mirror of https://github.com/vladmandic/automatic
43 lines
1.7 KiB
Python
43 lines
1.7 KiB
Python
import transformers
|
|
import diffusers
|
|
from huggingface_hub import file_exists
|
|
from modules import shared, devices, sd_models, model_quant, sd_hijack_te
|
|
from pipelines import generic
|
|
|
|
|
|
def load_pixart(checkpoint_info, diffusers_load_config=None):
|
|
if diffusers_load_config is None:
|
|
diffusers_load_config = {}
|
|
repo_id = sd_models.path_to_repo(checkpoint_info)
|
|
sd_models.hf_auth_check(checkpoint_info)
|
|
|
|
repo_id_tenc = repo_id
|
|
repo_id_pipe = repo_id
|
|
|
|
# PixArt-alpha/PixArt-Sigma-XL-2-2K-MS only holds transformer
|
|
if not file_exists(repo_id_tenc, "text_encoder/config.json"):
|
|
repo_id_tenc = "PixArt-alpha/PixArt-Sigma-XL-2-1024-MS"
|
|
if not file_exists(repo_id_pipe, "model_index.json"):
|
|
repo_id_pipe = "PixArt-alpha/PixArt-Sigma-XL-2-1024-MS"
|
|
|
|
load_args, _quant_args = model_quant.get_dit_args(diffusers_load_config, allow_quant=False)
|
|
shared.log.debug(f'Load model: type=PixArtSigma repo="{repo_id}" config={diffusers_load_config} offload={shared.opts.diffusers_offload_mode} dtype={devices.dtype} args={load_args}')
|
|
|
|
transformer = generic.load_transformer(repo_id, cls_name=diffusers.PixArtTransformer2DModel, load_config=diffusers_load_config)
|
|
text_encoder = generic.load_text_encoder(repo_id_tenc, cls_name=transformers.T5EncoderModel, load_config=diffusers_load_config)
|
|
|
|
pipe = diffusers.PixArtSigmaPipeline.from_pretrained(
|
|
repo_id_pipe,
|
|
transformer=transformer,
|
|
text_encoder=text_encoder,
|
|
cache_dir=shared.opts.diffusers_dir,
|
|
**load_args,
|
|
)
|
|
|
|
del text_encoder
|
|
del transformer
|
|
sd_hijack_te.init_hijack(pipe)
|
|
|
|
devices.torch_gc(force=True, reason='load')
|
|
return pipe
|