kokoro-tts-kazakh

Samples

All three were synthesized from held-out validation text not seen during training (see Why This Checkpoint).

Model Details

Training Configuration

Why This Checkpoint

StyleTTS2/GAN-style training can silently collapse the style encoder partway through a run, so checkpoints are not reliable to pick by filename or epoch number alone. Several km_m1 candidates produced during training were instead synthesized on 20 held-out validation sentences (not seen during training) and scored for word error rate (WER) using a Kazakh-fine-tuned Whisper ASR model with jiwer:

final matches the exact decoder weights shipped here (kokoro_kazakh.pth) — both were exported from the same training checkpoint, so there is no decoder/voicepack mismatch. Stage 2 validation loss also decreased monotonically across all 5 epochs of this run (0.402 → 0.385), with no sign of the divergence seen in the collapsed run.

Installation & Usage

Runs on CPU — no GPU needed.

pip install torch kokoro "misaki[kk]"

misaki[kk] pulls in espeak-ng bindings for Kazakh phonemization; see kikiri-tts for a ready-made inference script (inference_kazakh.py).

Note: kokoro_kazakh.pth is stored using the legacy weight_norm key format (weight_g / weight_v) so it loads cleanly with current torch/kokoro releases. If you re-export this checkpoint yourself from a different PyTorch version, double-check the resulting key names match what your installed kokoro expects — a silent weight_g/weight_v vs. parametrizations.weight.original0/original1 mismatch will make load_state_dict skip the decoder/predictor weights without raising an error, producing noise instead of speech.

import torch
from kokoro import KModel
from misaki import espeak

g2p = espeak.EspeakG2P(language="kk")
model = KModel(repo_id="hexgrad/Kokoro-82M", config="config.json", model="kokoro_kazakh.pth").eval()
voicepack = torch.load("km_m1.pt", map_location="cpu", weights_only=True)

text = "Сәлем! Бұл қазақ тіліндегі сөйлеу синтезі."
phonemes, _ = g2p(text)
ref_s = voicepack[min(len(phonemes) - 1, voicepack.shape[0] - 1)]

audio = model(phonemes, ref_s, speed=1.0)

Download the checkpoint files directly with huggingface_hub:

from huggingface_hub import snapshot_download
snapshot_download("AnuarSv/kokoro-tts-kazakh", local_dir="kokoro-tts-kazakh")

CPU Benchmarks

Measured on an AMD Ryzen 5 5500U (6 physical cores / 12 threads via SMT, up to 4.06 GHz, no GPU), synthesizing a 3-sentence, 22.45s-audio test set:

Throughput scales with core count up to the number of physical cores (6 here); beyond that, additional SMT threads add scheduling overhead without extra real compute, so speed slightly regresses. Even single-threaded, this model already runs faster than real time. Peak memory stays flat at ~1.4 GB regardless of thread count.

The table above measures single-request latency: how many threads to give one synthesis call so it finishes faster. That's a different question from throughput: how many simultaneous requests from different users the machine can serve. For that, each request should run single-threaded, with concurrency handled by running multiple worker processes in parallel instead of parallelizing one request across threads — this avoids the synchronization overhead that caps the table above at ~2.6×.

Concurrent single-thread