Fix decode.py

egs/librispeech/ASR/conformer_mmi_phone/decode.py

@@ -16,7 +16,6 @@ import torch.nn as nn
 from asr_datamodule import LibriSpeechAsrDataModule
 from conformer import Conformer
 
-from icefall.bpe_graph_compiler import BpeCtcTrainingGraphCompiler
 from icefall.checkpoint import average_checkpoints, load_checkpoint
 from icefall.decode import (
     get_lattice,
@@ -28,6 +27,7 @@ from icefall.decode import (
     rescore_with_whole_lattice,
 )
 from icefall.lexicon import Lexicon
+from icefall.mmi_graph_compiler import MmiTrainingGraphCompiler
 from icefall.utils import (
     AttributeDict,
     get_texts,
@@ -58,6 +58,11 @@ def get_parser():
         "'--epoch'. ",
     )
 
+    parser.add_argument(
+        "--method",
+        type=str,
+    )
+
     parser.add_argument(
         "--lattice-score-scale",
         type=float,
@@ -82,7 +87,7 @@ def get_params() -> AttributeDict:
             "nhead": 8,
             "attention_dim": 512,
             "subsampling_factor": 4,
-            "num_decoder_layers": 6,
+            "num_decoder_layers": 0,
             "vgg_frontend": False,
             "is_espnet_structure": True,
             "mmi_loss": False,
@@ -102,7 +107,7 @@ def get_params() -> AttributeDict:
             #  "method": "nbest",
             #  "method": "nbest-rescoring",
             #  "method": "whole-lattice-rescoring",
-            "method": "attention-decoder",
+            #  "method": "attention-decoder",
             #  "method": "nbest-oracle",
             # num_paths is used when method is "nbest", "nbest-rescoring",
             # attention-decoder, and nbest-oracle
@@ -118,8 +123,6 @@ def decode_one_batch(
     HLG: k2.Fsa,
     batch: dict,
     lexicon: Lexicon,
-    sos_id: int,
-    eos_id: int,
     G: Optional[k2.Fsa] = None,
 ) -> Dict[str, List[List[int]]]:
     """Decode one batch and return the result in a dict. The dict has the
@@ -153,10 +156,6 @@ def decode_one_batch(
         for the format of the `batch`.
       lexicon:
         It contains word symbol table.
-      sos_id:
-        The token ID of the SOS.
-      eos_id:
-        The token ID of the EOS.
       G:
         An LM. It is not None when params.method is "nbest-rescoring"
         or "whole-lattice-rescoring". In general, the G in HLG
@@ -234,7 +233,8 @@ def decode_one_batch(
         "attention-decoder",
     ]
 
-    lm_scale_list = [0.8, 0.9, 1.0, 1.1, 1.2, 1.3]
+    lm_scale_list = [0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7]
+    lm_scale_list += [0.8, 0.9, 1.0, 1.1, 1.2, 1.3]
     lm_scale_list += [1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2.0]
 
     if params.method == "nbest-rescoring":
@@ -261,8 +261,6 @@ def decode_one_batch(
             model=model,
             memory=memory,
             memory_key_padding_mask=memory_key_padding_mask,
-            sos_id=sos_id,
-            eos_id=eos_id,
             scale=params.lattice_score_scale,
         )
     else:
@@ -282,8 +280,6 @@ def decode_dataset(
     model: nn.Module,
     HLG: k2.Fsa,
     lexicon: Lexicon,
-    sos_id: int,
-    eos_id: int,
     G: Optional[k2.Fsa] = None,
 ) -> Dict[str, List[Tuple[List[int], List[int]]]]:
     """Decode dataset.
@@ -299,10 +295,6 @@ def decode_dataset(
         The decoding graph.
       lexicon:
         It contains word symbol table.
-      sos_id:
-        The token ID for SOS.
-      eos_id:
-        The token ID for EOS.
       G:
         An LM. It is not None when params.method is "nbest-rescoring"
         or "whole-lattice-rescoring". In general, the G in HLG
@@ -334,8 +326,6 @@ def decode_dataset(
             batch=batch,
             lexicon=lexicon,
             G=G,
-            sos_id=sos_id,
-            eos_id=eos_id,
         )
 
         for lm_scale, hyps in hyps_dict.items():
@@ -427,14 +417,10 @@ def main():
 
     logging.info(f"device: {device}")
 
-    graph_compiler = BpeCtcTrainingGraphCompiler(
+    graph_compiler = MmiTrainingGraphCompiler(
         params.lang_dir,
         device=device,
-        sos_token="<sos/eos>",
-        eos_token="<sos/eos>",
     )
-    sos_id = graph_compiler.sos_id
-    eos_id = graph_compiler.eos_id
 
     HLG = k2.Fsa.from_dict(
         torch.load(f"{params.lang_dir}/HLG.pt", map_location="cpu")
@@ -530,8 +516,6 @@ def main():
             HLG=HLG,
             lexicon=lexicon,
             G=G,
-            sos_id=sos_id,
-            eos_id=eos_id,
         )
 
         save_results(

pyproject.toml

@@ -4,8 +4,11 @@ profile = "black"
 [tool.black]
 line-length = 80
 exclude = '''
-/(
+(
+  /(
       \.git
     | \.github
-)/
+    | icefall/shared/*
+  )/
+)
 '''