Question about MASTER model for text recognition ?

[ThorPham]

decoder=dict( type='MasterDecoder', d_model=512, n_head=8, attn_drop=0., ffn_drop=0., d_inner=2048, n_layers=3, feat_pe_drop=0.2, feat_size=6 * 40, postprocessor=dict(type='AttentionPostprocessor'), module_loss=dict( type='CEModuleLoss', reduction='mean', ignore_first_char=True), max_seq_len=30, dictionary=dictionary), data_preprocessor=dict( type='TextRecogDataPreprocessor', mean=[127.5, 127.5, 127.5], std=[127.5, 127.5, 127.5]))

Anyone help me understand feat_size=6 * 40 in _base_master_resnet31.py . 6 and 40 come from, if i change size of image how to calculate fieat_size. Thank you so much