Mirrored_Repos/InvokeAI
1
0
Fork 0
mirror of https://github.com/invoke-ai/InvokeAI synced 2024-08-30 20:32:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

remove startup dependency on legacy models.yaml file

Browse Source
This commit is contained in:
Lincoln Stein 2024-02-21 10:18:30 -05:00 committed by Brandon Rising
parent c60cfdb285
commit 3312dcc58f
11 changed files with 11 additions and 468 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
invokeai/app/services/config/config_base.py
View File

@ -156,6 +156,7 @@ class InvokeAISettings(BaseSettings):
"lora_dir",
"embedding_dir",
"controlnet_dir",
"conf_path",
]
@classmethod

5
invokeai/app/services/config/config_default.py
View File

@ -30,7 +30,6 @@ InvokeAI:
lora_dir: null
embedding_dir: null
controlnet_dir: null
conf_path: configs/models.yaml
models_dir: models
legacy_conf_dir: configs/stable-diffusion
db_dir: databases
@ -123,7 +122,6 @@ a Path object:
root_path - path to InvokeAI root
output_path - path to default outputs directory
model_conf_path - path to models.yaml
conf - alias for the above
embedding_path - path to the embeddings directory
lora_path - path to the LoRA directory
@ -163,7 +161,6 @@ two configs are kept in separate sections of the config file:
InvokeAI:
Paths:
root: /home/lstein/invokeai-main
conf_path: configs/models.yaml
legacy_conf_dir: configs/stable-diffusion
outdir: outputs
...
@ -237,7 +234,6 @@ class InvokeAIAppConfig(InvokeAISettings):
# PATHS
root : Optional[Path] = Field(default=None, description='InvokeAI runtime root directory', json_schema_extra=Categories.Paths)
autoimport_dir : Path = Field(default=Path('autoimport'), description='Path to a directory of models files to be imported on startup.', json_schema_extra=Categories.Paths)
conf_path : Path = Field(default=Path('configs/models.yaml'), description='Path to models definition file', json_schema_extra=Categories.Paths)
models_dir : Path = Field(default=Path('models'), description='Path to the models directory', json_schema_extra=Categories.Paths)
convert_cache_dir : Path = Field(default=Path('models/.cache'), description='Path to the converted models cache directory', json_schema_extra=Categories.Paths)
legacy_conf_dir : Path = Field(default=Path('configs/stable-diffusion'), description='Path to directory of legacy checkpoint config files', json_schema_extra=Categories.Paths)
@ -301,6 +297,7 @@ class InvokeAIAppConfig(InvokeAISettings):
lora_dir : Optional[Path] = Field(default=None, description='Path to a directory of LoRA/LyCORIS models to be imported on startup.', json_schema_extra=Categories.Paths)
embedding_dir : Optional[Path] = Field(default=None, description='Path to a directory of Textual Inversion embeddings to be imported on startup.', json_schema_extra=Categories.Paths)
controlnet_dir : Optional[Path] = Field(default=None, description='Path to a directory of ControlNet embeddings to be imported on startup.', json_schema_extra=Categories.Paths)
conf_path : Path = Field(default=Path('configs/models.yaml'), description='Path to models definition file', json_schema_extra=Categories.Paths)
# this is not referred to in the source code and can be removed entirely
#free_gpu_mem : Optional[bool] = Field(default=None, description="If true, purge model from GPU after each generation.", json_schema_extra=Categories.MemoryPerformance)

1
invokeai/backend/install/check_root.py
View File

@ -8,7 +8,6 @@ from invokeai.app.services.config import InvokeAIAppConfig
def check_invokeai_root(config: InvokeAIAppConfig):
try:
assert config.model_conf_path.exists(), f"{config.model_conf_path} not found"
assert config.db_path.parent.exists(), f"{config.db_path.parent} not found"
assert config.models_path.exists(), f"{config.models_path} not found"
if not config.ignore_missing_core_models:

2
invokeai/backend/install/invokeai_configure.py
View File

@ -939,7 +939,7 @@ def main() -> None:
# run this unconditionally in case new directories need to be added
initialize_rootdir(config.root_path, opt.yes_to_all)
# this will initialize the models.yaml file if not present
# this will initialize and populate the models tables if not present
install_helper = InstallHelper(config, logger)
models_to_download = default_user_selections(opt, install_helper)

4
invokeai/backend/model_manager/merge.py
View File

@ -1,7 +1,7 @@
"""
invokeai.backend.model_manager.merge exports:
merge_diffusion_models() -- combine multiple models by location and return a pipeline object
merge_diffusion_models_and_commit() -- combine multiple models by ModelManager ID and write to models.yaml
merge_diffusion_models_and_commit() -- combine multiple models by ModelManager ID and write to the models tables
Copyright (c) 2023 Lincoln Stein and the InvokeAI Development Team
"""
@ -101,7 +101,7 @@ class ModelMerger(object):
**kwargs: Any,
) -> AnyModelConfig:
"""
:param models: up to three models, designated by their InvokeAI models.yaml model name
:param models: up to three models, designated by their registered InvokeAI model name
:param merged_model_name: name for new model
:param alpha: The interpolation parameter. Ranges from 0 to 1. It affects the ratio in which the checkpoints are merged. A 0.8 alpha
would mean that the first model checkpoints would affect the final result far less than an alpha of 0.2

2
invokeai/backend/training/textual_inversion_training.py
View File

@ -120,7 +120,7 @@ def parse_args() -> Namespace:
"--model",
type=str,
default="sd-1/main/stable-diffusion-v1-5",
help="Name of the diffusers model to train against, as defined in configs/models.yaml.",
help="Name of the diffusers model to train against.",
)
model_group.add_argument(
"--revision",

2
invokeai/frontend/install/model_install.py
View File

@ -455,7 +455,7 @@ class addModelsForm(CyclingForm, npyscreen.FormMultiPage):
selections = self.parentApp.install_selections
all_models = self.all_models
# Defined models (in INITIAL_CONFIG.yaml or models.yaml) to add/remove
# Defined models (in INITIAL_CONFIG.yaml or invokeai.db) to add/remove
ui_sections = [
self.starter_pipelines,
self.pipeline_models,

2
invokeai/frontend/merge/merge_diffusers.py
View File

@ -435,7 +435,7 @@ def main():
run_cli(args)
except widget.NotEnoughSpaceForWidget as e:
if str(e).startswith("Height of 1 allocated"):
logger.error("You need to have at least two diffusers models defined in models.yaml in order to merge")
logger.error("You need to have at least two diffusers models in order to merge")
else:
logger.error("Not enough room for the user interface. Try making this window larger.")
sys.exit(-1)

4
invokeai/frontend/training/textual_inversion.py Executable file → Normal file
View File

@ -261,7 +261,7 @@ class textualInversionForm(npyscreen.FormMultiPageAction):
def validate_field_values(self) -> bool:
bad_fields = []
if self.model.value is None:
bad_fields.append("Model Name must correspond to a known model in models.yaml")
bad_fields.append("Model Name must correspond to a known model in invokeai.db")
if not re.match("^[a-zA-Z0-9.-]+$", self.placeholder_token.value):
bad_fields.append("Trigger term must only contain alphanumeric characters, the dot and hyphen")
if self.train_data_dir.value is None:
@ -442,7 +442,7 @@ def main() -> None:
pass
except (widget.NotEnoughSpaceForWidget, Exception) as e:
if str(e).startswith("Height of 1 allocated"):
logger.error("You need to have at least one diffusers models defined in models.yaml in order to train")
logger.error("You need to have at least one diffusers models defined in invokeai.db in order to train")
elif str(e).startswith("addwstr"):
logger.error("Not enough window space for the interface. Please make your window larger and try again.")
else:

454
invokeai/frontend/training/textual_inversion2.py
View File

@ -1,454 +0,0 @@
#!/usr/bin/env python
"""
This is the frontend to "textual_inversion_training.py".
Copyright (c) 2023-24 Lincoln Stein and the InvokeAI Development Team
"""
import os
import re
import shutil
import sys
import traceback
from argparse import Namespace
from pathlib import Path
from typing import Dict, List, Optional, Tuple
import npyscreen
from npyscreen import widget
from omegaconf import OmegaConf
import invokeai.backend.util.logging as logger
from invokeai.app.services.config import InvokeAIAppConfig
from invokeai.backend.install.install_helper import initialize_installer
from invokeai.backend.model_manager import ModelType
from invokeai.backend.training import do_textual_inversion_training, parse_args
TRAINING_DATA = "text-inversion-training-data"
TRAINING_DIR = "text-inversion-output"
CONF_FILE = "preferences.conf"
config = None
class textualInversionForm(npyscreen.FormMultiPageAction):
resolutions = [512, 768, 1024]
lr_schedulers = [
"linear",
"cosine",
"cosine_with_restarts",
"polynomial",
"constant",
"constant_with_warmup",
]
precisions = ["no", "fp16", "bf16"]
learnable_properties = ["object", "style"]
def __init__(self, parentApp: npyscreen.NPSAppManaged, name: str, saved_args: Optional[Dict[str, str]] = None):
self.saved_args = saved_args or {}
super().__init__(parentApp, name)
def afterEditing(self) -> None:
self.parentApp.setNextForm(None)
def create(self) -> None:
self.model_names, default = self.get_model_names()
default_initializer_token = "★"
default_placeholder_token = ""
saved_args = self.saved_args
assert config is not None
try:
default = self.model_names.index(saved_args["model"])
except Exception:
pass
self.add_widget_intelligent(
npyscreen.FixedText,
value="Use ctrl-N and ctrl-P to move to the <N>ext and <P>revious fields, cursor arrows to make a selection, and space to toggle checkboxes.",
editable=False,
)
self.model = self.add_widget_intelligent(
npyscreen.TitleSelectOne,
name="Model Name:",
values=sorted(self.model_names),
value=default,
max_height=len(self.model_names) + 1,
scroll_exit=True,
)
self.placeholder_token = self.add_widget_intelligent(
npyscreen.TitleText,
name="Trigger Term:",
value="", # saved_args.get('placeholder_token',''), # to restore previous term
scroll_exit=True,
)
self.placeholder_token.when_value_edited = self.initializer_changed
self.nextrely -= 1
self.nextrelx += 30
self.prompt_token = self.add_widget_intelligent(
npyscreen.FixedText,
name="Trigger term for use in prompt",
value="",
editable=False,
scroll_exit=True,
)
self.nextrelx -= 30
self.initializer_token = self.add_widget_intelligent(
npyscreen.TitleText,
name="Initializer:",
value=saved_args.get("initializer_token", default_initializer_token),
scroll_exit=True,
)
self.resume_from_checkpoint = self.add_widget_intelligent(
npyscreen.Checkbox,
name="Resume from last saved checkpoint",
value=False,
scroll_exit=True,
)
self.learnable_property = self.add_widget_intelligent(
npyscreen.TitleSelectOne,
name="Learnable property:",
values=self.learnable_properties,
value=self.learnable_properties.index(saved_args.get("learnable_property", "object")),
max_height=4,
scroll_exit=True,
)
self.train_data_dir = self.add_widget_intelligent(
npyscreen.TitleFilename,
name="Data Training Directory:",
select_dir=True,
must_exist=False,
value=str(
saved_args.get(
"train_data_dir",
config.root_dir / TRAINING_DATA / default_placeholder_token,
)
),
scroll_exit=True,
)
self.output_dir = self.add_widget_intelligent(
npyscreen.TitleFilename,
name="Output Destination Directory:",
select_dir=True,
must_exist=False,
value=str(
saved_args.get(
"output_dir",
config.root_dir / TRAINING_DIR / default_placeholder_token,
)
),
scroll_exit=True,
)
self.resolution = self.add_widget_intelligent(
npyscreen.TitleSelectOne,
name="Image resolution (pixels):",
values=self.resolutions,
value=self.resolutions.index(saved_args.get("resolution", 512)),
max_height=4,
scroll_exit=True,
)
self.center_crop = self.add_widget_intelligent(
npyscreen.Checkbox,
name="Center crop images before resizing to resolution",
value=saved_args.get("center_crop", False),
scroll_exit=True,
)
self.mixed_precision = self.add_widget_intelligent(
npyscreen.TitleSelectOne,
name="Mixed Precision:",
values=self.precisions,
value=self.precisions.index(saved_args.get("mixed_precision", "fp16")),
max_height=4,
scroll_exit=True,
)
self.num_train_epochs = self.add_widget_intelligent(
npyscreen.TitleSlider,
name="Number of training epochs:",
out_of=1000,
step=50,
lowest=1,
value=saved_args.get("num_train_epochs", 100),
scroll_exit=True,
)
self.max_train_steps = self.add_widget_intelligent(
npyscreen.TitleSlider,
name="Max Training Steps:",
out_of=10000,
step=500,
lowest=1,
value=saved_args.get("max_train_steps", 3000),
scroll_exit=True,
)
self.train_batch_size = self.add_widget_intelligent(
npyscreen.TitleSlider,
name="Batch Size (reduce if you run out of memory):",
out_of=50,
step=1,
lowest=1,
value=saved_args.get("train_batch_size", 8),
scroll_exit=True,
)
self.gradient_accumulation_steps = self.add_widget_intelligent(
npyscreen.TitleSlider,
name="Gradient Accumulation Steps (may need to decrease this to resume from a checkpoint):",
out_of=10,
step=1,
lowest=1,
value=saved_args.get("gradient_accumulation_steps", 4),
scroll_exit=True,
)
self.lr_warmup_steps = self.add_widget_intelligent(
npyscreen.TitleSlider,
name="Warmup Steps:",
out_of=100,
step=1,
lowest=0,
value=saved_args.get("lr_warmup_steps", 0),
scroll_exit=True,
)
self.learning_rate = self.add_widget_intelligent(
npyscreen.TitleText,
name="Learning Rate:",
value=str(
saved_args.get("learning_rate", "5.0e-04"),
),
scroll_exit=True,
)
self.scale_lr = self.add_widget_intelligent(
npyscreen.Checkbox,
name="Scale learning rate by number GPUs, steps and batch size",
value=saved_args.get("scale_lr", True),
scroll_exit=True,
)
self.enable_xformers_memory_efficient_attention = self.add_widget_intelligent(
npyscreen.Checkbox,
name="Use xformers acceleration",
value=saved_args.get("enable_xformers_memory_efficient_attention", False),
scroll_exit=True,
)
self.lr_scheduler = self.add_widget_intelligent(
npyscreen.TitleSelectOne,
name="Learning rate scheduler:",
values=self.lr_schedulers,
max_height=7,
value=self.lr_schedulers.index(saved_args.get("lr_scheduler", "constant")),
scroll_exit=True,
)
self.model.editing = True
def initializer_changed(self) -> None:
placeholder = self.placeholder_token.value
self.prompt_token.value = f"(Trigger by using <{placeholder}> in your prompts)"
self.train_data_dir.value = str(config.root_dir / TRAINING_DATA / placeholder)
self.output_dir.value = str(config.root_dir / TRAINING_DIR / placeholder)
self.resume_from_checkpoint.value = Path(self.output_dir.value).exists()
def on_ok(self):
if self.validate_field_values():
self.parentApp.setNextForm(None)
self.editing = False
self.parentApp.ti_arguments = self.marshall_arguments()
npyscreen.notify("Launching textual inversion training. This will take a while...")
else:
self.editing = True
def ok_cancel(self):
sys.exit(0)
def validate_field_values(self) -> bool:
bad_fields = []
if self.model.value is None:
bad_fields.append("Model Name must correspond to a known model in models.yaml")
if not re.match("^[a-zA-Z0-9.-]+$", self.placeholder_token.value):
bad_fields.append("Trigger term must only contain alphanumeric characters, the dot and hyphen")
if self.train_data_dir.value is None:
bad_fields.append("Data Training Directory cannot be empty")
if self.output_dir.value is None:
bad_fields.append("The Output Destination Directory cannot be empty")
if len(bad_fields) > 0:
message = "The following problems were detected and must be corrected:"
for problem in bad_fields:
message += f"\n* {problem}"
npyscreen.notify_confirm(message)
return False
else:
return True
def get_model_names(self) -> Tuple[List[str], int]:
global config
assert config is not None
installer = initialize_installer(config)
store = installer.record_store
main_models = store.search_by_attr(model_type=ModelType.Main)
model_names = [f"{x.base.value}/{x.type.value}/{x.name}" for x in main_models if x.format == "diffusers"]
default = 0
return (model_names, default)
def marshall_arguments(self) -> dict:
args = {}
# the choices
args.update(
model=self.model_names[self.model.value[0]],
resolution=self.resolutions[self.resolution.value[0]],
lr_scheduler=self.lr_schedulers[self.lr_scheduler.value[0]],
mixed_precision=self.precisions[self.mixed_precision.value[0]],
learnable_property=self.learnable_properties[self.learnable_property.value[0]],
)
# all the strings and booleans
for attr in (
"initializer_token",
"placeholder_token",
"train_data_dir",
"output_dir",
"scale_lr",
"center_crop",
"enable_xformers_memory_efficient_attention",
):
args[attr] = getattr(self, attr).value
# all the integers
for attr in (
"train_batch_size",
"gradient_accumulation_steps",
"num_train_epochs",
"max_train_steps",
"lr_warmup_steps",
):
args[attr] = int(getattr(self, attr).value)
# the floats (just one)
args.update(learning_rate=float(self.learning_rate.value))
# a special case
if self.resume_from_checkpoint.value and Path(self.output_dir.value).exists():
args["resume_from_checkpoint"] = "latest"
return args
class MyApplication(npyscreen.NPSAppManaged):
def __init__(self, saved_args: Optional[Dict[str, str]] = None):
super().__init__()
self.ti_arguments = None
self.saved_args = saved_args
def onStart(self):
npyscreen.setTheme(npyscreen.Themes.DefaultTheme)
self.main = self.addForm(
"MAIN",
textualInversionForm,
name="Textual Inversion Settings",
saved_args=self.saved_args,
)
def copy_to_embeddings_folder(args: Dict[str, str]) -> None:
"""
Copy learned_embeds.bin into the embeddings folder, and offer to
delete the full model and checkpoints.
"""
assert config is not None
source = Path(args["output_dir"], "learned_embeds.bin")
dest_dir_name = args["placeholder_token"].strip("<>")
destination = config.root_dir / "embeddings" / dest_dir_name
os.makedirs(destination, exist_ok=True)
logger.info(f"Training completed. Copying learned_embeds.bin into {str(destination)}")
shutil.copy(source, destination)
if (input("Delete training logs and intermediate checkpoints? [y] ") or "y").startswith(("y", "Y")):
shutil.rmtree(Path(args["output_dir"]))
else:
logger.info(f'Keeping {args["output_dir"]}')
def save_args(args: dict) -> None:
"""
Save the current argument values to an omegaconf file
"""
assert config is not None
dest_dir = config.root_dir / TRAINING_DIR
os.makedirs(dest_dir, exist_ok=True)
conf_file = dest_dir / CONF_FILE
conf = OmegaConf.create(args)
OmegaConf.save(config=conf, f=conf_file)
def previous_args() -> dict:
"""
Get the previous arguments used.
"""
assert config is not None
conf_file = config.root_dir / TRAINING_DIR / CONF_FILE
try:
conf = OmegaConf.load(conf_file)
conf["placeholder_token"] = conf["placeholder_token"].strip("<>")
except Exception:
conf = None
return conf
def do_front_end() -> None:
global config
saved_args = previous_args()
myapplication = MyApplication(saved_args=saved_args)
myapplication.run()
if my_args := myapplication.ti_arguments:
os.makedirs(my_args["output_dir"], exist_ok=True)
# Automatically add angle brackets around the trigger
if not re.match("^<.+>$", my_args["placeholder_token"]):
my_args["placeholder_token"] = f"<{my_args['placeholder_token']}>"
my_args["only_save_embeds"] = True
save_args(my_args)
try:
print(my_args)
do_textual_inversion_training(config, **my_args)
copy_to_embeddings_folder(my_args)
except Exception as e:
logger.error("An exception occurred during training. The exception was:")
logger.error(str(e))
logger.error("DETAILS:")
logger.error(traceback.format_exc())
def main() -> None:
global config
args: Namespace = parse_args()
config = InvokeAIAppConfig.get_config()
config.parse_args([])
# change root if needed
if args.root_dir:
config.root = args.root_dir
try:
if args.front_end:
do_front_end()
else:
do_textual_inversion_training(config, **vars(args))
except AssertionError as e:
logger.error(e)
sys.exit(-1)
except KeyboardInterrupt:
pass
except (widget.NotEnoughSpaceForWidget, Exception) as e:
if str(e).startswith("Height of 1 allocated"):
logger.error("You need to have at least one diffusers models defined in models.yaml in order to train")
elif str(e).startswith("addwstr"):
logger.error("Not enough window space for the interface. Please make your window larger and try again.")
else:
logger.error(e)
sys.exit(-1)
if __name__ == "__main__":
main()

2
invokeai/frontend/web/public/locales/en.json
View File

@ -814,7 +814,7 @@
"simpleModelDesc": "Provide a path to a local Diffusers model, local checkpoint / safetensors model a HuggingFace Repo ID, or a checkpoint/diffusers model URL.",
"statusConverting": "Converting",
"syncModels": "Sync Models",
"syncModelsDesc": "If your models are out of sync with the backend, you can refresh them up using this option. This is generally handy in cases where you manually update your models.yaml file or add models to the InvokeAI root folder after the application has booted.",
"syncModelsDesc": "If your models are out of sync with the backend, you can refresh them up using this option. This is generally handy in cases where you add models to the InvokeAI root folder or autoimport directory after the application has booted.",
"updateModel": "Update Model",
"useCustomConfig": "Use Custom Config",
"v1": "v1",