add blank penalty

egs/librispeech/ASR/pruned_transducer_stateless2/beam_search.py

@@ -626,6 +626,7 @@ def greedy_search_batch(
     model: Transducer,
     encoder_out: torch.Tensor,
     encoder_out_lens: torch.Tensor,
+    blank_penalty: float = 0,
     return_timestamps: bool = False,
 ) -> Union[List[List[int]], DecodingResults]:
     """Greedy search in batch mode. It hardcodes --max-sym-per-frame=1.
@@ -701,6 +702,7 @@ def greedy_search_batch(
 
         logits = logits.squeeze(1).squeeze(1)  # (batch_size, vocab_size)
         assert logits.ndim == 2, logits.shape
+        logits[:, 0] -= blank_penalty
         y = logits.argmax(dim=1).tolist()
         emitted = False
         for i, v in enumerate(y):

egs/wenetspeech/ASR/pruned_transducer_stateless2/asr_datamodule.py

@@ -292,7 +292,7 @@ class WenetSpeechAsrDataModule:
                 max_duration=self.args.max_duration,
                 shuffle=self.args.shuffle,
                 num_buckets=self.args.num_buckets,
-                buffer_size=30000,
+                buffer_size=300000,
                 drop_last=True,
             )
         else:

egs/wenetspeech/ASR/pruned_transducer_stateless7/decode.py

@@ -125,6 +125,7 @@ from beam_search import (
 )
 from train import add_model_arguments, get_params, get_transducer_model
 
+from lhotse.cut import Cut
 from icefall.char_graph_compiler import CharCtcTrainingGraphCompiler
 from icefall.checkpoint import (
     average_checkpoints,
@@ -302,6 +303,13 @@ def get_parser():
         fast_beam_search_nbest_LG, and fast_beam_search_nbest_oracle""",
     )
 
+    parser.add_argument(
+        "--blank-penalty",
+        type=float,
+        default=0.0,
+        help="",
+    )
+
     add_model_arguments(parser)
 
     return parser
@@ -414,9 +422,7 @@ def decode_one_batch(
             hyps.append([lexicon.token_table[idx] for idx in hyp_tokens[i]])
     elif params.decoding_method == "greedy_search" and params.max_sym_per_frame == 1:
         hyp_tokens = greedy_search_batch(
-            model=model,
-            encoder_out=encoder_out,
-            encoder_out_lens=encoder_out_lens,
+            model=model, encoder_out=encoder_out, encoder_out_lens=encoder_out_lens, blank_penalty=params.blank_penalty,
         )
         for i in range(encoder_out.size(0)):
             hyps.append([lexicon.token_table[idx] for idx in hyp_tokens[i]])
@@ -444,9 +450,7 @@ def decode_one_batch(
                 )
             elif params.decoding_method == "beam_search":
                 hyp = beam_search(
-                    model=model,
-                    encoder_out=encoder_out_i,
-                    beam=params.beam_size,
+                    model=model, encoder_out=encoder_out_i, beam=params.beam_size,
                 )
             else:
                 raise ValueError(
@@ -625,6 +629,7 @@ def main():
     else:
         params.suffix += f"-context-{params.context_size}"
         params.suffix += f"-max-sym-per-frame-{params.max_sym_per_frame}"
+    params.suffix += f"-blank-penalty-{params.blank_penalty}"
 
     if params.use_averaged_model:
         params.suffix += "-use-averaged-model"
@@ -751,17 +756,30 @@ def main():
     args.return_cuts = True
     wenetspeech = WenetSpeechAsrDataModule(args)
 
+    def remove_short_utt(c: Cut):
+        T = ((c.num_frames - 7) // 2 + 1) // 2
+        if T <= 0:
+            logging.warning(
+                f"Exclude cut with ID {c.id} from decoding, num_frames : {c.num_frames}."
+            )
+        return T > 0
+
     dev_cuts = wenetspeech.valid_cuts()
+    dev_cuts = dev_cuts.filter(remove_short_utt)
     dev_dl = wenetspeech.valid_dataloaders(dev_cuts)
 
     test_net_cuts = wenetspeech.test_net_cuts()
+    test_net_cuts = test_net_cuts.filter(remove_short_utt)
     test_net_dl = wenetspeech.test_dataloaders(test_net_cuts)
 
     test_meeting_cuts = wenetspeech.test_meeting_cuts()
+    test_meeting_cuts = test_meeting_cuts.filter(remove_short_utt)
     test_meeting_dl = wenetspeech.test_dataloaders(test_meeting_cuts)
 
     test_sets = ["DEV", "TEST_NET", "TEST_MEETING"]
     test_dl = [dev_dl, test_net_dl, test_meeting_dl]
+    # test_sets = ["TEST_MEETING"]
+    # test_dl = [test_meeting_dl]
 
     for test_set, test_dl in zip(test_sets, test_dl):
         results_dict = decode_dataset(
@@ -774,9 +792,7 @@ def main():
         )
 
         save_results(
-            params=params,
-            test_set_name=test_set,
-            results_dict=results_dict,
+            params=params, test_set_name=test_set, results_dict=results_dict,
         )
 
     logging.info("Done!")