Turkish pocket-tts models, fine-tuned from the English 24-layer teacher and depth-distilled to 6 layers. Runs on CPU.
uvx pocket-tts generate --config hf://alperiox/pocket-tts-turkish/turkish_6l.yaml \
--voice your_voice.wav --text "Merhaba, bugün nasılsınız?"
| layers | file | CPU realtime | size | |
|---|---|---|---|---|
| student (recommended) | 6 | turkish_6l.yaml |
5.1x | 398 MB |
| teacher | 24 | turkish_24l.yaml |
1.6x | 1.3 GB |
60 held-out utterances, voices unseen in training. WER uses
[mpoyraz/wav2vec2-xls-r-300m-cv7-turkish]
(https://huggingface.co/mpoyraz/wav2vec2-xls-r-300m-cv7-turkish); UTMOS uses
utmos_pytorch; SIM-o is WavLM-base-plus-sv cosine against the voice prompt.
| WER | UTMOS | SIM-o | |
|---|---|---|---|
| real audio (same utterances) | 11.88% | 3.934 | — |
| teacher 24L | 13.38% | 3.863 | 0.964 |
| student 6L | 11.88% | 3.733 | 0.955 |
Read WER against the 11.88% floor, not against zero. That is the recognizer's own error rate on the ground-truth recordings of these utterances, so the student is statistically indistinguishable from real speech by this measure; the remaining errors belong to the ASR.
The student beats the teacher on WER (the regularisation effect depth distillation is known for) and trails it on naturalness by 0.13 UTMOS, which is the documented cost of 6 layers versus 24. UTMOS 3.733 matches the pocket-tts README's own target for this model size.
finetune_language.yaml from english_2026-04_24l, 30k steps.
Stopped when validation reached the corpus ceiling (UTMOS within 0.07 of the
training audio) rather than at a step count.depth_distill.yaml, 160k steps. WER improved 14.12% -> 11.88%
during the cosine tail alone, well after distill_mse had flattened.İ or I in the training text — the source corpus wrote those
lowercase. Apply Turkish-aware lowercasing to input text before synthesis.pocket-tts by Kyutai (MIT). Training data derived from Alania Turkish Synthetic Speech by PatientDesk AI (CC BY 4.0).
teacher.safetensors 1.3 GBmodel.safetensors 398 MBkyutai/pocket-tts