Merge branch '4.1-Latest' of https://github.com/svc-develop-team/so-vits-svc into 4.1-Latest

.github/workflows/reviewdog.yml

@@ -0,0 +1,18 @@
+name: Ruff Autofix
+on: [pull_request]
+jobs:
+  ruff:
+    permissions:
+      checks: write
+      contents: read
+      pull-requests: write
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v3
+      - uses: chartboost/ruff-action@v1
+        with:
+          args: --fix -e
+      - uses: reviewdog/action-suggester@v1
+        with:
+          tool_name: ruff
+

utils.py

@@ -6,12 +6,15 @@ import os
 import re
 import subprocess
 import sys
+import traceback
+from multiprocessing import cpu_count
 
 import faiss
 import librosa
 import numpy as np
 import torch
 from scipy.io.wavfile import read
+from sklearn.cluster import MiniBatchKMeans
 from torch.nn import functional as F
 
 MATPLOTLIB_FLAG = False
@@ -449,6 +452,7 @@ def change_rms(data1, sr1, data2, sr2, rate):  # 1是输入音频，2是输出
     return data2
 
 def train_index(spk_name,root_dir = "dataset/44k/"):  #from: RVC https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI
+    n_cpu = cpu_count()
     print("The feature index is constructing.")
     exp_dir = os.path.join(root_dir,spk_name)
     listdir_res = []
@@ -465,6 +469,25 @@ def train_index(spk_name,root_dir = "dataset/44k/"):  #from: RVC https://github.
     big_npy_idx = np.arange(big_npy.shape[0])
     np.random.shuffle(big_npy_idx)
     big_npy = big_npy[big_npy_idx]
+    if big_npy.shape[0] > 2e5:
+        # if(1):
+        info = "Trying doing kmeans %s shape to 10k centers." % big_npy.shape[0]
+        print(info)
+        try:
+            big_npy = (
+                MiniBatchKMeans(
+                    n_clusters=10000,
+                    verbose=True,
+                    batch_size=256 * n_cpu,
+                    compute_labels=False,
+                    init="random",
+                )
+                .fit(big_npy)
+                .cluster_centers_
+            )
+        except:  # noqa: E722
+            info = traceback.format_exc()
+            print(info)
     n_ivf = min(int(16 * np.sqrt(big_npy.shape[0])), big_npy.shape[0] // 39)
     index = faiss.index_factory(big_npy.shape[1] , "IVF%s,Flat" % n_ivf)
     index_ivf = faiss.extract_index_ivf(index)  #