turkish-openslr-24k-phase3 — placeholder (empty)

⚠️ This repository is empty

It contains no data files. It was created as a staging slot for a Turkish OpenSLR resample-to-24 kHz step in phase 3 of the TinyAya data pipeline, and was never populated.

Nothing here is usable. It is documented rather than left bare so the emptiness is obvious from the dataset page instead of only after a download.

Use instead: tr-subset-v0.1 for Turkish speech, or tr-hi-mimi-encoded for the corpus the released model actually trained on.

For the upstream corpora see OpenSLR, whose per-resource licences apply to any derivative.

Where this sits

The v0.3 speech-to-speech pipeline, end to end:

tr-hi-parallel-text          text triples (en pivot -> tr / hi)
        |  TTS
tr-hi-parallel-speech-v2     synthetic speech + QC signals
        |  Mimi encode
tr-hi-mimi-encoded           8-codebook tokens + word alignments
        |  Stage-2 training
tr-hi-s2st-v0.3              the released model

Model	tr-hi-s2st-v0.3
Text	tr-hi-parallel-text
Speech	tr-hi-parallel-speech-v2 · -v3
Encoded	tr-hi-mimi-encoded
Eval sets	fleurs-tr-hi-mimi-encoded · lahaja-eval · cv-tr-eval

Code

repo	what it does
codec-finetuning	fine-tunes Mimi / DualCodec / Kanade on Turkish + Hindi

Project

TinyAya Stage 2 — Turkish⇄Hindi speech-to-speech translation with a text inner-monologue: a LoRA-adapted Cohere2 backbone driving a frozen Moshi depth decoder over Mimi codes.

The v0.3 run covered 76,250 steps / 2.07 epochs on a Cloud TPU v6e-16 (best val composite 2.8199 @ step 76,000). Read honestly: the text inner-monologue learns to translate (free-run chrF++ ~25.7 / 25.1), while intelligible audio synthesis remains the frontier (ASR-chrF++ 3.7 / 9.6 against a 92.1 / 86.6 ground-truth-audio ceiling) — bounded by the frozen depth decoder, not by translation understanding.

• Results: v0.3 evaluation report
• Training run: W&B xzcb60bl · emergence report
• Blog: Adapting Moshi for Low-Resource Speech Translation

Compute for the v0.3 run was provided by Google's TPU Research Cloud (TRC).