mirrors/icefall
1
0
Fork 0
mirror of https://github.com/k2-fsa/icefall.git synced 2025-09-08 16:44:20 +00:00
Code Issues Packages Projects Releases Wiki Activity

minor updates

Browse Source
This commit is contained in:
JinZr 2023-09-02 16:09:19 +08:00
parent 0d10439a83
commit ca075f2c62
2 changed files with 19 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
egs/multi_zh-hans/ASR/zipformer/export.py
View File

@ -34,7 +34,7 @@ dataset, you should change the argument values according to your dataset.
./zipformer/export.py \
--exp-dir ./zipformer/exp \
--tokens data/lang_bpe_2000/tokens.txt \
--epoch 23 \
--epoch 20 \
--avg 1 \
--jit 1
@ -54,7 +54,7 @@ for how to use the exported models outside of icefall.
--chunk-size 16 \
--left-context-frames 128 \
--tokens data/lang_bpe_2000/tokens.txt \
--epoch 23 \
--epoch 20 \
--avg 1 \
--jit 1
@ -73,7 +73,7 @@ for how to use the exported models outside of icefall.
./zipformer/export.py \
--exp-dir ./zipformer/exp \
--tokens data/lang_bpe_2000/tokens.txt \
--epoch 23 \
--epoch 20 \
--avg 1
- For streaming model:
@ -82,7 +82,7 @@ for how to use the exported models outside of icefall.
--exp-dir ./zipformer/exp \
--causal 1 \
--tokens data/lang_bpe_2000/tokens.txt \
--epoch 23 \
--epoch 20 \
--avg 1
It will generate a file `pretrained.pt` in the given `exp_dir`. You can later
@ -203,7 +203,7 @@ def get_parser():
parser.add_argument(
"--epoch",
type=int,
default=23,
default=20,
help="""It specifies the checkpoint to use for decoding.
Note: Epoch counts from 1.
You can specify --avg to use more checkpoints for model averaging.""",

28
egs/multi_zh-hans/ASR/zipformer/multi_dataset.py
View File

@ -283,17 +283,17 @@ class MultiDataset:
self.fbank_dir / "kespeech" / "kespeech-asr_cuts_dev_phase2.jsonl.gz"
)
# # WeNetSpeech
# logging.info("Loading WeNetSpeech set in lazy mode")
# wenetspeech_test_meeting_cuts = load_manifest_lazy(
# self.fbank_dir / "wenetspeech" / "cuts_TEST_MEETING.jsonl.gz"
# )
# wenetspeech_test_net_cuts = load_manifest_lazy(
# self.fbank_dir / "wenetspeech" / "cuts_TEST_NET.jsonl.gz"
# )
# wenetspeech_dev_cuts = load_manifest_lazy(
# self.fbank_dir / "wenetspeech" / "cuts_DEV.jsonl.gz"
# )
# WeNetSpeech
logging.info("Loading WeNetSpeech set in lazy mode")
wenetspeech_test_meeting_cuts = load_manifest_lazy(
self.fbank_dir / "wenetspeech" / "cuts_TEST_MEETING.jsonl.gz"
)
wenetspeech_test_net_cuts = load_manifest_lazy(
self.fbank_dir / "wenetspeech" / "cuts_TEST_NET.jsonl.gz"
)
wenetspeech_dev_cuts = load_manifest_lazy(
self.fbank_dir / "wenetspeech" / "cuts_DEV.jsonl.gz"
)
return {
"aidatatang_test": aidatatang_test_cuts,
@ -310,7 +310,7 @@ class MultiDataset:
"kespeech-asr_test": kespeech_test_cuts,
"kespeech-asr_dev_phase1": kespeech_dev_phase1_cuts,
"kespeech-asr_dev_phase2": kespeech_dev_phase2_cuts,
# "wenetspeech-meeting_test": wenetspeech_test_meeting_cuts,
# "wenetspeech-net_test": wenetspeech_test_net_cuts,
# "wenetspeech_dev": wenetspeech_dev_cuts,
"wenetspeech-meeting_test": wenetspeech_test_meeting_cuts,
"wenetspeech-net_test": wenetspeech_test_net_cuts,
"wenetspeech_dev": wenetspeech_dev_cuts,
}