lucasnewman

import datetime
from pathlib import Path

import torch
import torchaudio
from torchaudio.transforms import MelSpectrogram

from einops import rearrange

from vocos import Vocos

lucasnewman

from __future__ import annotations

from typing import Any, Optional

import torch
from torch import nn

import torchaudio

import yaml

lucasnewman

from __future__ import annotations
from functools import reduce
from typing import List

import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.nn.utils.rnn import pad_sequence

import einx

lucasnewman

import torch

def mel_filterbank(
    n_freqs,
    f_min,
    f_max,
    n_mels,
    sample_rate,
    norm=None,
    mel_scale="htk"

lucasnewman

from functools import lru_cache

import librosa
import mlx.core as mx
import numpy as np

@lru_cache(maxsize=None)
def hanning(size):
    """
    Compute the Hanning window.

lucasnewman

import mlx.core as mx

def istft(
    x: mx.array,  # (freq_bins, num_frames)
    window: mx.array,
    hop_length: int = 256,
    win_length: int = 1024,
):
    num_frames = x.shape[1]
    total_len = (num_frames - 1) * hop_length + win_length

lucasnewman

diff --git a/mlx_audio/tts/models/spark/bicodec.py b/mlx_audio/tts/models/spark/bicodec.py
index 8de944b..1e2a320 100644
--- a/mlx_audio/tts/models/spark/bicodec.py
+++ b/mlx_audio/tts/models/spark/bicodec.py
@@ -1,5 +1,5 @@
 from pathlib import Path
-from typing import Any, Dict
+from typing import Any, Dict, Optional
 
 import mlx.core as mx