Transformer+CNN structure is added to vision transformer type

This commit is contained in:
vahidrezanezhad 2024-06-12 17:39:57 +02:00
parent f1fd74c7eb
commit 743f2e97d6
3 changed files with 176 additions and 39 deletions

View file

@ -2,9 +2,9 @@
"backbone_type" : "transformer",
"task": "binarization",
"n_classes" : 2,
"n_epochs" : 1,
"n_epochs" : 2,
"input_height" : 224,
"input_width" : 672,
"input_width" : 224,
"weight_decay" : 1e-6,
"n_batch" : 1,
"learning_rate": 1e-4,
@ -22,10 +22,14 @@
"scaling_flip" : false,
"rotation": false,
"rotation_not_90": false,
"transformer_num_patches_xy": [7, 7],
"transformer_patchsize_x": 3,
"transformer_patchsize_y": 1,
"transformer_projection_dim": 192,
"transformer_num_patches_xy": [56, 56],
"transformer_patchsize_x": 4,
"transformer_patchsize_y": 4,
"transformer_projection_dim": 64,
"transformer_mlp_head_units": [128, 64],
"transformer_layers": 1,
"transformer_num_heads": 1,
"transformer_cnn_first": false,
"blur_k" : ["blur","guass","median"],
"scales" : [0.6, 0.7, 0.8, 0.9, 1.1, 1.2, 1.4],
"brightness" : [1.3, 1.5, 1.7, 2],