2022-09-14 07:40:25 +00:00
|
|
|
import logging
|
2022-10-11 02:50:11 +00:00
|
|
|
import logging.config
|
2022-09-17 05:21:20 +00:00
|
|
|
import re
|
2022-10-11 02:50:11 +00:00
|
|
|
import warnings
|
2022-09-14 07:40:25 +00:00
|
|
|
|
|
|
|
import torch
|
2022-10-11 02:50:11 +00:00
|
|
|
from pytorch_lightning import _logger as pytorch_logger
|
2022-09-17 05:21:20 +00:00
|
|
|
from torchvision.transforms import ToPILImage
|
2022-10-11 02:50:11 +00:00
|
|
|
from transformers.modeling_utils import logger as modeling_logger
|
|
|
|
from transformers.utils.logging import _configure_library_root_logger
|
2022-09-24 05:58:48 +00:00
|
|
|
|
2022-09-14 07:40:25 +00:00
|
|
|
_CURRENT_LOGGING_CONTEXT = None
|
|
|
|
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
|
|
|
|
|
2022-09-17 05:21:20 +00:00
|
|
|
def log_conditioning(conditioning, description):
|
|
|
|
if _CURRENT_LOGGING_CONTEXT is None:
|
|
|
|
return
|
|
|
|
|
|
|
|
_CURRENT_LOGGING_CONTEXT.log_conditioning(conditioning, description)
|
|
|
|
|
|
|
|
|
2022-09-14 07:40:25 +00:00
|
|
|
def log_latent(latents, description):
|
|
|
|
if _CURRENT_LOGGING_CONTEXT is None:
|
|
|
|
return
|
2022-09-21 04:15:19 +00:00
|
|
|
|
2022-09-14 07:40:25 +00:00
|
|
|
_CURRENT_LOGGING_CONTEXT.log_latents(latents, description)
|
|
|
|
|
|
|
|
|
2022-09-18 13:07:07 +00:00
|
|
|
def log_img(img, description):
|
|
|
|
if _CURRENT_LOGGING_CONTEXT is None:
|
|
|
|
return
|
|
|
|
_CURRENT_LOGGING_CONTEXT.log_img(img, description)
|
|
|
|
|
|
|
|
|
2022-10-14 03:49:48 +00:00
|
|
|
def log_tensor(t, description=""):
|
|
|
|
if _CURRENT_LOGGING_CONTEXT is None:
|
|
|
|
return
|
|
|
|
_CURRENT_LOGGING_CONTEXT.log_img(t, description)
|
|
|
|
|
|
|
|
|
2022-09-17 05:21:20 +00:00
|
|
|
class ImageLoggingContext:
|
|
|
|
def __init__(self, prompt, model, img_callback=None, img_outdir=None):
|
2022-09-14 07:40:25 +00:00
|
|
|
self.prompt = prompt
|
|
|
|
self.model = model
|
|
|
|
self.step_count = 0
|
|
|
|
self.img_callback = img_callback
|
2022-09-17 05:21:20 +00:00
|
|
|
self.img_outdir = img_outdir
|
2022-09-14 07:40:25 +00:00
|
|
|
|
|
|
|
def __enter__(self):
|
2022-09-16 16:24:24 +00:00
|
|
|
global _CURRENT_LOGGING_CONTEXT # noqa
|
2022-09-14 07:40:25 +00:00
|
|
|
_CURRENT_LOGGING_CONTEXT = self
|
|
|
|
return self
|
|
|
|
|
|
|
|
def __exit__(self, exc_type, exc_val, exc_tb):
|
2022-09-16 16:24:24 +00:00
|
|
|
global _CURRENT_LOGGING_CONTEXT # noqa
|
2022-09-14 07:40:25 +00:00
|
|
|
_CURRENT_LOGGING_CONTEXT = None
|
|
|
|
|
2022-09-17 05:21:20 +00:00
|
|
|
def log_conditioning(self, conditioning, description):
|
|
|
|
if not self.img_callback:
|
|
|
|
return
|
|
|
|
img = conditioning_to_img(conditioning)
|
|
|
|
|
|
|
|
self.img_callback(img, description, self.step_count, self.prompt)
|
|
|
|
|
2022-09-15 02:40:50 +00:00
|
|
|
def log_latents(self, latents, description):
|
2022-10-11 02:50:11 +00:00
|
|
|
from imaginairy.img_utils import model_latents_to_pillow_imgs # noqa
|
|
|
|
|
2022-09-14 07:40:25 +00:00
|
|
|
if not self.img_callback:
|
|
|
|
return
|
2022-09-15 02:40:50 +00:00
|
|
|
if latents.shape[1] != 4:
|
2022-09-14 07:40:25 +00:00
|
|
|
# logger.info(f"Didn't save tensor of shape {samples.shape} for {description}")
|
|
|
|
return
|
|
|
|
self.step_count += 1
|
2022-10-14 03:49:48 +00:00
|
|
|
try:
|
|
|
|
shape_str = ",".join(tuple(latents.shape))
|
|
|
|
except TypeError:
|
|
|
|
shape_str = str(latents.shape)
|
|
|
|
description = f"{description}-{shape_str}"
|
2022-09-24 05:58:48 +00:00
|
|
|
for img in model_latents_to_pillow_imgs(latents):
|
2022-09-14 07:40:25 +00:00
|
|
|
self.img_callback(img, description, self.step_count, self.prompt)
|
2022-09-17 05:21:20 +00:00
|
|
|
|
2022-09-18 13:07:07 +00:00
|
|
|
def log_img(self, img, description):
|
|
|
|
if not self.img_callback:
|
|
|
|
return
|
|
|
|
self.step_count += 1
|
|
|
|
if isinstance(img, torch.Tensor):
|
|
|
|
img = ToPILImage()(img.squeeze().cpu().detach())
|
|
|
|
img = img.copy()
|
|
|
|
self.img_callback(img, description, self.step_count, self.prompt)
|
|
|
|
|
2022-10-14 03:49:48 +00:00
|
|
|
def log_tensor(self, t, description=""):
|
|
|
|
if not self.img_callback:
|
|
|
|
return
|
|
|
|
|
|
|
|
if len(t.shape) == 2:
|
|
|
|
self.log_img(t, description)
|
|
|
|
|
|
|
|
def log_indexed_graph_of_tensor(self):
|
|
|
|
pass
|
|
|
|
|
2022-09-17 05:21:20 +00:00
|
|
|
# def img_callback(self, img, description, step_count, prompt):
|
|
|
|
# steps_path = os.path.join(self.img_outdir, "steps", f"{self.file_num:08}_S{prompt.seed}")
|
|
|
|
# os.makedirs(steps_path, exist_ok=True)
|
|
|
|
# filename = f"{self.file_num:08}_S{prompt.seed}_step{step_count:04}_{filesafe_text(description)[:40]}.jpg"
|
|
|
|
# destination = os.path.join(steps_path, filename)
|
|
|
|
# draw = ImageDraw.Draw(img)
|
|
|
|
# draw.text((10, 10), str(description))
|
|
|
|
# img.save(destination)
|
|
|
|
|
|
|
|
|
|
|
|
def filesafe_text(t):
|
|
|
|
return re.sub(r"[^a-zA-Z0-9.,\[\]() -]+", "_", t)[:130]
|
|
|
|
|
|
|
|
|
|
|
|
def conditioning_to_img(conditioning):
|
|
|
|
return ToPILImage()(conditioning)
|
2022-10-11 02:50:11 +00:00
|
|
|
|
|
|
|
|
2022-10-11 03:13:32 +00:00
|
|
|
class IndentingFormatter(logging.Formatter):
|
|
|
|
def format(self, record):
|
|
|
|
s = super().format(record)
|
|
|
|
if _CURRENT_LOGGING_CONTEXT is not None:
|
|
|
|
s = f" {s}"
|
|
|
|
return s
|
|
|
|
|
|
|
|
|
2022-10-11 02:50:11 +00:00
|
|
|
def configure_logging(level="INFO"):
|
|
|
|
fmt = "%(message)s"
|
|
|
|
if level == "DEBUG":
|
|
|
|
fmt = "%(asctime)s [%(levelname)s] %(name)s:%(lineno)d: %(message)s"
|
|
|
|
|
|
|
|
LOGGING_CONFIG = {
|
|
|
|
"version": 1,
|
|
|
|
"disable_existing_loggers": True,
|
|
|
|
"formatters": {
|
2022-10-11 03:13:32 +00:00
|
|
|
"standard": {
|
|
|
|
"format": fmt,
|
|
|
|
"class": "imaginairy.log_utils.IndentingFormatter",
|
|
|
|
},
|
2022-10-11 02:50:11 +00:00
|
|
|
},
|
|
|
|
"handlers": {
|
|
|
|
"default": {
|
|
|
|
"level": "INFO",
|
|
|
|
"formatter": "standard",
|
|
|
|
"class": "logging.StreamHandler",
|
|
|
|
"stream": "ext://sys.stdout", # Default is stderr
|
|
|
|
},
|
|
|
|
},
|
|
|
|
"loggers": {
|
|
|
|
"": { # root logger
|
|
|
|
"handlers": ["default"],
|
|
|
|
"level": "WARNING",
|
|
|
|
"propagate": False,
|
|
|
|
},
|
|
|
|
"imaginairy": {"handlers": ["default"], "level": level, "propagate": False},
|
|
|
|
"transformers.modeling_utils": {
|
|
|
|
"handlers": ["default"],
|
|
|
|
"level": "ERROR",
|
|
|
|
"propagate": False,
|
|
|
|
},
|
|
|
|
},
|
|
|
|
}
|
|
|
|
suppress_annoying_logs_and_warnings()
|
|
|
|
logging.config.dictConfig(LOGGING_CONFIG)
|
|
|
|
|
|
|
|
|
|
|
|
def disable_transformers_custom_logging():
|
|
|
|
_configure_library_root_logger()
|
|
|
|
_logger = modeling_logger.parent
|
|
|
|
_logger.handlers = []
|
|
|
|
_logger.propagate = True
|
|
|
|
_logger.setLevel(logging.NOTSET)
|
2022-10-16 23:42:46 +00:00
|
|
|
modeling_logger.handlers = []
|
|
|
|
modeling_logger.propagate = True
|
|
|
|
modeling_logger.setLevel(logging.ERROR)
|
2022-10-11 02:50:11 +00:00
|
|
|
|
|
|
|
|
|
|
|
def disable_pytorch_lighting_custom_logging():
|
|
|
|
pytorch_logger.setLevel(logging.NOTSET)
|
|
|
|
|
|
|
|
|
|
|
|
def disable_common_warnings():
|
|
|
|
warnings.filterwarnings(
|
|
|
|
"ignore",
|
|
|
|
category=UserWarning,
|
|
|
|
message=r"The operator .*?is not currently supported.*",
|
|
|
|
)
|
|
|
|
warnings.filterwarnings(
|
|
|
|
"ignore", category=UserWarning, message=r"The parameter 'pretrained' is.*"
|
|
|
|
)
|
|
|
|
warnings.filterwarnings(
|
|
|
|
"ignore", category=UserWarning, message=r"Arguments other than a weight.*"
|
|
|
|
)
|
|
|
|
warnings.filterwarnings("ignore", category=DeprecationWarning)
|
|
|
|
|
|
|
|
|
|
|
|
def suppress_annoying_logs_and_warnings():
|
|
|
|
disable_transformers_custom_logging()
|
|
|
|
disable_pytorch_lighting_custom_logging()
|
|
|
|
disable_common_warnings()
|