Datasets:
jblitzar
/
github-python-corpus

Dataset card Data Studio Files
xet
 Community
text
stringlengths
1
93.6k
# blend the above tile and the left tile
# to the current tile and add the current tile to the result row
if i > 0:
tile = self.blend_v(
rows[i - 1][j], tile, blend_extent_height)
if j > 0:
tile = self.blend_h(row[j - 1], tile, blend_extent_width)
result_row.append(
tile[:, :, :, :row_limit_height, :row_limit_width])
result_rows.append(torch.cat(result_row, dim=4))
enc = torch.cat(result_rows, dim=3)
return enc
def _encode(
self, x: torch.Tensor, return_dict: bool = True
):
batch_size, num_channels, num_frames, height, width = x.shape
if self.use_encode_tiling and (width > self.encode_tile_sample_min_width or height > self.encode_tile_sample_min_height):
return self.tiled_encode(x)
if num_frames == 1:
h = self.encoder(x)
if self.quant_conv is not None:
h = self.quant_conv(h)
posterior = DiagonalGaussianDistribution(h)
else:
frame_batch_size = 4
h = []
for i in range(num_frames // frame_batch_size):
remaining_frames = num_frames % frame_batch_size
start_frame = frame_batch_size * i + \
(0 if i == 0 else remaining_frames)
end_frame = frame_batch_size * (i + 1) + remaining_frames
z_intermediate = x[:, :, start_frame:end_frame]
z_intermediate = self.encoder(z_intermediate)
if self.quant_conv is not None:
z_intermediate = self.quant_conv(z_intermediate)
h.append(z_intermediate)
self._clear_fake_context_parallel_cache()
h = torch.cat(h, dim=2)
return h
def enable_encode_tiling(
self,
tile_sample_min_height: Optional[int] = None,
tile_sample_min_width: Optional[int] = None,
tile_overlap_factor_height: Optional[float] = None,
tile_overlap_factor_width: Optional[float] = None,
) -> None:
r"""
Enable tiled VAE decoding. When this option is enabled, the VAE will split the input tensor into tiles to
compute decoding and encoding in several steps. This is useful for saving a large amount of memory and to allow
processing larger images.
Args:
tile_sample_min_height (`int`, *optional*):
The minimum height required for a sample to be separated into tiles across the height dimension.
tile_sample_min_width (`int`, *optional*):
The minimum width required for a sample to be separated into tiles across the width dimension.
tile_overlap_factor_height (`int`, *optional*):
The minimum amount of overlap between two consecutive vertical tiles. This is to ensure that there are
no tiling artifacts produced across the height dimension. Must be between 0 and 1. Setting a higher
value might cause more tiles to be processed leading to slow down of the decoding process.
tile_overlap_factor_width (`int`, *optional*):
The minimum amount of overlap between two consecutive horizontal tiles. This is to ensure that there
are no tiling artifacts produced across the width dimension. Must be between 0 and 1. Setting a higher
value might cause more tiles to be processed leading to slow down of the decoding process.
"""
self.use_encode_tiling = True
self.encode_tile_sample_min_height = tile_sample_min_height or self.tile_sample_min_height
print("encode_tile_sample_min_height: ", self.encode_tile_sample_min_height)
self.encode_tile_sample_min_width = tile_sample_min_width or self.tile_sample_min_width
print("encode_tile_sample_min_width: ", self.encode_tile_sample_min_width)
self.encode_tile_latent_min_height = int(
self.encode_tile_sample_min_height /
(2 ** (len(self.config.block_out_channels) - 1))
)
self.encode_tile_latent_min_width = int(
self.encode_tile_sample_min_width / (2 ** (len(self.config.block_out_channels) - 1)))
print("encode_tile_latent_min_height: ", self.encode_tile_latent_min_height)
print("encode_tile_latent_min_width: ", self.encode_tile_latent_min_width)
self.encode_tile_overlap_factor_height = tile_overlap_factor_height or self.tile_overlap_factor_height
print("encode_tile_overlap_factor_height: ", self.encode_tile_overlap_factor_height)
self.encode_tile_overlap_factor_width = tile_overlap_factor_width or self.tile_overlap_factor_width
print("encode_tile_overlap_factor_width: ", self.encode_tile_overlap_factor_width)
from types import MethodType
def enable_vae_encode_tiling(vae):
vae.encode = MethodType(encode, vae)
setattr(vae, "_encode", MethodType(_encode, vae))
setattr(vae, "tiled_encode", MethodType(tiled_encode, vae))
setattr(vae, "use_encode_tiling", True)
setattr(vae, "enable_encode_tiling", MethodType(enable_encode_tiling, vae))
vae.enable_encode_tiling()