app.py

import os
import shlex
import subprocess

subprocess.run(
    shlex.split("pip install flash-attn --no-build-isolation"),
    env=os.environ | {"FLASH_ATTENTION_SKIP_CUDA_BUILD": "TRUE"},
    check=True,
)
subprocess.run(
    shlex.split("pip install https://github.com/state-spaces/mamba/releases/download/v2.2.4/mamba_ssm-2.2.4+cu12torch2.4cxx11abiFALSE-cp310-cp310-linux_x86_64.whl"),
    check=True,
)
subprocess.run(
    shlex.split("pip install https://github.com/Dao-AILab/causal-conv1d/releases/download/v1.5.0.post8/causal_conv1d-1.5.0.post8+cu12torch2.4cxx11abiFALSE-cp310-cp310-linux_x86_64.whl"),
    check=True,
)

import spaces
import torch
import torchaudio
import gradio as gr
from os import getenv

from zonos.model import Zonos
from zonos.conditioning import make_cond_dict, supported_language_codes

# 1. hard-kill torch.compile / dynamo / inductor so they never run
os.environ["TORCH_COMPILE_DISABLE"] = "1"
os.environ["TORCHINDUCTOR_DISABLE"] = "1"
os.environ["TORCHDYNAMO_DISABLE"] = "1"          # <- the one that actually blocks torch._dynamo  
os.environ["TORCHDYNAMO_SUPPRESS_ERRORS"] = "True"  # fall back to eager if something still slips through  :contentReference[oaicite:1]{index=1}

torch._dynamo.disable()        # guard for older versions
torch.compile = lambda f,*_,**__: f   # no-op wrapper

device = "cuda"
MODEL_NAMES = ["Zyphra/Zonos-v0.1-transformer", "Zyphra/Zonos-v0.1-hybrid"]
MODELS = {name: Zonos.from_pretrained(name, device=device) for name in MODEL_NAMES}
for model in MODELS.values():
    model.requires_grad_(False).eval()

def _patch_cuda_props():
    if torch.cuda.is_available():
        for i in range(torch.cuda.device_count()):
            p = torch.cuda.get_device_properties(i)
            if not hasattr(p, "regs_per_multiprocessor"):
                setattr(p, "regs_per_multiprocessor", 65536)
            if not hasattr(p, "max_threads_per_multi_processor"):
                setattr(p, "max_threads_per_multi_processor", 2048)
                
_patch_cuda_props()

def update_ui(model_choice):
    """
    Dynamically show/hide UI elements based on the model's conditioners.
    We do NOT display 'language_id' or 'ctc_loss' even if they exist in the model.
    """
    model = MODELS[model_choice]
    cond_names = [c.name for c in model.prefix_conditioner.conditioners]
    print("Conditioners in this model:", cond_names)

    text_update = gr.update(visible=("espeak" in cond_names))
    language_update = gr.update(visible=("espeak" in cond_names))
    speaker_audio_update = gr.update(visible=("speaker" in cond_names))
    prefix_audio_update = gr.update(visible=True)
    emotion1_update = gr.update(visible=("emotion" in cond_names))
    emotion2_update = gr.update(visible=("emotion" in cond_names))
    emotion3_update = gr.update(visible=("emotion" in cond_names))
    emotion4_update = gr.update(visible=("emotion" in cond_names))
    emotion5_update = gr.update(visible=("emotion" in cond_names))
    emotion6_update = gr.update(visible=("emotion" in cond_names))
    emotion7_update = gr.update(visible=("emotion" in cond_names))
    emotion8_update = gr.update(visible=("emotion" in cond_names))
    vq_single_slider_update = gr.update(visible=("vqscore_8" in cond_names))
    fmax_slider_update = gr.update(visible=("fmax" in cond_names))
    pitch_std_slider_update = gr.update(visible=("pitch_std" in cond_names))
    speaking_rate_slider_update = gr.update(visible=("speaking_rate" in cond_names))
    dnsmos_slider_update = gr.update(visible=("dnsmos_ovrl" in cond_names))
    speaker_noised_checkbox_update = gr.update(visible=("speaker_noised" in cond_names))
    unconditional_keys_update = gr.update(
        choices=[name for name in cond_names if name not in ("espeak", "language_id")]
    )

    return (
        text_update,
        language_update,
        speaker_audio_update,
        prefix_audio_update,
        emotion1_update,
        emotion2_update,
        emotion3_update,
        emotion4_update,
        emotion5_update,
        emotion6_update,
        emotion7_update,
        emotion8_update,
        vq_single_slider_update,
        fmax_slider_update,
        pitch_std_slider_update,
        speaking_rate_slider_update,
        dnsmos_slider_update,
        speaker_noised_checkbox_update,
        unconditional_keys_update,
    )


@spaces.GPU(duration=120)
def generate_audio(
    model_choice,
    text,
    language,
    speaker_audio,
    prefix_audio,
    e1,
    e2,
    e3,
    e4,
    e5,
    e6,
    e7,
    e8,
    vq_single,
    fmax,
    pitch_std,
    speaking_rate,
    dnsmos_ovrl,
    speaker_noised,
    cfg_scale,
    min_p,
    seed,
    randomize_seed,
    unconditional_keys,
    progress=gr.Progress(),
):
    """
    Generates audio based on the provided UI parameters.
    We do NOT use language_id or ctc_loss even if the model has them.
    """
    selected_model = MODELS[model_choice]

    speaker_noised_bool = bool(speaker_noised)
    fmax = float(fmax)
    pitch_std = float(pitch_std)
    speaking_rate = float(speaking_rate)
    dnsmos_ovrl = float(dnsmos_ovrl)
    cfg_scale = float(cfg_scale)
    min_p = float(min_p)
    seed = int(seed)
    max_new_tokens = 86 * 30

    if randomize_seed:
        seed = torch.randint(0, 2**32 - 1, (1,)).item()
    torch.manual_seed(seed)

    speaker_embedding = None
    if speaker_audio is not None and "speaker" not in unconditional_keys:
        wav, sr = torchaudio.load(speaker_audio)
        speaker_embedding = selected_model.make_speaker_embedding(wav, sr)
        speaker_embedding = speaker_embedding.to(device, dtype=torch.bfloat16)

    audio_prefix_codes = None
    if prefix_audio is not None:
        wav_prefix, sr_prefix = torchaudio.load(prefix_audio)
        wav_prefix = wav_prefix.mean(0, keepdim=True)
        wav_prefix = torchaudio.functional.resample(wav_prefix, sr_prefix, selected_model.autoencoder.sampling_rate)
        wav_prefix = wav_prefix.to(device, dtype=torch.float32)
        with torch.autocast(device, dtype=torch.float32):
            audio_prefix_codes = selected_model.autoencoder.encode(wav_prefix.unsqueeze(0))

    emotion_tensor = torch.tensor(list(map(float, [e1, e2, e3, e4, e5, e6, e7, e8])), device=device)

    vq_val = float(vq_single)
    vq_tensor = torch.tensor([vq_val] * 8, device=device).unsqueeze(0)

    cond_dict = make_cond_dict(
        text=text,
        language=language,
        speaker=speaker_embedding,
        emotion=emotion_tensor,
        vqscore_8=vq_tensor,
        fmax=fmax,
        pitch_std=pitch_std,
        speaking_rate=speaking_rate,
        dnsmos_ovrl=dnsmos_ovrl,
        speaker_noised=speaker_noised_bool,
        device=device,
        unconditional_keys=unconditional_keys,
    )
    conditioning = selected_model.prepare_conditioning(cond_dict)

    estimated_generation_duration = 30 * len(text) / 400
    estimated_total_steps = int(estimated_generation_duration * 86)

    def update_progress(_frame: torch.Tensor, step: int, _total_steps: int) -> bool:
        progress((step, estimated_total_steps))
        return True

    codes = selected_model.generate(
        prefix_conditioning=conditioning,
        audio_prefix_codes=audio_prefix_codes,
        max_new_tokens=max_new_tokens,
        cfg_scale=cfg_scale,
        batch_size=1,
        sampling_params=dict(min_p=min_p),
        callback=update_progress,
    )

    wav_out = selected_model.autoencoder.decode(codes).cpu().detach()
    sr_out = selected_model.autoencoder.sampling_rate
    if wav_out.dim() == 2 and wav_out.size(0) > 1:
        wav_out = wav_out[0:1, :]
    return (sr_out, wav_out.squeeze().numpy()), seed

def build_interface():
    # Build interface with enhanced visual elements and layout
    with gr.Blocks() as demo:
        # Header section
        with gr.Column(elem_classes="app-header"):
            gr.Markdown("# ✨ Zonos Text-to-Speech Generator ✨")
            gr.Markdown("Create natural-sounding speech with customizable voice characteristics")
        
        # Main content container 
        with gr.Column(elem_classes="container"):
            # First panel - Text & Model Selection
            with gr.Column(elem_classes="panel"):
                gr.Markdown('<div class="title">💬 Text & Model Configuration</div>')
                with gr.Row():
                    with gr.Column(scale=2):
                        model_choice = gr.Dropdown(
                            choices=MODEL_NAMES,
                            value="Zyphra/Zonos-v0.1-transformer",
                            label="Zonos Model Type",
                            info="Select the model variant to use.",
                        )
                        text = gr.Textbox(
                            label="Text to Synthesize",
                            value="Zonos uses eSpeak for text to phoneme conversion!",
                            lines=4,
                            max_length=500,
                        )
                        language = gr.Dropdown(
                            choices=supported_language_codes,
                            value="en-us",
                            label="Language Code",
                            info="Select a language code.",
                        )
                    with gr.Column(scale=1):
                        prefix_audio = gr.Audio(
                            value="assets/silence_100ms.wav",
                            label="Optional Prefix Audio (continue from this audio)",
                            type="filepath",
                        )
            
            # Second panel - Voice Characteristics
            with gr.Column(elem_classes="panel"):
                gr.Markdown('<div class="title">🎤 Voice Characteristics</div>')
                with gr.Row():
                    with gr.Column(scale=1):
                        speaker_audio = gr.Audio(
                            label="Optional Speaker Audio (for voice cloning)",
                            type="filepath",
                        )
                        speaker_noised_checkbox = gr.Checkbox(label="Denoise Speaker?", value=False)
                    
                    with gr.Column(scale=2):
                        with gr.Row():
                            with gr.Column():
                                dnsmos_slider = gr.Slider(1.0, 5.0, value=4.0, step=0.1, label="Voice Quality", elem_classes="slider-container")
                                fmax_slider = gr.Slider(0, 24000, value=24000, step=1, label="Frequency Max (Hz)", elem_classes="slider-container")
                                vq_single_slider = gr.Slider(0.5, 0.8, 0.78, 0.01, label="Voice Clarity", elem_classes="slider-container")
                            with gr.Column():
                                pitch_std_slider = gr.Slider(0.0, 300.0, value=45.0, step=1, label="Pitch Variation", elem_classes="slider-container")
                                speaking_rate_slider = gr.Slider(5.0, 30.0, value=15.0, step=0.5, label="Speaking Rate", elem_classes="slider-container")
            
            # Third panel - Generation Parameters
            with gr.Column(elem_classes="panel"):
                gr.Markdown('<div class="title">⚙️ Generation Parameters</div>')
                with gr.Row():
                    with gr.Column():
                        cfg_scale_slider = gr.Slider(1.0, 5.0, 2.0, 0.1, label="Guidance Scale", elem_classes="slider-container")
                        min_p_slider = gr.Slider(0.0, 1.0, 0.15, 0.01, label="Min P (Randomness)", elem_classes="slider-container")
                    with gr.Column():
                        seed_number = gr.Number(label="Seed", value=420, precision=0)
                        randomize_seed_toggle = gr.Checkbox(label="Randomize Seed (before generation)", value=True)
            
            # Emotion Panel with Tabbed Interface
            with gr.Accordion("🎭 Emotion Settings", open=False, elem_classes="panel"):
                gr.Markdown(
                    "Adjust these sliders to control the emotional tone of the generated speech.\n"
                    "For a neutral voice, keep 'Neutral' high and other emotions low."
                )
                with gr.Row(elem_classes="emotion-grid"):
                    emotion1 = gr.Slider(0.0, 1.0, 1.0, 0.05, label="Happiness", elem_classes="slider-container")
                    emotion2 = gr.Slider(0.0, 1.0, 0.05, 0.05, label="Sadness", elem_classes="slider-container")
                    emotion3 = gr.Slider(0.0, 1.0, 0.05, 0.05, label="Disgust", elem_classes="slider-container")
                    emotion4 = gr.Slider(0.0, 1.0, 0.05, 0.05, label="Fear", elem_classes="slider-container")
                with gr.Row(elem_classes="emotion-grid"):
                    emotion5 = gr.Slider(0.0, 1.0, 0.05, 0.05, label="Surprise", elem_classes="slider-container")
                    emotion6 = gr.Slider(0.0, 1.0, 0.05, 0.05, label="Anger", elem_classes="slider-container")
                    emotion7 = gr.Slider(0.0, 1.0, 0.1, 0.05, label="Other", elem_classes="slider-container")
                    emotion8 = gr.Slider(0.0, 1.0, 0.2, 0.05, label="Neutral", elem_classes="slider-container")
            
            # Advanced Settings Panel
            with gr.Accordion("⚡ Advanced Settings", open=False, elem_classes="panel"):
                gr.Markdown(
                    "### Unconditional Toggles\n"
                    "Checking a box will make the model ignore the corresponding conditioning value and make it unconditional.\n"
                    'Practically this means the given conditioning feature will be unconstrained and "filled in automatically".'
                )
                unconditional_keys = gr.CheckboxGroup(
                    [
                        "speaker",
                        "emotion",
                        "vqscore_8",
                        "fmax",
                        "pitch_std",
                        "speaking_rate",
                        "dnsmos_ovrl",
                        "speaker_noised",
                    ],
                    value=["emotion"],
                    label="Unconditional Keys",
                )
            
            # Generate Button and Output Area
            with gr.Column(elem_classes="panel output-container"):
                gr.Markdown('<div class="title">🔊 Generate & Output</div>')
                generate_button = gr.Button("Generate Audio", elem_classes="generate-button")
                output_audio = gr.Audio(label="Generated Audio", type="numpy", autoplay=True, elem_classes="audio-output")

        model_choice.change(
            fn=update_ui,
            inputs=[model_choice],
            outputs=[
                text,
                language,
                speaker_audio,
                prefix_audio,
                emotion1,
                emotion2,
                emotion3,
                emotion4,
                emotion5,
                emotion6,
                emotion7,
                emotion8,
                vq_single_slider,
                fmax_slider,
                pitch_std_slider,
                speaking_rate_slider,
                dnsmos_slider,
                speaker_noised_checkbox,
                unconditional_keys,
            ],
        )

        # On page load, trigger the same UI refresh
        demo.load(
            fn=update_ui,
            inputs=[model_choice],
            outputs=[
                text,
                language,
                speaker_audio,
                prefix_audio,
                emotion1,
                emotion2,
                emotion3,
                emotion4,
                emotion5,
                emotion6,
                emotion7,
                emotion8,
                vq_single_slider,
                fmax_slider,
                pitch_std_slider,
                speaking_rate_slider,
                dnsmos_slider,
                speaker_noised_checkbox,
                unconditional_keys,
            ],
        )

        # Generate audio on button click
        generate_button.click(
            fn=generate_audio,
            inputs=[
                model_choice,
                text,
                language,
                speaker_audio,
                prefix_audio,
                emotion1,
                emotion2,
                emotion3,
                emotion4,
                emotion5,
                emotion6,
                emotion7,
                emotion8,
                vq_single_slider,
                fmax_slider,
                pitch_std_slider,
                speaking_rate_slider,
                dnsmos_slider,
                speaker_noised_checkbox,
                cfg_scale_slider,
                min_p_slider,
                seed_number,
                randomize_seed_toggle,
                unconditional_keys,
            ],
            outputs=[output_audio, seed_number],
        )

    return demo


if __name__ == "__main__":
    demo = build_interface()
    demo.launch()