muyleanging commited on
Commit
c22eb0f
·
verified ·
1 Parent(s): 51d5f97

Add planned architecture figure

Browse files
Files changed (3) hide show
  1. .gitattributes +1 -0
  2. README.md +3 -1
  3. architecture.png +3 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ architecture.png filter=lfs diff=lfs merge=lfs -text
README.md CHANGED
@@ -19,7 +19,9 @@ tags:
19
  **Vacana** ("words") will clean Khmer text in both directions: numbers, dates and money into Khmer words (for speech), and speech-recognition output back into readable text with punctuation and digits.
20
  It belongs to the same family as [Sota](https://huggingface.co/muyleanging/Sota-ASR-Khmer-ONNX) (ASR, listens), [Chakkhu](https://muyleanging.github.io/chakkhu-web/) (OCR, reads), [Tossana](https://huggingface.co/muyleanging/Tossana-Caption-Khmer) (captioning, sees), [Akkhara](https://muyleanging.github.io/akkhara-web/) (keyboard, types) and [VoxPocket](https://huggingface.co/muyleanging/VoxPocket-TTS-Khmer) (TTS, speaks).
21
 
22
- ## Planned design
 
 
23
  - Words → text: punctuation and digits for Sota transcripts
24
  - Text → words: numbers, dates, money and units read out in Khmer, for VoxPocket
25
  - Starts from the rule-based number speller used to build Sota's data
 
19
  **Vacana** ("words") will clean Khmer text in both directions: numbers, dates and money into Khmer words (for speech), and speech-recognition output back into readable text with punctuation and digits.
20
  It belongs to the same family as [Sota](https://huggingface.co/muyleanging/Sota-ASR-Khmer-ONNX) (ASR, listens), [Chakkhu](https://muyleanging.github.io/chakkhu-web/) (OCR, reads), [Tossana](https://huggingface.co/muyleanging/Tossana-Caption-Khmer) (captioning, sees), [Akkhara](https://muyleanging.github.io/akkhara-web/) (keyboard, types) and [VoxPocket](https://huggingface.co/muyleanging/VoxPocket-TTS-Khmer) (TTS, speaks).
21
 
22
+ ![Vacana architecture](architecture.png)
23
+
24
+ ## Planned design (≈4M parameters, estimate)
25
  - Words → text: punctuation and digits for Sota transcripts
26
  - Text → words: numbers, dates, money and units read out in Khmer, for VoxPocket
27
  - Starts from the rule-based number speller used to build Sota's data
architecture.png ADDED

Git LFS Details

  • SHA256: 7f2cc37f166c005e56c7b085492d6db8be667618c2ba85399f558dc0982280fd
  • Pointer size: 131 Bytes
  • Size of remote file: 421 kB