2022-10-14 13:00:38 +08:00
|
|
|
import collections
|
2022-09-17 17:05:04 +08:00
|
|
|
import os.path
|
|
|
|
import sys
|
|
|
|
from collections import namedtuple
|
|
|
|
import torch
|
|
|
|
from omegaconf import OmegaConf
|
|
|
|
|
|
|
|
from ldm.util import instantiate_from_config
|
|
|
|
|
2022-10-02 20:03:39 +08:00
|
|
|
from modules import shared, modelloader, devices
|
2022-09-28 00:01:13 +08:00
|
|
|
from modules.paths import models_path
|
2022-10-20 04:47:45 +08:00
|
|
|
from modules.sd_hijack_inpainting import do_inpainting_hijack, should_hijack_inpainting
|
2022-09-28 00:01:13 +08:00
|
|
|
|
|
|
|
model_dir = "Stable-diffusion"
|
2022-09-30 16:42:40 +08:00
|
|
|
model_path = os.path.abspath(os.path.join(models_path, model_dir))
|
2022-09-17 17:05:04 +08:00
|
|
|
|
2022-10-09 04:26:48 +08:00
|
|
|
CheckpointInfo = namedtuple("CheckpointInfo", ['filename', 'title', 'hash', 'model_name', 'config'])
|
2022-09-17 17:05:04 +08:00
|
|
|
checkpoints_list = {}
|
2022-10-14 13:00:38 +08:00
|
|
|
checkpoints_loaded = collections.OrderedDict()
|
2022-09-17 17:05:04 +08:00
|
|
|
|
|
|
|
try:
|
|
|
|
# this silences the annoying "Some weights of the model checkpoint were not used when initializing..." message at start.
|
|
|
|
|
|
|
|
from transformers import logging
|
|
|
|
|
|
|
|
logging.set_verbosity_error()
|
|
|
|
except Exception:
|
|
|
|
pass
|
|
|
|
|
|
|
|
|
2022-10-03 02:09:10 +08:00
|
|
|
def setup_model():
|
2022-09-28 00:01:13 +08:00
|
|
|
if not os.path.exists(model_path):
|
|
|
|
os.makedirs(model_path)
|
2022-10-03 02:09:10 +08:00
|
|
|
|
2022-09-30 08:59:36 +08:00
|
|
|
list_models()
|
|
|
|
|
|
|
|
|
2022-09-29 05:59:44 +08:00
|
|
|
def checkpoint_tiles():
|
|
|
|
return sorted([x.title for x in checkpoints_list.values()])
|
|
|
|
|
|
|
|
|
2022-09-17 17:05:04 +08:00
|
|
|
def list_models():
|
|
|
|
checkpoints_list.clear()
|
2022-10-03 02:09:10 +08:00
|
|
|
model_list = modelloader.load_models(model_path=model_path, command_path=shared.cmd_opts.ckpt_dir, ext_filter=[".ckpt"])
|
2022-09-17 17:05:04 +08:00
|
|
|
|
2022-09-30 16:42:40 +08:00
|
|
|
def modeltitle(path, shorthash):
|
2022-09-17 17:05:04 +08:00
|
|
|
abspath = os.path.abspath(path)
|
|
|
|
|
2022-10-03 02:22:20 +08:00
|
|
|
if shared.cmd_opts.ckpt_dir is not None and abspath.startswith(shared.cmd_opts.ckpt_dir):
|
|
|
|
name = abspath.replace(shared.cmd_opts.ckpt_dir, '')
|
2022-09-30 16:42:40 +08:00
|
|
|
elif abspath.startswith(model_path):
|
|
|
|
name = abspath.replace(model_path, '')
|
2022-09-17 17:05:04 +08:00
|
|
|
else:
|
|
|
|
name = os.path.basename(path)
|
|
|
|
|
|
|
|
if name.startswith("\\") or name.startswith("/"):
|
|
|
|
name = name[1:]
|
|
|
|
|
2022-09-29 05:59:44 +08:00
|
|
|
shortname = os.path.splitext(name.replace("/", "_").replace("\\", "_"))[0]
|
|
|
|
|
2022-09-30 16:42:40 +08:00
|
|
|
return f'{name} [{shorthash}]', shortname
|
2022-09-17 17:05:04 +08:00
|
|
|
|
|
|
|
cmd_ckpt = shared.cmd_opts.ckpt
|
|
|
|
if os.path.exists(cmd_ckpt):
|
|
|
|
h = model_hash(cmd_ckpt)
|
2022-09-30 16:42:40 +08:00
|
|
|
title, short_model_name = modeltitle(cmd_ckpt, h)
|
2022-10-09 04:26:48 +08:00
|
|
|
checkpoints_list[title] = CheckpointInfo(cmd_ckpt, title, h, short_model_name, shared.cmd_opts.config)
|
2022-10-02 22:24:50 +08:00
|
|
|
shared.opts.data['sd_model_checkpoint'] = title
|
2022-09-17 17:05:04 +08:00
|
|
|
elif cmd_ckpt is not None and cmd_ckpt != shared.default_sd_model_file:
|
2022-09-28 00:01:13 +08:00
|
|
|
print(f"Checkpoint in --ckpt argument not found (Possible it was moved to {model_path}: {cmd_ckpt}", file=sys.stderr)
|
|
|
|
for filename in model_list:
|
|
|
|
h = model_hash(filename)
|
2022-09-30 16:42:40 +08:00
|
|
|
title, short_model_name = modeltitle(filename, h)
|
2022-10-09 04:26:48 +08:00
|
|
|
|
|
|
|
basename, _ = os.path.splitext(filename)
|
|
|
|
config = basename + ".yaml"
|
|
|
|
if not os.path.exists(config):
|
|
|
|
config = shared.cmd_opts.config
|
|
|
|
|
|
|
|
checkpoints_list[title] = CheckpointInfo(filename, title, h, short_model_name, config)
|
2022-09-30 16:42:40 +08:00
|
|
|
|
2022-09-17 17:05:04 +08:00
|
|
|
|
2022-09-29 05:30:09 +08:00
|
|
|
def get_closet_checkpoint_match(searchString):
|
2022-09-30 02:08:03 +08:00
|
|
|
applicable = sorted([info for info in checkpoints_list.values() if searchString in info.title], key = lambda x:len(x.title))
|
2022-09-30 16:42:40 +08:00
|
|
|
if len(applicable) > 0:
|
2022-09-29 05:30:09 +08:00
|
|
|
return applicable[0]
|
|
|
|
return None
|
2022-09-17 17:05:04 +08:00
|
|
|
|
2022-09-30 16:42:40 +08:00
|
|
|
|
2022-09-17 17:05:04 +08:00
|
|
|
def model_hash(filename):
|
|
|
|
try:
|
|
|
|
with open(filename, "rb") as file:
|
|
|
|
import hashlib
|
|
|
|
m = hashlib.sha256()
|
|
|
|
|
|
|
|
file.seek(0x100000)
|
|
|
|
m.update(file.read(0x10000))
|
|
|
|
return m.hexdigest()[0:8]
|
|
|
|
except FileNotFoundError:
|
|
|
|
return 'NOFILE'
|
|
|
|
|
|
|
|
|
|
|
|
def select_checkpoint():
|
|
|
|
model_checkpoint = shared.opts.sd_model_checkpoint
|
|
|
|
checkpoint_info = checkpoints_list.get(model_checkpoint, None)
|
|
|
|
if checkpoint_info is not None:
|
|
|
|
return checkpoint_info
|
|
|
|
|
|
|
|
if len(checkpoints_list) == 0:
|
2022-09-19 04:52:01 +08:00
|
|
|
print(f"No checkpoints found. When searching for checkpoints, looked at:", file=sys.stderr)
|
2022-10-03 02:09:10 +08:00
|
|
|
if shared.cmd_opts.ckpt is not None:
|
|
|
|
print(f" - file {os.path.abspath(shared.cmd_opts.ckpt)}", file=sys.stderr)
|
|
|
|
print(f" - directory {model_path}", file=sys.stderr)
|
|
|
|
if shared.cmd_opts.ckpt_dir is not None:
|
|
|
|
print(f" - directory {os.path.abspath(shared.cmd_opts.ckpt_dir)}", file=sys.stderr)
|
2022-09-19 04:52:01 +08:00
|
|
|
print(f"Can't run without a checkpoint. Find and place a .ckpt file into any of those locations. The program will exit.", file=sys.stderr)
|
|
|
|
exit(1)
|
2022-09-17 17:05:04 +08:00
|
|
|
|
|
|
|
checkpoint_info = next(iter(checkpoints_list.values()))
|
|
|
|
if model_checkpoint is not None:
|
|
|
|
print(f"Checkpoint {model_checkpoint} not found; loading fallback {checkpoint_info.title}", file=sys.stderr)
|
|
|
|
|
|
|
|
return checkpoint_info
|
|
|
|
|
|
|
|
|
2022-10-19 13:42:22 +08:00
|
|
|
chckpoint_dict_replacements = {
|
|
|
|
'cond_stage_model.transformer.embeddings.': 'cond_stage_model.transformer.text_model.embeddings.',
|
|
|
|
'cond_stage_model.transformer.encoder.': 'cond_stage_model.transformer.text_model.encoder.',
|
|
|
|
'cond_stage_model.transformer.final_layer_norm.': 'cond_stage_model.transformer.text_model.final_layer_norm.',
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
def transform_checkpoint_dict_key(k):
|
|
|
|
for text, replacement in chckpoint_dict_replacements.items():
|
|
|
|
if k.startswith(text):
|
|
|
|
k = replacement + k[len(text):]
|
|
|
|
|
|
|
|
return k
|
|
|
|
|
|
|
|
|
2022-10-09 15:23:31 +08:00
|
|
|
def get_state_dict_from_checkpoint(pl_sd):
|
|
|
|
if "state_dict" in pl_sd:
|
2022-10-19 13:42:22 +08:00
|
|
|
pl_sd = pl_sd["state_dict"]
|
|
|
|
|
|
|
|
sd = {}
|
|
|
|
for k, v in pl_sd.items():
|
|
|
|
new_key = transform_checkpoint_dict_key(k)
|
|
|
|
|
|
|
|
if new_key is not None:
|
|
|
|
sd[new_key] = v
|
2022-10-09 15:23:31 +08:00
|
|
|
|
2022-10-19 17:45:30 +08:00
|
|
|
pl_sd.clear()
|
|
|
|
pl_sd.update(sd)
|
|
|
|
|
|
|
|
return pl_sd
|
2022-10-09 15:23:31 +08:00
|
|
|
|
|
|
|
|
2022-10-09 04:26:48 +08:00
|
|
|
def load_model_weights(model, checkpoint_info):
|
|
|
|
checkpoint_file = checkpoint_info.filename
|
|
|
|
sd_model_hash = checkpoint_info.hash
|
|
|
|
|
2022-10-14 13:00:38 +08:00
|
|
|
if checkpoint_info not in checkpoints_loaded:
|
|
|
|
print(f"Loading weights [{sd_model_hash}] from {checkpoint_file}")
|
2022-09-17 17:05:04 +08:00
|
|
|
|
2022-10-15 15:35:18 +08:00
|
|
|
pl_sd = torch.load(checkpoint_file, map_location=shared.weight_load_location)
|
2022-10-14 13:00:38 +08:00
|
|
|
if "global_step" in pl_sd:
|
|
|
|
print(f"Global Step: {pl_sd['global_step']}")
|
2022-10-09 15:23:31 +08:00
|
|
|
|
2022-10-14 13:00:38 +08:00
|
|
|
sd = get_state_dict_from_checkpoint(pl_sd)
|
2022-10-19 13:42:22 +08:00
|
|
|
missing, extra = model.load_state_dict(sd, strict=False)
|
2022-09-17 17:05:04 +08:00
|
|
|
|
2022-10-14 13:00:38 +08:00
|
|
|
if shared.cmd_opts.opt_channelslast:
|
|
|
|
model.to(memory_format=torch.channels_last)
|
2022-09-17 17:05:04 +08:00
|
|
|
|
2022-10-14 13:00:38 +08:00
|
|
|
if not shared.cmd_opts.no_half:
|
|
|
|
model.half()
|
2022-09-17 17:05:04 +08:00
|
|
|
|
2022-10-14 13:00:38 +08:00
|
|
|
devices.dtype = torch.float32 if shared.cmd_opts.no_half else torch.float16
|
|
|
|
devices.dtype_vae = torch.float32 if shared.cmd_opts.no_half or shared.cmd_opts.no_half_vae else torch.float16
|
2022-09-17 17:05:04 +08:00
|
|
|
|
2022-10-14 13:00:38 +08:00
|
|
|
vae_file = os.path.splitext(checkpoint_file)[0] + ".vae.pt"
|
2022-10-02 20:03:39 +08:00
|
|
|
|
2022-10-14 13:00:38 +08:00
|
|
|
if not os.path.exists(vae_file) and shared.cmd_opts.vae_path is not None:
|
|
|
|
vae_file = shared.cmd_opts.vae_path
|
2022-10-11 01:46:55 +08:00
|
|
|
|
2022-10-14 13:00:38 +08:00
|
|
|
if os.path.exists(vae_file):
|
|
|
|
print(f"Loading VAE weights from: {vae_file}")
|
2022-10-15 15:35:18 +08:00
|
|
|
vae_ckpt = torch.load(vae_file, map_location=shared.weight_load_location)
|
2022-10-14 13:00:38 +08:00
|
|
|
vae_dict = {k: v for k, v in vae_ckpt["state_dict"].items() if k[0:4] != "loss"}
|
|
|
|
model.first_stage_model.load_state_dict(vae_dict)
|
2022-10-07 15:40:22 +08:00
|
|
|
|
2022-10-14 13:00:38 +08:00
|
|
|
model.first_stage_model.to(devices.dtype_vae)
|
2022-10-07 15:40:22 +08:00
|
|
|
|
2022-10-14 13:00:38 +08:00
|
|
|
checkpoints_loaded[checkpoint_info] = model.state_dict().copy()
|
|
|
|
while len(checkpoints_loaded) > shared.opts.sd_checkpoint_cache:
|
|
|
|
checkpoints_loaded.popitem(last=False) # LRU
|
|
|
|
else:
|
|
|
|
print(f"Loading weights [{sd_model_hash}] from cache")
|
|
|
|
checkpoints_loaded.move_to_end(checkpoint_info)
|
|
|
|
model.load_state_dict(checkpoints_loaded[checkpoint_info])
|
2022-10-10 21:11:14 +08:00
|
|
|
|
2022-09-17 17:05:04 +08:00
|
|
|
model.sd_model_hash = sd_model_hash
|
2022-10-09 03:12:24 +08:00
|
|
|
model.sd_model_checkpoint = checkpoint_file
|
2022-10-09 04:26:48 +08:00
|
|
|
model.sd_checkpoint_info = checkpoint_info
|
2022-09-17 17:05:04 +08:00
|
|
|
|
|
|
|
|
2022-10-21 07:01:27 +08:00
|
|
|
def load_model(checkpoint_info=None):
|
2022-09-17 17:05:04 +08:00
|
|
|
from modules import lowvram, sd_hijack
|
2022-10-21 07:01:27 +08:00
|
|
|
checkpoint_info = checkpoint_info or select_checkpoint()
|
2022-09-17 17:05:04 +08:00
|
|
|
|
2022-10-09 04:26:48 +08:00
|
|
|
if checkpoint_info.config != shared.cmd_opts.config:
|
2022-10-09 15:31:47 +08:00
|
|
|
print(f"Loading config from: {checkpoint_info.config}")
|
2022-10-09 04:26:48 +08:00
|
|
|
|
|
|
|
sd_config = OmegaConf.load(checkpoint_info.config)
|
2022-10-20 04:47:45 +08:00
|
|
|
|
|
|
|
if should_hijack_inpainting(checkpoint_info):
|
|
|
|
# Hardcoded config for now...
|
|
|
|
sd_config.model.target = "ldm.models.diffusion.ddpm.LatentInpaintDiffusion"
|
|
|
|
sd_config.model.params.use_ema = False
|
|
|
|
sd_config.model.params.conditioning_key = "hybrid"
|
|
|
|
sd_config.model.params.unet_config.params.in_channels = 9
|
|
|
|
|
|
|
|
# Create a "fake" config with a different name so that we know to unload it when switching models.
|
|
|
|
checkpoint_info = checkpoint_info._replace(config=checkpoint_info.config.replace(".yaml", "-inpainting.yaml"))
|
|
|
|
|
2022-10-21 04:28:43 +08:00
|
|
|
do_inpainting_hijack()
|
2022-09-17 17:05:04 +08:00
|
|
|
sd_model = instantiate_from_config(sd_config.model)
|
2022-10-09 04:26:48 +08:00
|
|
|
load_model_weights(sd_model, checkpoint_info)
|
2022-09-17 17:05:04 +08:00
|
|
|
|
|
|
|
if shared.cmd_opts.lowvram or shared.cmd_opts.medvram:
|
|
|
|
lowvram.setup_for_low_vram(sd_model, shared.cmd_opts.medvram)
|
|
|
|
else:
|
|
|
|
sd_model.to(shared.device)
|
|
|
|
|
|
|
|
sd_hijack.model_hijack.hijack(sd_model)
|
|
|
|
|
|
|
|
sd_model.eval()
|
|
|
|
|
|
|
|
print(f"Model loaded.")
|
|
|
|
return sd_model
|
|
|
|
|
|
|
|
|
2022-09-17 18:49:36 +08:00
|
|
|
def reload_model_weights(sd_model, info=None):
|
2022-09-29 20:40:28 +08:00
|
|
|
from modules import lowvram, devices, sd_hijack
|
2022-09-17 18:49:36 +08:00
|
|
|
checkpoint_info = info or select_checkpoint()
|
2022-09-17 17:05:04 +08:00
|
|
|
|
2022-10-09 03:12:24 +08:00
|
|
|
if sd_model.sd_model_checkpoint == checkpoint_info.filename:
|
2022-09-17 17:05:04 +08:00
|
|
|
return
|
|
|
|
|
2022-10-20 04:47:45 +08:00
|
|
|
if sd_model.sd_checkpoint_info.config != checkpoint_info.config or should_hijack_inpainting(checkpoint_info) != should_hijack_inpainting(sd_model.sd_checkpoint_info):
|
2022-10-14 13:00:38 +08:00
|
|
|
checkpoints_loaded.clear()
|
2022-10-21 07:01:27 +08:00
|
|
|
shared.sd_model = load_model(checkpoint_info)
|
2022-10-09 18:23:30 +08:00
|
|
|
return shared.sd_model
|
2022-10-09 04:26:48 +08:00
|
|
|
|
2022-09-17 17:05:04 +08:00
|
|
|
if shared.cmd_opts.lowvram or shared.cmd_opts.medvram:
|
|
|
|
lowvram.send_everything_to_cpu()
|
|
|
|
else:
|
|
|
|
sd_model.to(devices.cpu)
|
|
|
|
|
2022-09-29 20:40:28 +08:00
|
|
|
sd_hijack.model_hijack.undo_hijack(sd_model)
|
|
|
|
|
2022-10-09 04:26:48 +08:00
|
|
|
load_model_weights(sd_model, checkpoint_info)
|
2022-09-17 17:05:04 +08:00
|
|
|
|
2022-09-29 20:40:28 +08:00
|
|
|
sd_hijack.model_hijack.hijack(sd_model)
|
|
|
|
|
2022-09-17 17:05:04 +08:00
|
|
|
if not shared.cmd_opts.lowvram and not shared.cmd_opts.medvram:
|
|
|
|
sd_model.to(devices.device)
|
|
|
|
|
|
|
|
print(f"Weights loaded.")
|
|
|
|
return sd_model
|