Method __init__

LDPS_Graph/layers/PatchTST_backbone.py:131–156

(self, c_in, patch_num, patch_len, max_seq_len=1024,
                 n_layers=3, d_model=128, n_heads=16, d_k=None, d_v=None,
                 d_ff=256, norm='BatchNorm', attn_dropout=0., dropout=0., act="gelu", store_attn=False,
                 key_padding_mask='auto', padding_var=None, attn_mask=None, res_attention=True, pre_norm=False,
                 pe='zeros', learn_pe=True, verbose=False, **kwargs)

    
class TSTiEncoder(nn.Module):  #i means channel-independent
    def __init__(self, c_in, patch_num, patch_len, max_seq_len=1024,
                 n_layers=3, d_model=128, n_heads=16, d_k=None, d_v=None,
                 d_ff=256, norm='BatchNorm', attn_dropout=0., dropout=0., act="gelu", store_attn=False,
                 key_padding_mask='auto', padding_var=None, attn_mask=None, res_attention=True, pre_norm=False,
                 pe='zeros', learn_pe=True, verbose=False, **kwargs):
        
        
        super().__init__()
        
        self.patch_num = patch_num
        self.patch_len = patch_len
        
        # Input encoding
        q_len = patch_num
        self.W_P = nn.Linear(patch_len, d_model)        # Eq 1: projection of feature vectors onto a d-dim vector space
        self.seq_len = q_len

        # Positional encoding
        self.W_pos = positional_encoding(pe, learn_pe, q_len, d_model)

        # Residual dropout
        self.dropout = nn.Dropout(dropout)

        # Encoder
        self.encoder = TSTEncoder(q_len, d_model, n_heads, d_k=d_k, d_v=d_v, d_ff=d_ff, norm=norm, attn_dropout=attn_dropout, dropout=dropout,
                                   pre_norm=pre_norm, activation=act, res_attention=res_attention, n_layers=n_layers, store_attn=store_attn)

        
    def forward(self, x,