripts dir.
+
+            sys.path.insert(0, d)
+            import sgm  # noqa: F401
+            sys.path.pop(0)
         else:
             sys.path.append(d)
         paths[what] = d

modules/processing.py

@@ -330,8 +330,21 @@ class StableDiffusionProcessing:
 
         caches is a list with items described above.
         """
+
+        cached_params = (
+            required_prompts,
+            steps,
+            opts.CLIP_stop_at_last_layers,
+            shared.sd_model.sd_checkpoint_info,
+            extra_network_data,
+            opts.sdxl_crop_left,
+            opts.sdxl_crop_top,
+            self.width,
+            self.height,
+        )
+
         for cache in caches:
-            if cache[0] is not None and (required_prompts, steps, opts.CLIP_stop_at_last_layers, shared.sd_model.sd_checkpoint_info, extra_network_data) == cache[0]:
+            if cache[0] is not None and cached_params == cache[0]:
                 return cache[1]
 
         cache = caches[0]
@@ -339,14 +352,17 @@ class StableDiffusionProcessing:
         with devices.autocast():
             cache[1] = function(shared.sd_model, required_prompts, steps)
 
-        cache[0] = (required_prompts, steps, opts.CLIP_stop_at_last_layers, shared.sd_model.sd_checkpoint_info, extra_network_data)
+        cache[0] = cached_params
         return cache[1]
 
     def setup_conds(self):
+        prompts = prompt_parser.SdConditioning(self.prompts, width=self.width, height=self.height)
+        negative_prompts = prompt_parser.SdConditioning(self.negative_prompts, width=self.width, height=self.height, is_negative_prompt=True)
+
         sampler_config = sd_samplers.find_sampler_config(self.sampler_name)
         self.step_multiplier = 2 if sampler_config and sampler_config.options.get("second_order", False) else 1
-        self.uc = self.get_conds_with_caching(prompt_parser.get_learned_conditioning, self.negative_prompts, self.steps * self.step_multiplier, [self.cached_uc], self.extra_network_data)
-        self.c = self.get_conds_with_caching(prompt_parser.get_multicond_learned_conditioning, self.prompts, self.steps * self.step_multiplier, [self.cached_c], self.extra_network_data)
+        self.uc = self.get_conds_with_caching(prompt_parser.get_learned_conditioning, negative_prompts, self.steps * self.step_multiplier, [self.cached_uc], self.extra_network_data)
+        self.c = self.get_conds_with_caching(prompt_parser.get_multicond_learned_conditioning, prompts, self.steps * self.step_multiplier, [self.cached_c], self.extra_network_data)
 
     def parse_extra_network_prompts(self):
         self.prompts, self.extra_network_data = extra_networks.parse_prompts(self.prompts)
@@ -523,8 +539,7 @@ def create_random_tensors(shape, seeds, subseeds=None, subseed_strength=0.0, see
 
 
 def decode_first_stage(model, x):
-    with devices.autocast(disable=x.dtype == devices.dtype_vae):
-        x = model.decode_first_stage(x)
+    x = model.decode_first_stage(x.to(devices.dtype_vae))
 
     return x
 

modules/prompt_parser.py

+from __future__ import annotations
+
 import re
 from collections import namedtuple
 from typing import List
@@ -109,7 +111,25 @@ def get_learned_conditioning_prompt_schedules(prompts, steps):
 ScheduledPromptConditioning = namedtuple("ScheduledPromptConditioning", ["end_at_step", "cond"])
 
 
-def get_learned_conditioning(model, prompts, steps):
+class SdConditioning(list):
+    """
+    A list with prompts for stable diffusion's conditioner model.
+    Can also specify width and height of created image - SDXL needs it.
+    """
+    def __init__(self, prompts, is_negative_prompt=False, width=None, height=None, copy_from=None):
+        super().__init__()
+        self.extend(prompts)
+
+        if copy_from is None:
+            copy_from = prompts
+
+        self.is_negative_prompt = is_negative_prompt or getattr(copy_from, 'is_negative_prompt', False)
+        self.width = width or getattr(copy_from, 'width', None)
+        self.height = height or getattr(copy_from, 'height', None)
+
+
+
+def get_learned_conditioning(model, prompts: SdConditioning | list[str], steps):
     """converts a list of prompts into a list of prompt schedules - each schedule is a list of ScheduledPromptConditioning, specifying the comdition (cond),
     and the sampling step at which this condition is to be replaced by the next one.
 
@@ -139,12 +159,17 @@ def get_learned_conditioning(model, prompts, steps):
             res.append(cached)
             continue
 
-        texts = [x[1] for x in prompt_schedule]
+        texts = SdConditioning([x[1] for x in prompt_schedule], copy_from=prompts)
         conds = model.get_learned_conditioning(texts)
 
         cond_schedule = []
         for i, (end_at_step, _) in enumerate(prompt_schedule):
-            cond_schedule.append(ScheduledPromptConditioning(end_at_step, conds[i]))
+            if isinstance(conds, dict):
+                cond = {k: v[i] for k, v in conds.items()}
+            else:
+                cond = conds[i]
+
+            cond_schedule.append(ScheduledPromptConditioning(end_at_step, cond))
 
         cache[prompt] = cond_schedule
         res.append(cond_schedule)
@@ -155,11 +180,13 @@ def get_learned_conditioning(model, prompts, steps):
 re_AND = re.compile(r"\bAND\b")
 re_weight = re.compile(r"^(.*?)(?:\s*:\s*([-+]?(?:\d+\.?|\d*\.\d+)))?\s*$")
 
-def get_multicond_prompt_list(prompts):
+
+def get_multicond_prompt_list(prompts: SdConditioning | list[str]):
     res_indexes = []
 
-    prompt_flat_list = []
     prompt_indexes = {}
+    prompt_flat_list = SdConditioning(prompts)
+    prompt_flat_list.clear()
 
     for prompt in prompts:
         subprompts = re_AND.split(prompt)
@@ -196,6 +223,7 @@ class MulticondLearnedConditioning:
         self.shape: tuple = shape  # the shape field is needed to send this object to DDIM/PLMS
         self.batch: List[List[ComposableScheduledPromptConditioning]] = batch
 
+
 def get_multicond_learned_conditioning(model, prompts, steps) -> MulticondLearnedConditioning:
     """same as get_learned_conditioning, but returns a list of ScheduledPromptConditioning along with the weight objects for each prompt.
     For each prompt, the list is obtained by splitting the prompt using the AND separator.
@@ -214,20 +242,57 @@ def get_multicond_learned_conditioning(model, prompts, steps) -> MulticondLearne
     return MulticondLearnedConditioning(shape=(len(prompts),), batch=res)
 
 
+class DictWithShape(dict):
+    def __init__(self, x, shape):
+        super().__init__()
+        self.update(x)
+
+    @property
+    def shape(self):
+        return self["crossattn"].shape
+
+
 def reconstruct_cond_batch(c: List[List[ScheduledPromptConditioning]], current_step):
     param = c[0][0].cond
-    res = torch.zeros((len(c),) + param.shape, device=param.device, dtype=param.dtype)
+    is_dict = isinstance(param, dict)
+
+    if is_dict:
+        dict_cond = param
+        res = {k: torch.zeros((len(c),) + param.shape, device=param.device, dtype=param.dtype) for k, param in dict_cond.items()}
+        res = DictWithShape(res, (len(c),) + dict_cond['crossattn'].shape)
+    else:
+        res = torch.zeros((len(c),) + param.shape, device=param.device, dtype=param.dtype)
+
     for i, cond_schedule in enumerate(c):
         target_index = 0
         for current, entry in enumerate(cond_schedule):
             if current_step <= entry.end_at_step:
                 target_index = current
                 break
-        res[i] = cond_schedule[target_index].cond
+
+        if is_dict:
+            for k, param in cond_schedule[target_index].cond.items():
+                res[k][i] = param
+        else:
+            res[i] = cond_schedule[target_index].cond
 
     return res
 
 
+def stack_conds(tensors):
+    # if prompts have wildly different lengths above the limit we'll get tensors of different shapes
+    # and won't be able to torch.stack them. So this fixes that.
+    token_count = max([x.shape[0] for x in tensors])
+    for i in range(len(tensors)):
+        if tensors[i].shape[0] != token_count:
+            last_vector = tensors[i][-1:]
+            last_vector_repeated = last_vector.repeat([token_count - tensors[i].shape[0], 1])
+            tensors[i] = torch.vstack([tensors[i], last_vector_repeated])
+
+    return torch.stack(tensors)
+
+
+
 def reconstruct_multicond_batch(c: MulticondLearnedConditioning, current_step):
     param = c.batch[0][0].schedules[0].cond
 
@@ -249,16 +314,14 @@ def reconstruct_multicond_batch(c: MulticondLearnedConditioning, current_step):
 
         conds_list.append(conds_for_batch)
 
-    # if prompts have wildly different lengths above the limit we'll get tensors fo different shapes
-    # and won't be able to torch.stack them. So this fixes that.
-    token_count = max([x.shape[0] for x in tensors])
-    for i in range(len(tensors)):
-        if tensors[i].shape[0] != token_count:
-            last_vector = tensors[i][-1:]
-            last_vector_repeated = last_vector.repeat([token_count - tensors[i].shape[0], 1])
-            tensors[i] = torch.vstack([tensors[i], last_vector_repeated])
+    if isinstance(tensors[0], dict):
+        keys = list(tensors[0].keys())
+        stacked = {k: stack_conds([x[k] for x in tensors]) for k in keys}
+        stacked = DictWithShape(stacked, stacked['crossattn'].shape)
+    else:
+        stacked = stack_conds(tensors).to(device=param.device, dtype=param.dtype)
 
-    return conds_list, torch.stack(tensors).to(device=param.device, dtype=param.dtype)
+    return conds_list, stacked
 
 
 re_attention = re.compile(r"""

modules/sd_hijack.py

@@ -15,6 +15,11 @@ import ldm.models.diffusion.ddim
 import ldm.models.diffusion.plms
 import ldm.modules.encoders.modules
 
+import sgm.modules.attention
+import sgm.modules.diffusionmodules.model
+import sgm.modules.diffusionmodules.openaimodel
+import sgm.modules.encoders.modules
+
 attention_CrossAttention_forward = ldm.modules.attention.CrossAttention.forward
 diffusionmodules_model_nonlinearity = ldm.modules.diffusionmodules.model.nonlinearity
 diffusionmodules_model_AttnBlock_forward = ldm.modules.diffusionmodules.model.AttnBlock.forward
@@ -56,6 +61,9 @@ def apply_optimizations(option=None):
     ldm.modules.diffusionmodules.model.nonlinearity = silu
     ldm.modules.diffusionmodules.openaimodel.th = sd_hijack_unet.th
 
+    sgm.modules.diffusionmodules.model.nonlinearity = silu
+    sgm.modules.diffusionmodules.openaimodel.th = sd_hijack_unet.th
+
     if current_optimizer is not None:
         current_optimizer.undo()
         current_optimizer = None
@@ -89,6 +97,10 @@ def undo_optimizations():
     ldm.modules.attention.CrossAttention.forward = hypernetwork.attention_CrossAttention_forward
     ldm.modules.diffusionmodules.model.AttnBlock.forward = diffusionmodules_model_AttnBlock_forward
 
+    sgm.modules.diffusionmodules.model.nonlinearity = diffusionmodules_model_nonlinearity
+    sgm.modules.attention.CrossAttention.forward = hypernetwork.attention_CrossAttention_forward
+    sgm.modules.diffusionmodules.model.AttnBlock.forward = diffusionmodules_model_AttnBlock_forward
+
 
 def fix_checkpoint():
     """checkpoints are now added and removed in embedding/hypernet code, since torch doesn't want
@@ -168,6 +180,32 @@ class StableDiffusionModelHijack:
             undo_optimizations()
 
     def hijack(self, m):
+        conditioner = getattr(m, 'conditioner', None)
+        if conditioner:
+            text_cond_models = []
+
+            for i in range(len(conditioner.embedders)):
+                embedder = conditioner.embedders[i]
+                typename = type(embedder).__name__
+                if typename == 'FrozenOpenCLIPEmbedder':
+                    embedder.model.token_embedding = EmbeddingsWithFixes(embedder.model.token_embedding, self)
+                    conditioner.embedders[i] = sd_hijack_open_clip.FrozenOpenCLIPEmbedderWithCustomWords(embedder, self)
+                    text_cond_models.append(conditioner.embedders[i])
+                if typename == 'FrozenCLIPEmbedder':
+                    model_embeddings = embedder.transformer.text_model.embeddings
+                    model_embeddings.token_embedding = EmbeddingsWithFixes(model_embeddings.token_embedding, self)
+                    conditioner.embedders[i] = sd_hijack_clip.FrozenCLIPEmbedderForSDXLWithCustomWords(embedder, self)
+                    text_cond_models.append(conditioner.embedders[i])
+                if typename == 'FrozenOpenCLIPEmbedder2':
+                    embedder.model.token_embedding = EmbeddingsWithFixes(embedder.model.token_embedding, self)
+                    conditioner.embedders[i] = sd_hijack_open_clip.FrozenOpenCLIPEmbedder2WithCustomWords(embedder, self)
+                    text_cond_models.append(conditioner.embedders[i])
+
+            if len(text_cond_models) == 1:
+                m.cond_stage_model = text_cond_models[0]
+            else:
+                m.cond_stage_model = conditioner
+
         if type(m.cond_stage_model) == xlmr.BertSeriesModelWithTransformation:
             model_embeddings = m.cond_stage_model.roberta.embeddings
             model_embeddings.token_embedding = EmbeddingsWithFixes(model_embeddings.word_embeddings, self)

modules/sd_hijack_clip.py

@@ -42,6 +42,10 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module):
         self.hijack: sd_hijack.StableDiffusionModelHijack = hijack
         self.chunk_length = 75
 
+        self.is_trainable = getattr(wrapped, 'is_trainable', False)
+        self.input_key = getattr(wrapped, 'input_key', 'txt')
+        self.legacy_ucg_val = None
+
     def empty_chunk(self):
         """creates an empty PromptChunk and returns it"""
 
@@ -199,8 +203,9 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module):
         """
         Accepts an array of texts; Passes texts through transformers network to create a tensor with numerical representation of those texts.
         Returns a tensor with shape of (B, T, C), where B is length of the array; T is length, in tokens, of texts (including padding) - T will
-        be a multiple of 77; and C is dimensionality of each token - for SD1 it's 768, and for SD2 it's 1024.
+        be a multiple of 77; and C is dimensionality of each token - for SD1 it's 768, for SD2 it's 1024, and for SDXL it's 1280.
         An example shape returned by this function can be: (2, 77, 768).
+        For SDXL, instead of returning one tensor avobe, it returns a tuple with two: the other one with shape (B, 1280) with pooled values.
         Webui usually sends just one text at a time through this function - the only time when texts is an array with more than one elemenet
         is when you do prompt editing: "a picture of a [cat:dog:0.4] eating ice cream"
         """
@@ -242,7 +247,10 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module):
             if hashes:
                 self.hijack.extra_generation_params["TI hashes"] = ", ".join(hashes)
 
-        return torch.hstack(zs)
+        if getattr(self.wrapped, 'return_pooled', False):
+            return torch.hstack(zs), zs[0].pooled
+        else:
+            return torch.hstack(zs)
 
     def process_tokens(self, remade_batch_tokens, batch_multipliers):
         """
@@ -265,9 +273,9 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module):
         # restoring original mean is likely not correct, but it seems to work well to prevent artifacts that happen otherwise
         batch_multipliers = torch.asarray(batch_multipliers).to(devices.device)
         original_mean = z.mean()
-        z = z * batch_multipliers.reshape(batch_multipliers.shape + (1,)).expand(z.shape)
+        z *= batch_multipliers.reshape(batch_multipliers.shape + (1,)).expand(z.shape)
         new_mean = z.mean()
-        z = z * (original_mean / new_mean)
+        z *= (original_mean / new_mean)
 
         return z
 
@@ -324,3 +332,18 @@ class FrozenCLIPEmbedderWithCustomWords(FrozenCLIPEmbedderWithCustomWordsBase):
         embedded = embedding_layer.token_embedding.wrapped(ids.to(embedding_layer.token_embedding.wrapped.weight.device)).squeeze(0)
 
         return embedded
+
+
+class FrozenCLIPEmbedderForSDXLWithCustomWords(FrozenCLIPEmbedderWithCustomWords):
+    def __init__(self, wrapped, hijack):
+        super().__init__(wrapped, hijack)
+
+    def encode_with_transformers(self, tokens):
+        outputs = self.wrapped.transformer(input_ids=tokens, output_hidden_states=self.wrapped.layer == "hidden")
+
+        if self.wrapped.layer == "last":
+            z = outputs.last_hidden_state
+        else:
+            z = outputs.hidden_states[self.wrapped.layer_idx]
+
+        return z

modules/sd_hijack_open_clip.py

@@ -32,6 +32,40 @@ class FrozenOpenCLIPEmbedderWithCustomWords(sd_hijack_clip.FrozenCLIPEmbedderWit
     def encode_embedding_init_text(self, init_text, nvpt):
         ids = tokenizer.encode(init_text)
         ids = torch.asarray([ids], device=devices.device, dtype=torch.int)
-        embedded = self.wrapped.model.token_embedding.wrapped(ids).squeeze(0)
+        embedded = self.wrapped.model.token_embedding.wrapped(ids.to(self.wrapped.model.token_embedding.wrapped.weight.device)).squeeze(0)
+
+        return embedded
+
+
+class FrozenOpenCLIPEmbedder2WithCustomWords(sd_hijack_clip.FrozenCLIPEmbedderWithCustomWordsBase):
+    def __init__(self, wrapped, hijack):
+        super().__init__(wrapped, hijack)
+
+        self.comma_token = [v for k, v in tokenizer.encoder.items() if k == ',</w>'][0]
+        self.id_start = tokenizer.encoder["<start_of_text>"]
+        self.id_end = tokenizer.encoder["<end_of_text>"]
+        self.id_pad = 0
+
+    def tokenize(self, texts):
+        assert not opts.use_old_emphasis_implementation, 'Old emphasis implementation not supported for Open Clip'
+
+        tokenized = [tokenizer.encode(text) for text in texts]
+
+        return tokenized
+
+    def encode_with_transformers(self, tokens):
+        d = self.wrapped.encode_with_transformer(tokens)
+        z = d[self.wrapped.layer]
+
+        pooled = d.get("pooled")
+        if pooled is not None:
+            z.pooled = pooled
+
+        return z
+
+    def encode_embedding_init_text(self, init_text, nvpt):
+        ids = tokenizer.encode(init_text)
+        ids = torch.asarray([ids], device=devices.device, dtype=torch.int)
+        embedded = self.wrapped.model.token_embedding.wrapped(ids.to(self.wrapped.model.token_embedding.wrapped.weight.device)).squeeze(0)
 
         return embedded

modules/sd_hijack_optimizations.py

@@ -14,7 +14,11 @@ from modules.hypernetworks import hypernetwork
 import ldm.modules.attention
 import ldm.modules.diffusionmodules.model
 
+import sgm.modules.attention
+import sgm.modules.diffusionmodules.model
+
 diffusionmodules_model_AttnBlock_forward = ldm.modules.diffusionmodules.model.AttnBlock.forward
+sgm_diffusionmodules_model_AttnBlock_forward = sgm.modules.diffusionmodules.model.AttnBlock.forward
 
 
 class SdOptimization:
@@ -39,6 +43,9 @@ class SdOptimization:
         ldm.modules.attention.CrossAttention.forward = hypernetwork.attention_CrossAttention_forward
         ldm.modules.diffusionmodules.model.AttnBlock.forward = diffusionmodules_model_AttnBlock_forward
 
+        sgm.modules.attention.CrossAttention.forward = hypernetwork.attention_CrossAttention_forward
+        sgm.modules.diffusionmodules.model.AttnBlock.forward = sgm_diffusionmodules_model_AttnBlock_forward
+
 
 class SdOptimizationXformers(SdOptimization):
     name = "xformers"
@@ -51,6 +58,8 @@ class SdOptimizationXformers(SdOptimization):
     def apply(self):
         ldm.modules.attention.CrossAttention.forward = xformers_attention_forward
         ldm.modules.diffusionmodules.model.AttnBlock.forward = xformers_attnblock_forward
+        sgm.modules.attention.CrossAttention.forward = xformers_attention_forward
+        sgm.modules.diffusionmodules.model.AttnBlock.forward = xformers_attnblock_forward
 
 
 class SdOptimizationSdpNoMem(SdOptimization):
@@ -65,6 +74,8 @@ class SdOptimizationSdpNoMem(SdOptimization):
     def apply(self):
         ldm.modules.attention.CrossAttention.forward = scaled_dot_product_no_mem_attention_forward
         ldm.modules.diffusionmodules.model.AttnBlock.forward = sdp_no_mem_attnblock_forward
+        sgm.modules.attention.CrossAttention.forward = scaled_dot_product_no_mem_attention_forward
+        sgm.modules.diffusionmodules.model.AttnBlock.forward = sdp_no_mem_attnblock_forward
 
 
 class SdOptimizationSdp(SdOptimizationSdpNoMem):
@@ -76,6 +87,8 @@ class SdOptimizationSdp(SdOptimizationSdpNoMem):
     def apply(self):
         ldm.modules.attention.CrossAttention.forward = scaled_dot_product_attention_forward
         ldm.modules.diffusionmodules.model.AttnBlock.forward = sdp_attnblock_forward
+        sgm.modules.attention.CrossAttention.forward = scaled_dot_product_attention_forward
+        sgm.modules.diffusionmodules.model.AttnBlock.forward = sdp_attnblock_forward
 
 
 class SdOptimizationSubQuad(SdOptimization):
@@ -86,6 +99,8 @@ class SdOptimizationSubQuad(SdOptimization):
     def apply(self):
         ldm.modules.attention.CrossAttention.forward = sub_quad_attention_forward
         ldm.modules.diffusionmodules.model.AttnBlock.forward = sub_quad_attnblock_forward
+        sgm.modules.attention.CrossAttention.forward = sub_quad_attention_forward
+        sgm.modules.diffusionmodules.model.AttnBlock.forward = sub_quad_attnblock_forward
 
 
 class SdOptimizationV1(SdOptimization):
@@ -94,9 +109,9 @@ class SdOptimizationV1(SdOptimization):
     cmd_opt = "opt_split_attention_v1"
     priority = 10
 
-
     def apply(self):
         ldm.modules.attention.CrossAttention.forward = split_cross_attention_forward_v1
+        sgm.modules.attention.CrossAttention.forward = split_cross_attention_forward_v1
 
 
 class SdOptimizationInvokeAI(SdOptimization):
@@ -109,6 +124,7 @@ class SdOptimizationInvokeAI(SdOptimization):
 
     def apply(self):
         ldm.modules.attention.CrossAttention.forward = split_cross_attention_forward_invokeAI
+        sgm.modules.attention.CrossAttention.forward = split_cross_attention_forward_invokeAI
 
 
 class SdOptimizationDoggettx(SdOptimization):
@@ -119,6 +135,8 @@ class SdOptimizationDoggettx(SdOptimization):
     def apply(self):
         ldm.modules.attention.CrossAttention.forward = split_cross_attention_forward
         ldm.modules.diffusionmodules.model.AttnBlock.forward = cross_attention_attnblock_forward
+        sgm.modules.attention.CrossAttention.forward = split_cross_attention_forward
+        sgm.modules.diffusionmodules.model.AttnBlock.forward = cross_attention_attnblock_forward
 
 
 def list_optimizers(res):
@@ -155,7 +173,7 @@ def get_available_vram():
 
 
 # see https://github.com/basujindal/stable-diffusion/pull/117 for discussion
-def split_cross_attention_forward_v1(self, x, context=None, mask=None):
+def split_cross_attention_forward_v1(self, x, context=None, mask=None, **kwargs):
     h = self.heads
 
     q_in = self.to_q(x)
@@ -196,7 +214,7 @@ def split_cross_attention_forward_v1(self, x, context=None, mask=None):
 
 
 # taken from https://github.com/Doggettx/stable-diffusion and modified
-def split_cross_attention_forward(self, x, context=None, mask=None):
+def split_cross_attention_forward(self, x, context=None, mask=None, **kwargs):
     h = self.heads
 
     q_in = self.to_q(x)
@@ -262,11 +280,13 @@ def split_cross_attention_forward(self, x, context=None, mask=None):
 # -- Taken from https://github.com/invoke-ai/InvokeAI and modified --
 mem_total_gb = psutil.virtual_memory().total // (1 << 30)
 
+
 def einsum_op_compvis(q, k, v):
     s = einsum('b i d, b j d -> b i j', q, k)
     s = s.softmax(dim=-1, dtype=s.dtype)
     return einsum('b i j, b j d -> b i d', s, v)
 
+
 def einsum_op_slice_0(q, k, v, slice_size):
     r = torch.zeros(q.shape[0], q.shape[1], v.shape[2], device=q.device, dtype=q.dtype)
     for i in range(0, q.shape[0], slice_size):
@@ -274,6 +294,7 @@ def einsum_op_slice_0(q, k, v, slice_size):
         r[i:end] = einsum_op_compvis(q[i:end], k[i:end], v[i:end])
     return r
 
+
 def einsum_op_slice_1(q, k, v, slice_size):
     r = torch.zeros(q.shape[0], q.shape[1], v.shape[2], device=q.device, dtype=q.dtype)
     for i in range(0, q.shape[1], slice_size):
@@ -281,6 +302,7 @@ def einsum_op_slice_1(q, k, v, slice_size):
         r[:, i:end] = einsum_op_compvis(q[:, i:end], k, v)
     return r
 
+
 def einsum_op_mps_v1(q, k, v):
     if q.shape[0] * q.shape[1] <= 2**16: # (512x512) max q.shape[1]: 4096
         return einsum_op_compvis(q, k, v)
@@ -290,12 +312,14 @@ def einsum_op_mps_v1(q, k, v):
             slice_size -= 1
         return einsum_op_slice_1(q, k, v, slice_size)
 
+
 def einsum_op_mps_v2(q, k, v):
     if mem_total_gb > 8 and q.shape[0] * q.shape[1] <= 2**16:
         return einsum_op_compvis(q, k, v)
     else:
         return einsum_op_slice_0(q, k, v, 1)
 
+
 def einsum_op_tensor_mem(q, k, v, max_tensor_mb):
     size_mb = q.shape[0] * q.shape[1] * k.shape[1] * q.element_size() // (1 << 20)
     if size_mb <= max_tensor_mb:
@@ -305,6 +329,7 @@ def einsum_op_tensor_mem(q, k, v, max_tensor_mb):
         return einsum_op_slice_0(q, k, v, q.shape[0] // div)
     return einsum_op_slice_1(q, k, v, max(q.shape[1] // div, 1))
 
+
 def einsum_op_cuda(q, k, v):
     stats = torch.cuda.memory_stats(q.device)
     mem_active = stats['active_bytes.all.current']
@@ -315,6 +340,7 @@ def einsum_op_cuda(q, k, v):
     # Divide factor of safety as there's copying and fragmentation
     return einsum_op_tensor_mem(q, k, v, mem_free_total / 3.3 / (1 << 20))
 
+
 def einsum_op(q, k, v):
     if q.device.type == 'cuda':
         return einsum_op_cuda(q, k, v)
@@ -328,7 +354,8 @@ def einsum_op(q, k, v):
     # Tested on i7 with 8MB L3 cache.
     return einsum_op_tensor_mem(q, k, v, 32)
 
-def split_cross_attention_forward_invokeAI(self, x, context=None, mask=None):
+
+def split_cross_attention_forward_invokeAI(self, x, context=None, mask=None, **kwargs):
     h = self.heads
 
     q = self.to_q(x)
@@ -356,7 +383,7 @@ def split_cross_attention_forward_invokeAI(self, x, context=None, mask=None):
 
 # Based on Birch-san's modified implementation of sub-quadratic attention from https://github.com/Birch-san/diffusers/pull/1
 # The sub_quad_attention_forward function is under the MIT License listed under Memory Efficient Attention in the Licenses section of the web UI interface
-def sub_quad_attention_forward(self, x, context=None, mask=None):
+def sub_quad_attention_forward(self, x, context=None, mask=None, **kwargs):
     assert mask is None, "attention-mask not currently implemented for SubQuadraticCrossAttnProcessor."
 
     h = self.heads
@@ -392,6 +419,7 @@ def sub_quad_attention_forward(self, x, context=None, mask=None):
 
     return x
 
+
 def sub_quad_attention(q, k, v, q_chunk_size=1024, kv_chunk_size=None, kv_chunk_size_min=None, chunk_threshold=None, use_checkpoint=True):
     bytes_per_token = torch.finfo(q.dtype).bits//8
     batch_x_heads, q_tokens, _ = q.shape
@@ -442,7 +470,7 @@ def get_xformers_flash_attention_op(q, k, v):
     return None
 
 
-def xformers_attention_forward(self, x, context=None, mask=None):
+def xformers_attention_forward(self, x, context=None, mask=None, **kwargs):
     h = self.heads
     q_in = self.to_q(x)
     context = default(context, x)
@@ -465,9 +493,10 @@ def xformers_attention_forward(self, x, context=None, mask=None):
     out = rearrange(out, 'b n h d -> b n (h d)', h=h)
     return self.to_out(out)
 
+
 # Based on Diffusers usage of scaled dot product attention from https://github.com/huggingface/diffusers/blob/c7da8fd23359a22d0df2741688b5b4f33c26df21/src/diffusers/models/cross_attention.py
 # The scaled_dot_product_attention_forward function contains parts of code under Apache-2.0 license listed under Scaled Dot Product Attention in the Licenses section of the web UI interface
-def scaled_dot_product_attention_forward(self, x, context=None, mask=None):
+def scaled_dot_product_attention_forward(self, x, context=None, mask=None, **kwargs):
     batch_size, sequence_length, inner_dim = x.shape
 
     if mask is not None:
@@ -507,10 +536,12 @@ def scaled_dot_product_attention_forward(self, x, context=None, mask=None):
     hidden_states = self.to_out[1](hidden_states)
     return hidden_states
 
-def scaled_dot_product_no_mem_attention_forward(self, x, context=None, mask=None):
+
+def scaled_dot_product_no_mem_attention_forward(self, x, context=None, mask=None, **kwargs):
     with torch.backends.cuda.sdp_kernel(enable_flash=True, enable_math=True, enable_mem_efficient=False):
         return scaled_dot_product_attention_forward(self, x, context, mask)
 
+
 def cross_attention_attnblock_forward(self, x):
         h_ = x
         h_ = self.norm(h_)
@@ -569,6 +600,7 @@ def cross_attention_attnblock_forward(self, x):
 
         return h3
 
+
 def xformers_attnblock_forward(self, x):
     try:
         h_ = x
@@ -592,6 +624,7 @@ def xformers_attnblock_forward(self, x):
     except NotImplementedError:
         return cross_attention_attnblock_forward(self, x)
 
+
 def sdp_attnblock_forward(self, x):
     h_ = x
     h_ = self.norm(h_)
@@ -612,10 +645,12 @@ def sdp_attnblock_forward(self, x):
     out = self.proj_out(out)
     return x + out
 
+
 def sdp_no_mem_attnblock_forward(self, x):
     with torch.backends.cuda.sdp_kernel(enable_flash=True, enable_math=True, enable_mem_efficient=False):
         return sdp_attnblock_forward(self, x)
 
+
 def sub_quad_attnblock_forward(self, x):
     h_ = x
     h_ = self.norm(h_)

modules/sd_models.py

@@ -14,7 +14,7 @@ import ldm.modules.midas as midas
 
 from ldm.util import instantiate_from_config
 
-from modules import paths, shared, modelloader, devices, script_callbacks, sd_vae, sd_disable_initialization, errors, hashes, sd_models_config, sd_unet
+from modules import paths, shared, modelloader, devices, script_callbacks, sd_vae, sd_disable_initialization, errors, hashes, sd_models_config, sd_unet, sd_models_xl
 from modules.sd_hijack_inpainting import do_inpainting_hijack
 from modules.timer import Timer
 import tomesd
@@ -289,6 +289,10 @@ def load_model_weights(model, checkpoint_info: CheckpointInfo, state_dict, timer
     if state_dict is None:
         state_dict = get_checkpoint_state_dict(checkpoint_info, timer)
 
+    model.is_sdxl = hasattr(model, 'conditioner')
+    if model.is_sdxl:
+        sd_models_xl.extend_sdxl(model)
+
     model.load_state_dict(state_dict, strict=False)
     del state_dict
     timer.record("apply weights to model")
@@ -334,7 +338,8 @@ def load_model_weights(model, checkpoint_info: CheckpointInfo, state_dict, timer
     model.sd_checkpoint_info = checkpoint_info
     shared.opts.data["sd_checkpoint_hash"] = checkpoint_info.sha256
 
-    model.logvar = model.logvar.to(devices.device)  # fix for training
+    if hasattr(model, 'logvar'):
+        model.logvar = model.logvar.to(devices.device)  # fix for training
 
     sd_vae.delete_base_vae()
     sd_vae.clear_loaded_vae()
@@ -391,10 +396,11 @@ def repair_config(sd_config):
     if not hasattr(sd_config.model.params, "use_ema"):
         sd_config.model.params.use_ema = False
 
-    if shared.cmd_opts.no_half:
-        sd_config.model.params.unet_config.params.use_fp16 = False
-    elif shared.cmd_opts.upcast_sampling:
-        sd_config.model.params.unet_config.params.use_fp16 = True
+    if hasattr(sd_config.model.params, 'unet_config'):
+        if shared.cmd_opts.no_half:
+            sd_config.model.params.unet_config.params.use_fp16 = False
+        elif shared.cmd_opts.upcast_sampling:
+            sd_config.model.params.unet_config.params.use_fp16 = True
 
     if getattr(sd_config.model.params.first_stage_config.params.ddconfig, "attn_type", None) == "vanilla-xformers" and not shared.xformers_available:
         sd_config.model.params.first_stage_config.params.ddconfig.attn_type = "vanilla"
@@ -407,6 +413,8 @@ def repair_config(sd_config):
 
 sd1_clip_weight = 'cond_stage_model.transformer.text_model.embeddings.token_embedding.weight'
 sd2_clip_weight = 'cond_stage_model.model.transformer.resblocks.0.attn.in_proj_weight'
+sdxl_clip_weight = 'conditioner.embedders.1.model.ln_final.weight'
+sdxl_refiner_clip_weight = 'conditioner.embedders.0.model.ln_final.weight'
 
 
 class SdModelData:
@@ -441,6 +449,15 @@ class SdModelData:
 model_data = SdModelData()
 
 
+def get_empty_cond(sd_model):
+    if hasattr(sd_model, 'conditioner'):
+        d = sd_model.get_learned_conditioning([""])
+        return d['crossattn']
+    else:
+        return sd_model.cond_stage_model([""])
+
+
+
 def load_model(checkpoint_info=None, already_loaded_state_dict=None):
     from modules import lowvram, sd_hijack
     checkpoint_info = checkpoint_info or select_checkpoint()
@@ -461,7 +478,7 @@ def load_model(checkpoint_info=None, already_loaded_state_dict=None):
         state_dict = get_checkpoint_state_dict(checkpoint_info, timer)
 
     checkpoint_config = sd_models_config.find_checkpoint_config(state_dict, checkpoint_info)
-    clip_is_included_into_sd = sd1_clip_weight in state_dict or sd2_clip_weight in state_dict
+    clip_is_included_into_sd = any(x for x in [sd1_clip_weight, sd2_clip_weight, sdxl_clip_weight, sdxl_refiner_clip_weight] if x in state_dict)
 
     timer.record("find config")
 
@@ -513,7 +530,7 @@ def load_model(checkpoint_info=None, already_loaded_state_dict=None):
     timer.record("scripts callbacks")
 
     with devices.autocast(), torch.no_grad():
-        sd_model.cond_stage_model_empty_prompt = sd_model.cond_stage_model([""])
+        sd_model.cond_stage_model_empty_prompt = get_empty_cond(sd_model)
 
     timer.record("calculate empty prompt")
 

modules/sd_models_config.py

@@ -6,12 +6,15 @@ from modules import shared, paths, sd_disable_initialization
 
 sd_configs_path = shared.sd_configs_path
 sd_repo_configs_path = os.path.join(paths.paths['Stable Diffusion'], "configs", "stable-diffusion")
+sd_xl_repo_configs_path = os.path.join(paths.paths['Stable Diffusion XL'], "configs", "inference")
 
 
 config_default = shared.sd_default_config
 config_sd2 = os.path.join(sd_repo_configs_path, "v2-inference.yaml")
 config_sd2v = os.path.join(sd_repo_configs_path, "v2-inference-v.yaml")
 config_sd2_inpainting = os.path.join(sd_repo_configs_path, "v2-inpainting-inference.yaml")
+config_sdxl = os.path.join(sd_xl_repo_configs_path, "sd_xl_base.yaml")
+config_sdxl_refiner = os.path.join(sd_xl_repo_configs_path, "sd_xl_refiner.yaml")
 config_depth_model = os.path.join(sd_repo_configs_path, "v2-midas-inference.yaml")
 config_unclip = os.path.join(sd_repo_configs_path, "v2-1-stable-unclip-l-inference.yaml")
 config_unopenclip = os.path.join(sd_repo_configs_path, "v2-1-stable-unclip-h-inference.yaml")
@@ -68,7 +71,11 @@ def guess_model_config_from_state_dict(sd, filename):
     diffusion_model_input = sd.get('model.diffusion_model.input_blocks.0.0.weight', None)
     sd2_variations_weight = sd.get('embedder.model.ln_final.weight', None)
 
-    if sd.get('depth_model.model.pretrained.act_postprocess3.0.project.0.bias', None) is not None:
+    if sd.get('conditioner.embedders.1.model.ln_final.weight', None) is not None:
+        return config_sdxl
+    if sd.get('conditioner.embedders.0.model.ln_final.weight', None) is not None:
+        return config_sdxl_refiner
+    elif sd.get('depth_model.model.pretrained.act_postprocess3.0.project.0.bias', None) is not None:
         return config_depth_model
     elif sd2_variations_weight is not None and sd2_variations_weight.shape[0] == 768:
         return config_unclip

modules/sd_models_xl.py

+from __future__ import annotations
+
+import torch
+
+import sgm.models.diffusion
+import sgm.modules.diffusionmodules.denoiser_scaling
+import sgm.modules.diffusionmodules.discretizer
+from modules import devices, shared, prompt_parser
+
+
+def get_learned_conditioning(self: sgm.models.diffusion.DiffusionEngine, batch: prompt_parser.SdConditioning | list[str]):
+    for embedder in self.conditioner.embedders:
+        embedder.ucg_rate = 0.0
+
+    width = getattr(self, 'target_width', 1024)
+    height = getattr(self, 'target_height', 1024)
+    is_negative_prompt = getattr(batch, 'is_negative_prompt', False)
+    aesthetic_score = shared.opts.sdxl_refiner_low_aesthetic_score if is_negative_prompt else shared.opts.sdxl_refiner_high_aesthetic_score
+
+    devices_args = dict(device=devices.device, dtype=devices.dtype)
+
+    sdxl_conds = {
+        "txt": batch,
+        "original_size_as_tuple": torch.tensor([height, width], **devices_args).repeat(len(batch), 1),
+        "crop_coords_top_left": torch.tensor([shared.opts.sdxl_crop_top, shared.opts.sdxl_crop_left], **devices_args).repeat(len(batch), 1),
+        "target_size_as_tuple": torch.tensor([height, width], **devices_args).repeat(len(batch), 1),
+        "aesthetic_score": torch.tensor([aesthetic_score], **devices_args).repeat(len(batch), 1),
+    }
+
+    force_zero_negative_prompt = is_negative_prompt and all(x == '' for x in batch)
+    c = self.conditioner(sdxl_conds, force_zero_embeddings=['txt'] if force_zero_negative_prompt else [])
+
+    return c
+
+
+def apply_model(self: sgm.models.diffusion.DiffusionEngine, x, t, cond):
+    return self.model(x, t, cond)
+
+
+def get_first_stage_encoding(self, x):  # SDXL's encode_first_stage does everything so get_first_stage_encoding is just there for compatibility
+    return x
+
+
+sgm.models.diffusion.DiffusionEngine.get_learned_conditioning = get_learned_conditioning
+sgm.models.diffusion.DiffusionEngine.apply_model = apply_model
+sgm.models.diffusion.DiffusionEngine.get_first_stage_encoding = get_first_stage_encoding
+
+
+def encode_embedding_init_text(self: sgm.modules.GeneralConditioner, init_text, nvpt):
+    res = []
+
+    for embedder in [embedder for embedder in self.embedders if hasattr(embedder, 'encode_embedding_init_text')]:
+        encoded = embedder.encode_embedding_init_text(init_text, nvpt)
+        res.append(encoded)
+
+    return torch.cat(res, dim=1)
+
+
+def process_texts(self, texts):
+    for embedder in [embedder for embedder in self.embedders if hasattr(embedder, 'process_texts')]:
+        return embedder.process_texts(texts)
+
+
+def get_target_prompt_token_count(self, token_count):
+    for embedder in [embedder for embedder in self.embedders if hasattr(embedder, 'get_target_prompt_token_count')]:
+        return embedder.get_target_prompt_token_count(token_count)
+
+
+# those additions to GeneralConditioner make it possible to use it as model.cond_stage_model from SD1.5 in exist
+sgm.modules.GeneralConditioner.encode_embedding_init_text = encode_embedding_init_text
+sgm.modules.GeneralConditioner.process_texts = process_texts
+sgm.modules.GeneralConditioner.get_target_prompt_token_count = get_target_prompt_token_count
+
+
+def extend_sdxl(model):
+    """this adds a bunch of parameters to make SDXL model look a bit more like SD1.5 to the rest of the codebase."""
+
+    dtype = next(model.model.diffusion_model.parameters()).dtype
+    model.model.diffusion_model.dtype = dtype
+    model.model.conditioning_key = 'crossattn'
+    model.cond_stage_key = 'txt'
+    # model.cond_stage_model will be set in sd_hijack
+
+    model.parameterization = "v" if isinstance(model.denoiser.scaling, sgm.modules.diffusionmodules.denoiser_scaling.VScaling) else "eps"
+
+    discretization = sgm.modules.diffusionmodules.discretizer.LegacyDDPMDiscretization()
+    model.alphas_cumprod = torch.asarray(discretization.alphas_cumprod, device=devices.device, dtype=dtype)
+
+    model.conditioner.wrapped = torch.nn.Module()
+
+
+sgm.modules.attention.print = lambda *args: None
+sgm.modules.diffusionmodules.model.print = lambda *args: None
+sgm.modules.diffusionmodules.openaimodel.print = lambda *args: None
+sgm.modules.encoders.modules.print = lambda *args: None
+
+# this gets the code to load the vanilla attention that we override
+sgm.modules.attention.SDP_IS_AVAILABLE = True
+sgm.modules.attention.XFORMERS_IS_AVAILABLE = False

modules/sd_samplers.py

@@ -28,6 +28,9 @@ def create_sampler(name, model):
 
     assert config is not None, f'bad sampler name: {name}'
 
+    if model.is_sdxl and config.options.get("no_sdxl", False):
+        raise Exception(f"Sampler {config.name} is not supported for SDXL")
+
     sampler = config.constructor(model)
     sampler.config = config
 

modules/sd_samplers_compvis.py

@@ -11,9 +11,9 @@ import modules.models.diffusion.uni_pc
 
 
 samplers_data_compvis = [
-    sd_samplers_common.SamplerData('DDIM', lambda model: VanillaStableDiffusionSampler(ldm.models.diffusion.ddim.DDIMSampler, model), [], {"default_eta_is_0": True, "uses_ensd": True}),
-    sd_samplers_common.SamplerData('PLMS', lambda model: VanillaStableDiffusionSampler(ldm.models.diffusion.plms.PLMSSampler, model), [], {}),
-    sd_samplers_common.SamplerData('UniPC', lambda model: VanillaStableDiffusionSampler(modules.models.diffusion.uni_pc.UniPCSampler, model), [], {}),
+    sd_samplers_common.SamplerData('DDIM', lambda model: VanillaStableDiffusionSampler(ldm.models.diffusion.ddim.DDIMSampler, model), [], {"default_eta_is_0": True, "uses_ensd": True, "no_sdxl": True}),
+    sd_samplers_common.SamplerData('PLMS', lambda model: VanillaStableDiffusionSampler(ldm.models.diffusion.plms.PLMSSampler, model), [], {"no_sdxl": True}),
+    sd_samplers_common.SamplerData('UniPC', lambda model: VanillaStableDiffusionSampler(modules.models.diffusion.uni_pc.UniPCSampler, model), [], {"no_sdxl": True}),
 ]
 
 

modules/sd_samplers_kdiffusion.py

@@ -53,6 +53,28 @@ k_diffusion_scheduler = {
 }
 
 
+def catenate_conds(conds):
+    if not isinstance(conds[0], dict):
+        return torch.cat(conds)
+
+    return {key: torch.cat([x[key] for x in conds]) for key in conds[0].keys()}
+
+
+def subscript_cond(cond, a, b):
+    if not isinstance(cond, dict):
+        return cond[a:b]
+
+    return {key: vec[a:b] for key, vec in cond.items()}
+
+
+def pad_cond(tensor, repeats, empty):
+    if not isinstance(tensor, dict):
+        return torch.cat([tensor, empty.repeat((tensor.shape[0], repeats, 1))], axis=1)
+
+    tensor['crossattn'] = pad_cond(tensor['crossattn'], repeats, empty)
+    return tensor
+
+
 class CFGDenoiser(torch.nn.Module):
     """
     Classifier free guidance denoiser. A wrapper for stable diffusion model (specifically for unet)
@@ -105,10 +127,13 @@ class CFGDenoiser(torch.nn.Module):
 
         if shared.sd_model.model.conditioning_key == "crossattn-adm":
             image_uncond = torch.zeros_like(image_cond)
-            make_condition_dict = lambda c_crossattn, c_adm: {"c_crossattn": c_crossattn, "c_adm": c_adm}
+            make_condition_dict = lambda c_crossattn, c_adm: {"c_crossattn": [c_crossattn], "c_adm": c_adm}
         else:
             image_uncond = image_cond
-            make_condition_dict = lambda c_crossattn, c_concat: {"c_crossattn": c_crossattn, "c_concat": [c_concat]}
+            if isinstance(uncond, dict):
+                make_condition_dict = lambda c_crossattn, c_concat: {**c_crossattn, "c_concat": [c_concat]}
+            else:
+                make_condition_dict = lambda c_crossattn, c_concat: {"c_crossattn": [c_crossattn], "c_concat": [c_concat]}
 
         if not is_edit_model:
             x_in = torch.cat([torch.stack([x[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [x])
@@ -140,28 +165,28 @@ class CFGDenoiser(torch.nn.Module):
             num_repeats = (tensor.shape[1] - uncond.shape[1]) // empty.shape[1]
 
             if num_repeats < 0:
-                tensor = torch.cat([tensor, empty.repeat((tensor.shape[0], -num_repeats, 1))], axis=1)
+                tensor = pad_cond(tensor, -num_repeats, empty)
                 self.padded_cond_uncond = True
             elif num_repeats > 0:
-                uncond = torch.cat([uncond, empty.repeat((uncond.shape[0], num_repeats, 1))], axis=1)
+                uncond = pad_cond(uncond, num_repeats, empty)
                 self.padded_cond_uncond = True
 
         if tensor.shape[1] == uncond.shape[1] or skip_uncond:
             if is_edit_model:
-                cond_in = torch.cat([tensor, uncond, uncond])
+                cond_in = catenate_conds([tensor, uncond, uncond])
             elif skip_uncond:
                 cond_in = tensor
             else:
-                cond_in = torch.cat([tensor, uncond])
+                cond_in = catenate_conds([tensor, uncond])
 
             if shared.batch_cond_uncond:
-                x_out = self.inner_model(x_in, sigma_in, cond=make_condition_dict([cond_in], image_cond_in))
+                x_out = self.inner_model(x_in, sigma_in, cond=make_condition_dict(cond_in, image_cond_in))
             else:
                 x_out = torch.zeros_like(x_in)
                 for batch_offset in range(0, x_out.shape[0], batch_size):
                     a = batch_offset
                     b = a + batch_size
-                    x_out[a:b] = self.inner_model(x_in[a:b], sigma_in[a:b], cond=make_condition_dict([cond_in[a:b]], image_cond_in[a:b]))
+                    x_out[a:b] = self.inner_model(x_in[a:b], sigma_in[a:b], cond=make_condition_dict(subscript_cond(cond_in, a, b), image_cond_in[a:b]))
         else:
             x_out = torch.zeros_like(x_in)
             batch_size = batch_size*2 if shared.batch_cond_uncond else batch_size
@@ -170,14 +195,14 @@ class CFGDenoiser(torch.nn.Module):
                 b = min(a + batch_size, tensor.shape[0])
 
                 if not is_edit_model:
-                    c_crossattn = [tensor[a:b]]
+                    c_crossattn = subscript_cond(tensor, a, b)
                 else:
                     c_crossattn = torch.cat([tensor[a:b]], uncond)
 
                 x_out[a:b] = self.inner_model(x_in[a:b], sigma_in[a:b], cond=make_condition_dict(c_crossattn, image_cond_in[a:b]))
 
             if not skip_uncond:
-                x_out[-uncond.shape[0]:] = self.inner_model(x_in[-uncond.shape[0]:], sigma_in[-uncond.shape[0]:], cond=make_condition_dict([uncond], image_cond_in[-uncond.shape[0]:]))
+                x_out[-uncond.shape[0]:] = self.inner_model(x_in[-uncond.shape[0]:], sigma_in[-uncond.shape[0]:], cond=make_condition_dict(uncond, image_cond_in[-uncond.shape[0]:]))
 
         denoised_image_indexes = [x[0][0] for x in conds_list]
         if skip_uncond:

modules/sd_vae_approx.py

@@ -2,9 +2,9 @@ import os
 
 import torch
 from torch import nn
-from modules import devices, paths
+from modules import devices, paths, shared
 
-sd_vae_approx_model = None
+sd_vae_approx_models = {}
 
 
 class VAEApprox(nn.Module):
@@ -31,30 +31,55 @@ class VAEApprox(nn.Module):
         return x
 
 
+def download_model(model_path, model_url):
+    if not os.path.exists(model_path):
+        os.makedirs(os.path.dirname(model_path), exist_ok=True)
+
+        print(f'Downloading VAEApprox model to: {model_path}')
+        torch.hub.download_url_to_file(model_url, model_path)
+
+
 def model():
-    global sd_vae_approx_model
+    model_name = "vaeapprox-sdxl.pt" if getattr(shared.sd_model, 'is_sdxl', False) else "model.pt"
+    loaded_model = sd_vae_approx_models.get(model_name)
 
-    if sd_vae_approx_model is None:
-        model_path = os.path.join(paths.models_path, "VAE-approx", "model.pt")
-        sd_vae_approx_model = VAEApprox()
+    if loaded_model is None:
+        model_path = os.path.join(paths.models_path, "VAE-approx", model_name)
         if not os.path.exists(model_path):
-            model_path = os.path.join(paths.script_path, "models", "VAE-approx", "model.pt")
-        sd_vae_approx_model.load_state_dict(torch.load(model_path, map_location='cpu' if devices.device.type != 'cuda' else None))
-        sd_vae_approx_model.eval()
-        sd_vae_approx_model.to(devices.device, devices.dtype)
+            model_path = os.path.join(paths.script_path, "models", "VAE-approx", model_name)
+
+        if not os.path.exists(model_path):
+            model_path = os.path.join(paths.models_path, "VAE-approx", model_name)
+            download_model(model_path, 'https://github.com/AUTOMATIC1111/stable-diffusion-webui/releases/download/v1.0.0-pre/' + model_name)
+
+        loaded_model = VAEApprox()
+        loaded_model.load_state_dict(torch.load(model_path, map_location='cpu' if devices.device.type != 'cuda' else None))
+        loaded_model.eval()
+        loaded_model.to(devices.device, devices.dtype)
+        sd_vae_approx_models[model_name] = loaded_model
 
-    return sd_vae_approx_model
+    return loaded_model
 
 
 def cheap_approximation(sample):
     # https://discuss.huggingface.co/t/decoding-latents-to-rgb-without-upscaling/23204/2
 
-    coefs = torch.tensor([
-        [0.298, 0.207, 0.208],
-        [0.187, 0.286, 0.173],
-        [-0.158, 0.189, 0.264],
-        [-0.184, -0.271, -0.473],
-    ]).to(sample.device)
+    if shared.sd_model.is_sdxl:
+        coeffs = [
+            [ 0.3448,  0.4168,  0.4395],
+            [-0.1953, -0.0290,  0.0250],
+            [ 0.1074,  0.0886, -0.0163],
+            [-0.3730, -0.2499, -0.2088],
+        ]
+    else:
+        coeffs = [
+            [ 0.298,  0.207,  0.208],
+            [ 0.187,  0.286,  0.173],
+            [-0.158,  0.189,  0.264],
+            [-0.184, -0.271, -0.473],
+        ]
+
+    coefs = torch.tensor(coeffs).to(sample.device)
 
     x_sample = torch.einsum("lxy,lr -> rxy", sample, coefs)
 

modules/sd_vae_taesd.py

@@ -8,9 +8,9 @@ import os
 import torch
 import torch.nn as nn
 
-from modules import devices, paths_internal
+from modules import devices, paths_internal, shared
 
-sd_vae_taesd = None
+sd_vae_taesd_models = {}
 
 
 def conv(n_in, n_out, **kwargs):
@@ -61,9 +61,7 @@ class TAESD(nn.Module):
         return x.sub(TAESD.latent_shift).mul(2 * TAESD.latent_magnitude)
 
 
-def download_model(model_path):
-    model_url = 'https://github.com/madebyollin/taesd/raw/main/taesd_decoder.pth'
-
+def download_model(model_path, model_url):
     if not os.path.exists(model_path):
         os.makedirs(os.path.dirname(model_path), exist_ok=True)
 
@@ -72,17 +70,19 @@ def download_model(model_path):
 
 
 def model():
-    global sd_vae_taesd
+    model_name = "taesdxl_decoder.pth" if getattr(shared.sd_model, 'is_sdxl', False) else "taesd_decoder.pth"
+    loaded_model = sd_vae_taesd_models.get(model_name)
 
-    if sd_vae_taesd is None:
-        model_path = os.path.join(paths_internal.models_path, "VAE-taesd", "taesd_decoder.pth")
-        download_model(model_path)
+    if loaded_model is None:
+        model_path = os.path.join(paths_internal.models_path, "VAE-taesd", model_name)
+        download_model(model_path, 'https://github.com/madebyollin/taesd/raw/main/' + model_name)
 
         if os.path.exists(model_path):
-            sd_vae_taesd = TAESD(model_path)
-            sd_vae_taesd.eval()
-            sd_vae_taesd.to(devices.device, devices.dtype)
+            loaded_model = TAESD(model_path)
+            loaded_model.eval()
+            loaded_model.to(devices.device, devices.dtype)
+            sd_vae_taesd_models[model_name] = loaded_model
         else:
             raise FileNotFoundError('TAESD model not found')
 
-    return sd_vae_taesd.decoder
+    return loaded_model.decoder

modules/shared.py

@@ -429,9 +429,16 @@ options_templates.update(options_section(('sd', "Stable Diffusion"), {
     "randn_source": OptionInfo("GPU", "Random number generator source.", gr.Radio, {"choices": ["GPU", "CPU"]}).info("changes seeds drastically; use CPU to produce the same picture across different videocard vendors"),
 }))
 
+options_templates.update(options_section(('sdxl', "Stable Diffusion XL"), {
+    "sdxl_crop_top": OptionInfo(0, "crop top coordinate"),
+    "sdxl_crop_left": OptionInfo(0, "crop left coordinate"),
+    "sdxl_refiner_low_aesthetic_score": OptionInfo(2.5, "SDXL low aesthetic score", gr.Number).info("used for refiner model negative prompt"),
+    "sdxl_refiner_high_aesthetic_score": OptionInfo(6.0, "SDXL high aesthetic score", gr.Number).info("used for refiner model prompt"),
+}))
+
 options_templates.update(options_section(('optimizations', "Optimizations"), {
     "cross_attention_optimization": OptionInfo("Automatic", "Cross attention optimization", gr.Dropdown, lambda: {"choices": shared_items.cross_attention_optimizations()}),
-    "s_min_uncond": OptionInfo(0.0, "Negative Guidance minimum sigma", gr.Slider, {"minimum": 0.0, "maximum": 4.0, "step": 0.01}).link("PR", "https://github.com/AUTOMATIC1111/stable-diffusion-webui/pull/9177").info("skip negative prompt for some steps when the image is almost ready; 0=disable, higher=faster"),
+    "s_min_uncond": OptionInfo(0.0, "Negative Guidance minimum sigma", gr.Slider, {"minimum": 0.0, "maximum": 15.0, "step": 0.01}).link("PR", "https://github.com/AUTOMATIC1111/stable-diffusion-webui/pull/9177").info("skip negative prompt for some steps when the image is almost ready; 0=disable, higher=faster"),
     "token_merging_ratio": OptionInfo(0.0, "Token merging ratio", gr.Slider, {"minimum": 0.0, "maximum": 0.9, "step": 0.1}).link("PR", "https://github.com/AUTOMATIC1111/stable-diffusion-webui/pull/9256").info("0=disable, higher=faster"),
     "token_merging_ratio_img2img": OptionInfo(0.0, "Token merging ratio for img2img", gr.Slider, {"minimum": 0.0, "maximum": 0.9, "step": 0.1}).info("only applies if non-zero and overrides above"),
     "token_merging_ratio_hr": OptionInfo(0.0, "Token merging ratio for high-res pass", gr.Slider, {"minimum": 0.0, "maximum": 0.9, "step": 0.1}).info("only applies if non-zero and overrides above"),

requirements.txt

@@ -14,6 +14,7 @@ kornia
 lark
 numpy
 omegaconf
+open-clip-torch
 
 piexif
 psutil

requirements_versions.txt

@@ -15,6 +15,7 @@ kornia==0.6.7
 lark==1.1.2
 numpy==1.23.5
 omegaconf==2.2.3
+open-clip-torch==2.20.0
 piexif==1.1.3
 psutil~=5.9.5
 pytorch_lightning==1.9.4