YuCeong-May Claude Code commited on
Commit
422aea4
·
verified ·
1 Parent(s): d962636

Use the inference package and WakeCueASR class from GitHub

Browse files

The inference code is the `inference` package and `infer.py` in github.com/YuCeong-May/WakeCue-ASR, whose main class is now WakeCueASR.

Co-Authored-By: Claude Code <noreply@anthropic.com>

Files changed (1) hide show
  1. README.md +3 -3
README.md CHANGED
@@ -27,7 +27,7 @@ transcribing the wrong person.
27
  - Paper: coming soon
28
  - Code: [github.com/YuCeong-May/WakeCue-ASR](https://github.com/YuCeong-May/WakeCue-ASR)
29
  - Demo: [WakeCue-ASR Space](https://huggingface.co/spaces/YuCeong-May/WakeCue-ASR)
30
- - Inference code: the `tcg_tsasr` package in the WakeCue-ASR Space repository.
31
  - Language: Mandarin Chinese, 16 kHz audio. Audio within 20 seconds is recommended.
32
  - Weights: `model.pt` (1.6 GB), SHA-256
33
  `870aef70a8a8c2b292b3eb959ea4a0907edc751848729070aea04465b449fcf5`.
@@ -66,9 +66,9 @@ Qwen input embedding and output head are tied and, since LoRA is applied to
66
  makes the file 1.6 GB; it is not an extra trained module.
67
 
68
  ```python
69
- from tcg_tsasr import TextCueGuidedTSASR # from the Space repository
70
 
71
- asr = TextCueGuidedTSASR.from_pretrained("YuCeong-May/WakeCue-ASR")
72
  out = asr.transcribe("examples/midea_pos_2465.wav", "小美小美")
73
  print(out["text"], out["rejected"])
74
  ```
 
27
  - Paper: coming soon
28
  - Code: [github.com/YuCeong-May/WakeCue-ASR](https://github.com/YuCeong-May/WakeCue-ASR)
29
  - Demo: [WakeCue-ASR Space](https://huggingface.co/spaces/YuCeong-May/WakeCue-ASR)
30
+ - Inference code: the `inference` package and `infer.py` in the GitHub repository.
31
  - Language: Mandarin Chinese, 16 kHz audio. Audio within 20 seconds is recommended.
32
  - Weights: `model.pt` (1.6 GB), SHA-256
33
  `870aef70a8a8c2b292b3eb959ea4a0907edc751848729070aea04465b449fcf5`.
 
66
  makes the file 1.6 GB; it is not an extra trained module.
67
 
68
  ```python
69
+ from inference import WakeCueASR # from the GitHub repository
70
 
71
+ asr = WakeCueASR.from_pretrained("YuCeong-May/WakeCue-ASR")
72
  out = asr.transcribe("examples/midea_pos_2465.wav", "小美小美")
73
  print(out["text"], out["rejected"])
74
  ```