Functions — github.com/TimSeizinger/Bokehlicious

CodeHub MCPcopy Create free account

hub / github.com/TimSeizinger/Bokehlicious / functions

Functions80 in github.com/TimSeizinger/Bokehlicious

⨍ Functions 80 ◇ Types & classes 25

↓ 13 callersMethod__init__

(self, dim, kernel_size, stride, padding)

method/nn_util.py:70

↓ 5 callersFunctionbokehlicious_config_builder

( in_chans = 3, im_range = 1., positional_enc = True, positional_dfe = True,

method/config.py:26

↓ 4 callersMethod__init__

Modified block from NAFNet with optional CoordConv on th first convolution. Args: channels: Number of input channels

method/blocks.py:154

↓ 3 callersFunctionappend_av

(av_dict, key, value)

↓ 3 callersFunctionbokehlicious_size_builder

method/config.py:1

↓ 3 callersFunctionbuild_input_dict

(maps: dict, aperture_embedding: float, target_name: str, device='cuda')

dataset/util.py:80

↓ 3 callersFunctioncalculate_aperture_embedding

(tgt_av: float, max_av: float = 2.0)

dataset/util.py:47

↓ 3 callersFunctioncrop_to_divisible

(img: Image, divisor: int)

dataset/util.py:93

↓ 3 callersFunctiongenerate_maps

(source: Image, aperture_embedding, target: Image = None)

dataset/util.py:50

↓ 3 callersFunctionget_base_parser

util/parser.py:4

↓ 2 callersFunctionadd_network_args

(parser: ArgumentParser)

util/parser.py:10

↓ 2 callersMethodgenerate_1d_decay

generate 1d decay mask, the result is l*l

method/nn_util.py:268

↓ 1 callersFunctionadd_img_args

(parser: ArgumentParser)

util/parser.py:14

↓ 1 callersFunctioncenter_crop

(img:Image, new_width, new_height)

dataset/util.py:97

↓ 1 callersFunctiondownsample

(img: Image, max_dim: int, interpolation)

dataset/util.py:112

↓ 1 callersMethodforward_features

(self, x, pos_map=None, bokeh_strength=None)

method/model.py:235

↓ 1 callersFunctionget_activation

Get the activation function from the activation type. :param activation_type: Activation type :return: Activation function

method/util.py:9

↓ 1 callersFunctionget_cnn_attention

(attention_type: Attention_Mechanism = None)

method/util.py:33

↓ 1 callersFunctionget_eval_parser

util/parser.py:26

↓ 1 callersFunctionget_map

(w: int, h: int, bokeh_strength: float)

dataset/util.py:90

↓ 1 callersFunctionget_ntire_parser

util/parser.py:33

↓ 1 callersFunctionget_pos_map

(w: int, h: int)

dataset/util.py:64

↓ 1 callersFunctionget_predict_parser

util/parser.py:20

↓ 1 callersFunctionget_resolution

(height, width, longest_side)

dataset/util.py:121

↓ 1 callersFunctionload_image

Loads an image and prepares it for input to the model :param img_path: path to the image :param target_av: target aperture value between

dataset/util.py:17

↓ 1 callersFunctionpreprocess_batch

↓ 1 callersMethodreset_parameters

method/nn_util.py:150

↓ 1 callersFunctionunsqueeeze_batch

submit_ntire.py:33

Method__getitem__

(self, index: int)

dataset/loader.py:70

Method__getitem__

(self, index: int)

dataset/loader.py:130

(self, data_path: Union[str, Path], mode: Mode, binary_bokeh: bool = False,

dataset/loader.py:19

(self, data_path: Union[str, Path], mode: Mode, device: str = 'cuda')

dataset/loader.py:106

(self, embed_dim, embed_dim_next, depth, num_heads, heads_range, init_value, ffn_dim=96, laye

method/blocks.py:31

(self, embed_dim, depth, num_heads, init_value: float, heads_range: float, f

method/blocks.py:88

(self, embed_dim: int, num_heads: int, ffn_dim: int, drop_path=0., layerscale=False, norm_lay

method/blocks.py:125

( self, embed_dim, ffn_dim, activation_fn=F.gelu,

method/blocks.py:249

method/nn_util.py:13

method/nn_util.py:20

method/nn_util.py:27

method/nn_util.py:34

(self, embed_dim, embed_dim_next)

method/nn_util.py:41

(self, in_channels, out_channels, kernel_size, padding='same', bias=True)

method/nn_util.py:54

:param num_channel: Number of channels in the input tensor :param apply_att_weights: Should attention weights be applied in the forw

method/nn_util.py:87

:param num_channel: Number of channels in the input tensor :param apply_att_weights: Should attention weights be applied in the forw

method/nn_util.py:111

(self, embed_dim, num_heads, value_factor=1)

method/nn_util.py:133

(self, channels, eps=1e-6)

method/nn_util.py:238

recurrent_chunk_size: (clh clw) num_chunks: (nch ncw) clh * clw == cl nch * ncw == nc default: clh==clw, clh

method/nn_util.py:251

(self, embed_dim=96, norm_layer=None)

method/nn_util.py:329

(self, embed_dim=96)

method/nn_util.py:348

(self, # General Args in_chans=3, img_range=1., dynamic_conv_k=4,

method/model.py:14

dataset/loader.py:62

dataset/loader.py:127

Method_init_weights

method/model.py:208

(ctx, grad_output)

method/nn_util.py:223

Methodcheck_image_size

method/model.py:228

(self, x, x_size, pos_map=None, att_range_factor=None, **kwargs)

method/blocks.py:72

(self, x, att_range_factor=None)

method/blocks.py:108

(self, x: torch.Tensor, attention_rel_pos=None)

method/blocks.py:141

(self, source: Tensor, pos_map: Tensor = None)

method/blocks.py:220

method/blocks.py:276

(self, x=None, *args, **kwargs)

method/nn_util.py:16

(self, x: Tensor, y: Tensor)

method/nn_util.py:23

(self, x: Tensor, skip: Tensor)

method/nn_util.py:30

(self, x: Tensor, weights: Tensor)

method/nn_util.py:37

(self, x, x_size)

method/nn_util.py:47

method/nn_util.py:63

method/nn_util.py:74

(self, x: Tensor)

method/nn_util.py:105

(self, x: Tensor, att_weights: Tensor = None)

method/nn_util.py:127

(self, x: torch.Tensor, rel_pos)

method/nn_util.py:157

method/nn_util.py:205

(ctx, x, weight, bias, eps)

method/nn_util.py:212

method/nn_util.py:245

(self, slen: Tuple[int], range_factor: Tensor)

method/nn_util.py:316

method/nn_util.py:333

(self, x, x_size)

method/nn_util.py:352

(self, source: Tensor, bokeh_strength: Tensor = None, pos_map: Tensor = None, bokeh_strength_m

method/model.py:252

Methodno_weight_decay

method/model.py:221

Methodno_weight_decay_keywords

method/model.py:225

Methodreset_parameters

method/blocks.py:270