Kokoro 82M β€” LiteRT

Kokoro-82M exported as a staged FP32 LiteRT bundle for CPU text-to-speech. It produces 24 kHz Float32 audio with the included af_heart voice.

The current bundle (v2.0.0) has a 120-frame (3 s) contract and packed token LSTMs, so its durations match upstream Kokoro. The v1.0.0 60-frame files stay in the repository for older releases.

Model

Property Value
Parameters 82M
Architecture Non-autoregressive, single pass
Format LiteRT / TFLite
Precision FP32
Model files 334,176,844 bytes total (319 MiB)
Sample rate 24 kHz
Public text input 128 token slots
Active text encoder/recurrent window 32 token slots
Output tensor 120 frames / 72,000 samples / 3.0 seconds
Runtime safety boundary 116 frames / 69,600 samples / 2.9 seconds
Included voice af_heart

Every token LSTM runs over its valid prefix, as upstream Kokoro's packed sequences do: the backward direction starts at the last phoneme, so padding never reaches a valid output. Only the 32 token slots ALBERT fills are evaluated, and each input projection runs once per sequence outside compact LiteRT loops.

Longer text is supported by host-side chunking, up to 32 tokens per invocation at speed 1.0. The model-reported duration is checked after the recurrent stage. Output above 116 frames must be discarded and retried as smaller text; the final four tensor frames are reserved as right context.

v1.0.0's recurrent graph walked the zero padding before reaching the last phoneme and predicted durations too short, most for short replies. With corrected durations its 60-frame tensor holds only a word or two per piece.

Files

File Bytes SHA-256
kokoro-encoder.tflite 47,877,656 60ac2065c78cc5ec724b3a43beb6789c084a03121018f664c561c131ce65ecaa
kokoro-recurrent-packed32-120.tflite 45,057,700 01eabd16c2b9b16376442bbc2ff9ff709e47948833cbc5060e45f55a1a32b514
kokoro-vocoder-120.tflite 241,241,488 55adae7d19fd95ca998476e03157bb88bd38f805063ce780de0984eceb98e8af
config.json <2 KB Runtime geometry and artifact hashes
vocab_index.json 1,822 IPA symbol-to-token mapping
us_gold.json 3,000,469 Primary English pronunciation dictionary
us_silver.json 3,099,517 Fallback English pronunciation dictionary
voices/af_heart.bin 1,024 Default 256-value Float32 voice embedding

v1.0.0 files, used with the same encoder:

File Bytes SHA-256
kokoro-recurrent-equivalent32.tflite 45,004,644 560b6318be2ac91dbeba928a634097fc6f1c30229b016f9dedacc0a6a64d2865
kokoro-vocoder.tflite 240,647,344 d8cc32713beeab8fadfb4190d5a7ada4d832ae2ecd16b82570c891f02d0e6ff6

Validation

The chained bundle was compared with FP32 PyTorch using packed sequences, on identical token, voice, speed, and harmonic-phase inputs.

Check Result
Waveform correlation 0.99924
Magnitude-spectrogram correlation 0.99982
Duration maximum absolute error 0
Duration mismatch count 0
Recurrent equivalence matrix 0 mismatches across 42 length/speed cases

Through the speech-core wrapper, a physical Galaxy S23 Ultra and an arm64 emulator produce bit-identical audio.

Quality and performance

Samsung Galaxy S23 Ultra (SM-S918B, Snapdragon 8 Gen 2), Android 16, LiteRT 2.1.5, XNNPACK, eight threads (taskset ff), both bundles on the same day. Timing uses one warm-up and ten measured runs per bundle, each batch started from thermal status 0, for the reply "Sure, I can help with that. The meeting has been moved to Friday afternoon at two o'clock, and I will remind you an hour before it starts." Quality is Parakeet TDT v3 over 21 assistant replies synthesized on the phone.

Bundle 24-word reply Pieces Peak RSS Replies transcribed exactly Word errors
v2.0.0, 120 frames 6.7 s 5 1,062 MiB 19 of 21 1.9%
v1.0.0, 60 frames 9.4 s 14 1,003 MiB 15 of 21 6.2%

Usage with speech-core

#include <speech_core/models/litert_kokoro_tts.h>

speech_core::LiteRTKokoroTts tts(
    "kokoro-encoder.tflite",
    "kokoro-recurrent-packed32-120.tflite",
    "kokoro-vocoder-120.tflite",
    "voices",
    ".",
    /*hw_accel=*/false,
    /*num_threads=*/8);

tts.set_speed(1.0f);
tts.set_seed(1234);
tts.synthesize("Hello world.", "en",
    [](const float* pcm, size_t samples, bool is_final) {
        // 24 kHz mono Float32 PCM
    });

The wrapper reads the frame contract from the graphs and includes the English phonemizer, proactive token chunking, the duration guard, retry splitting, waveform checks, and final audio assembly. Older speech-core releases load only the v1.0.0 files.

Build and run the physical benchmark:

cmake --build speech-core/build --target speech_kokoro_litert_bench
speech_kokoro_litert_bench /path/to/this/bundle \
  --variant packed32-120 --threads 8 --warmup 1 --runs 10

Source

Converted from hexgrad/Kokoro-82M. The bundle uses explicit harmonic phases, compact FP32 recurrent loops, and a LiteRT-compatible iSTFT path.

Downloads last month
160
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Model tree for soniqo/Kokoro-82M-LiteRT

Finetuned
(84)
this model