'''mcibiplusplus_upernet_swinlarge_pascalcontext59''' import os import copy from .base_cfg import SEGMENTOR_CFG from .._base_ import DATASET_CFG_PASCALCONTEXT59_480x480, DATALOADER_CFG_BS16 # deepcopy SEGMENTOR_CFG = copy.deepcopy(SEGMENTOR_CFG) # modify dataset config SEGMENTOR_CFG['dataset'] = DATASET_CFG_PASCALCONTEXT59_480x480.copy() # modify dataloader config SEGMENTOR_CFG['dataloader'] = DATALOADER_CFG_BS16.copy() # modify scheduler config SEGMENTOR_CFG['scheduler']['max_epochs'] = 260 SEGMENTOR_CFG['scheduler']['min_lr'] = 0.0 SEGMENTOR_CFG['scheduler']['power'] = 1.0 SEGMENTOR_CFG['scheduler']['warmup_cfg'] = {'type': 'linear', 'ratio': 1e-6, 'iters': 1500} SEGMENTOR_CFG['scheduler']['optimizer'] = { 'type': 'AdamW', 'lr': 0.00006, 'betas': (0.9, 0.999), 'weight_decay': 0.01, 'params_rules': { 'absolute_pos_embed': dict(wd_multiplier=0.), 'relative_position_bias_table': dict(wd_multiplier=0.), 'norm': dict(wd_multiplier=0.), }, } # modify other segmentor configs SEGMENTOR_CFG['num_classes'] = 59 SEGMENTOR_CFG['backbone'] = { 'type': 'SwinTransformer', 'structure_type': 'swin_large_patch4_window12_384_22k', 'pretrained': True, 'selected_indices': (0, 1, 2, 3), 'norm_cfg': {'type': 'LayerNorm'}, 'pretrain_img_size': 384, 'in_channels': 3, 'embed_dims': 192, 'patch_size': 4, 'window_size': 12, 'mlp_ratio': 4, 'depths': [2, 2, 18, 2], 'num_heads': [6, 12, 24, 48], 'qkv_bias': True, 'qk_scale': None, 'patch_norm': True, 'drop_rate': 0., 'attn_drop_rate': 0., 'drop_path_rate': 0.3, 'use_abs_pos_embed': False, } SEGMENTOR_CFG['head']['fpn'] = { 'in_channels_list': [192, 384, 768, 1024], 'feats_channels': 1024, 'out_channels': 512, } SEGMENTOR_CFG['head']['decoder'] = { 'pr': {'in_channels': 512, 'out_channels': 512, 'dropout': 0.1}, 'cwi': {'in_channels': 512, 'out_channels': 512, 'dropout': 0.1}, 'cls': {'in_channels': 2560, 'out_channels': 512, 'dropout': 0.1, 'kernel_size': 3, 'padding': 1}, } SEGMENTOR_CFG['head']['in_channels'] = 1536 SEGMENTOR_CFG['head']['context_within_image']['is_on'] = True SEGMENTOR_CFG['head']['context_within_image']['use_self_attention'] = False SEGMENTOR_CFG['auxiliary'] = {'in_channels': 768, 'out_channels': 512, 'dropout': 0.1} SEGMENTOR_CFG['inference'] = { 'forward': {'mode': 'slide', 'cropsize': (480, 480), 'stride': (320, 320)}, 'tta': {'multiscale': [1], 'flip': False, 'use_probs_before_resize': True}, 'evaluate': {'metric_list': ['iou', 'miou']}, } SEGMENTOR_CFG['work_dir'] = os.path.split(__file__)[-1].split('.')[0] SEGMENTOR_CFG['logger_handle_cfg']['logfilepath'] = os.path.join(SEGMENTOR_CFG['work_dir'], f"{os.path.split(__file__)[-1].split('.')[0]}.log") # modify inference config # --single scale SEGMENTOR_CFG['inference'] = SEGMENTOR_CFG['inference'].copy() # --multi scale ''' SEGMENTOR_CFG['inference'] = { 'forward': {'mode': 'slide', 'cropsize': (480, 480), 'stride': (320, 320)}, 'tta': {'multiscale': [0.5, 0.75, 1.0, 1.25, 1.5, 1.75], 'flip': True, 'use_probs_before_resize': True}, 'evaluate': {'metric_list': ['iou', 'miou']}, } '''