do some changes

egs/wenetspeech/ASR/pruned_transducer_stateless3/asr_datamodule.py

@@ -27,6 +27,7 @@ from lhotse import (
     CutSet,
     Fbank,
     FbankConfig,
+    load_manifest,
     load_manifest_lazy,
     set_caching_enabled,
 )
@@ -191,13 +192,6 @@ class WenetSpeechAsrDataModule:
             "with training dataset. ",
         )
 
-        group.add_argument(
-            "--lazy-load",
-            type=str2bool,
-            default=True,
-            help="lazily open CutSets to avoid OOM (for L|XL subset)",
-        )
-
         group.add_argument(
             "--training-subset",
             type=str,
@@ -218,7 +212,7 @@ class WenetSpeechAsrDataModule:
             The state dict for the training sampler.
         """
         logging.info("About to get Musan cuts")
-        cuts_musan = load_manifest_lazy(
+        cuts_musan = load_manifest(
             self.args.manifest_dir / "musan_cuts.jsonl.gz"
         )
 
@@ -419,18 +413,10 @@ class WenetSpeechAsrDataModule:
     @lru_cache()
     def train_cuts(self) -> CutSet:
         logging.info("About to get train cuts")
-        if self.args.lazy_load:
-            logging.info("use lazy cuts")
-            cuts_train = CutSet.from_jsonl_lazy(
-                self.args.manifest_dir
-                / f"cuts_{self.args.training_subset}.jsonl.gz"
-            )
-        else:
-            cuts_train = CutSet.from_file(
-                self.args.manifest_dir
-                / f"cuts_{self.args.training_subset}.jsonl.gz"
-            )
-        return cuts_train
+        return load_manifest_lazy(
+            self.args.manifest_dir
+            / f"cuts_{self.args.training_subset}.jsonl.gz"
+        )
 
     @lru_cache()
     def valid_cuts(self) -> CutSet:

egs/wenetspeech/ASR/pruned_transducer_stateless3/train.py

@@ -883,7 +883,7 @@ def run(rank, world_size, args):
         dst_state_dict.update(pretrained_dict)
         model.load_state_dict(dst_state_dict)
 
-        initial_lr = 1.5e-4
+        initial_lr = 1e-3
 
     checkpoints = load_checkpoint_if_available(params=params, model=model)
 
@@ -1012,6 +1012,7 @@ def scan_pessimistic_batches_for_oom(
     graph_compiler: CharCtcTrainingGraphCompiler,
     params: AttributeDict,
 ):
+    return
     from lhotse.dataset import find_pessimistic_batches
 
     logging.info(