KohyaSS/library/model_util.py

1166 lines
47 KiB
Python
Raw Permalink Normal View History

2022-12-05 15:49:02 +00:00
# v1: split from train_db_fixed.py.
# v2: support safetensors
import math
2022-12-05 15:49:02 +00:00
import os
import torch
2023-03-09 16:06:59 +00:00
from transformers import CLIPTextModel, CLIPTokenizer, CLIPTextConfig, logging
from diffusers import AutoencoderKL, DDIMScheduler, StableDiffusionPipeline, UNet2DConditionModel
2022-12-05 15:49:02 +00:00
from safetensors.torch import load_file, save_file
2022-12-02 17:48:43 +00:00
# DiffUsers版StableDiffusionのモデルパラメータ
NUM_TRAIN_TIMESTEPS = 1000
BETA_START = 0.00085
BETA_END = 0.0120
UNET_PARAMS_MODEL_CHANNELS = 320
UNET_PARAMS_CHANNEL_MULT = [1, 2, 4, 4]
UNET_PARAMS_ATTENTION_RESOLUTIONS = [4, 2, 1]
2023-01-29 16:10:06 +00:00
UNET_PARAMS_IMAGE_SIZE = 64 # fixed from old invalid value `32`
UNET_PARAMS_IN_CHANNELS = 4
UNET_PARAMS_OUT_CHANNELS = 4
UNET_PARAMS_NUM_RES_BLOCKS = 2
UNET_PARAMS_CONTEXT_DIM = 768
UNET_PARAMS_NUM_HEADS = 8
VAE_PARAMS_Z_CHANNELS = 4
VAE_PARAMS_RESOLUTION = 256
VAE_PARAMS_IN_CHANNELS = 3
VAE_PARAMS_OUT_CH = 3
VAE_PARAMS_CH = 128
VAE_PARAMS_CH_MULT = [1, 2, 4, 4]
VAE_PARAMS_NUM_RES_BLOCKS = 2
2022-12-02 17:48:43 +00:00
# V2
V2_UNET_PARAMS_ATTENTION_HEAD_DIM = [5, 10, 20, 20]
V2_UNET_PARAMS_CONTEXT_DIM = 1024
2022-12-18 01:36:31 +00:00
# Diffusersの設定を読み込むための参照モデル
2023-03-09 16:06:59 +00:00
DIFFUSERS_REF_MODEL_ID_V1 = "runwayml/stable-diffusion-v1-5"
DIFFUSERS_REF_MODEL_ID_V2 = "stabilityai/stable-diffusion-2-1"
2022-12-18 01:36:31 +00:00
# region StableDiffusion->Diffusersの変換コード
2022-12-05 15:49:02 +00:00
# convert_original_stable_diffusion_to_diffusers をコピーして修正しているASL 2.0
2022-12-02 17:48:43 +00:00
def shave_segments(path, n_shave_prefix_segments=1):
2023-04-01 11:14:25 +00:00
"""
Removes segments. Positive values shave the first segments, negative shave the last segments.
"""
if n_shave_prefix_segments >= 0:
return ".".join(path.split(".")[n_shave_prefix_segments:])
else:
return ".".join(path.split(".")[:n_shave_prefix_segments])
def renew_resnet_paths(old_list, n_shave_prefix_segments=0):
2023-04-01 11:14:25 +00:00
"""
Updates paths inside resnets to the new naming scheme (local renaming)
"""
mapping = []
for old_item in old_list:
new_item = old_item.replace("in_layers.0", "norm1")
new_item = new_item.replace("in_layers.2", "conv1")
2023-03-09 16:06:59 +00:00
2023-04-01 11:14:25 +00:00
new_item = new_item.replace("out_layers.0", "norm2")
new_item = new_item.replace("out_layers.3", "conv2")
2023-01-01 18:10:32 +00:00
2023-04-01 11:14:25 +00:00
new_item = new_item.replace("emb_layers.1", "time_emb_proj")
new_item = new_item.replace("skip_connection", "conv_shortcut")
2023-01-01 18:10:32 +00:00
2023-04-01 11:14:25 +00:00
new_item = shave_segments(new_item, n_shave_prefix_segments=n_shave_prefix_segments)
2023-03-09 16:06:59 +00:00
2023-04-01 11:14:25 +00:00
mapping.append({"old": old_item, "new": new_item})
2023-03-09 16:06:59 +00:00
2023-04-01 11:14:25 +00:00
return mapping
def renew_vae_resnet_paths(old_list, n_shave_prefix_segments=0):
2023-04-01 11:14:25 +00:00
"""
Updates paths inside resnets to the new naming scheme (local renaming)
"""
mapping = []
for old_item in old_list:
new_item = old_item
2023-04-01 11:14:25 +00:00
new_item = new_item.replace("nin_shortcut", "conv_shortcut")
new_item = shave_segments(new_item, n_shave_prefix_segments=n_shave_prefix_segments)
2023-01-01 18:10:32 +00:00
2023-04-01 11:14:25 +00:00
mapping.append({"old": old_item, "new": new_item})
2023-03-09 16:06:59 +00:00
2023-04-01 11:14:25 +00:00
return mapping
def renew_attention_paths(old_list, n_shave_prefix_segments=0):
2023-04-01 11:14:25 +00:00
"""
Updates paths inside attentions to the new naming scheme (local renaming)
"""
mapping = []
for old_item in old_list:
new_item = old_item
2023-04-01 11:14:25 +00:00
# new_item = new_item.replace('norm.weight', 'group_norm.weight')
# new_item = new_item.replace('norm.bias', 'group_norm.bias')
2023-04-01 11:14:25 +00:00
# new_item = new_item.replace('proj_out.weight', 'proj_attn.weight')
# new_item = new_item.replace('proj_out.bias', 'proj_attn.bias')
2023-04-01 11:14:25 +00:00
# new_item = shave_segments(new_item, n_shave_prefix_segments=n_shave_prefix_segments)
2023-04-01 11:14:25 +00:00
mapping.append({"old": old_item, "new": new_item})
2023-04-01 11:14:25 +00:00
return mapping
def renew_vae_attention_paths(old_list, n_shave_prefix_segments=0):
2023-04-01 11:14:25 +00:00
"""
Updates paths inside attentions to the new naming scheme (local renaming)
"""
mapping = []
for old_item in old_list:
new_item = old_item
2023-04-01 11:14:25 +00:00
new_item = new_item.replace("norm.weight", "group_norm.weight")
new_item = new_item.replace("norm.bias", "group_norm.bias")
2023-04-01 11:14:25 +00:00
new_item = new_item.replace("q.weight", "query.weight")
new_item = new_item.replace("q.bias", "query.bias")
2023-04-01 11:14:25 +00:00
new_item = new_item.replace("k.weight", "key.weight")
new_item = new_item.replace("k.bias", "key.bias")
2023-04-01 11:14:25 +00:00
new_item = new_item.replace("v.weight", "value.weight")
new_item = new_item.replace("v.bias", "value.bias")
2023-04-01 11:14:25 +00:00
new_item = new_item.replace("proj_out.weight", "proj_attn.weight")
new_item = new_item.replace("proj_out.bias", "proj_attn.bias")
2023-04-01 11:14:25 +00:00
new_item = shave_segments(new_item, n_shave_prefix_segments=n_shave_prefix_segments)
2023-04-01 11:14:25 +00:00
mapping.append({"old": old_item, "new": new_item})
2023-04-01 11:14:25 +00:00
return mapping
def assign_to_checkpoint(
2023-03-09 16:06:59 +00:00
paths, checkpoint, old_checkpoint, attention_paths_to_split=None, additional_replacements=None, config=None
):
2023-04-01 11:14:25 +00:00
"""
This does the final conversion step: take locally converted weights and apply a global renaming
to them. It splits attention layers, and takes into account additional replacements
that may arise.
2023-04-01 11:14:25 +00:00
Assigns the weights to the new checkpoint.
"""
assert isinstance(paths, list), "Paths should be a list of dicts containing 'old' and 'new' keys."
2023-04-01 11:14:25 +00:00
# Splits the attention layers into three variables.
if attention_paths_to_split is not None:
for path, path_map in attention_paths_to_split.items():
old_tensor = old_checkpoint[path]
channels = old_tensor.shape[0] // 3
2023-04-01 11:14:25 +00:00
target_shape = (-1, channels) if len(old_tensor.shape) == 3 else (-1)
2023-04-01 11:14:25 +00:00
num_heads = old_tensor.shape[0] // config["num_head_channels"] // 3
2023-04-01 11:14:25 +00:00
old_tensor = old_tensor.reshape((num_heads, 3 * channels // num_heads) + old_tensor.shape[1:])
query, key, value = old_tensor.split(channels // num_heads, dim=1)
2022-12-02 17:48:43 +00:00
2023-04-01 11:14:25 +00:00
checkpoint[path_map["query"]] = query.reshape(target_shape)
checkpoint[path_map["key"]] = key.reshape(target_shape)
checkpoint[path_map["value"]] = value.reshape(target_shape)
2022-12-02 17:48:43 +00:00
2023-04-01 11:14:25 +00:00
for path in paths:
new_path = path["new"]
2023-04-01 11:14:25 +00:00
# These have already been assigned
if attention_paths_to_split is not None and new_path in attention_paths_to_split:
continue
2023-04-01 11:14:25 +00:00
# Global renaming happens here
new_path = new_path.replace("middle_block.0", "mid_block.resnets.0")
new_path = new_path.replace("middle_block.1", "mid_block.attentions.0")
new_path = new_path.replace("middle_block.2", "mid_block.resnets.1")
2023-04-01 11:14:25 +00:00
if additional_replacements is not None:
for replacement in additional_replacements:
new_path = new_path.replace(replacement["old"], replacement["new"])
2023-04-01 11:14:25 +00:00
# proj_attn.weight has to be converted from conv 1D to linear
if "proj_attn.weight" in new_path:
checkpoint[new_path] = old_checkpoint[path["old"]][:, :, 0]
else:
checkpoint[new_path] = old_checkpoint[path["old"]]
2023-03-02 00:24:11 +00:00
2023-03-09 16:06:59 +00:00
def conv_attn_to_linear(checkpoint):
2023-04-01 11:14:25 +00:00
keys = list(checkpoint.keys())
attn_keys = ["query.weight", "key.weight", "value.weight"]
for key in keys:
if ".".join(key.split(".")[-2:]) in attn_keys:
if checkpoint[key].ndim > 2:
checkpoint[key] = checkpoint[key][:, :, 0, 0]
elif "proj_attn.weight" in key:
if checkpoint[key].ndim > 2:
checkpoint[key] = checkpoint[key][:, :, 0]
2023-03-02 00:24:11 +00:00
2023-03-09 16:06:59 +00:00
def linear_transformer_to_conv(checkpoint):
2023-04-01 11:14:25 +00:00
keys = list(checkpoint.keys())
tf_keys = ["proj_in.weight", "proj_out.weight"]
for key in keys:
if ".".join(key.split(".")[-2:]) in tf_keys:
if checkpoint[key].ndim == 2:
checkpoint[key] = checkpoint[key].unsqueeze(2).unsqueeze(2)
2023-01-01 18:10:32 +00:00
2023-03-02 00:24:11 +00:00
2023-03-09 16:06:59 +00:00
def convert_ldm_unet_checkpoint(v2, checkpoint, config):
2023-04-01 11:14:25 +00:00
"""
Takes a state dict and a config, and returns a converted checkpoint.
"""
# extract state_dict for UNet
unet_state_dict = {}
unet_key = "model.diffusion_model."
keys = list(checkpoint.keys())
for key in keys:
if key.startswith(unet_key):
unet_state_dict[key.replace(unet_key, "")] = checkpoint.pop(key)
new_checkpoint = {}
new_checkpoint["time_embedding.linear_1.weight"] = unet_state_dict["time_embed.0.weight"]
new_checkpoint["time_embedding.linear_1.bias"] = unet_state_dict["time_embed.0.bias"]
new_checkpoint["time_embedding.linear_2.weight"] = unet_state_dict["time_embed.2.weight"]
new_checkpoint["time_embedding.linear_2.bias"] = unet_state_dict["time_embed.2.bias"]
new_checkpoint["conv_in.weight"] = unet_state_dict["input_blocks.0.0.weight"]
new_checkpoint["conv_in.bias"] = unet_state_dict["input_blocks.0.0.bias"]
new_checkpoint["conv_norm_out.weight"] = unet_state_dict["out.0.weight"]
new_checkpoint["conv_norm_out.bias"] = unet_state_dict["out.0.bias"]
new_checkpoint["conv_out.weight"] = unet_state_dict["out.2.weight"]
new_checkpoint["conv_out.bias"] = unet_state_dict["out.2.bias"]
# Retrieves the keys for the input blocks only
num_input_blocks = len({".".join(layer.split(".")[:2]) for layer in unet_state_dict if "input_blocks" in layer})
input_blocks = {
layer_id: [key for key in unet_state_dict if f"input_blocks.{layer_id}." in key] for layer_id in range(num_input_blocks)
}
# Retrieves the keys for the middle blocks only
num_middle_blocks = len({".".join(layer.split(".")[:2]) for layer in unet_state_dict if "middle_block" in layer})
middle_blocks = {
layer_id: [key for key in unet_state_dict if f"middle_block.{layer_id}." in key] for layer_id in range(num_middle_blocks)
}
# Retrieves the keys for the output blocks only
num_output_blocks = len({".".join(layer.split(".")[:2]) for layer in unet_state_dict if "output_blocks" in layer})
output_blocks = {
layer_id: [key for key in unet_state_dict if f"output_blocks.{layer_id}." in key] for layer_id in range(num_output_blocks)
}
for i in range(1, num_input_blocks):
block_id = (i - 1) // (config["layers_per_block"] + 1)
layer_in_block_id = (i - 1) % (config["layers_per_block"] + 1)
resnets = [key for key in input_blocks[i] if f"input_blocks.{i}.0" in key and f"input_blocks.{i}.0.op" not in key]
attentions = [key for key in input_blocks[i] if f"input_blocks.{i}.1" in key]
if f"input_blocks.{i}.0.op.weight" in unet_state_dict:
new_checkpoint[f"down_blocks.{block_id}.downsamplers.0.conv.weight"] = unet_state_dict.pop(
f"input_blocks.{i}.0.op.weight"
)
new_checkpoint[f"down_blocks.{block_id}.downsamplers.0.conv.bias"] = unet_state_dict.pop(f"input_blocks.{i}.0.op.bias")
paths = renew_resnet_paths(resnets)
meta_path = {"old": f"input_blocks.{i}.0", "new": f"down_blocks.{block_id}.resnets.{layer_in_block_id}"}
assign_to_checkpoint(paths, new_checkpoint, unet_state_dict, additional_replacements=[meta_path], config=config)
if len(attentions):
paths = renew_attention_paths(attentions)
meta_path = {"old": f"input_blocks.{i}.1", "new": f"down_blocks.{block_id}.attentions.{layer_in_block_id}"}
assign_to_checkpoint(paths, new_checkpoint, unet_state_dict, additional_replacements=[meta_path], config=config)
resnet_0 = middle_blocks[0]
attentions = middle_blocks[1]
resnet_1 = middle_blocks[2]
resnet_0_paths = renew_resnet_paths(resnet_0)
assign_to_checkpoint(resnet_0_paths, new_checkpoint, unet_state_dict, config=config)
resnet_1_paths = renew_resnet_paths(resnet_1)
assign_to_checkpoint(resnet_1_paths, new_checkpoint, unet_state_dict, config=config)
attentions_paths = renew_attention_paths(attentions)
meta_path = {"old": "middle_block.1", "new": "mid_block.attentions.0"}
assign_to_checkpoint(attentions_paths, new_checkpoint, unet_state_dict, additional_replacements=[meta_path], config=config)
for i in range(num_output_blocks):
block_id = i // (config["layers_per_block"] + 1)
layer_in_block_id = i % (config["layers_per_block"] + 1)
output_block_layers = [shave_segments(name, 2) for name in output_blocks[i]]
output_block_list = {}
for layer in output_block_layers:
layer_id, layer_name = layer.split(".")[0], shave_segments(layer, 1)
if layer_id in output_block_list:
output_block_list[layer_id].append(layer_name)
else:
output_block_list[layer_id] = [layer_name]
if len(output_block_list) > 1:
resnets = [key for key in output_blocks[i] if f"output_blocks.{i}.0" in key]
attentions = [key for key in output_blocks[i] if f"output_blocks.{i}.1" in key]
resnet_0_paths = renew_resnet_paths(resnets)
paths = renew_resnet_paths(resnets)
meta_path = {"old": f"output_blocks.{i}.0", "new": f"up_blocks.{block_id}.resnets.{layer_in_block_id}"}
assign_to_checkpoint(paths, new_checkpoint, unet_state_dict, additional_replacements=[meta_path], config=config)
# オリジナル:
# if ["conv.weight", "conv.bias"] in output_block_list.values():
# index = list(output_block_list.values()).index(["conv.weight", "conv.bias"])
# biasとweightの順番に依存しないようにするもっといいやり方がありそうだが
for l in output_block_list.values():
l.sort()
if ["conv.bias", "conv.weight"] in output_block_list.values():
index = list(output_block_list.values()).index(["conv.bias", "conv.weight"])
new_checkpoint[f"up_blocks.{block_id}.upsamplers.0.conv.bias"] = unet_state_dict[
f"output_blocks.{i}.{index}.conv.bias"
]
new_checkpoint[f"up_blocks.{block_id}.upsamplers.0.conv.weight"] = unet_state_dict[
f"output_blocks.{i}.{index}.conv.weight"
]
# Clear attentions as they have been attributed above.
if len(attentions) == 2:
attentions = []
if len(attentions):
paths = renew_attention_paths(attentions)
meta_path = {
"old": f"output_blocks.{i}.1",
"new": f"up_blocks.{block_id}.attentions.{layer_in_block_id}",
}
assign_to_checkpoint(paths, new_checkpoint, unet_state_dict, additional_replacements=[meta_path], config=config)
else:
resnet_0_paths = renew_resnet_paths(output_block_layers, n_shave_prefix_segments=1)
for path in resnet_0_paths:
old_path = ".".join(["output_blocks", str(i), path["old"]])
new_path = ".".join(["up_blocks", str(block_id), "resnets", str(layer_in_block_id), path["new"]])
new_checkpoint[new_path] = unet_state_dict[old_path]
# SDのv2では1*1のconv2dがlinearに変わっているので、linear->convに変換する
if v2:
linear_transformer_to_conv(new_checkpoint)
2023-04-01 11:14:25 +00:00
return new_checkpoint
2023-03-09 16:06:59 +00:00
def convert_ldm_vae_checkpoint(checkpoint, config):
2023-04-01 11:14:25 +00:00
# extract state dict for VAE
vae_state_dict = {}
vae_key = "first_stage_model."
keys = list(checkpoint.keys())
for key in keys:
if key.startswith(vae_key):
vae_state_dict[key.replace(vae_key, "")] = checkpoint.get(key)
# if len(vae_state_dict) == 0:
# # 渡されたcheckpointは.ckptから読み込んだcheckpointではなくvaeのstate_dict
# vae_state_dict = checkpoint
new_checkpoint = {}
new_checkpoint["encoder.conv_in.weight"] = vae_state_dict["encoder.conv_in.weight"]
new_checkpoint["encoder.conv_in.bias"] = vae_state_dict["encoder.conv_in.bias"]
new_checkpoint["encoder.conv_out.weight"] = vae_state_dict["encoder.conv_out.weight"]
new_checkpoint["encoder.conv_out.bias"] = vae_state_dict["encoder.conv_out.bias"]
new_checkpoint["encoder.conv_norm_out.weight"] = vae_state_dict["encoder.norm_out.weight"]
new_checkpoint["encoder.conv_norm_out.bias"] = vae_state_dict["encoder.norm_out.bias"]
new_checkpoint["decoder.conv_in.weight"] = vae_state_dict["decoder.conv_in.weight"]
new_checkpoint["decoder.conv_in.bias"] = vae_state_dict["decoder.conv_in.bias"]
new_checkpoint["decoder.conv_out.weight"] = vae_state_dict["decoder.conv_out.weight"]
new_checkpoint["decoder.conv_out.bias"] = vae_state_dict["decoder.conv_out.bias"]
new_checkpoint["decoder.conv_norm_out.weight"] = vae_state_dict["decoder.norm_out.weight"]
new_checkpoint["decoder.conv_norm_out.bias"] = vae_state_dict["decoder.norm_out.bias"]
new_checkpoint["quant_conv.weight"] = vae_state_dict["quant_conv.weight"]
new_checkpoint["quant_conv.bias"] = vae_state_dict["quant_conv.bias"]
new_checkpoint["post_quant_conv.weight"] = vae_state_dict["post_quant_conv.weight"]
new_checkpoint["post_quant_conv.bias"] = vae_state_dict["post_quant_conv.bias"]
# Retrieves the keys for the encoder down blocks only
num_down_blocks = len({".".join(layer.split(".")[:3]) for layer in vae_state_dict if "encoder.down" in layer})
down_blocks = {layer_id: [key for key in vae_state_dict if f"down.{layer_id}" in key] for layer_id in range(num_down_blocks)}
# Retrieves the keys for the decoder up blocks only
num_up_blocks = len({".".join(layer.split(".")[:3]) for layer in vae_state_dict if "decoder.up" in layer})
up_blocks = {layer_id: [key for key in vae_state_dict if f"up.{layer_id}" in key] for layer_id in range(num_up_blocks)}
for i in range(num_down_blocks):
resnets = [key for key in down_blocks[i] if f"down.{i}" in key and f"down.{i}.downsample" not in key]
if f"encoder.down.{i}.downsample.conv.weight" in vae_state_dict:
new_checkpoint[f"encoder.down_blocks.{i}.downsamplers.0.conv.weight"] = vae_state_dict.pop(
f"encoder.down.{i}.downsample.conv.weight"
)
new_checkpoint[f"encoder.down_blocks.{i}.downsamplers.0.conv.bias"] = vae_state_dict.pop(
f"encoder.down.{i}.downsample.conv.bias"
)
paths = renew_vae_resnet_paths(resnets)
meta_path = {"old": f"down.{i}.block", "new": f"down_blocks.{i}.resnets"}
assign_to_checkpoint(paths, new_checkpoint, vae_state_dict, additional_replacements=[meta_path], config=config)
mid_resnets = [key for key in vae_state_dict if "encoder.mid.block" in key]
num_mid_res_blocks = 2
for i in range(1, num_mid_res_blocks + 1):
resnets = [key for key in mid_resnets if f"encoder.mid.block_{i}" in key]
paths = renew_vae_resnet_paths(resnets)
meta_path = {"old": f"mid.block_{i}", "new": f"mid_block.resnets.{i - 1}"}
assign_to_checkpoint(paths, new_checkpoint, vae_state_dict, additional_replacements=[meta_path], config=config)
mid_attentions = [key for key in vae_state_dict if "encoder.mid.attn" in key]
paths = renew_vae_attention_paths(mid_attentions)
meta_path = {"old": "mid.attn_1", "new": "mid_block.attentions.0"}
2023-03-09 16:06:59 +00:00
assign_to_checkpoint(paths, new_checkpoint, vae_state_dict, additional_replacements=[meta_path], config=config)
2023-04-01 11:14:25 +00:00
conv_attn_to_linear(new_checkpoint)
for i in range(num_up_blocks):
block_id = num_up_blocks - 1 - i
resnets = [key for key in up_blocks[block_id] if f"up.{block_id}" in key and f"up.{block_id}.upsample" not in key]
if f"decoder.up.{block_id}.upsample.conv.weight" in vae_state_dict:
new_checkpoint[f"decoder.up_blocks.{i}.upsamplers.0.conv.weight"] = vae_state_dict[
f"decoder.up.{block_id}.upsample.conv.weight"
]
new_checkpoint[f"decoder.up_blocks.{i}.upsamplers.0.conv.bias"] = vae_state_dict[
f"decoder.up.{block_id}.upsample.conv.bias"
]
paths = renew_vae_resnet_paths(resnets)
meta_path = {"old": f"up.{block_id}.block", "new": f"up_blocks.{i}.resnets"}
assign_to_checkpoint(paths, new_checkpoint, vae_state_dict, additional_replacements=[meta_path], config=config)
mid_resnets = [key for key in vae_state_dict if "decoder.mid.block" in key]
num_mid_res_blocks = 2
for i in range(1, num_mid_res_blocks + 1):
resnets = [key for key in mid_resnets if f"decoder.mid.block_{i}" in key]
paths = renew_vae_resnet_paths(resnets)
meta_path = {"old": f"mid.block_{i}", "new": f"mid_block.resnets.{i - 1}"}
assign_to_checkpoint(paths, new_checkpoint, vae_state_dict, additional_replacements=[meta_path], config=config)
mid_attentions = [key for key in vae_state_dict if "decoder.mid.attn" in key]
paths = renew_vae_attention_paths(mid_attentions)
meta_path = {"old": "mid.attn_1", "new": "mid_block.attentions.0"}
2023-03-09 16:06:59 +00:00
assign_to_checkpoint(paths, new_checkpoint, vae_state_dict, additional_replacements=[meta_path], config=config)
2023-04-01 11:14:25 +00:00
conv_attn_to_linear(new_checkpoint)
return new_checkpoint
2023-03-02 00:24:11 +00:00
2023-03-09 16:06:59 +00:00
def create_unet_diffusers_config(v2):
2023-04-01 11:14:25 +00:00
"""
Creates a config for the diffusers based on the config of the LDM model.
"""
# unet_params = original_config.model.params.unet_config.params
block_out_channels = [UNET_PARAMS_MODEL_CHANNELS * mult for mult in UNET_PARAMS_CHANNEL_MULT]
down_block_types = []
resolution = 1
for i in range(len(block_out_channels)):
block_type = "CrossAttnDownBlock2D" if resolution in UNET_PARAMS_ATTENTION_RESOLUTIONS else "DownBlock2D"
down_block_types.append(block_type)
if i != len(block_out_channels) - 1:
resolution *= 2
up_block_types = []
for i in range(len(block_out_channels)):
block_type = "CrossAttnUpBlock2D" if resolution in UNET_PARAMS_ATTENTION_RESOLUTIONS else "UpBlock2D"
up_block_types.append(block_type)
resolution //= 2
config = dict(
sample_size=UNET_PARAMS_IMAGE_SIZE,
in_channels=UNET_PARAMS_IN_CHANNELS,
out_channels=UNET_PARAMS_OUT_CHANNELS,
down_block_types=tuple(down_block_types),
up_block_types=tuple(up_block_types),
block_out_channels=tuple(block_out_channels),
layers_per_block=UNET_PARAMS_NUM_RES_BLOCKS,
cross_attention_dim=UNET_PARAMS_CONTEXT_DIM if not v2 else V2_UNET_PARAMS_CONTEXT_DIM,
attention_head_dim=UNET_PARAMS_NUM_HEADS if not v2 else V2_UNET_PARAMS_ATTENTION_HEAD_DIM,
)
return config
def create_vae_diffusers_config():
2023-04-01 11:14:25 +00:00
"""
Creates a config for the diffusers based on the config of the LDM model.
"""
# vae_params = original_config.model.params.first_stage_config.params.ddconfig
# _ = original_config.model.params.first_stage_config.params.embed_dim
block_out_channels = [VAE_PARAMS_CH * mult for mult in VAE_PARAMS_CH_MULT]
down_block_types = ["DownEncoderBlock2D"] * len(block_out_channels)
up_block_types = ["UpDecoderBlock2D"] * len(block_out_channels)
config = dict(
sample_size=VAE_PARAMS_RESOLUTION,
in_channels=VAE_PARAMS_IN_CHANNELS,
out_channels=VAE_PARAMS_OUT_CH,
down_block_types=tuple(down_block_types),
up_block_types=tuple(up_block_types),
block_out_channels=tuple(block_out_channels),
latent_channels=VAE_PARAMS_Z_CHANNELS,
layers_per_block=VAE_PARAMS_NUM_RES_BLOCKS,
)
return config
2022-12-02 17:48:43 +00:00
def convert_ldm_clip_checkpoint_v1(checkpoint):
2023-04-01 11:14:25 +00:00
keys = list(checkpoint.keys())
text_model_dict = {}
for key in keys:
if key.startswith("cond_stage_model.transformer"):
text_model_dict[key[len("cond_stage_model.transformer.") :]] = checkpoint[key]
return text_model_dict
2022-12-02 17:48:43 +00:00
def convert_ldm_clip_checkpoint_v2(checkpoint, max_length):
2023-04-01 11:14:25 +00:00
# 嫌になるくらい違うぞ!
def convert_key(key):
if not key.startswith("cond_stage_model"):
return None
# common conversion
key = key.replace("cond_stage_model.model.transformer.", "text_model.encoder.")
key = key.replace("cond_stage_model.model.", "text_model.")
if "resblocks" in key:
# resblocks conversion
key = key.replace(".resblocks.", ".layers.")
if ".ln_" in key:
key = key.replace(".ln_", ".layer_norm")
elif ".mlp." in key:
key = key.replace(".c_fc.", ".fc1.")
key = key.replace(".c_proj.", ".fc2.")
elif ".attn.out_proj" in key:
key = key.replace(".attn.out_proj.", ".self_attn.out_proj.")
elif ".attn.in_proj" in key:
key = None # 特殊なので後で処理する
else:
raise ValueError(f"unexpected key in SD: {key}")
elif ".positional_embedding" in key:
key = key.replace(".positional_embedding", ".embeddings.position_embedding.weight")
elif ".text_projection" in key:
key = None # 使われない???
elif ".logit_scale" in key:
key = None # 使われない???
elif ".token_embedding" in key:
key = key.replace(".token_embedding.weight", ".embeddings.token_embedding.weight")
elif ".ln_final" in key:
key = key.replace(".ln_final", ".final_layer_norm")
return key
keys = list(checkpoint.keys())
new_sd = {}
for key in keys:
# remove resblocks 23
if ".resblocks.23." in key:
continue
new_key = convert_key(key)
if new_key is None:
continue
new_sd[new_key] = checkpoint[key]
# attnの変換
for key in keys:
if ".resblocks.23." in key:
continue
if ".resblocks" in key and ".attn.in_proj_" in key:
# 三つに分割
values = torch.chunk(checkpoint[key], 3)
key_suffix = ".weight" if "weight" in key else ".bias"
key_pfx = key.replace("cond_stage_model.model.transformer.resblocks.", "text_model.encoder.layers.")
key_pfx = key_pfx.replace("_weight", "")
key_pfx = key_pfx.replace("_bias", "")
key_pfx = key_pfx.replace(".attn.in_proj", ".self_attn.")
new_sd[key_pfx + "q_proj" + key_suffix] = values[0]
new_sd[key_pfx + "k_proj" + key_suffix] = values[1]
new_sd[key_pfx + "v_proj" + key_suffix] = values[2]
# rename or add position_ids
ANOTHER_POSITION_IDS_KEY = "text_model.encoder.text_model.embeddings.position_ids"
if ANOTHER_POSITION_IDS_KEY in new_sd:
# waifu diffusion v1.4
position_ids = new_sd[ANOTHER_POSITION_IDS_KEY]
del new_sd[ANOTHER_POSITION_IDS_KEY]
else:
position_ids = torch.Tensor([list(range(max_length))]).to(torch.int64)
new_sd["text_model.embeddings.position_ids"] = position_ids
return new_sd
# endregion
# region Diffusers->StableDiffusion の変換コード
2022-12-05 15:49:02 +00:00
# convert_diffusers_to_original_stable_diffusion をコピーして修正しているASL 2.0
2023-04-01 11:14:25 +00:00
2022-12-02 17:48:43 +00:00
def conv_transformer_to_linear(checkpoint):
2023-04-01 11:14:25 +00:00
keys = list(checkpoint.keys())
tf_keys = ["proj_in.weight", "proj_out.weight"]
for key in keys:
if ".".join(key.split(".")[-2:]) in tf_keys:
if checkpoint[key].ndim > 2:
checkpoint[key] = checkpoint[key][:, :, 0, 0]
2022-12-02 17:48:43 +00:00
def convert_unet_state_dict_to_sd(v2, unet_state_dict):
2023-04-01 11:14:25 +00:00
unet_conversion_map = [
# (stable-diffusion, HF Diffusers)
("time_embed.0.weight", "time_embedding.linear_1.weight"),
("time_embed.0.bias", "time_embedding.linear_1.bias"),
("time_embed.2.weight", "time_embedding.linear_2.weight"),
("time_embed.2.bias", "time_embedding.linear_2.bias"),
("input_blocks.0.0.weight", "conv_in.weight"),
("input_blocks.0.0.bias", "conv_in.bias"),
("out.0.weight", "conv_norm_out.weight"),
("out.0.bias", "conv_norm_out.bias"),
("out.2.weight", "conv_out.weight"),
("out.2.bias", "conv_out.bias"),
]
unet_conversion_map_resnet = [
# (stable-diffusion, HF Diffusers)
("in_layers.0", "norm1"),
("in_layers.2", "conv1"),
("out_layers.0", "norm2"),
("out_layers.3", "conv2"),
("emb_layers.1", "time_emb_proj"),
("skip_connection", "conv_shortcut"),
]
unet_conversion_map_layer = []
for i in range(4):
# loop over downblocks/upblocks
for j in range(2):
# loop over resnets/attentions for downblocks
hf_down_res_prefix = f"down_blocks.{i}.resnets.{j}."
sd_down_res_prefix = f"input_blocks.{3*i + j + 1}.0."
unet_conversion_map_layer.append((sd_down_res_prefix, hf_down_res_prefix))
if i < 3:
# no attention layers in down_blocks.3
hf_down_atn_prefix = f"down_blocks.{i}.attentions.{j}."
sd_down_atn_prefix = f"input_blocks.{3*i + j + 1}.1."
unet_conversion_map_layer.append((sd_down_atn_prefix, hf_down_atn_prefix))
for j in range(3):
# loop over resnets/attentions for upblocks
hf_up_res_prefix = f"up_blocks.{i}.resnets.{j}."
sd_up_res_prefix = f"output_blocks.{3*i + j}.0."
unet_conversion_map_layer.append((sd_up_res_prefix, hf_up_res_prefix))
if i > 0:
# no attention layers in up_blocks.0
hf_up_atn_prefix = f"up_blocks.{i}.attentions.{j}."
sd_up_atn_prefix = f"output_blocks.{3*i + j}.1."
unet_conversion_map_layer.append((sd_up_atn_prefix, hf_up_atn_prefix))
if i < 3:
# no downsample in down_blocks.3
hf_downsample_prefix = f"down_blocks.{i}.downsamplers.0.conv."
sd_downsample_prefix = f"input_blocks.{3*(i+1)}.0.op."
unet_conversion_map_layer.append((sd_downsample_prefix, hf_downsample_prefix))
# no upsample in up_blocks.3
hf_upsample_prefix = f"up_blocks.{i}.upsamplers.0."
sd_upsample_prefix = f"output_blocks.{3*i + 2}.{1 if i == 0 else 2}."
unet_conversion_map_layer.append((sd_upsample_prefix, hf_upsample_prefix))
hf_mid_atn_prefix = "mid_block.attentions.0."
sd_mid_atn_prefix = "middle_block.1."
unet_conversion_map_layer.append((sd_mid_atn_prefix, hf_mid_atn_prefix))
for j in range(2):
2023-04-01 11:14:25 +00:00
hf_mid_res_prefix = f"mid_block.resnets.{j}."
sd_mid_res_prefix = f"middle_block.{2*j}."
unet_conversion_map_layer.append((sd_mid_res_prefix, hf_mid_res_prefix))
# buyer beware: this is a *brittle* function,
# and correct output requires that all of these pieces interact in
# the exact order in which I have arranged them.
mapping = {k: k for k in unet_state_dict.keys()}
for sd_name, hf_name in unet_conversion_map:
mapping[hf_name] = sd_name
for k, v in mapping.items():
if "resnets" in k:
for sd_part, hf_part in unet_conversion_map_resnet:
v = v.replace(hf_part, sd_part)
mapping[k] = v
for k, v in mapping.items():
for sd_part, hf_part in unet_conversion_map_layer:
v = v.replace(hf_part, sd_part)
mapping[k] = v
new_state_dict = {v: unet_state_dict[k] for k, v in mapping.items()}
if v2:
conv_transformer_to_linear(new_state_dict)
return new_state_dict
2022-12-05 15:49:02 +00:00
# ================#
# VAE Conversion #
# ================#
2023-04-01 11:14:25 +00:00
2022-12-05 15:49:02 +00:00
def reshape_weight_for_sd(w):
# convert HF linear weights to SD conv2d weights
2023-04-01 11:14:25 +00:00
return w.reshape(*w.shape, 1, 1)
2022-12-05 15:49:02 +00:00
def convert_vae_state_dict(vae_state_dict):
2023-04-01 11:14:25 +00:00
vae_conversion_map = [
# (stable-diffusion, HF Diffusers)
("nin_shortcut", "conv_shortcut"),
("norm_out", "conv_norm_out"),
("mid.attn_1.", "mid_block.attentions.0."),
]
for i in range(4):
# down_blocks have two resnets
for j in range(2):
hf_down_prefix = f"encoder.down_blocks.{i}.resnets.{j}."
sd_down_prefix = f"encoder.down.{i}.block.{j}."
vae_conversion_map.append((sd_down_prefix, hf_down_prefix))
if i < 3:
hf_downsample_prefix = f"down_blocks.{i}.downsamplers.0."
sd_downsample_prefix = f"down.{i}.downsample."
vae_conversion_map.append((sd_downsample_prefix, hf_downsample_prefix))
hf_upsample_prefix = f"up_blocks.{i}.upsamplers.0."
sd_upsample_prefix = f"up.{3-i}.upsample."
vae_conversion_map.append((sd_upsample_prefix, hf_upsample_prefix))
# up_blocks have three resnets
# also, up blocks in hf are numbered in reverse from sd
for j in range(3):
hf_up_prefix = f"decoder.up_blocks.{i}.resnets.{j}."
sd_up_prefix = f"decoder.up.{3-i}.block.{j}."
vae_conversion_map.append((sd_up_prefix, hf_up_prefix))
# this part accounts for mid blocks in both the encoder and the decoder
for i in range(2):
hf_mid_res_prefix = f"mid_block.resnets.{i}."
sd_mid_res_prefix = f"mid.block_{i+1}."
vae_conversion_map.append((sd_mid_res_prefix, hf_mid_res_prefix))
vae_conversion_map_attn = [
# (stable-diffusion, HF Diffusers)
("norm.", "group_norm."),
("q.", "query."),
("k.", "key."),
("v.", "value."),
("proj_out.", "proj_attn."),
]
mapping = {k: k for k in vae_state_dict.keys()}
for k, v in mapping.items():
for sd_part, hf_part in vae_conversion_map:
v = v.replace(hf_part, sd_part)
mapping[k] = v
for k, v in mapping.items():
if "attentions" in k:
for sd_part, hf_part in vae_conversion_map_attn:
v = v.replace(hf_part, sd_part)
mapping[k] = v
new_state_dict = {v: vae_state_dict[k] for k, v in mapping.items()}
weights_to_convert = ["q", "k", "v", "proj_out"]
for k, v in new_state_dict.items():
for weight_name in weights_to_convert:
if f"mid.attn_1.{weight_name}.weight" in k:
# print(f"Reshaping {k} for SD format")
new_state_dict[k] = reshape_weight_for_sd(v)
return new_state_dict
2022-12-05 15:49:02 +00:00
# endregion
2022-12-13 18:49:14 +00:00
# region 自作のモデル読み書きなど
2022-12-05 15:49:02 +00:00
2023-04-01 11:14:25 +00:00
2022-12-05 15:49:02 +00:00
def is_safetensors(path):
2023-04-01 11:14:25 +00:00
return os.path.splitext(path)[1].lower() == ".safetensors"
def load_checkpoint_with_text_encoder_conversion(ckpt_path, device="cpu"):
# text encoderの格納形式が違うモデルに対応する ('text_model'がない)
TEXT_ENCODER_KEY_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."),
]
if is_safetensors(ckpt_path):
checkpoint = None
state_dict = load_file(ckpt_path) # , device) # may causes error
2022-12-05 15:49:02 +00:00
else:
2023-04-01 11:14:25 +00:00
checkpoint = torch.load(ckpt_path, map_location=device)
if "state_dict" in checkpoint:
state_dict = checkpoint["state_dict"]
else:
state_dict = checkpoint
checkpoint = None
2023-04-01 11:14:25 +00:00
key_reps = []
for rep_from, rep_to in TEXT_ENCODER_KEY_REPLACEMENTS:
for key in state_dict.keys():
if key.startswith(rep_from):
new_key = rep_to + key[len(rep_from) :]
key_reps.append((key, new_key))
2023-04-01 11:14:25 +00:00
for key, new_key in key_reps:
state_dict[new_key] = state_dict[key]
del state_dict[key]
2023-04-01 11:14:25 +00:00
return checkpoint, state_dict
2022-12-05 15:49:02 +00:00
# TODO dtype指定の動作が怪しいので確認する text_encoderを指定形式で作れるか未確認
2023-04-01 11:14:25 +00:00
def load_models_from_stable_diffusion_checkpoint(v2, ckpt_path, device="cpu", dtype=None):
_, state_dict = load_checkpoint_with_text_encoder_conversion(ckpt_path, device)
# Convert the UNet2DConditionModel model.
unet_config = create_unet_diffusers_config(v2)
converted_unet_checkpoint = convert_ldm_unet_checkpoint(v2, state_dict, unet_config)
unet = UNet2DConditionModel(**unet_config).to(device)
info = unet.load_state_dict(converted_unet_checkpoint)
print("loading u-net:", info)
# Convert the VAE model.
vae_config = create_vae_diffusers_config()
converted_vae_checkpoint = convert_ldm_vae_checkpoint(state_dict, vae_config)
vae = AutoencoderKL(**vae_config).to(device)
info = vae.load_state_dict(converted_vae_checkpoint)
print("loading vae:", info)
# convert text_model
if v2:
converted_text_encoder_checkpoint = convert_ldm_clip_checkpoint_v2(state_dict, 77)
cfg = CLIPTextConfig(
vocab_size=49408,
hidden_size=1024,
intermediate_size=4096,
num_hidden_layers=23,
num_attention_heads=16,
max_position_embeddings=77,
hidden_act="gelu",
layer_norm_eps=1e-05,
dropout=0.0,
attention_dropout=0.0,
initializer_range=0.02,
initializer_factor=1.0,
pad_token_id=1,
bos_token_id=0,
eos_token_id=2,
model_type="clip_text_model",
projection_dim=512,
torch_dtype="float32",
transformers_version="4.25.0.dev0",
)
text_model = CLIPTextModel._from_config(cfg)
info = text_model.load_state_dict(converted_text_encoder_checkpoint)
else:
converted_text_encoder_checkpoint = convert_ldm_clip_checkpoint_v1(state_dict)
2023-03-09 16:06:59 +00:00
2023-04-01 11:14:25 +00:00
logging.set_verbosity_error() # don't show annoying warning
text_model = CLIPTextModel.from_pretrained("openai/clip-vit-large-patch14").to(device)
logging.set_verbosity_warning()
2023-03-09 16:06:59 +00:00
2023-04-01 11:14:25 +00:00
info = text_model.load_state_dict(converted_text_encoder_checkpoint)
print("loading text encoder:", info)
2023-03-09 16:06:59 +00:00
2023-04-01 11:14:25 +00:00
return text_model, vae, unet
2023-03-09 16:06:59 +00:00
def convert_text_encoder_state_dict_to_sd_v2(checkpoint, make_dummy_weights=False):
2023-04-01 11:14:25 +00:00
def convert_key(key):
# position_idsの除去
if ".position_ids" in key:
return None
# common
key = key.replace("text_model.encoder.", "transformer.")
key = key.replace("text_model.", "")
if "layers" in key:
# resblocks conversion
key = key.replace(".layers.", ".resblocks.")
if ".layer_norm" in key:
key = key.replace(".layer_norm", ".ln_")
elif ".mlp." in key:
key = key.replace(".fc1.", ".c_fc.")
key = key.replace(".fc2.", ".c_proj.")
elif ".self_attn.out_proj" in key:
key = key.replace(".self_attn.out_proj.", ".attn.out_proj.")
elif ".self_attn." in key:
key = None # 特殊なので後で処理する
else:
raise ValueError(f"unexpected key in DiffUsers model: {key}")
elif ".position_embedding" in key:
key = key.replace("embeddings.position_embedding.weight", "positional_embedding")
elif ".token_embedding" in key:
key = key.replace("embeddings.token_embedding.weight", "token_embedding.weight")
elif "final_layer_norm" in key:
key = key.replace("final_layer_norm", "ln_final")
return key
keys = list(checkpoint.keys())
new_sd = {}
2022-12-05 15:49:02 +00:00
for key in keys:
2023-04-01 11:14:25 +00:00
new_key = convert_key(key)
if new_key is None:
continue
new_sd[new_key] = checkpoint[key]
2022-12-05 15:49:02 +00:00
2023-04-01 11:14:25 +00:00
# attnの変換
for key in keys:
if "layers" in key and "q_proj" in key:
# 三つを結合
key_q = key
key_k = key.replace("q_proj", "k_proj")
key_v = key.replace("q_proj", "v_proj")
value_q = checkpoint[key_q]
value_k = checkpoint[key_k]
value_v = checkpoint[key_v]
value = torch.cat([value_q, value_k, value_v])
new_key = key.replace("text_model.encoder.layers.", "transformer.resblocks.")
new_key = new_key.replace(".self_attn.q_proj.", ".attn.in_proj_")
new_sd[new_key] = value
2022-12-05 15:49:02 +00:00
2023-04-01 11:14:25 +00:00
# 最後の層などを捏造するか
if make_dummy_weights:
print("make dummy weights for resblock.23, text_projection and logit scale.")
keys = list(new_sd.keys())
for key in keys:
if key.startswith("transformer.resblocks.22."):
new_sd[key.replace(".22.", ".23.")] = new_sd[key].clone() # copyしないとsafetensorsの保存で落ちる
# Diffusersに含まれない重みを作っておく
new_sd["text_projection"] = torch.ones((1024, 1024), dtype=new_sd[keys[0]].dtype, device=new_sd[keys[0]].device)
new_sd["logit_scale"] = torch.tensor(1)
return new_sd
2022-12-02 17:48:43 +00:00
2023-03-09 16:06:59 +00:00
def save_stable_diffusion_checkpoint(v2, output_file, text_encoder, unet, ckpt_path, epochs, steps, save_dtype=None, vae=None):
2023-04-01 11:14:25 +00:00
if ckpt_path is not None:
# epoch/stepを参照する。またVAEがメモリ上にないときなど、もう一度VAEを含めて読み込む
checkpoint, state_dict = load_checkpoint_with_text_encoder_conversion(ckpt_path)
if checkpoint is None: # safetensors または state_dictのckpt
checkpoint = {}
strict = False
else:
strict = True
if "state_dict" in state_dict:
del state_dict["state_dict"]
2022-12-05 15:49:02 +00:00
else:
2023-04-01 11:14:25 +00:00
# 新しく作る
assert vae is not None, "VAE is required to save a checkpoint without a given checkpoint"
checkpoint = {}
state_dict = {}
strict = False
def update_sd(prefix, sd):
for k, v in sd.items():
key = prefix + k
assert not strict or key in state_dict, f"Illegal key in save SD: {key}"
if save_dtype is not None:
v = v.detach().clone().to("cpu").to(save_dtype)
state_dict[key] = v
# Convert the UNet model
unet_state_dict = convert_unet_state_dict_to_sd(v2, unet.state_dict())
update_sd("model.diffusion_model.", unet_state_dict)
# Convert the text encoder model
if v2:
make_dummy = ckpt_path is None # 参照元のcheckpointがない場合は最後の層を前の層から複製して作るなどダミーの重みを入れる
text_enc_dict = convert_text_encoder_state_dict_to_sd_v2(text_encoder.state_dict(), make_dummy)
update_sd("cond_stage_model.model.", text_enc_dict)
else:
text_enc_dict = text_encoder.state_dict()
update_sd("cond_stage_model.transformer.", text_enc_dict)
2023-03-09 16:06:59 +00:00
2023-04-01 11:14:25 +00:00
# Convert the VAE
if vae is not None:
vae_dict = convert_vae_state_dict(vae.state_dict())
update_sd("first_stage_model.", vae_dict)
2023-03-09 16:06:59 +00:00
2023-04-01 11:14:25 +00:00
# Put together new checkpoint
key_count = len(state_dict.keys())
new_ckpt = {"state_dict": state_dict}
# epoch and global_step are sometimes not int
try:
if "epoch" in checkpoint:
epochs += checkpoint["epoch"]
if "global_step" in checkpoint:
steps += checkpoint["global_step"]
except:
pass
new_ckpt["epoch"] = epochs
new_ckpt["global_step"] = steps
if is_safetensors(output_file):
# TODO Tensor以外のdictの値を削除したほうがいいか
save_file(state_dict, output_file)
2022-12-18 01:36:31 +00:00
else:
2023-04-01 11:14:25 +00:00
torch.save(new_ckpt, output_file)
return key_count
def save_diffusers_checkpoint(v2, output_dir, text_encoder, unet, pretrained_model_name_or_path, vae=None, use_safetensors=False):
if pretrained_model_name_or_path is None:
# load default settings for v1/v2
if v2:
pretrained_model_name_or_path = DIFFUSERS_REF_MODEL_ID_V2
else:
pretrained_model_name_or_path = DIFFUSERS_REF_MODEL_ID_V1
scheduler = DDIMScheduler.from_pretrained(pretrained_model_name_or_path, subfolder="scheduler")
tokenizer = CLIPTokenizer.from_pretrained(pretrained_model_name_or_path, subfolder="tokenizer")
if vae is None:
vae = AutoencoderKL.from_pretrained(pretrained_model_name_or_path, subfolder="vae")
pipeline = StableDiffusionPipeline(
unet=unet,
text_encoder=text_encoder,
vae=vae,
scheduler=scheduler,
tokenizer=tokenizer,
safety_checker=None,
feature_extractor=None,
requires_safety_checker=None,
)
pipeline.save_pretrained(output_dir, safe_serialization=use_safetensors)
2022-12-05 15:49:02 +00:00
2023-03-09 16:06:59 +00:00
VAE_PREFIX = "first_stage_model."
2022-12-05 15:49:02 +00:00
2023-01-01 18:10:32 +00:00
def load_vae(vae_id, dtype):
2023-04-01 11:14:25 +00:00
print(f"load VAE: {vae_id}")
if os.path.isdir(vae_id) or not os.path.isfile(vae_id):
# Diffusers local/remote
try:
vae = AutoencoderKL.from_pretrained(vae_id, subfolder=None, torch_dtype=dtype)
except EnvironmentError as e:
print(f"exception occurs in loading vae: {e}")
print("retry with subfolder='vae'")
vae = AutoencoderKL.from_pretrained(vae_id, subfolder="vae", torch_dtype=dtype)
return vae
# local
vae_config = create_vae_diffusers_config()
if vae_id.endswith(".bin"):
# SD 1.5 VAE on Huggingface
converted_vae_checkpoint = torch.load(vae_id, map_location="cpu")
else:
# StableDiffusion
vae_model = load_file(vae_id, "cpu") if is_safetensors(vae_id) else torch.load(vae_id, map_location="cpu")
vae_sd = vae_model["state_dict"] if "state_dict" in vae_model else vae_model
# vae only or full model
full_model = False
for vae_key in vae_sd:
if vae_key.startswith(VAE_PREFIX):
full_model = True
break
if not full_model:
sd = {}
for key, value in vae_sd.items():
sd[VAE_PREFIX + key] = value
vae_sd = sd
del sd
# Convert the VAE model.
converted_vae_checkpoint = convert_ldm_vae_checkpoint(vae_sd, vae_config)
vae = AutoencoderKL(**vae_config)
vae.load_state_dict(converted_vae_checkpoint)
2023-03-09 16:06:59 +00:00
return vae
2022-12-05 15:49:02 +00:00
# endregion
2023-03-09 16:06:59 +00:00
def make_bucket_resolutions(max_reso, min_size=256, max_size=1024, divisible=64):
2023-04-01 11:14:25 +00:00
max_width, max_height = max_reso
max_area = (max_width // divisible) * (max_height // divisible)
resos = set()
size = int(math.sqrt(max_area)) * divisible
resos.add((size, size))
size = min_size
while size <= max_size:
width = size
height = min(max_size, (max_area // (width // divisible)) * divisible)
resos.add((width, height))
resos.add((height, width))
# # make additional resos
# if width >= height and width - divisible >= min_size:
# resos.add((width - divisible, height))
# resos.add((height, width - divisible))
# if height >= width and height - divisible >= min_size:
# resos.add((width, height - divisible))
# resos.add((height - divisible, width))
size += divisible
resos = list(resos)
resos.sort()
return resos
if __name__ == "__main__":
resos = make_bucket_resolutions((512, 768))
print(len(resos))
print(resos)
aspect_ratios = [w / h for w, h in resos]
print(aspect_ratios)
ars = set()
for ar in aspect_ratios:
if ar in ars:
print("error! duplicate ar:", ar)
ars.add(ar)