speech-nano/README.md

---
license: apache-2.0
language:
- en
tags:
- text-to-speech
- tts
- speech-synthesis
- pytorch
- tiny-tts
- experimental
pipeline_tag: text-to-speech
library_name: pytorch
---

# Inflect-Nano-v1

Inflect-Nano-v1 is an experimental ultra-small English TTS stack. It is built to test how far a sub-5M-parameter text-to-speech system can be pushed with a compact non-autoregressive acoustic model and a small neural vocoder.

This is **not** a production-quality or SOTA TTS model. It is a research/demo release: small, local, and runnable, but still audibly limited.

## Quick Facts

| Item | Value |
|---|---:|
| Total inference parameters | **4.632M** |
| Acoustic model | **3.465M** |
| Vocoder generator | **1.167M** |
| Language | English |
| Voice | single Mark-style synthetic male voice |
| Sample rate | 24 kHz |
| Acoustic output | 80-bin mel spectrogram |
| Vocoder | custom Snake-activation HiFi-GAN-style generator |
| Training source | synthetic Qwen3-TTS Mark-style teacher data |

## Audio Examples

These are unseen/OOD stress prompts, not hand-picked training rows.

| Prompt | Audio |
|---|---|
| Wait, are you actually being for real now? I can't believe it! | <audio controls src="examples/example_01.wav"></audio> |
| Sophia sent me 43 pictures of her doing stuff... interesting. | <audio controls src="examples/example_02.wav"></audio> |
| Please say chrysanthemum, thoroughly, proprietary, and rural without rushing through the middle syllables. | <audio controls src="examples/example_03.wav"></audio> |
| No, seriously, did Jordan leave the receipt in Albuquerque, or did Priya move it to Worcester? | <audio controls src="examples/example_04.wav"></audio> |
| The Wi-Fi password is Q7-Delta-9921, but please do not say the dash like a minus sign. | <audio controls src="examples/example_05.wav"></audio> |
| I appreciate the honesty, but that explanation sounded weirdly dramatic for a Tuesday morning. | <audio controls src="examples/example_06.wav"></audio> |
| Could you whisper the first part, then brighten up when you say, 'we finally solved it'? | <audio controls src="examples/example_07.wav"></audio> |
| The dermatologist, the anesthesiologist, and the statistician all disagreed about February. | <audio controls src="examples/example_08.wav"></audio> |

## Install

```bash
git clone https://huggingface.co/owensong/Inflect-Nano-v1
cd Inflect-Nano-v1
pip install -r requirements.txt
```

The text frontend uses TinyTTS-style English G2P and may download `bert-base-uncased` tokenizer files on first run.

## Usage

```bash
python inference.py \
  --text "Wait, are you actually being for real now? I can't believe it!" \
  --out sample.wav
```

CPU example:

```bash
python inference.py \
  --device cpu \
  --text "Please say chrysanthemum, thoroughly, proprietary, and rural clearly." \
  --out sample_cpu.wav
```

Optional controls:

```bash
python inference.py \
  --text "No, seriously, did Jordan leave the receipt in Albuquerque?" \
  --length-scale 1.03 \
  --pitch-scale 1.00 \
  --energy-scale 1.00 \
  --out sample_controlled.wav
```

Gradio demo:

```bash
python app.py
```

## Architecture

Inflect-Nano-v1 is a two-part TTS stack:

```text
text
-> TinyTTS-style normalization + G2P phoneme/tone/lang IDs
-> compact FastSpeech-style acoustic model
-> 80-bin mel spectrogram
-> Snake V2Mid HiFi-GAN-style vocoder
-> 24 kHz waveform
```

### Acoustic Model

The acoustic model is a small non-autoregressive FastSpeech-style network. It predicts duration, energy, brightness, and pitch, then expands token states into frame states and decodes mels.

Main config:

```json
{
  "hidden": 168,
  "encoder_layers": 5,
  "decoder_layers": 6,
  "decoder_ff_mult": 3,
  "kernel_size": 7,
  "speaker_dim": 64,
  "dropout": 0.08,
  "n_mels": 80,
  "sample_rate": 24000,
  "max_frames": 1400,
  "postnet_scale": 0.1,
  "use_frame_pitch": true,
  "abs_frame_bins": 512
}
```

Acoustic parameter split:

```text
total acoustic: 3.465M
encoder:        1.292M
decoder:        1.211M
postnet:        0.276M
local context:  0.226M
frame GRU:      0.128M
heads/embeds/projections: remainder
```

### Vocoder

The vocoder is a custom Snake-activation HiFi-GAN-style generator.

Main config:

```json
{
  "variant": "snake_v2mid",
  "sample_rate": 24000,
  "n_fft": 1024,
  "hop_size": 256,
  "win_size": 1024,
  "num_mels": 80,
  "fmax": 12000.0,
  "upsample_rates": [8, 8, 2, 2],
  "upsample_kernel_sizes": [16, 16, 4, 4],
  "upsample_initial_channel": 144,
  "resblock_kernel_sizes": [3, 7, 11],
  "resblock_dilation_sizes": [[1, 3, 5], [1, 3, 5], [1, 3, 5]],
  "activation": "snake"
}
```

The vocoder was trained with HiFi-GAN-style adversarial losses and multi-resolution spectral pressure. Discriminators are training-only and are not included in inference.

## Training Data

The final acoustic model was trained primarily on synthetic Mark-style teacher speech.

Known final broad training mix:

```text
mixed_80k_v2_40_v1_20_old_20
rows: 80,000
estimated duration: ~125.2 hours
composition:
  40k newer generalization v2 examples
  20k generalization v1 examples
  20k older Mark anchor examples
```

The practical release voice is one speaker:

```text
qwen3_mark
```

## Training Recipe

The final F checkpoint is a staged recovery candidate, not a single monolithic run.

High-level acoustic lineage:

```text
1. Mark-focused acoustic base
2. mixed 80k generalization training
3. predictor-exposure heads training
4. robust prosody bridge
5. recovery phases A/B/C
6. short predictor-tail cleanup
```

Known acoustic continuation after the earlier Mark base:

```text
mixed80k broad training:      ~32k selected checkpoint
predictor exposure:            9k steps
robust prosody bridge:       1.2k steps
recovery phase A:             3k steps
recovery phase B:             3k steps
recovery phase C:           2.5k steps
predictor tail:             0.8k steps
known continuation total:   ~51.5k steps
```

Acoustic losses included mel reconstruction, MSE, delta/acceleration losses, duration loss, energy loss, brightness loss, pitch loss, predicted-prosody exposure, and robust-prosody exposure.

Final selected files in this repo:

```text
weights/inflect_nano_v1_acoustic.pt
weights/inflect_nano_v1_vocoder.pt
```

## Limitations

This model is intentionally tiny and has clear quality limits:

- Unseen text can stumble or sound unstable.
- The voice can sound robotic, buzzy, or artifacted.
- Long or unusual prompts are less reliable.
- It inherits habits from synthetic Qwen3-TTS teacher data.
- It is not a voice cloning model.
- It is not multilingual.
- It is not suitable for production accessibility, safety, or high-quality narration use.

## Recommended Framing

Use this as:

> An experimental 4.63M-parameter English TTS model exploring the quality/size tradeoff for ultra-small local speech synthesis.

Do not present it as SOTA or production-quality.

## License

Apache-2.0. The repo includes TinyTTS text frontend code; its license is included as `TINY_TTS_LICENSE`.