MCPcopy Create free account
hub / github.com/CompVis/zigma / allocate_inference_cache

Method allocate_inference_cache

model_zigma.py:462–465  ·  view source on GitHub ↗
(self, batch_size, max_seqlen, dtype=None, **kwargs)

Source from the content-addressed store, hash-verified

460 return x, residual
461
462 def allocate_inference_cache(self, batch_size, max_seqlen, dtype=None, **kwargs):
463 return self.mixer.allocate_inference_cache(
464 batch_size, max_seqlen, dtype=dtype, **kwargs
465 )
466
467
468def create_block(

Callers

nothing calls this directly

Calls

no outgoing calls

Tested by

no test coverage detected