mirror of
https://github.com/vladmandic/sdnext.git
synced 2026-01-27 15:02:48 +03:00
60 lines
2.3 KiB
Python
60 lines
2.3 KiB
Python
import diffusers
|
|
import transformers
|
|
from modules import shared, devices, sd_models, model_quant, sd_hijack_te, sd_hijack_vae
|
|
from pipelines import generic
|
|
|
|
|
|
def postprocess(p, result): # pylint: disable=unused-argument
|
|
shared.log.debug('Postprocess: model=ChronoEdit')
|
|
if result is not None and hasattr(result, 'images'):
|
|
result.images = result.images[-1]
|
|
return result
|
|
|
|
|
|
def load_chrono(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)
|
|
|
|
load_args, _quant_args = model_quant.get_dit_args(diffusers_load_config, allow_quant=False)
|
|
shared.log.debug(f'Load model: type=ChronoEdit 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.ChronoEditTransformer3DModel, load_config=diffusers_load_config, subfolder="transformer")
|
|
text_encoder = generic.load_text_encoder(repo_id, cls_name=transformers.UMT5EncoderModel, load_config=diffusers_load_config, subfolder="text_encoder")
|
|
|
|
try:
|
|
pipe = diffusers.ChronoEditPipeline.from_pretrained(
|
|
repo_id,
|
|
transformer=transformer,
|
|
text_encoder=text_encoder,
|
|
cache_dir=shared.opts.diffusers_dir,
|
|
**load_args,
|
|
)
|
|
except Exception as e:
|
|
import os
|
|
from modules import errors
|
|
errors.display(e, 'Chrono')
|
|
os._exit(1)
|
|
pipe.postprocess = postprocess
|
|
pipe.task_args = {
|
|
'num_temporal_reasoning_steps': shared.opts.model_chrono_temporal_steps,
|
|
'output_type': 'np',
|
|
}
|
|
# reference: <https://github.com/nv-tlabs/ChronoEdit/blob/main/scripts/run_inference_diffusers.py>
|
|
if shared.opts.model_chrono_temporal_steps > 0:
|
|
pipe.task_args['num_frames'] = 29
|
|
pipe.task_args['enable_temporal_reasoning'] = True
|
|
else:
|
|
pipe.task_args['num_frames'] = 5
|
|
pipe.task_args['enable_temporal_reasoning'] = False
|
|
|
|
del text_encoder
|
|
del transformer
|
|
|
|
sd_hijack_te.init_hijack(pipe)
|
|
sd_hijack_vae.init_hijack(pipe)
|
|
|
|
devices.torch_gc()
|
|
return pipe
|