github/ai-toolkit
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Big refactor of SD runner and added image generator

Browse Source
This commit is contained in:
Jaret Burkett
2023-08-03 14:51:25 -06:00
parent 75ec5d9292
commit 66c6f0f6f7
16 changed files with 923 additions and 430 deletions
Download Patch File Download Diff File Expand all files Collapse all files

443
jobs/process/BaseSDTrainProcess.py
View File

@@ -1,34 +1,23 @@
import glob
import time
from collections import OrderedDict
import os
from diffusers.pipelines.stable_diffusion_xl.pipeline_stable_diffusion_xl import rescale_noise_cfg
from toolkit.lora_special import LoRASpecialNetwork
from toolkit.optimizer import get_optimizer
from toolkit.paths import REPOS_ROOT
import sys
from toolkit.pipelines import CustomStableDiffusionXLPipeline, CustomStableDiffusionPipeline
sys.path.append(REPOS_ROOT)
sys.path.append(os.path.join(REPOS_ROOT, 'leco'))
from diffusers import StableDiffusionPipeline, StableDiffusionXLPipeline, KDPM2DiscreteScheduler, PNDMScheduler, \
DDIMScheduler, DDPMScheduler
from toolkit.scheduler import get_lr_scheduler
from toolkit.stable_diffusion_model import StableDiffusion
from jobs.process import BaseTrainProcess
from toolkit.metadata import get_meta_for_safetensors, load_metadata_from_safetensors, add_base_model_info_to_meta
from toolkit.train_tools import get_torch_dtype, apply_noise_offset
from toolkit.train_tools import get_torch_dtype
import gc
import torch
from tqdm import tqdm
from leco import train_util, model_util
from toolkit.config_modules import SaveConfig, LogingConfig, SampleConfig, NetworkConfig, TrainConfig, ModelConfig
from toolkit.stable_diffusion_model import StableDiffusion, PromptEmbeds
from toolkit.config_modules import SaveConfig, LogingConfig, SampleConfig, NetworkConfig, TrainConfig, ModelConfig, \
GenerateImageConfig
def flush():
@@ -36,11 +25,9 @@ def flush():
gc.collect()
UNET_IN_CHANNELS = 4 # Stable Diffusion の in_channels は 4 で固定。XLも同じ。
VAE_SCALE_FACTOR = 8 # 2 ** (len(vae.config.block_out_channels) - 1) = 8
class BaseSDTrainProcess(BaseTrainProcess):
sd: StableDiffusion
def __init__(self, process_id: int, job, config: OrderedDict, custom_pipeline=None):
super().__init__(process_id, job, config)
self.custom_pipeline = custom_pipeline
@@ -64,177 +51,52 @@ class BaseSDTrainProcess(BaseTrainProcess):
self.logging_config = LogingConfig(**self.get_conf('logging', {}))
self.optimizer = None
self.lr_scheduler = None
self.sd: 'StableDiffusion' = None
# sdxl stuff
self.logit_scale = None
self.ckppt_info = None
self.sd = StableDiffusion(
device=self.device,
model_config=self.model_config,
dtype=self.train_config.dtype,
custom_pipeline=self.custom_pipeline,
)
# added later
# to hold network if there is one
self.network = None
def sample(self, step=None, is_first=False):
sample_folder = os.path.join(self.save_root, 'samples')
if not os.path.exists(sample_folder):
os.makedirs(sample_folder, exist_ok=True)
if self.network is not None:
self.network.eval()
# save current seed state for training
rng_state = torch.get_rng_state()
cuda_rng_state = torch.cuda.get_rng_state() if torch.cuda.is_available() else None
original_device_dict = {
'vae': self.sd.vae.device,
'unet': self.sd.unet.device,
# 'tokenizer': self.sd.tokenizer.device,
}
# handle sdxl text encoder
if isinstance(self.sd.text_encoder, list):
for encoder, i in zip(self.sd.text_encoder, range(len(self.sd.text_encoder))):
original_device_dict[f'text_encoder_{i}'] = encoder.device
encoder.to(self.device_torch)
else:
original_device_dict['text_encoder'] = self.sd.text_encoder.device
self.sd.text_encoder.to(self.device_torch)
self.sd.vae.to(self.device_torch)
self.sd.unet.to(self.device_torch)
# self.sd.text_encoder.to(self.device_torch)
# self.sd.tokenizer.to(self.device_torch)
# TODO add clip skip
if self.sd.is_xl:
pipeline = StableDiffusionXLPipeline(
vae=self.sd.vae,
unet=self.sd.unet,
text_encoder=self.sd.text_encoder[0],
text_encoder_2=self.sd.text_encoder[1],
tokenizer=self.sd.tokenizer[0],
tokenizer_2=self.sd.tokenizer[1],
scheduler=self.sd.noise_scheduler,
).to(self.device_torch)
else:
pipeline = StableDiffusionPipeline(
vae=self.sd.vae,
unet=self.sd.unet,
text_encoder=self.sd.text_encoder,
tokenizer=self.sd.tokenizer,
scheduler=self.sd.noise_scheduler,
safety_checker=None,
feature_extractor=None,
requires_safety_checker=False,
).to(self.device_torch)
# disable progress bar
pipeline.set_progress_bar_config(disable=True)
gen_img_config_list = []
sample_config = self.first_sample_config if is_first else self.sample_config
start_seed = sample_config.seed
start_multiplier = self.network.multiplier
current_seed = start_seed
for i in range(len(sample_config.prompts)):
if sample_config.walk_seed:
current_seed = start_seed + i
pipeline.to(self.device_torch)
with self.network:
with torch.no_grad():
if self.network is not None:
assert self.network.is_active
if self.logging_config.verbose:
print("network_state", {
'is_active': self.network.is_active,
'multiplier': self.network.multiplier,
})
step_num = ''
if step is not None:
# zero-pad 9 digits
step_num = f"_{str(step).zfill(9)}"
for i in tqdm(range(len(sample_config.prompts)), desc=f"Generating Samples - step: {step}",
leave=False):
raw_prompt = sample_config.prompts[i]
filename = f"[time]_{step_num}_[count].png"
neg = sample_config.neg
multiplier = sample_config.network_multiplier
p_split = raw_prompt.split('--')
prompt = p_split[0].strip()
height = sample_config.height
width = sample_config.width
output_path = os.path.join(sample_folder, filename)
if len(p_split) > 1:
for split in p_split:
flag = split[:1]
content = split[1:].strip()
if flag == 'n':
neg = content
elif flag == 'm':
# multiplier
multiplier = float(content)
elif flag == 'w':
# multiplier
width = int(content)
elif flag == 'h':
# multiplier
height = int(content)
gen_img_config_list.append(GenerateImageConfig(
prompt=sample_config.prompts[i], # it will autoparse the prompt
width=sample_config.width,
height=sample_config.height,
negative_prompt=sample_config.neg,
seed=current_seed,
guidance_scale=sample_config.guidance_scale,
guidance_rescale=sample_config.guidance_rescale,
num_inference_steps=sample_config.sample_steps,
network_multiplier=sample_config.network_multiplier,
output_path=output_path,
))
height = max(64, height - height % 8) # round to divisible by 8
width = max(64, width - width % 8) # round to divisible by 8
if sample_config.walk_seed:
current_seed += i
if self.network is not None:
self.network.multiplier = multiplier
torch.manual_seed(current_seed)
torch.cuda.manual_seed(current_seed)
if self.sd.is_xl:
img = pipeline(
prompt,
height=height,
width=width,
num_inference_steps=sample_config.sample_steps,
guidance_scale=sample_config.guidance_scale,
negative_prompt=neg,
guidance_rescale=0.7,
).images[0]
else:
img = pipeline(
prompt,
height=height,
width=width,
num_inference_steps=sample_config.sample_steps,
guidance_scale=sample_config.guidance_scale,
negative_prompt=neg,
).images[0]
step_num = ''
if step is not None:
# zero-pad 9 digits
step_num = f"_{str(step).zfill(9)}"
seconds_since_epoch = int(time.time())
# zero-pad 2 digits
i_str = str(i).zfill(2)
filename = f"{seconds_since_epoch}{step_num}_{i_str}.png"
output_path = os.path.join(sample_folder, filename)
img.save(output_path)
# clear pipeline and cache to reduce vram usage
del pipeline
torch.cuda.empty_cache()
# restore training state
torch.set_rng_state(rng_state)
if cuda_rng_state is not None:
torch.cuda.set_rng_state(cuda_rng_state)
self.sd.vae.to(original_device_dict['vae'])
self.sd.unet.to(original_device_dict['unet'])
if isinstance(self.sd.text_encoder, list):
for encoder, i in zip(self.sd.text_encoder, range(len(self.sd.text_encoder))):
encoder.to(original_device_dict[f'text_encoder_{i}'])
else:
self.sd.text_encoder.to(original_device_dict['text_encoder'])
if self.network is not None:
self.network.train()
self.network.multiplier = start_multiplier
# self.sd.tokenizer.to(original_device_dict['tokenizer'])
# send to be generated
self.sd.generate_images(gen_img_config_list)
def update_training_metadata(self):
o_dict = OrderedDict({
@@ -328,148 +190,10 @@ class BaseSDTrainProcess(BaseTrainProcess):
def hook_before_train_loop(self):
pass
def get_latent_noise(
self,
height=None,
width=None,
pixel_height=None,
pixel_width=None,
):
if height is None and pixel_height is None:
raise ValueError("height or pixel_height must be specified")
if width is None and pixel_width is None:
raise ValueError("width or pixel_width must be specified")
if height is None:
height = pixel_height // VAE_SCALE_FACTOR
if width is None:
width = pixel_width // VAE_SCALE_FACTOR
noise = torch.randn(
(
self.train_config.batch_size,
UNET_IN_CHANNELS,
height,
width,
),
device="cpu",
)
noise = apply_noise_offset(noise, self.train_config.noise_offset)
return noise
def hook_train_loop(self):
# return loss
return 0.0
def get_time_ids_from_latents(self, latents):
bs, ch, h, w = list(latents.shape)
height = h * VAE_SCALE_FACTOR
width = w * VAE_SCALE_FACTOR
dtype = get_torch_dtype(self.train_config.dtype)
if self.sd.is_xl:
prompt_ids = train_util.get_add_time_ids(
height,
width,
dynamic_crops=False, # look into this
dtype=dtype,
).to(self.device_torch, dtype=dtype)
return train_util.concat_embeddings(
prompt_ids, prompt_ids, bs
)
else:
return None
def predict_noise(
self,
latents: torch.FloatTensor,
text_embeddings: PromptEmbeds,
timestep: int,
guidance_scale=7.5,
guidance_rescale=0, # 0.7
add_time_ids=None,
**kwargs,
):
if self.sd.is_xl:
if add_time_ids is None:
add_time_ids = self.get_time_ids_from_latents(latents)
latent_model_input = torch.cat([latents] * 2)
latent_model_input = self.sd.noise_scheduler.scale_model_input(latent_model_input, timestep)
added_cond_kwargs = {
"text_embeds": text_embeddings.pooled_embeds,
"time_ids": add_time_ids,
}
# predict the noise residual
noise_pred = self.sd.unet(
latent_model_input,
timestep,
encoder_hidden_states=text_embeddings.text_embeds,
added_cond_kwargs=added_cond_kwargs,
).sample
# perform guidance
noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
noise_pred = noise_pred_uncond + guidance_scale * (
noise_pred_text - noise_pred_uncond
)
# https://github.com/huggingface/diffusers/blob/7a91ea6c2b53f94da930a61ed571364022b21044/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl.py#L775
if guidance_rescale > 0.0:
# Based on 3.4. in https://arxiv.org/pdf/2305.08891.pdf
noise_pred = rescale_noise_cfg(noise_pred, noise_pred_text, guidance_rescale=guidance_rescale)
else:
# if we are doing classifier free guidance, need to double up
latent_model_input = torch.cat([latents] * 2)
latent_model_input = self.sd.noise_scheduler.scale_model_input(latent_model_input, timestep)
# predict the noise residual
noise_pred = self.sd.unet(
latent_model_input,
timestep,
encoder_hidden_states=text_embeddings.text_embeds,
).sample
noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
noise_pred = noise_pred_uncond + guidance_scale * (
noise_pred_text - noise_pred_uncond
)
return noise_pred
# ref: https://github.com/huggingface/diffusers/blob/0bab447670f47c28df60fbd2f6a0f833f75a16f5/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion.py#L746
def diffuse_some_steps(
self,
latents: torch.FloatTensor,
text_embeddings: PromptEmbeds,
total_timesteps: int = 1000,
start_timesteps=0,
guidance_scale=1,
add_time_ids=None,
**kwargs,
):
for timestep in tqdm(self.sd.noise_scheduler.timesteps[start_timesteps:total_timesteps], leave=False):
noise_pred = self.predict_noise(
latents,
text_embeddings,
timestep,
guidance_scale=guidance_scale,
add_time_ids=add_time_ids,
**kwargs,
)
latents = self.sd.noise_scheduler.step(noise_pred, timestep, latents).prev_sample
# return latents_steps
return latents
def get_latest_save_path(self):
# get latest saved step
if os.path.exists(self.save_root):
@@ -497,92 +221,33 @@ class BaseSDTrainProcess(BaseTrainProcess):
print("load_weights not implemented for non-network models")
def run(self):
super().run()
# run base process run
BaseTrainProcess.run(self)
### HOOK ###
self.hook_before_model_load()
# run base sd process run
self.sd.load_model()
dtype = get_torch_dtype(self.train_config.dtype)
# TODO handle other schedulers
# sch = KDPM2DiscreteScheduler
sch = DDPMScheduler
# do our own scheduler
prediction_type = "v_prediction" if self.model_config.is_v_pred else "epsilon"
scheduler = sch(
num_train_timesteps=1000,
beta_start=0.00085,
beta_end=0.0120,
beta_schedule="scaled_linear",
clip_sample=False,
prediction_type=prediction_type,
)
if self.model_config.is_xl:
if self.custom_pipeline is not None:
pipln = self.custom_pipeline
else:
pipln = CustomStableDiffusionXLPipeline
pipe = pipln.from_single_file(
self.model_config.name_or_path,
dtype=dtype,
scheduler_type='ddpm',
device=self.device_torch,
).to(self.device_torch)
# model is loaded from BaseSDProcess
unet = self.sd.unet
vae = self.sd.vae
tokenizer = self.sd.tokenizer
text_encoder = self.sd.text_encoder
noise_scheduler = self.sd.noise_scheduler
text_encoders = [pipe.text_encoder, pipe.text_encoder_2]
tokenizer = [pipe.tokenizer, pipe.tokenizer_2]
for text_encoder in text_encoders:
text_encoder.to(self.device_torch, dtype=dtype)
text_encoder.requires_grad_(False)
text_encoder.eval()
text_encoder = text_encoders
else:
if self.custom_pipeline is not None:
pipln = self.custom_pipeline
else:
pipln = CustomStableDiffusionPipeline
pipe = pipln.from_single_file(
self.model_config.name_or_path,
dtype=dtype,
scheduler_type='dpm',
device=self.device_torch,
load_safety_checker=False,
).to(self.device_torch)
pipe.register_to_config(requires_safety_checker=False)
text_encoder = pipe.text_encoder
text_encoder.to(self.device_torch, dtype=dtype)
text_encoder.requires_grad_(False)
text_encoder.eval()
tokenizer = pipe.tokenizer
# scheduler doesn't get set sometimes, so we set it here
pipe.scheduler = scheduler
unet = pipe.unet
noise_scheduler = pipe.scheduler
vae = pipe.vae.to('cpu', dtype=dtype)
vae.eval()
vae.requires_grad_(False)
flush()
self.sd = StableDiffusion(
vae,
tokenizer,
text_encoder,
unet,
noise_scheduler,
is_xl=self.model_config.is_xl,
pipeline=pipe
)
unet.to(self.device_torch, dtype=dtype)
if self.train_config.xformers:
vae.set_use_memory_efficient_attention_xformers(True)
unet.enable_xformers_memory_efficient_attention()
if self.train_config.gradient_checkpointing:
unet.enable_gradient_checkpointing()
unet.to(self.device_torch, dtype=dtype)
unet.requires_grad_(False)
unet.eval()
vae = vae.to(torch.device('cpu'), dtype=dtype)
vae.requires_grad_(False)
vae.eval()
if self.network_config is not None:
self.network = LoRASpecialNetwork(
@@ -598,6 +263,8 @@ class BaseSDTrainProcess(BaseTrainProcess):
)
self.network.force_to(self.device_torch, dtype=dtype)
# give network to sd so it can use it
self.sd.network = self.network
self.network.apply_to(
text_encoder,
@@ -650,7 +317,7 @@ class BaseSDTrainProcess(BaseTrainProcess):
optimizer_params=self.train_config.optimizer_params)
self.optimizer = optimizer
lr_scheduler = train_util.get_lr_scheduler(
lr_scheduler = get_lr_scheduler(
self.train_config.lr_scheduler,
optimizer,
max_iterations=self.train_config.steps,