|
|
f852e3e97d
|
reduce maxTokens to 45 for 3s bucket compatibility
|
2026-02-19 12:56:07 +01:00 |
|
|
|
def76b02b5
|
add more debug output for token count
|
2026-02-19 12:38:46 +01:00 |
|
|
|
2774fb8379
|
add per-channel min-max normalization for ASR features
|
2026-02-19 12:36:29 +01:00 |
|
|
|
2b24d520fa
|
add debug output for chunk being played
|
2026-02-19 11:30:44 +01:00 |
|
|
|
ef32087be7
|
add debug output for sample count
|
2026-02-19 11:03:10 +01:00 |
|
|
|
2c975007ca
|
revert tokenizer to character-by-character, add debug output
|
2026-02-19 10:39:54 +01:00 |
|
|
|
4c40e93542
|
reduce chunk size to 80 chars to fit within 128 token model limit
|
2026-02-19 10:35:36 +01:00 |
|
|
|
217eb78e87
|
fix tokenizer to split on whitespace for proper phoneme tokens
|
2026-02-19 10:33:46 +01:00 |
|
|
|
7098f318e9
|
increase maxTokens from 96 to 512 to fix text truncation
|
2026-02-19 10:28:16 +01:00 |
|
|
|
1dd0676fca
|
snapshot current state before gitea sync
|
2026-02-18 10:59:55 +01:00 |
|
