10 Commits

Author	SHA1	Message	Date
Zengwei Yao	9ceffa4db1	Merge 3dc33515c0bcba749a775cf08b8aba546763fb66 into 3199058194a48d45aeee740f2aa9bdbef0bec29d	2023-09-12 08:25:12 -07:00
yaozengwei	3dc33515c0	split utterance over 512 frames into overlapping chunks	2023-08-04 10:26:52 +08:00
Fangjun Kuang	1dbbd7759e	Add tests for subsample.py and fix typos (#1180)	2023-07-25 14:46:18 +08:00
yaozengwei	215541c7c5	Do block-wise attention when seq_len is larger than 512, with block_size <= 512	2023-07-23 16:12:57 +08:00
yaozengwei	ee485c02fc	modify attn_offsets	2023-07-21 15:38:22 +08:00
yaozengwei	6aaa971b34	make block-size be a list	2023-07-21 11:34:19 +08:00
yaozengwei	80a14f93d3	Use block-wise attention	2023-07-20 19:38:03 +08:00
Fangjun Kuang	947f0614c9	Fix running exported model on GPU. (#1131)	2023-06-15 12:25:15 +08:00
danfu	0cb71ad3bc	add updated zipformer onnx export (#1108) ... Co-authored-by: Fangjun Kuang <csukuangfj@gmail.com>	2023-06-12 14:02:23 +08:00
Zengwei Yao	f18b539fbc	Add the upgraded Zipformer model (#1058) ... * add the zipformer codes, copied from branch from_dan_scaled_adam_exp1119 * support model export with torch.jit.script * update RESULTS.md * support exporting streaming model with torch.jit.script * add results of streaming models, with some minor changes * update README.md * add CI test * update k2 version in requirements-ci.txt * update pyproject.toml	2023-05-19 16:47:59 +08:00