stable-diffusion-webui/modules/sd_vae.py

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

283 lines
8.3 KiB
Python
Raw Normal View History

2022-10-30 17:54:31 +03:00
import os
2022-12-25 15:49:25 +03:00
import collections
from dataclasses import dataclass
2023-08-10 16:22:10 +03:00
from modules import paths, shared, devices, script_callbacks, sd_models, extra_networks, lowvram, sd_hijack, hashes
2022-10-30 17:54:31 +03:00
import glob
2022-11-17 14:04:10 +03:00
from copy import deepcopy
2022-10-30 17:54:31 +03:00
vae_path = os.path.abspath(os.path.join(paths.models_path, "VAE"))
2022-10-30 17:54:31 +03:00
vae_ignore_keys = {"model_ema.decay", "model_ema.num_updates"}
vae_dict = {}
2022-10-30 17:54:31 +03:00
base_vae = None
loaded_vae_file = None
checkpoint_info = None
2022-12-25 15:49:25 +03:00
checkpoints_loaded = collections.OrderedDict()
2023-08-10 16:22:10 +03:00
def get_loaded_vae_name():
if loaded_vae_file is None:
return None
return os.path.basename(loaded_vae_file)
def get_loaded_vae_hash():
if loaded_vae_file is None:
return None
sha256 = hashes.sha256(loaded_vae_file, 'vae')
return sha256[0:10] if sha256 else None
2023-08-10 16:22:10 +03:00
def get_base_vae(model):
if base_vae is not None and checkpoint_info == model.sd_checkpoint_info and model:
return base_vae
return None
def store_base_vae(model):
global base_vae, checkpoint_info
if checkpoint_info != model.sd_checkpoint_info:
assert not loaded_vae_file, "Trying to store non-base VAE!"
2022-11-17 14:04:10 +03:00
base_vae = deepcopy(model.first_stage_model.state_dict())
checkpoint_info = model.sd_checkpoint_info
def delete_base_vae():
global base_vae, checkpoint_info
base_vae = None
checkpoint_info = None
def restore_base_vae(model):
global loaded_vae_file
if base_vae is not None and checkpoint_info == model.sd_checkpoint_info:
print("Restoring base VAE")
_load_vae_dict(model, base_vae)
loaded_vae_file = None
delete_base_vae()
2022-10-30 17:54:31 +03:00
def get_filename(filepath):
return os.path.basename(filepath)
def refresh_vae_list():
vae_dict.clear()
paths = [
os.path.join(sd_models.model_path, '**/*.vae.ckpt'),
os.path.join(sd_models.model_path, '**/*.vae.pt'),
os.path.join(sd_models.model_path, '**/*.vae.safetensors'),
os.path.join(vae_path, '**/*.ckpt'),
os.path.join(vae_path, '**/*.pt'),
os.path.join(vae_path, '**/*.safetensors'),
2022-10-30 17:54:31 +03:00
]
if shared.cmd_opts.ckpt_dir is not None and os.path.isdir(shared.cmd_opts.ckpt_dir):
paths += [
os.path.join(shared.cmd_opts.ckpt_dir, '**/*.vae.ckpt'),
os.path.join(shared.cmd_opts.ckpt_dir, '**/*.vae.pt'),
os.path.join(shared.cmd_opts.ckpt_dir, '**/*.vae.safetensors'),
]
2023-01-17 17:50:41 +03:00
if shared.cmd_opts.vae_dir is not None and os.path.isdir(shared.cmd_opts.vae_dir):
paths += [
os.path.join(shared.cmd_opts.vae_dir, '**/*.ckpt'),
os.path.join(shared.cmd_opts.vae_dir, '**/*.pt'),
os.path.join(shared.cmd_opts.vae_dir, '**/*.safetensors'),
]
candidates = []
for path in paths:
candidates += glob.iglob(path, recursive=True)
2022-10-30 17:54:31 +03:00
for filepath in candidates:
name = get_filename(filepath)
vae_dict[name] = filepath
2023-08-19 08:31:06 +03:00
vae_dict.update(dict(sorted(vae_dict.items(), key=lambda item: shared.natural_sort_key(item[0]))))
2023-08-04 06:01:28 +03:00
def find_vae_near_checkpoint(checkpoint_file):
checkpoint_path = os.path.basename(checkpoint_file).rsplit('.', 1)[0]
for vae_file in vae_dict.values():
if os.path.basename(vae_file).startswith(checkpoint_path):
return vae_file
return None
@dataclass
class VaeResolution:
vae: str = None
source: str = None
resolved: bool = True
def tuple(self):
return self.vae, self.source
def is_automatic():
return shared.opts.sd_vae in {"Automatic", "auto"} # "auto" for people with old config
def resolve_vae_from_setting() -> VaeResolution:
if shared.opts.sd_vae == "None":
return VaeResolution()
vae_from_options = vae_dict.get(shared.opts.sd_vae, None)
if vae_from_options is not None:
return VaeResolution(vae_from_options, 'specified in settings')
if not is_automatic():
print(f"Couldn't find VAE named {shared.opts.sd_vae}; using None instead")
return VaeResolution(resolved=False)
def resolve_vae_from_user_metadata(checkpoint_file) -> VaeResolution:
metadata = extra_networks.get_user_metadata(checkpoint_file)
vae_metadata = metadata.get("vae", None)
if vae_metadata is not None and vae_metadata != "Automatic":
if vae_metadata == "None":
return VaeResolution()
vae_from_metadata = vae_dict.get(vae_metadata, None)
if vae_from_metadata is not None:
return VaeResolution(vae_from_metadata, "from user metadata")
return VaeResolution(resolved=False)
def resolve_vae_near_checkpoint(checkpoint_file) -> VaeResolution:
vae_near_checkpoint = find_vae_near_checkpoint(checkpoint_file)
if vae_near_checkpoint is not None and (not shared.opts.sd_vae_overrides_per_model_preferences or is_automatic()):
return VaeResolution(vae_near_checkpoint, 'found near the checkpoint')
return VaeResolution(resolved=False)
def resolve_vae(checkpoint_file) -> VaeResolution:
if shared.cmd_opts.vae_path is not None:
return VaeResolution(shared.cmd_opts.vae_path, 'from commandline argument')
if shared.opts.sd_vae_overrides_per_model_preferences and not is_automatic():
return resolve_vae_from_setting()
res = resolve_vae_from_user_metadata(checkpoint_file)
if res.resolved:
return res
res = resolve_vae_near_checkpoint(checkpoint_file)
if res.resolved:
return res
res = resolve_vae_from_setting()
return res
def load_vae_dict(filename, map_location):
vae_ckpt = sd_models.read_state_dict(filename, map_location=map_location)
vae_dict_1 = {k: v for k, v in vae_ckpt.items() if k[0:4] != "loss" and k not in vae_ignore_keys}
return vae_dict_1
def load_vae(model, vae_file=None, vae_source="from unknown source"):
global vae_dict, base_vae, loaded_vae_file
# save_settings = False
2022-12-25 15:49:25 +03:00
cache_enabled = shared.opts.sd_vae_checkpoint_cache > 0
2022-10-30 17:54:31 +03:00
if vae_file:
2022-12-25 15:49:25 +03:00
if cache_enabled and vae_file in checkpoints_loaded:
# use vae checkpoint cache
print(f"Loading VAE weights {vae_source}: cached {get_filename(vae_file)}")
2022-12-25 15:49:25 +03:00
store_base_vae(model)
_load_vae_dict(model, checkpoints_loaded[vae_file])
else:
assert os.path.isfile(vae_file), f"VAE {vae_source} doesn't exist: {vae_file}"
print(f"Loading VAE weights {vae_source}: {vae_file}")
2022-12-25 15:49:25 +03:00
store_base_vae(model)
vae_dict_1 = load_vae_dict(vae_file, map_location=shared.weight_load_location)
2022-12-25 15:49:25 +03:00
_load_vae_dict(model, vae_dict_1)
if cache_enabled:
# cache newly loaded vae
checkpoints_loaded[vae_file] = vae_dict_1.copy()
# clean up cache if limit is reached
if cache_enabled:
while len(checkpoints_loaded) > shared.opts.sd_vae_checkpoint_cache + 1: # we need to count the current model
checkpoints_loaded.popitem(last=False) # LRU
2022-10-30 17:54:31 +03:00
# If vae used is not in dict, update it
# It will be removed on refresh though
2022-10-30 17:54:31 +03:00
vae_opt = get_filename(vae_file)
if vae_opt not in vae_dict:
vae_dict[vae_opt] = vae_file
2022-11-13 07:11:14 +03:00
elif loaded_vae_file:
restore_base_vae(model)
2022-10-30 17:54:31 +03:00
loaded_vae_file = vae_file
model.base_vae = base_vae
model.loaded_vae_file = loaded_vae_file
# don't call this from outside
def _load_vae_dict(model, vae_dict_1):
model.first_stage_model.load_state_dict(vae_dict_1)
2022-10-30 17:54:31 +03:00
model.first_stage_model.to(devices.dtype_vae)
def clear_loaded_vae():
global loaded_vae_file
loaded_vae_file = None
unspecified = object()
def reload_vae_weights(sd_model=None, vae_file=unspecified):
if not sd_model:
sd_model = shared.sd_model
checkpoint_info = sd_model.sd_checkpoint_info
checkpoint_file = checkpoint_info.filename
if vae_file == unspecified:
vae_file, vae_source = resolve_vae(checkpoint_file).tuple()
else:
vae_source = "from function argument"
if loaded_vae_file == vae_file:
return
2023-08-22 18:49:08 +03:00
if sd_model.lowvram:
lowvram.send_everything_to_cpu()
else:
sd_model.to(devices.cpu)
sd_hijack.model_hijack.undo_hijack(sd_model)
load_vae(sd_model, vae_file, vae_source)
sd_hijack.model_hijack.hijack(sd_model)
script_callbacks.model_loaded_callback(sd_model)
2023-08-22 18:49:08 +03:00
if not sd_model.lowvram:
sd_model.to(devices.device)
print("VAE weights loaded.")
return sd_model