ComfyUI/comfy/sd.py

1168 lines
46 KiB
Python
Raw Normal View History

2023-01-03 06:53:32 +00:00
import torch
2023-02-17 20:31:38 +00:00
import contextlib
import copy
import inspect
2023-01-03 06:53:32 +00:00
2023-05-04 22:07:41 +00:00
from . import sd1_clip
from . import sd2_clip
from comfy import model_management
2023-03-07 16:00:35 +00:00
from .ldm.util import instantiate_from_config
from .ldm.models.autoencoder import AutoencoderKL
import yaml
2023-02-16 15:38:08 +00:00
from .cldm import cldm
from .t2i_adapter import adapter
2023-02-16 15:38:08 +00:00
from . import utils
from . import clip_vision
2023-04-19 13:36:19 +00:00
from . import gligen
2023-05-28 06:02:09 +00:00
from . import diffusers_convert
2023-06-09 16:24:24 +00:00
from . import model_base
def load_model_weights(model, sd, verbose=False, load_state_dict_to=[]):
2023-06-09 16:24:24 +00:00
replace_prefix = {"model.diffusion_model.": "diffusion_model."}
for rp in replace_prefix:
replace = list(map(lambda a: (a, "{}{}".format(replace_prefix[rp], a[len(rp):])), filter(lambda a: a.startswith(rp), sd.keys())))
for x in replace:
sd[x[1]] = sd.pop(x[0])
2023-01-03 06:53:32 +00:00
m, u = model.load_state_dict(sd, strict=False)
k = list(sd.keys())
for x in k:
# print(x)
if x.startswith("cond_stage_model.transformer.") and not x.startswith("cond_stage_model.transformer.text_model."):
y = x.replace("cond_stage_model.transformer.", "cond_stage_model.transformer.text_model.")
sd[y] = sd.pop(x)
2023-01-28 07:14:22 +00:00
if 'cond_stage_model.transformer.text_model.embeddings.position_ids' in sd:
ids = sd['cond_stage_model.transformer.text_model.embeddings.position_ids']
if ids.dtype == torch.float32:
sd['cond_stage_model.transformer.text_model.embeddings.position_ids'] = ids.round()
sd = utils.transformers_convert(sd, "cond_stage_model.model", "cond_stage_model.transformer.text_model", 24)
2023-01-03 06:53:32 +00:00
for x in load_state_dict_to:
x.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)
model.eval()
return model
LORA_CLIP_MAP = {
"mlp.fc1": "mlp_fc1",
"mlp.fc2": "mlp_fc2",
"self_attn.k_proj": "self_attn_k_proj",
"self_attn.q_proj": "self_attn_q_proj",
"self_attn.v_proj": "self_attn_v_proj",
"self_attn.out_proj": "self_attn_out_proj",
}
2023-03-10 02:41:24 +00:00
LORA_UNET_MAP_ATTENTIONS = {
"proj_in": "proj_in",
"proj_out": "proj_out",
"transformer_blocks.0.attn1.to_q": "transformer_blocks_0_attn1_to_q",
"transformer_blocks.0.attn1.to_k": "transformer_blocks_0_attn1_to_k",
"transformer_blocks.0.attn1.to_v": "transformer_blocks_0_attn1_to_v",
"transformer_blocks.0.attn1.to_out.0": "transformer_blocks_0_attn1_to_out_0",
"transformer_blocks.0.attn2.to_q": "transformer_blocks_0_attn2_to_q",
"transformer_blocks.0.attn2.to_k": "transformer_blocks_0_attn2_to_k",
"transformer_blocks.0.attn2.to_v": "transformer_blocks_0_attn2_to_v",
"transformer_blocks.0.attn2.to_out.0": "transformer_blocks_0_attn2_to_out_0",
"transformer_blocks.0.ff.net.0.proj": "transformer_blocks_0_ff_net_0_proj",
"transformer_blocks.0.ff.net.2": "transformer_blocks_0_ff_net_2",
}
2023-03-10 02:41:24 +00:00
LORA_UNET_MAP_RESNET = {
"in_layers.2": "resnets_{}_conv1",
"emb_layers.1": "resnets_{}_time_emb_proj",
"out_layers.3": "resnets_{}_conv2",
"skip_connection": "resnets_{}_conv_shortcut"
}
def load_lora(path, to_load):
2023-06-13 14:12:03 +00:00
lora = utils.load_torch_file(path, safe_load=True)
patch_dict = {}
loaded_keys = set()
for x in to_load:
2023-03-23 07:40:12 +00:00
alpha_name = "{}.alpha".format(x)
alpha = None
if alpha_name in lora.keys():
alpha = lora[alpha_name].item()
loaded_keys.add(alpha_name)
A_name = "{}.lora_up.weight".format(x)
B_name = "{}.lora_down.weight".format(x)
2023-03-21 18:51:51 +00:00
mid_name = "{}.lora_mid.weight".format(x)
2023-03-23 07:40:12 +00:00
if A_name in lora.keys():
2023-03-21 18:51:51 +00:00
mid = None
if mid_name in lora.keys():
mid = lora[mid_name]
loaded_keys.add(mid_name)
patch_dict[to_load[x]] = (lora[A_name], lora[B_name], alpha, mid)
loaded_keys.add(A_name)
loaded_keys.add(B_name)
2023-03-23 07:40:12 +00:00
2023-05-01 22:11:58 +00:00
######## loha
2023-03-23 07:40:12 +00:00
hada_w1_a_name = "{}.hada_w1_a".format(x)
hada_w1_b_name = "{}.hada_w1_b".format(x)
hada_w2_a_name = "{}.hada_w2_a".format(x)
hada_w2_b_name = "{}.hada_w2_b".format(x)
hada_t1_name = "{}.hada_t1".format(x)
hada_t2_name = "{}.hada_t2".format(x)
2023-03-23 07:40:12 +00:00
if hada_w1_a_name in lora.keys():
hada_t1 = None
hada_t2 = None
if hada_t1_name in lora.keys():
hada_t1 = lora[hada_t1_name]
hada_t2 = lora[hada_t2_name]
loaded_keys.add(hada_t1_name)
loaded_keys.add(hada_t2_name)
patch_dict[to_load[x]] = (lora[hada_w1_a_name], lora[hada_w1_b_name], alpha, lora[hada_w2_a_name], lora[hada_w2_b_name], hada_t1, hada_t2)
2023-03-23 07:40:12 +00:00
loaded_keys.add(hada_w1_a_name)
loaded_keys.add(hada_w1_b_name)
loaded_keys.add(hada_w2_a_name)
loaded_keys.add(hada_w2_b_name)
2023-05-01 22:11:58 +00:00
######## lokr
lokr_w1_name = "{}.lokr_w1".format(x)
lokr_w2_name = "{}.lokr_w2".format(x)
lokr_w1_a_name = "{}.lokr_w1_a".format(x)
lokr_w1_b_name = "{}.lokr_w1_b".format(x)
lokr_t2_name = "{}.lokr_t2".format(x)
lokr_w2_a_name = "{}.lokr_w2_a".format(x)
lokr_w2_b_name = "{}.lokr_w2_b".format(x)
lokr_w1 = None
if lokr_w1_name in lora.keys():
lokr_w1 = lora[lokr_w1_name]
loaded_keys.add(lokr_w1_name)
lokr_w2 = None
if lokr_w2_name in lora.keys():
lokr_w2 = lora[lokr_w2_name]
loaded_keys.add(lokr_w2_name)
lokr_w1_a = None
if lokr_w1_a_name in lora.keys():
lokr_w1_a = lora[lokr_w1_a_name]
loaded_keys.add(lokr_w1_a_name)
lokr_w1_b = None
if lokr_w1_b_name in lora.keys():
lokr_w1_b = lora[lokr_w1_b_name]
loaded_keys.add(lokr_w1_b_name)
lokr_w2_a = None
if lokr_w2_a_name in lora.keys():
lokr_w2_a = lora[lokr_w2_a_name]
loaded_keys.add(lokr_w2_a_name)
lokr_w2_b = None
if lokr_w2_b_name in lora.keys():
lokr_w2_b = lora[lokr_w2_b_name]
loaded_keys.add(lokr_w2_b_name)
lokr_t2 = None
if lokr_t2_name in lora.keys():
lokr_t2 = lora[lokr_t2_name]
loaded_keys.add(lokr_t2_name)
if (lokr_w1 is not None) or (lokr_w2 is not None) or (lokr_w1_a is not None) or (lokr_w2_a is not None):
patch_dict[to_load[x]] = (lokr_w1, lokr_w2, alpha, lokr_w1_a, lokr_w1_b, lokr_w2_a, lokr_w2_b, lokr_t2)
for x in lora.keys():
if x not in loaded_keys:
print("lora key not loaded", x)
return patch_dict
def model_lora_keys(model, key_map={}):
sdk = model.state_dict().keys()
counter = 0
for b in range(12):
2023-06-09 16:24:24 +00:00
tk = "diffusion_model.input_blocks.{}.1".format(b)
up_counter = 0
2023-03-10 02:41:24 +00:00
for c in LORA_UNET_MAP_ATTENTIONS:
k = "{}.{}.weight".format(tk, c)
if k in sdk:
2023-03-10 02:41:24 +00:00
lora_key = "lora_unet_down_blocks_{}_attentions_{}_{}".format(counter // 2, counter % 2, LORA_UNET_MAP_ATTENTIONS[c])
key_map[lora_key] = k
up_counter += 1
if up_counter >= 4:
counter += 1
2023-03-10 02:41:24 +00:00
for c in LORA_UNET_MAP_ATTENTIONS:
2023-06-09 16:24:24 +00:00
k = "diffusion_model.middle_block.1.{}.weight".format(c)
if k in sdk:
2023-03-10 02:41:24 +00:00
lora_key = "lora_unet_mid_block_attentions_0_{}".format(LORA_UNET_MAP_ATTENTIONS[c])
key_map[lora_key] = k
counter = 3
for b in range(12):
2023-06-09 16:24:24 +00:00
tk = "diffusion_model.output_blocks.{}.1".format(b)
up_counter = 0
2023-03-10 02:41:24 +00:00
for c in LORA_UNET_MAP_ATTENTIONS:
k = "{}.{}.weight".format(tk, c)
if k in sdk:
2023-03-10 02:41:24 +00:00
lora_key = "lora_unet_up_blocks_{}_attentions_{}_{}".format(counter // 3, counter % 3, LORA_UNET_MAP_ATTENTIONS[c])
key_map[lora_key] = k
up_counter += 1
if up_counter >= 4:
counter += 1
counter = 0
2023-02-05 06:54:09 +00:00
text_model_lora_key = "lora_te_text_model_encoder_layers_{}_{}"
for b in range(24):
for c in LORA_CLIP_MAP:
k = "transformer.text_model.encoder.layers.{}.{}.weight".format(b, c)
if k in sdk:
2023-02-05 06:54:09 +00:00
lora_key = text_model_lora_key.format(b, LORA_CLIP_MAP[c])
key_map[lora_key] = k
2023-02-05 06:54:09 +00:00
2023-03-10 02:41:24 +00:00
#Locon stuff
ds_counter = 0
counter = 0
for b in range(12):
2023-06-09 16:24:24 +00:00
tk = "diffusion_model.input_blocks.{}.0".format(b)
2023-03-10 02:41:24 +00:00
key_in = False
for c in LORA_UNET_MAP_RESNET:
k = "{}.{}.weight".format(tk, c)
if k in sdk:
lora_key = "lora_unet_down_blocks_{}_{}".format(counter // 2, LORA_UNET_MAP_RESNET[c].format(counter % 2))
key_map[lora_key] = k
key_in = True
for bb in range(3):
k = "{}.{}.op.weight".format(tk[:-2], bb)
if k in sdk:
lora_key = "lora_unet_down_blocks_{}_downsamplers_0_conv".format(ds_counter)
key_map[lora_key] = k
ds_counter += 1
if key_in:
counter += 1
counter = 0
for b in range(3):
2023-06-09 16:24:24 +00:00
tk = "diffusion_model.middle_block.{}".format(b)
2023-03-10 02:41:24 +00:00
key_in = False
for c in LORA_UNET_MAP_RESNET:
k = "{}.{}.weight".format(tk, c)
if k in sdk:
lora_key = "lora_unet_mid_block_{}".format(LORA_UNET_MAP_RESNET[c].format(counter))
key_map[lora_key] = k
key_in = True
if key_in:
counter += 1
counter = 0
us_counter = 0
for b in range(12):
2023-06-09 16:24:24 +00:00
tk = "diffusion_model.output_blocks.{}.0".format(b)
2023-03-10 02:41:24 +00:00
key_in = False
for c in LORA_UNET_MAP_RESNET:
k = "{}.{}.weight".format(tk, c)
if k in sdk:
lora_key = "lora_unet_up_blocks_{}_{}".format(counter // 3, LORA_UNET_MAP_RESNET[c].format(counter % 3))
key_map[lora_key] = k
key_in = True
for bb in range(3):
k = "{}.{}.conv.weight".format(tk[:-2], bb)
if k in sdk:
lora_key = "lora_unet_up_blocks_{}_upsamplers_0_conv".format(us_counter)
key_map[lora_key] = k
us_counter += 1
if key_in:
counter += 1
return key_map
class ModelPatcher:
def __init__(self, model, size=0):
self.size = size
self.model = model
self.patches = []
self.backup = {}
self.model_options = {"transformer_options":{}}
self.model_size()
def model_size(self):
if self.size > 0:
return self.size
model_sd = self.model.state_dict()
size = 0
for k in model_sd:
t = model_sd[k]
size += t.nelement() * t.element_size()
self.size = size
return size
def clone(self):
n = ModelPatcher(self.model, self.size)
n.patches = self.patches[:]
n.model_options = copy.deepcopy(self.model_options)
return n
def set_model_tomesd(self, ratio):
self.model_options["transformer_options"]["tomesd"] = {"ratio": ratio}
def set_model_sampler_cfg_function(self, sampler_cfg_function):
if len(inspect.signature(sampler_cfg_function).parameters) == 3:
self.model_options["sampler_cfg_function"] = lambda args: sampler_cfg_function(args["cond"], args["uncond"], args["cond_scale"]) #Old way
else:
self.model_options["sampler_cfg_function"] = sampler_cfg_function
def set_model_patch(self, patch, name):
to = self.model_options["transformer_options"]
if "patches" not in to:
to["patches"] = {}
to["patches"][name] = to["patches"].get(name, []) + [patch]
def set_model_attn1_patch(self, patch):
self.set_model_patch(patch, "attn1_patch")
def set_model_attn2_patch(self, patch):
self.set_model_patch(patch, "attn2_patch")
def model_patches_to(self, device):
to = self.model_options["transformer_options"]
if "patches" in to:
patches = to["patches"]
for name in patches:
patch_list = patches[name]
for i in range(len(patch_list)):
if hasattr(patch_list[i], "to"):
patch_list[i] = patch_list[i].to(device)
def model_dtype(self):
2023-06-09 16:24:24 +00:00
return self.model.get_dtype()
def add_patches(self, patches, strength=1.0):
p = {}
model_sd = self.model.state_dict()
for k in patches:
if k in model_sd:
p[k] = patches[k]
self.patches += [(strength, p)]
return p.keys()
def patch_model(self):
model_sd = self.model.state_dict()
for p in self.patches:
for k in p[1]:
v = p[1][k]
key = k
2023-02-05 06:54:09 +00:00
if key not in model_sd:
print("could not patch. key doesn't exist in model:", k)
continue
2023-02-05 06:54:09 +00:00
weight = model_sd[key]
if key not in self.backup:
self.backup[key] = weight.clone()
alpha = p[0]
2023-03-23 07:40:12 +00:00
if len(v) == 4: #lora/locon
mat1 = v[0]
mat2 = v[1]
if v[2] is not None:
alpha *= v[2] / mat2.shape[0]
if v[3] is not None:
#locon mid weights, hopefully the math is fine because I didn't properly test it
final_shape = [mat2.shape[1], mat2.shape[0], v[3].shape[2], v[3].shape[3]]
mat2 = torch.mm(mat2.transpose(0, 1).flatten(start_dim=1).float(), v[3].transpose(0, 1).flatten(start_dim=1).float()).reshape(final_shape).transpose(0, 1)
weight += (alpha * torch.mm(mat1.flatten(start_dim=1).float(), mat2.flatten(start_dim=1).float())).reshape(weight.shape).type(weight.dtype).to(weight.device)
2023-05-01 22:11:58 +00:00
elif len(v) == 8: #lokr
w1 = v[0]
w2 = v[1]
w1_a = v[3]
w1_b = v[4]
w2_a = v[5]
w2_b = v[6]
t2 = v[7]
dim = None
if w1 is None:
dim = w1_b.shape[0]
w1 = torch.mm(w1_a.float(), w1_b.float())
if w2 is None:
dim = w2_b.shape[0]
if t2 is None:
w2 = torch.mm(w2_a.float(), w2_b.float())
else:
w2 = torch.einsum('i j k l, j r, i p -> p r k l', t2.float(), w2_b.float(), w2_a.float())
if len(w2.shape) == 4:
w1 = w1.unsqueeze(2).unsqueeze(2)
if v[2] is not None and dim is not None:
alpha *= v[2] / dim
weight += alpha * torch.kron(w1.float(), w2.float()).reshape(weight.shape).type(weight.dtype).to(weight.device)
2023-03-23 07:40:12 +00:00
else: #loha
w1a = v[0]
w1b = v[1]
if v[2] is not None:
alpha *= v[2] / w1b.shape[0]
w2a = v[3]
w2b = v[4]
if v[5] is not None: #cp decomposition
t1 = v[5]
t2 = v[6]
m1 = torch.einsum('i j k l, j r, i p -> p r k l', t1.float(), w1b.float(), w1a.float())
m2 = torch.einsum('i j k l, j r, i p -> p r k l', t2.float(), w2b.float(), w2a.float())
else:
m1 = torch.mm(w1a.float(), w1b.float())
m2 = torch.mm(w2a.float(), w2b.float())
weight += (alpha * m1 * m2).reshape(weight.shape).type(weight.dtype).to(weight.device)
return self.model
def unpatch_model(self):
model_sd = self.model.state_dict()
2023-02-28 17:27:43 +00:00
keys = list(self.backup.keys())
for k in keys:
model_sd[k][:] = self.backup[k]
2023-02-28 17:27:43 +00:00
del self.backup[k]
self.backup = {}
def load_lora_for_models(model, clip, lora_path, strength_model, strength_clip):
key_map = model_lora_keys(model.model)
key_map = model_lora_keys(clip.cond_stage_model, key_map)
loaded = load_lora(lora_path, key_map)
new_modelpatcher = model.clone()
k = new_modelpatcher.add_patches(loaded, strength_model)
new_clip = clip.clone()
k1 = new_clip.add_patches(loaded, strength_clip)
k = set(k)
k1 = set(k1)
for x in loaded:
if (x not in k) and (x not in k1):
print("NOT LOADED", x)
return (new_modelpatcher, new_clip)
2023-01-03 06:53:32 +00:00
class CLIP:
def __init__(self, config={}, embedding_directory=None, no_init=False):
if no_init:
return
2023-01-03 06:53:32 +00:00
self.target_clip = config["target"]
if "params" in config:
params = config["params"]
else:
params = {}
2023-05-04 22:07:41 +00:00
if self.target_clip.endswith("FrozenOpenCLIPEmbedder"):
2023-01-03 06:53:32 +00:00
clip = sd2_clip.SD2ClipModel
tokenizer = sd2_clip.SD2Tokenizer
2023-05-04 22:07:41 +00:00
elif self.target_clip.endswith("FrozenCLIPEmbedder"):
2023-01-03 06:53:32 +00:00
clip = sd1_clip.SD1ClipModel
tokenizer = sd1_clip.SD1Tokenizer
self.device = model_management.text_encoder_device()
params["device"] = self.device
self.cond_stage_model = clip(**(params))
self.cond_stage_model = self.cond_stage_model.to(self.device)
self.tokenizer = tokenizer(embedding_directory=embedding_directory)
self.patcher = ModelPatcher(self.cond_stage_model)
self.layer_idx = None
def clone(self):
n = CLIP(no_init=True)
n.target_clip = self.target_clip
n.patcher = self.patcher.clone()
n.cond_stage_model = self.cond_stage_model
n.tokenizer = self.tokenizer
n.layer_idx = self.layer_idx
return n
def load_from_state_dict(self, sd):
self.cond_stage_model.transformer.load_state_dict(sd, strict=False)
def add_patches(self, patches, strength=1.0):
return self.patcher.add_patches(patches, strength)
2023-01-03 06:53:32 +00:00
def clip_layer(self, layer_idx):
self.layer_idx = layer_idx
def tokenize(self, text, return_word_ids=False):
return self.tokenizer.tokenize_with_weights(text, return_word_ids)
2023-04-13 20:06:50 +00:00
2023-04-19 13:36:19 +00:00
def encode_from_tokens(self, tokens, return_pooled=False):
if self.layer_idx is not None:
self.cond_stage_model.clip_layer(self.layer_idx)
try:
self.patcher.patch_model()
cond = self.cond_stage_model.encode_token_weights(tokens)
self.patcher.unpatch_model()
except Exception as e:
self.patcher.unpatch_model()
raise e
2023-04-19 13:36:19 +00:00
if return_pooled:
eos_token_index = max(range(len(tokens[0])), key=tokens[0].__getitem__)
pooled = cond[:, eos_token_index]
return cond, pooled
2023-01-03 06:53:32 +00:00
return cond
def encode(self, text):
tokens = self.tokenize(text)
return self.encode_from_tokens(tokens)
2023-01-03 06:53:32 +00:00
class VAE:
def __init__(self, ckpt_path=None, scale_factor=0.18215, device=None, config=None):
2023-01-03 06:53:32 +00:00
if config is None:
#default SD1.x/SD2.x VAE parameters
ddconfig = {'double_z': True, 'z_channels': 4, 'resolution': 256, 'in_channels': 3, 'out_ch': 3, 'ch': 128, 'ch_mult': [1, 2, 4, 4], 'num_res_blocks': 2, 'attn_resolutions': [], 'dropout': 0.0}
2023-05-28 06:02:09 +00:00
self.first_stage_model = AutoencoderKL(ddconfig, {'target': 'torch.nn.Identity'}, 4, monitor="val/rec_loss")
2023-01-03 06:53:32 +00:00
else:
2023-05-28 06:02:09 +00:00
self.first_stage_model = AutoencoderKL(**(config['params']))
2023-01-03 06:53:32 +00:00
self.first_stage_model = self.first_stage_model.eval()
2023-05-28 06:02:09 +00:00
if ckpt_path is not None:
sd = utils.load_torch_file(ckpt_path)
if 'decoder.up_blocks.0.resnets.0.norm1.weight' in sd.keys(): #diffusers format
sd = diffusers_convert.convert_vae_state_dict(sd)
self.first_stage_model.load_state_dict(sd, strict=False)
2023-01-03 06:53:32 +00:00
self.scale_factor = scale_factor
if device is None:
device = model_management.get_torch_device()
2023-01-03 06:53:32 +00:00
self.device = device
def decode_tiled_(self, samples, tile_x=64, tile_y=64, overlap = 16):
2023-05-03 16:33:19 +00:00
steps = samples.shape[0] * utils.get_tiled_scale_steps(samples.shape[3], samples.shape[2], tile_x, tile_y, overlap)
2023-05-03 21:48:35 +00:00
steps += samples.shape[0] * utils.get_tiled_scale_steps(samples.shape[3], samples.shape[2], tile_x // 2, tile_y * 2, overlap)
steps += samples.shape[0] * utils.get_tiled_scale_steps(samples.shape[3], samples.shape[2], tile_x * 2, tile_y // 2, overlap)
2023-05-03 17:19:22 +00:00
pbar = utils.ProgressBar(steps)
2023-04-24 10:55:44 +00:00
decode_fn = lambda a: (self.first_stage_model.decode(1. / self.scale_factor * a.to(self.device)) + 1.0)
output = torch.clamp((
2023-04-24 10:55:44 +00:00
(utils.tiled_scale(samples, decode_fn, tile_x // 2, tile_y * 2, overlap, upscale_amount = 8, pbar = pbar) +
utils.tiled_scale(samples, decode_fn, tile_x * 2, tile_y // 2, overlap, upscale_amount = 8, pbar = pbar) +
utils.tiled_scale(samples, decode_fn, tile_x, tile_y, overlap, upscale_amount = 8, pbar = pbar))
/ 3.0) / 2.0, min=0.0, max=1.0)
return output
def encode_tiled_(self, pixel_samples, tile_x=512, tile_y=512, overlap = 64):
steps = pixel_samples.shape[0] * utils.get_tiled_scale_steps(pixel_samples.shape[3], pixel_samples.shape[2], tile_x, tile_y, overlap)
steps += pixel_samples.shape[0] * utils.get_tiled_scale_steps(pixel_samples.shape[3], pixel_samples.shape[2], tile_x // 2, tile_y * 2, overlap)
steps += pixel_samples.shape[0] * utils.get_tiled_scale_steps(pixel_samples.shape[3], pixel_samples.shape[2], tile_x * 2, tile_y // 2, overlap)
pbar = utils.ProgressBar(steps)
encode_fn = lambda a: self.first_stage_model.encode(2. * a.to(self.device) - 1.).sample() * self.scale_factor
samples = utils.tiled_scale(pixel_samples, encode_fn, tile_x, tile_y, overlap, upscale_amount = (1/8), out_channels=4, pbar=pbar)
samples += utils.tiled_scale(pixel_samples, encode_fn, tile_x * 2, tile_y // 2, overlap, upscale_amount = (1/8), out_channels=4, pbar=pbar)
samples += utils.tiled_scale(pixel_samples, encode_fn, tile_x // 2, tile_y * 2, overlap, upscale_amount = (1/8), out_channels=4, pbar=pbar)
samples /= 3.0
return samples
def decode(self, samples_in):
model_management.unload_model()
2023-01-03 06:53:32 +00:00
self.first_stage_model = self.first_stage_model.to(self.device)
try:
free_memory = model_management.get_free_memory(self.device)
batch_number = int((free_memory * 0.7) / (2562 * samples_in.shape[2] * samples_in.shape[3] * 64))
batch_number = max(1, batch_number)
pixel_samples = torch.empty((samples_in.shape[0], 3, round(samples_in.shape[2] * 8), round(samples_in.shape[3] * 8)), device="cpu")
for x in range(0, samples_in.shape[0], batch_number):
samples = samples_in[x:x+batch_number].to(self.device)
pixel_samples[x:x+batch_number] = torch.clamp((self.first_stage_model.decode(1. / self.scale_factor * samples) + 1.0) / 2.0, min=0.0, max=1.0).cpu()
except model_management.OOM_EXCEPTION as e:
print("Warning: Ran out of memory when regular VAE decoding, retrying with tiled VAE decoding.")
pixel_samples = self.decode_tiled_(samples_in)
2023-01-03 06:53:32 +00:00
self.first_stage_model = self.first_stage_model.cpu()
pixel_samples = pixel_samples.cpu().movedim(1,-1)
return pixel_samples
def decode_tiled(self, samples, tile_x=64, tile_y=64, overlap = 16):
model_management.unload_model()
self.first_stage_model = self.first_stage_model.to(self.device)
output = self.decode_tiled_(samples, tile_x, tile_y, overlap)
self.first_stage_model = self.first_stage_model.cpu()
return output.movedim(1,-1)
2023-01-03 06:53:32 +00:00
def encode(self, pixel_samples):
model_management.unload_model()
2023-01-03 06:53:32 +00:00
self.first_stage_model = self.first_stage_model.to(self.device)
pixel_samples = pixel_samples.movedim(-1,1)
try:
free_memory = model_management.get_free_memory(self.device)
batch_number = int((free_memory * 0.7) / (2078 * pixel_samples.shape[2] * pixel_samples.shape[3])) #NOTE: this constant along with the one in the decode above are estimated from the mem usage for the VAE and could change.
batch_number = max(1, batch_number)
samples = torch.empty((pixel_samples.shape[0], 4, round(pixel_samples.shape[2] // 8), round(pixel_samples.shape[3] // 8)), device="cpu")
for x in range(0, pixel_samples.shape[0], batch_number):
pixels_in = (2. * pixel_samples[x:x+batch_number] - 1.).to(self.device)
samples[x:x+batch_number] = self.first_stage_model.encode(pixels_in).sample().cpu() * self.scale_factor
except model_management.OOM_EXCEPTION as e:
print("Warning: Ran out of memory when regular VAE encoding, retrying with tiled VAE encoding.")
samples = self.encode_tiled_(pixel_samples)
2023-01-03 06:53:32 +00:00
self.first_stage_model = self.first_stage_model.cpu()
return samples
2023-03-11 20:28:15 +00:00
def encode_tiled(self, pixel_samples, tile_x=512, tile_y=512, overlap = 64):
model_management.unload_model()
self.first_stage_model = self.first_stage_model.to(self.device)
pixel_samples = pixel_samples.movedim(-1,1)
samples = self.encode_tiled_(pixel_samples, tile_x=tile_x, tile_y=tile_y, overlap=overlap)
2023-03-11 20:28:15 +00:00
self.first_stage_model = self.first_stage_model.cpu()
return samples
2023-05-12 21:49:09 +00:00
def broadcast_image_to(tensor, target_batch_size, batched_number):
current_batch_size = tensor.shape[0]
2023-05-12 21:57:40 +00:00
#print(current_batch_size, target_batch_size)
if current_batch_size == 1:
return tensor
per_batch = target_batch_size // batched_number
tensor = tensor[:per_batch]
if per_batch > tensor.shape[0]:
tensor = torch.cat([tensor] * (per_batch // tensor.shape[0]) + [tensor[:(per_batch % tensor.shape[0])]], dim=0)
current_batch_size = tensor.shape[0]
if current_batch_size == target_batch_size:
return tensor
else:
return torch.cat([tensor] * batched_number, dim=0)
2023-02-16 15:38:08 +00:00
class ControlNet:
def __init__(self, control_model, global_average_pooling=False, device=None):
2023-02-16 15:38:08 +00:00
self.control_model = control_model
self.cond_hint_original = None
self.cond_hint = None
self.strength = 1.0
if device is None:
device = model_management.get_torch_device()
2023-02-17 20:31:38 +00:00
self.device = device
self.previous_controlnet = None
self.global_average_pooling = global_average_pooling
2023-02-16 15:38:08 +00:00
def get_control(self, x_noisy, t, cond_txt, batched_number):
control_prev = None
if self.previous_controlnet is not None:
2023-02-26 03:12:22 +00:00
control_prev = self.previous_controlnet.get_control(x_noisy, t, cond_txt, batched_number)
2023-02-17 20:31:38 +00:00
output_dtype = x_noisy.dtype
2023-02-16 15:38:08 +00:00
if self.cond_hint is None or x_noisy.shape[2] * 8 != self.cond_hint.shape[2] or x_noisy.shape[3] * 8 != self.cond_hint.shape[3]:
if self.cond_hint is not None:
del self.cond_hint
self.cond_hint = None
2023-05-12 21:49:09 +00:00
self.cond_hint = utils.common_upscale(self.cond_hint_original, x_noisy.shape[3] * 8, x_noisy.shape[2] * 8, 'nearest-exact', "center").to(self.control_model.dtype).to(self.device)
if x_noisy.shape[0] != self.cond_hint.shape[0]:
self.cond_hint = broadcast_image_to(self.cond_hint, x_noisy.shape[0], batched_number)
2023-02-17 20:31:38 +00:00
if self.control_model.dtype == torch.float16:
precision_scope = torch.autocast
else:
precision_scope = contextlib.nullcontext
with precision_scope(model_management.get_autocast_device(self.device)):
2023-02-17 20:45:29 +00:00
self.control_model = model_management.load_if_low_vram(self.control_model)
2023-02-17 20:31:38 +00:00
control = self.control_model(x=x_noisy, hint=self.cond_hint, timesteps=t, context=cond_txt)
2023-02-17 20:45:29 +00:00
self.control_model = model_management.unload_if_low_vram(self.control_model)
out = {'middle':[], 'output': []}
2023-02-17 20:31:38 +00:00
autocast_enabled = torch.is_autocast_enabled()
for i in range(len(control)):
2023-02-25 04:36:17 +00:00
if i == (len(control) - 1):
key = 'middle'
index = 0
else:
key = 'output'
index = i
x = control[i]
if self.global_average_pooling:
x = torch.mean(x, dim=(2, 3), keepdim=True).repeat(1, 1, x.shape[2], x.shape[3])
x *= self.strength
2023-02-17 20:31:38 +00:00
if x.dtype != output_dtype and not autocast_enabled:
x = x.to(output_dtype)
2023-02-25 04:36:17 +00:00
if control_prev is not None and key in control_prev:
prev = control_prev[key][index]
if prev is not None:
x += prev
out[key].append(x)
if control_prev is not None and 'input' in control_prev:
out['input'] = control_prev['input']
2023-02-17 20:31:38 +00:00
return out
2023-02-16 15:38:08 +00:00
def set_cond_hint(self, cond_hint, strength=1.0):
2023-02-16 15:38:08 +00:00
self.cond_hint_original = cond_hint
self.strength = strength
2023-02-16 15:38:08 +00:00
return self
def set_previous_controlnet(self, controlnet):
self.previous_controlnet = controlnet
return self
2023-02-16 15:38:08 +00:00
def cleanup(self):
if self.previous_controlnet is not None:
self.previous_controlnet.cleanup()
2023-02-16 15:38:08 +00:00
if self.cond_hint is not None:
del self.cond_hint
self.cond_hint = None
def copy(self):
c = ControlNet(self.control_model, global_average_pooling=self.global_average_pooling)
2023-02-16 15:38:08 +00:00
c.cond_hint_original = self.cond_hint_original
c.strength = self.strength
2023-02-16 15:38:08 +00:00
return c
2023-04-19 13:36:19 +00:00
def get_models(self):
out = []
if self.previous_controlnet is not None:
2023-04-19 13:36:19 +00:00
out += self.previous_controlnet.get_models()
out.append(self.control_model)
return out
2023-02-23 04:22:03 +00:00
def load_controlnet(ckpt_path, model=None):
2023-06-13 14:12:03 +00:00
controlnet_data = utils.load_torch_file(ckpt_path, safe_load=True)
2023-02-16 15:38:08 +00:00
pth_key = 'control_model.input_blocks.1.1.transformer_blocks.0.attn2.to_k.weight'
pth = False
sd2 = False
key = 'input_blocks.1.1.transformer_blocks.0.attn2.to_k.weight'
if pth_key in controlnet_data:
pth = True
key = pth_key
elif key in controlnet_data:
pass
else:
net = load_t2i_adapter(controlnet_data)
if net is None:
print("error checkpoint does not contain controlnet or t2i adapter data", ckpt_path)
return net
2023-02-16 15:38:08 +00:00
context_dim = controlnet_data[key].shape[1]
2023-02-17 20:31:38 +00:00
use_fp16 = False
if model_management.should_use_fp16() and controlnet_data[key].dtype == torch.float16:
2023-02-17 20:31:38 +00:00
use_fp16 = True
2023-03-08 06:13:38 +00:00
if context_dim == 768:
#SD1.x
control_model = cldm.ControlNet(image_size=32,
in_channels=4,
hint_channels=3,
model_channels=320,
attention_resolutions=[ 4, 2, 1 ],
num_res_blocks=2,
channel_mult=[ 1, 2, 4, 4 ],
num_heads=8,
use_spatial_transformer=True,
transformer_depth=1,
context_dim=context_dim,
2023-05-31 17:03:24 +00:00
use_checkpoint=False,
2023-03-08 06:13:38 +00:00
legacy=False,
use_fp16=use_fp16)
else:
#SD2.x
control_model = cldm.ControlNet(image_size=32,
in_channels=4,
hint_channels=3,
model_channels=320,
attention_resolutions=[ 4, 2, 1 ],
num_res_blocks=2,
channel_mult=[ 1, 2, 4, 4 ],
num_head_channels=64,
use_spatial_transformer=True,
use_linear_in_transformer=True,
transformer_depth=1,
context_dim=context_dim,
2023-05-31 17:03:24 +00:00
use_checkpoint=False,
2023-03-08 06:13:38 +00:00
legacy=False,
use_fp16=use_fp16)
2023-02-16 15:38:08 +00:00
if pth:
2023-02-23 04:22:03 +00:00
if 'difference' in controlnet_data:
if model is not None:
m = model.patch_model()
model_sd = m.state_dict()
for x in controlnet_data:
c_m = "control_model."
if x.startswith(c_m):
2023-06-09 16:24:24 +00:00
sd_key = "diffusion_model.{}".format(x[len(c_m):])
2023-02-23 04:22:03 +00:00
if sd_key in model_sd:
cd = controlnet_data[x]
cd += model_sd[sd_key].type(cd.dtype).to(cd.device)
model.unpatch_model()
else:
print("WARNING: Loaded a diff controlnet without a model. It will very likely not work.")
2023-02-16 15:38:08 +00:00
class WeightsLoader(torch.nn.Module):
pass
w = WeightsLoader()
w.control_model = control_model
w.load_state_dict(controlnet_data, strict=False)
else:
control_model.load_state_dict(controlnet_data, strict=False)
if use_fp16:
control_model = control_model.half()
global_average_pooling = False
if ckpt_path.endswith("_shuffle.pth") or ckpt_path.endswith("_shuffle.safetensors") or ckpt_path.endswith("_shuffle_fp16.safetensors"): #TODO: smarter way of enabling global_average_pooling
global_average_pooling = True
control = ControlNet(control_model, global_average_pooling=global_average_pooling)
2023-02-16 15:38:08 +00:00
return control
class T2IAdapter:
def __init__(self, t2i_model, channels_in, device=None):
self.t2i_model = t2i_model
self.channels_in = channels_in
self.strength = 1.0
if device is None:
device = model_management.get_torch_device()
self.device = device
self.previous_controlnet = None
self.control_input = None
self.cond_hint_original = None
self.cond_hint = None
def get_control(self, x_noisy, t, cond_txt, batched_number):
control_prev = None
if self.previous_controlnet is not None:
2023-02-26 03:12:22 +00:00
control_prev = self.previous_controlnet.get_control(x_noisy, t, cond_txt, batched_number)
if self.cond_hint is None or x_noisy.shape[2] * 8 != self.cond_hint.shape[2] or x_noisy.shape[3] * 8 != self.cond_hint.shape[3]:
if self.cond_hint is not None:
del self.cond_hint
2023-05-12 21:49:09 +00:00
self.control_input = None
self.cond_hint = None
2023-05-12 21:49:09 +00:00
self.cond_hint = utils.common_upscale(self.cond_hint_original, x_noisy.shape[3] * 8, x_noisy.shape[2] * 8, 'nearest-exact', "center").float().to(self.device)
if self.channels_in == 1 and self.cond_hint.shape[1] > 1:
self.cond_hint = torch.mean(self.cond_hint, 1, keepdim=True)
2023-05-12 21:49:09 +00:00
if x_noisy.shape[0] != self.cond_hint.shape[0]:
self.cond_hint = broadcast_image_to(self.cond_hint, x_noisy.shape[0], batched_number)
if self.control_input is None:
self.t2i_model.to(self.device)
self.control_input = self.t2i_model(self.cond_hint)
self.t2i_model.cpu()
output_dtype = x_noisy.dtype
out = {'input':[]}
2023-02-25 17:19:03 +00:00
autocast_enabled = torch.is_autocast_enabled()
for i in range(len(self.control_input)):
key = 'input'
x = self.control_input[i] * self.strength
if x.dtype != output_dtype and not autocast_enabled:
x = x.to(output_dtype)
if control_prev is not None and key in control_prev:
index = len(control_prev[key]) - i * 3 - 3
prev = control_prev[key][index]
if prev is not None:
x += prev
out[key].insert(0, None)
out[key].insert(0, None)
out[key].insert(0, x)
if control_prev is not None and 'input' in control_prev:
for i in range(len(out['input'])):
if out['input'][i] is None:
out['input'][i] = control_prev['input'][i]
if control_prev is not None and 'middle' in control_prev:
out['middle'] = control_prev['middle']
if control_prev is not None and 'output' in control_prev:
out['output'] = control_prev['output']
return out
def set_cond_hint(self, cond_hint, strength=1.0):
self.cond_hint_original = cond_hint
self.strength = strength
return self
def set_previous_controlnet(self, controlnet):
self.previous_controlnet = controlnet
return self
def copy(self):
c = T2IAdapter(self.t2i_model, self.channels_in)
c.cond_hint_original = self.cond_hint_original
c.strength = self.strength
return c
def cleanup(self):
if self.previous_controlnet is not None:
self.previous_controlnet.cleanup()
if self.cond_hint is not None:
del self.cond_hint
self.cond_hint = None
2023-04-19 13:36:19 +00:00
def get_models(self):
out = []
if self.previous_controlnet is not None:
2023-04-19 13:36:19 +00:00
out += self.previous_controlnet.get_models()
return out
def load_t2i_adapter(t2i_data):
keys = t2i_data.keys()
if "body.0.in_conv.weight" in keys:
cin = t2i_data['body.0.in_conv.weight'].shape[1]
model_ad = adapter.Adapter_light(cin=cin, channels=[320, 640, 1280, 1280], nums_rb=4)
elif 'conv_in.weight' in keys:
cin = t2i_data['conv_in.weight'].shape[1]
model_ad = adapter.Adapter(cin=cin, channels=[320, 640, 1280, 1280][:4], nums_rb=2, ksize=1, sk=True, use_conv=False)
else:
return None
model_ad.load_state_dict(t2i_data)
return T2IAdapter(model_ad, cin // 64)
2023-02-16 15:38:08 +00:00
class StyleModel:
def __init__(self, model, device="cpu"):
self.model = model
def get_cond(self, input):
return self.model(input.last_hidden_state)
def load_style_model(ckpt_path):
2023-06-13 14:12:03 +00:00
model_data = utils.load_torch_file(ckpt_path, safe_load=True)
keys = model_data.keys()
if "style_embedding" in keys:
model = adapter.StyleAdapter(width=1024, context_dim=768, num_head=8, n_layes=3, num_token=8)
else:
raise Exception("invalid style model {}".format(ckpt_path))
model.load_state_dict(model_data)
return StyleModel(model)
def load_clip(ckpt_path, embedding_directory=None):
2023-06-13 14:12:03 +00:00
clip_data = utils.load_torch_file(ckpt_path, safe_load=True)
config = {}
if "text_model.encoder.layers.22.mlp.fc1.weight" in clip_data:
2023-05-04 22:07:41 +00:00
config['target'] = 'comfy.ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder'
else:
2023-05-04 22:07:41 +00:00
config['target'] = 'comfy.ldm.modules.encoders.modules.FrozenCLIPEmbedder'
clip = CLIP(config=config, embedding_directory=embedding_directory)
clip.load_from_state_dict(clip_data)
return clip
2023-01-03 06:53:32 +00:00
2023-04-19 13:36:19 +00:00
def load_gligen(ckpt_path):
2023-06-13 14:12:03 +00:00
data = utils.load_torch_file(ckpt_path, safe_load=True)
2023-04-19 13:36:19 +00:00
model = gligen.load_gligen(data)
if model_management.should_use_fp16():
model = model.half()
return model
2023-06-09 16:24:24 +00:00
def load_checkpoint(config_path=None, ckpt_path=None, output_vae=True, output_clip=True, embedding_directory=None, state_dict=None, config=None):
if config is None:
with open(config_path, 'r') as stream:
config = yaml.safe_load(stream)
2023-01-03 06:53:32 +00:00
model_config_params = config['model']['params']
clip_config = model_config_params['cond_stage_config']
scale_factor = model_config_params['scale_factor']
vae_config = model_config_params['first_stage_config']
fp16 = False
if "unet_config" in model_config_params:
if "params" in model_config_params["unet_config"]:
2023-06-09 16:24:24 +00:00
unet_config = model_config_params["unet_config"]["params"]
if "use_fp16" in unet_config:
fp16 = unet_config["use_fp16"]
noise_aug_config = None
if "noise_aug_config" in model_config_params:
noise_aug_config = model_config_params["noise_aug_config"]
v_prediction = False
if "parameterization" in model_config_params:
if model_config_params["parameterization"] == "v":
v_prediction = True
2023-01-03 06:53:32 +00:00
clip = None
vae = None
class WeightsLoader(torch.nn.Module):
pass
w = WeightsLoader()
load_state_dict_to = []
if output_vae:
vae = VAE(scale_factor=scale_factor, config=vae_config)
w.first_stage_model = vae.first_stage_model
load_state_dict_to = [w]
if output_clip:
clip = CLIP(config=clip_config, embedding_directory=embedding_directory)
2023-01-03 06:53:32 +00:00
w.cond_stage_model = clip.cond_stage_model
load_state_dict_to = [w]
2023-06-09 16:24:24 +00:00
if config['model']["target"].endswith("LatentInpaintDiffusion"):
model = model_base.SDInpaint(unet_config, v_prediction=v_prediction)
elif config['model']["target"].endswith("ImageEmbeddingConditionedLatentDiffusion"):
model = model_base.SD21UNCLIP(unet_config, noise_aug_config["params"], v_prediction=v_prediction)
else:
model = model_base.BaseModel(unet_config, v_prediction=v_prediction)
if state_dict is None:
state_dict = utils.load_torch_file(ckpt_path)
model = load_model_weights(model, state_dict, verbose=False, load_state_dict_to=load_state_dict_to)
if fp16:
model = model.half()
return (ModelPatcher(model), clip, vae)
def load_checkpoint_guess_config(ckpt_path, output_vae=True, output_clip=True, output_clipvision=False, embedding_directory=None):
sd = utils.load_torch_file(ckpt_path)
sd_keys = sd.keys()
clip = None
clipvision = None
vae = None
fp16 = model_management.should_use_fp16()
class WeightsLoader(torch.nn.Module):
pass
w = WeightsLoader()
load_state_dict_to = []
if output_vae:
vae = VAE()
w.first_stage_model = vae.first_stage_model
load_state_dict_to = [w]
if output_clip:
clip_config = {}
if "cond_stage_model.model.transformer.resblocks.22.attn.out_proj.weight" in sd_keys:
2023-05-04 22:07:41 +00:00
clip_config['target'] = 'comfy.ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder'
else:
2023-05-04 22:07:41 +00:00
clip_config['target'] = 'comfy.ldm.modules.encoders.modules.FrozenCLIPEmbedder'
clip = CLIP(config=clip_config, embedding_directory=embedding_directory)
w.cond_stage_model = clip.cond_stage_model
load_state_dict_to = [w]
clipvision_key = "embedder.model.visual.transformer.resblocks.0.attn.in_proj_weight"
noise_aug_config = None
if clipvision_key in sd_keys:
size = sd[clipvision_key].shape[1]
if output_clipvision:
clipvision = clip_vision.load_clipvision_from_sd(sd)
noise_aug_key = "noise_augmentor.betas"
if noise_aug_key in sd_keys:
noise_aug_config = {}
params = {}
noise_schedule_config = {}
noise_schedule_config["timesteps"] = sd[noise_aug_key].shape[0]
noise_schedule_config["beta_schedule"] = "squaredcos_cap_v2"
params["noise_schedule_config"] = noise_schedule_config
2023-05-04 22:07:41 +00:00
noise_aug_config['target'] = "comfy.ldm.modules.encoders.noise_aug_modules.CLIPEmbeddingNoiseAugmentation"
if size == 1280: #h
params["timestep_dim"] = 1024
elif size == 1024: #l
params["timestep_dim"] = 768
noise_aug_config['params'] = params
sd_config = {
"linear_start": 0.00085,
"linear_end": 0.012,
"num_timesteps_cond": 1,
"log_every_t": 200,
"timesteps": 1000,
"first_stage_key": "jpg",
"cond_stage_key": "txt",
"image_size": 64,
"channels": 4,
"cond_stage_trainable": False,
"monitor": "val/loss_simple_ema",
"scale_factor": 0.18215,
"use_ema": False,
}
unet_config = {
2023-05-31 17:03:24 +00:00
"use_checkpoint": False,
"image_size": 32,
"out_channels": 4,
"attention_resolutions": [
4,
2,
1
],
"num_res_blocks": 2,
"channel_mult": [
1,
2,
4,
4
],
"use_spatial_transformer": True,
"transformer_depth": 1,
"legacy": False
}
2023-06-06 07:25:49 +00:00
if len(sd['model.diffusion_model.input_blocks.4.1.proj_in.weight'].shape) == 2:
unet_config['use_linear_in_transformer'] = True
unet_config["use_fp16"] = fp16
unet_config["model_channels"] = sd['model.diffusion_model.input_blocks.0.0.weight'].shape[0]
unet_config["in_channels"] = sd['model.diffusion_model.input_blocks.0.0.weight'].shape[1]
2023-06-06 07:25:49 +00:00
unet_config["context_dim"] = sd['model.diffusion_model.input_blocks.4.1.transformer_blocks.0.attn2.to_k.weight'].shape[1]
2023-05-04 22:07:41 +00:00
sd_config["unet_config"] = {"target": "comfy.ldm.modules.diffusionmodules.openaimodel.UNetModel", "params": unet_config}
2023-06-09 16:24:24 +00:00
unclip_model = False
inpaint_model = False
if noise_aug_config is not None: #SD2.x unclip model
sd_config["noise_aug_config"] = noise_aug_config
sd_config["image_size"] = 96
sd_config["embedding_dropout"] = 0.25
sd_config["conditioning_key"] = 'crossattn-adm'
2023-06-09 16:24:24 +00:00
unclip_model = True
elif unet_config["in_channels"] > 4: #inpainting model
sd_config["conditioning_key"] = "hybrid"
sd_config["finetune_keys"] = None
2023-06-09 16:24:24 +00:00
inpaint_model = True
else:
sd_config["conditioning_key"] = "crossattn"
2023-06-06 07:25:49 +00:00
if unet_config["context_dim"] == 768:
unet_config["num_heads"] = 8 #SD1.x
2023-06-06 07:25:49 +00:00
else:
unet_config["num_head_channels"] = 64 #SD2.x
unclip = 'model.diffusion_model.label_emb.0.0.weight'
if unclip in sd_keys:
unet_config["num_classes"] = "sequential"
unet_config["adm_in_channels"] = sd[unclip].shape[1]
2023-06-09 16:24:24 +00:00
v_prediction = False
if unet_config["context_dim"] == 1024 and unet_config["in_channels"] == 4: #only SD2.x non inpainting models are v prediction
k = "model.diffusion_model.output_blocks.11.1.transformer_blocks.0.norm1.bias"
out = sd[k]
if torch.std(out, unbiased=False) > 0.09: # not sure how well this will actually work. I guess we will find out.
2023-06-09 16:24:24 +00:00
v_prediction = True
sd_config["parameterization"] = 'v'
2023-06-09 16:24:24 +00:00
if inpaint_model:
model = model_base.SDInpaint(unet_config, v_prediction=v_prediction)
elif unclip_model:
model = model_base.SD21UNCLIP(unet_config, noise_aug_config["params"], v_prediction=v_prediction)
else:
model = model_base.BaseModel(unet_config, v_prediction=v_prediction)
if fp16:
model = model.half()
model = load_model_weights(model, sd, verbose=False, load_state_dict_to=load_state_dict_to)
return (ModelPatcher(model), clip, vae, clipvision)