mirror of https://github.com/vladmandic/automatic
56 lines
2.4 KiB
Python
56 lines
2.4 KiB
Python
import transformers
|
|
import diffusers
|
|
from modules import shared, sd_models, sd_hijack_te, devices, model_quant
|
|
from modules.logger import log
|
|
from pipelines import generic
|
|
|
|
|
|
def load_lumina(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_config, _quant_config = model_quant.get_dit_args(diffusers_load_config, allow_quant=False)
|
|
log.debug(f'Load model: type=LuminaSFT repo="{repo_id}" config={diffusers_load_config} offload={shared.opts.diffusers_offload_mode} dtype={devices.dtype} args={diffusers_load_config}')
|
|
pipe = diffusers.LuminaText2ImgPipeline.from_pretrained(
|
|
'Alpha-VLLM/Lumina-Next-SFT-diffusers',
|
|
cache_dir = shared.opts.diffusers_dir,
|
|
**load_config,
|
|
)
|
|
sd_hijack_te.init_hijack(pipe)
|
|
devices.torch_gc(force=True, reason='load')
|
|
return pipe
|
|
|
|
|
|
def load_lumina2(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)
|
|
|
|
if shared.opts.teacache_enabled:
|
|
from modules import teacache
|
|
log.debug(f'Transformers cache: type=teacache patch=forward cls={diffusers.Lumina2Transformer2DModel.__name__}')
|
|
diffusers.Lumina2Transformer2DModel.forward = teacache.teacache_lumina2_forward # patch must be done before transformer is loaded
|
|
|
|
log.debug(f'Load model: type=Lumina2 repo="{repo_id}" config={diffusers_load_config} offload={shared.opts.diffusers_offload_mode} dtype={devices.dtype} args={diffusers_load_config}')
|
|
transformer = generic.load_transformer(repo_id, cls_name=diffusers.Lumina2Transformer2DModel, load_config=diffusers_load_config)
|
|
text_encoder = generic.load_text_encoder(repo_id, cls_name=transformers.Gemma2Model, load_config=diffusers_load_config)
|
|
|
|
load_config, _quant_args = model_quant.get_dit_args(diffusers_load_config, allow_quant=False)
|
|
pipe = diffusers.Lumina2Pipeline.from_pretrained(
|
|
repo_id,
|
|
cache_dir=shared.opts.diffusers_dir,
|
|
text_encoder=text_encoder,
|
|
transformer=transformer,
|
|
**load_config,
|
|
)
|
|
|
|
del transformer
|
|
del text_encoder
|
|
sd_hijack_te.init_hijack(pipe)
|
|
|
|
devices.torch_gc(force=True, reason='load')
|
|
return pipe
|