MCPcopy Create free account
hub / github.com/NVIDIA/TensorRT / get_minmax_dims

Method get_minmax_dims

demo/Diffusion/models.py:179–192  ·  view source on GitHub ↗
(self, batch_size, image_height, image_width, static_batch, static_shape)

Source from the content-addressed store, hash-verified

177 return (latent_height, latent_width)
178
179 def get_minmax_dims(self, batch_size, image_height, image_width, static_batch, static_shape):
180 min_batch = batch_size if static_batch else self.min_batch
181 max_batch = batch_size if static_batch else self.max_batch
182 latent_height = image_height // 8
183 latent_width = image_width // 8
184 min_image_height = image_height if static_shape else self.min_image_shape
185 max_image_height = image_height if static_shape else self.max_image_shape
186 min_image_width = image_width if static_shape else self.min_image_shape
187 max_image_width = image_width if static_shape else self.max_image_shape
188 min_latent_height = latent_height if static_shape else self.min_latent_shape
189 max_latent_height = latent_height if static_shape else self.max_latent_shape
190 min_latent_width = latent_width if static_shape else self.min_latent_shape
191 max_latent_width = latent_width if static_shape else self.max_latent_shape
192 return (min_batch, max_batch, min_image_height, max_image_height, min_image_width, max_image_width, min_latent_height, max_latent_height, min_latent_width, max_latent_width)
193
194class CLIP(BaseModel):
195 def __init__(self,

Callers 4

get_input_profileMethod · 0.80
get_input_profileMethod · 0.80
get_input_profileMethod · 0.80
get_input_profileMethod · 0.80

Calls

no outgoing calls

Tested by

no test coverage detected