cog.py

# Prediction interface for Cog ⚙️
# https://cog.run/python

from cog import BasePredictor, Input, Path

import os
import re
import torch
import torchaudio
import gradio as gr
import numpy as np
import tempfile
from einops import rearrange
from ema_pytorch import EMA
from vocos import Vocos
from pydub import AudioSegment
from model import CFM, UNetT, DiT, MMDiT
from cached_path import cached_path
from model.utils import (
    get_tokenizer,
    convert_char_to_pinyin,
    save_spectrogram,
)
from transformers import pipeline
import librosa

device = "cuda" if torch.cuda.is_available() else "mps" if torch.backends.mps.is_available() else "cpu"

target_sample_rate = 24000
n_mel_channels = 100
hop_length = 256
target_rms = 0.1
nfe_step = 32  # 16, 32
cfg_strength = 2.0
ode_method = 'euler'
sway_sampling_coef = -1.0
speed = 1.0
# fix_duration = 27  # None or float (duration in seconds)
fix_duration = None


class Predictor(BasePredictor):
    def load_model(exp_name, model_cls, model_cfg, ckpt_step):
        checkpoint = torch.load(str(cached_path(f"hf://SWivid/F5-TTS/{exp_name}/model_{ckpt_step}.pt")), map_location=device)
        vocab_char_map, vocab_size = get_tokenizer("Emilia_ZH_EN", "pinyin")
        model = CFM(
            transformer=model_cls(
                **model_cfg,
                text_num_embeds=vocab_size,
                mel_dim=n_mel_channels
            ),
            mel_spec_kwargs=dict(
                target_sample_rate=target_sample_rate,
                n_mel_channels=n_mel_channels,
                hop_length=hop_length,
            ),
            odeint_kwargs=dict(
                method=ode_method,
            ),
            vocab_char_map=vocab_char_map,
        ).to(device)

        ema_model = EMA(model, include_online_model=False).to(device)
        ema_model.load_state_dict(checkpoint['ema_model_state_dict'])
        ema_model.copy_params_from_ema_to_model()

        return ema_model, model
    def setup(self) -> None:
        """Load the model into memory to make running multiple predictions efficient"""
        # self.model = torch.load("./weights.pth")
        print("Loading Whisper model...")
        self.pipe = pipeline(
            "automatic-speech-recognition",
            model="openai/whisper-large-v3-turbo",
            torch_dtype=torch.float16,
            device=device,
        )
        print("Loading F5-TTS model...")

        F5TTS_model_cfg = dict(dim=1024, depth=22, heads=16, ff_mult=2, text_dim=512, conv_layers=4)
        self.F5TTS_ema_model, self.F5TTS_base_model = self.load_model("F5TTS_Base", DiT, F5TTS_model_cfg, 1200000)


    def predict(
        self,
        gen_text: str = Input(description="Text to generate"),
        ref_audio_orig: Path = Input(description="Reference audio"),
        remove_silence: bool = Input(description="Remove silences", default=True),
    ) -> Path:
        """Run a single prediction on the model"""
        model_choice = "F5-TTS"
        print(gen_text)
        if len(gen_text) > 200:
            raise gr.Error("Please keep your text under 200 chars.")
        gr.Info("Converting audio...")
        with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as f:
            aseg = AudioSegment.from_file(ref_audio_orig)
            audio_duration = len(aseg)
            if audio_duration > 15000:
                gr.Warning("Audio is over 15s, clipping to only first 15s.")
                aseg = aseg[:15000]
            aseg.export(f.name, format="wav")
            ref_audio = f.name
        ema_model = self.F5TTS_ema_model
        base_model = self.F5TTS_base_model

        if not ref_text.strip():
            gr.Info("No reference text provided, transcribing reference audio...")
            ref_text = outputs = self.pipe(
                ref_audio,
                chunk_length_s=30,
                batch_size=128,
                generate_kwargs={"task": "transcribe"},
                return_timestamps=False,
            )['text'].strip()
            gr.Info("Finished transcription")
        else:
            gr.Info("Using custom reference text...")
        audio, sr = torchaudio.load(ref_audio)

        rms = torch.sqrt(torch.mean(torch.square(audio)))
        if rms < target_rms:
            audio = audio * target_rms / rms
        if sr != target_sample_rate:
            resampler = torchaudio.transforms.Resample(sr, target_sample_rate)
            audio = resampler(audio)
        audio = audio.to(device)

        # Prepare the text
        text_list = [ref_text + gen_text]
        final_text_list = convert_char_to_pinyin(text_list)

        # Calculate duration
        ref_audio_len = audio.shape[-1] // hop_length
        # if fix_duration is not None:
        #     duration = int(fix_duration * target_sample_rate / hop_length)
        # else:
        zh_pause_punc = r"。，、；：？！"
        ref_text_len = len(ref_text) + len(re.findall(zh_pause_punc, ref_text))
        gen_text_len = len(gen_text) + len(re.findall(zh_pause_punc, gen_text))
        duration = ref_audio_len + int(ref_audio_len / ref_text_len * gen_text_len / speed)

        # inference
        gr.Info(f"Generating audio using F5-TTS")
        with torch.inference_mode():
            generated, _ = base_model.sample(
                cond=audio,
                text=final_text_list,
                duration=duration,
                steps=nfe_step,
                cfg_strength=cfg_strength,
                sway_sampling_coef=sway_sampling_coef,
            )

        generated = generated[:, ref_audio_len:, :]
        generated_mel_spec = rearrange(generated, '1 n d -> 1 d n')
        gr.Info("Running vocoder")
        vocos = Vocos.from_pretrained("charactr/vocos-mel-24khz")
        generated_wave = vocos.decode(generated_mel_spec.cpu())
        if rms < target_rms:
            generated_wave = generated_wave * rms / target_rms

        # wav -> numpy
        generated_wave = generated_wave.squeeze().cpu().numpy()

        if remove_silence:
            gr.Info("Removing audio silences... This may take a moment")
            non_silent_intervals = librosa.effects.split(generated_wave, top_db=30)
            non_silent_wave = np.array([])
            for interval in non_silent_intervals:
                start, end = interval
                non_silent_wave = np.concatenate([non_silent_wave, generated_wave[start:end]])
            generated_wave = non_silent_wave


        # spectogram
        with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as tmp_wav:
            wav_path = tmp_wav.name
            torchaudio.save(wav_path, torch.tensor(generated_wave), target_sample_rate)

        return wav_path