mirrors/icefall
1
0
Fork 0
mirror of https://github.com/k2-fsa/icefall.git synced 2025-08-08 09:32:20 +00:00
Code Issues Packages Projects Releases Wiki Activity
icefall/egs/librispeech/ASR/pruned_transducer_stateless7_streaming
History
Li Peng 3e4da5f781
Replace deprecated pytorch methods (#1814) 
* Replace deprecated pytorch methods

- torch.cuda.amp.GradScaler(...) => torch.amp.GradScaler("cuda", ...)
- torch.cuda.amp.autocast(...) => torch.amp.autocast("cuda", ...)

* Replace `with autocast(...)` with `with autocast("cuda", ...)`


Co-authored-by: Li Peng <lipeng@unisound.ai>
2024-12-16 10:24:16 +08:00
..
asr_datamodule.py
Add streaming zipformer (#787)
2022-12-30 10:52:18 +08:00
beam_search.py
Add streaming zipformer (#787)
2022-12-30 10:52:18 +08:00
decode_stream.py
fix initial states (#1398)
2023-12-04 22:29:42 +08:00
decode.py
Decode zipformer with external LMs (#1193)
2023-08-03 15:50:35 +08:00
decoder.py
Add streaming zipformer (#787)
2022-12-30 10:52:18 +08:00
do_not_use_it_directly.py
Replace deprecated pytorch methods (#1814)
2024-12-16 10:24:16 +08:00
encoder_interface.py
Add streaming zipformer (#787)
2022-12-30 10:52:18 +08:00
export-for-ncnn-zh.py
Rename train2.py to avoid confusion (#1386)
2023-11-17 18:12:59 +08:00
export-for-ncnn.py
Rename train2.py to avoid confusion (#1386)
2023-11-17 18:12:59 +08:00
export-onnx-zh.py
Fix exporting decoder model to onnx (#1264)
2023-09-22 09:57:15 +08:00
export-onnx.py
Fix exporting decoder model to onnx (#1264)
2023-09-22 09:57:15 +08:00
export.py
Fix torchscript export to use tokens.txt instead of lang_dir (#1475)
2024-01-26 19:18:33 +08:00
jit_pretrained.py
Use high_freq -400 in computing fbank features. (#1447)
2024-01-04 13:59:32 +08:00
jit_trace_export.py
Fix typo (#1114)
2023-06-02 14:12:42 +08:00
jit_trace_pretrained.py
Use high_freq -400 in computing fbank features. (#1447)
2024-01-04 13:59:32 +08:00
joiner.py
Add streaming zipformer (#787)
2022-12-30 10:52:18 +08:00
model.py
Add streaming zipformer (#787)
2022-12-30 10:52:18 +08:00
ncnn_custom_layer.py
Strengthened style constraints (#1527)
2024-03-04 23:28:04 +08:00
onnx_check.py
Streaming Zipformer with multi-dataset (#984)
2023-04-21 15:43:28 +08:00
onnx_model_wrapper.py
Refactor onnx export for streaming zipformer (#879)
2023-02-07 12:12:26 +08:00
onnx_pretrained.py
Use high_freq -400 in computing fbank features. (#1447)
2024-01-04 13:59:32 +08:00
optim.py
Add streaming zipformer (#787)
2022-12-30 10:52:18 +08:00
pretrained.py
Use high_freq -400 in computing fbank features. (#1447)
2024-01-04 13:59:32 +08:00
README.md
Rename train2.py to avoid confusion (#1386)
2023-11-17 18:12:59 +08:00
scaling_converter.py
Add streaming zipformer (#787)
2022-12-30 10:52:18 +08:00
scaling.py
Add streaming zipformer (#787)
2022-12-30 10:52:18 +08:00
streaming_beam_search.py
Add streaming zipformer (#787)
2022-12-30 10:52:18 +08:00
streaming_decode.py
streaming_decode.py, relax the audio range from [-1,+1] to [-10,+10] (#1448)
2024-01-05 10:21:27 +08:00
streaming-ncnn-decode.py
Strengthened style constraints (#1527)
2024-03-04 23:28:04 +08:00
test_model.py
Streaming Zipformer with multi-dataset (#984)
2023-04-21 15:43:28 +08:00
train.py
Replace deprecated pytorch methods (#1814)
2024-12-16 10:24:16 +08:00
zipformer_for_ncnn_export_only.py
Replace deprecated pytorch methods (#1814)
2024-12-16 10:24:16 +08:00
zipformer.py
Replace deprecated pytorch methods (#1814)
2024-12-16 10:24:16 +08:00

README.md

This recipe implements Streaming Zipformer-Transducer model.

See https://k2-fsa.github.io/icefall/recipes/Streaming-ASR/librispeech/zipformer_transducer.html for detailed tutorials.

./emformer.py and ./train.py are basically the same as ./emformer2.py and ./do_not_use_it_directly.py. The only purpose of ./emformer2.py and ./do_not_use_it_directly.py is for exporting to sherpa-ncnn.