update rec_sar_head

ppocr/losses/rec_sar_loss.py

@@ -9,7 +9,7 @@ from paddle import nn
 class SARLoss(nn.Layer):
     def __init__(self, **kwargs):
         super(SARLoss, self).__init__()
-        self.loss_func = paddle.nn.loss.CrossEntropyLoss(reduction="mean", ignore_index=92)
+        self.loss_func = paddle.nn.loss.CrossEntropyLoss(reduction="mean", ignore_index=96)
 
     def forward(self, predicts, batch):
         predict = predicts[:, :-1, :] # ignore last index of outputs to be in same seq_len with targets

ppocr/modeling/heads/rec_sar_head.py

@@ -118,8 +118,7 @@ class BaseDecoder(nn.Layer):
 class ParallelSARDecoder(BaseDecoder):
     """
     Args:
-        num_classes (int): Output class number.
-        channels (list[int]): Network layer channels.
+        out_channels (int): Output class number.
         enc_bi_rnn (bool): If True, use bidirectional RNN in encoder.
         dec_bi_rnn (bool): If True, use bidirectional RNN in decoder.
         dec_drop_rnn (float): Dropout of RNN layer in decoder.
@@ -137,7 +136,7 @@ class ParallelSARDecoder(BaseDecoder):
     """
 
     def __init__(self,
-                num_classes=93, # 90 + unknown + start + padding
+                out_channels, # 90 + unknown + start + padding
                 enc_bi_rnn=False,
                 dec_bi_rnn=False,
                 dec_drop_rnn=0.0,
@@ -148,8 +147,6 @@ class ParallelSARDecoder(BaseDecoder):
                 pred_dropout=0.1,
                 max_text_length=30,
                 mask=True,
-                start_idx=91,
-                padding_idx=92, # 92
                 pred_concat=True,
                 **kwargs):
         super().__init__()
@@ -157,7 +154,8 @@ class ParallelSARDecoder(BaseDecoder):
         self.num_classes = num_classes
         self.enc_bi_rnn = enc_bi_rnn
         self.d_k = d_k
-        self.start_idx = start_idx
+        self.start_idx = out_channels - 2
+	self.padding_idx = out_channels - 1
         self.max_seq_len = max_text_length
         self.mask = mask
         self.pred_concat = pred_concat
@@ -191,7 +189,7 @@ class ParallelSARDecoder(BaseDecoder):
 
         # Decoder input embedding
         self.embedding = nn.Embedding(
-            self.num_classes, encoder_rnn_out_size, padding_idx=padding_idx)
+            self.num_classes, encoder_rnn_out_size, padding_idx=self.padding_idx)
         
         # Prediction layer
         self.pred_dropout = nn.Dropout(pred_dropout)
@@ -330,6 +328,7 @@ class ParallelSARDecoder(BaseDecoder):
 
 class SARHead(nn.Layer):
     def __init__(self, 
+		out_channels,
                 enc_bi_rnn=False,
                 enc_drop_rnn=0.1,
                 enc_gru=False,
@@ -351,7 +350,8 @@ class SARHead(nn.Layer):
 
         # decoder module
         self.decoder = ParallelSARDecoder(
-            enc_bi_rnn=enc_bi_rnn,
+            out_channels=out_channels,
+	    enc_bi_rnn=enc_bi_rnn,
             dec_bi_rnn=dec_bi_rnn,
             dec_drop_rnn=dec_drop_rnn,
             dec_gru=dec_gru,
@@ -375,4 +375,4 @@ class SARHead(nn.Layer):
             # (bsz, seq_len, num_classes)
         
         return final_out
-        
\ No newline at end of file
+