MCPcopy Create free account
hub / github.com/OpenImagingLab/FlashVSR / attn_processor

Method attn_processor

diffsynth/models/stepvideo_dit.py:330–336  ·  view source on GitHub ↗
(self, attn_type)

Source from the content-addressed store, hash-verified

328 super().__init__()
329
330 def attn_processor(self, attn_type):
331 if attn_type == 'torch':
332 return self.torch_attn_func
333 elif attn_type == 'parallel':
334 return self.parallel_attn_func
335 else:
336 raise Exception('Not supported attention type...')
337
338 def torch_attn_func(
339 self,

Callers 2

__init__Method · 0.80
__init__Method · 0.80

Calls

no outgoing calls

Tested by

no test coverage detected