Update because training works

core.py

@@ -168,8 +168,8 @@ def run_preprocess_script(model_name, dataset_path, sampling_rate):
         str(sampling_rate),
         str(per),
     ]
-    os.makedirs(os.path.join(logs_path, str(model_name)), exist_ok=True)
 
+    os.mkdir(os.path.join(logs_path, str(model_name)))
     subprocess.run(command)
     return f"Model {model_name} preprocessed successfully."
 

rvc/train/index_generator.py

@@ -45,24 +45,39 @@
 
     n_ivf = min(int(16 * np.sqrt(big_npy.shape[0])), big_npy.shape[0] // 39)
 
-    index = faiss.index_factory(256 if version == "v1" else 768, f"IVF{n_ivf},Flat")
+    # index_trained
+    index_trained = faiss.index_factory(
+        256 if version == "v1" else 768, f"IVF{n_ivf},Flat"
+    )
+    index_ivf_trained = faiss.extract_index_ivf(index_trained)
+    index_ivf_trained.nprobe = 1
+    index_trained.train(big_npy)
+
+    index_filename_trained = (
+        f"trained_IVF{n_ivf}_Flat_nprobe_{index_ivf_trained.nprobe}_{version}.index"
+    )
+    index_filepath_trained = os.path.join(exp_dir, index_filename_trained)
 
-    index_ivf = faiss.extract_index_ivf(index)
-    index_ivf.nprobe = 1
-    index.train(big_npy)
+    faiss.write_index(index_trained, index_filepath_trained)
 
-    index_filename = (
-        f"trained_IVF{n_ivf}_Flat_nprobe_{index_ivf.nprobe}_{version}.index"
+    # index_added
+    index_added = faiss.index_factory(
+        256 if version == "v1" else 768, f"IVF{n_ivf},Flat"
     )
-    index_filepath = os.path.join(exp_dir, index_filename)
+    index_ivf_added = faiss.extract_index_ivf(index_added)
+    index_ivf_added.nprobe = 1
+    index_added.train(big_npy)
 
-    faiss.write_index(index, index_filepath)
+    index_filename_added = (
+        f"added_IVF{n_ivf}_Flat_nprobe_{index_ivf_added.nprobe}_{version}.index"
+    )
+    index_filepath_added = os.path.join(exp_dir, index_filename_added)
 
     batch_size_add = 8192
     for i in range(0, big_npy.shape[0], batch_size_add):
-        index.add(big_npy[i : i + batch_size_add])
+        index_added.add(big_npy[i : i + batch_size_add])
 
-    faiss.write_index(index, index_filepath)
+    faiss.write_index(index_added, index_filepath_added)
 
 except Exception as error:
     print(f"Failed to train index: {error}")

rvc/train/losses.py

@@ -1,5 +1,6 @@
 import torch
 
+
 def feature_loss(fmap_r, fmap_g):
     loss = 0
     for dr, dg in zip(fmap_r, fmap_g):

rvc/train/slicer.py

@@ -4,6 +4,7 @@
 import soundfile
 import numpy as np
 
+
 class Slicer:
     def __init__(
         self,

rvc/train/train.py

@@ -90,7 +90,7 @@ def record(self):
         elapsed_time = round(elapsed_time, 1)
         elapsed_time_str = str(datetime.timedelta(seconds=int(elapsed_time)))
         current_time = datetime.datetime.now().strftime("%H:%M:%S")
-        return f"[{current_time}] | ({elapsed_time_str})"
+        return f"time={current_time} | training_speed={elapsed_time_str}"
 
 
 def main():
@@ -201,6 +201,7 @@ def run(
         net_d = DDP(net_d)
 
     try:
+        print("Starting training...")
         _, _, _, epoch_str = load_checkpoint(
             latest_checkpoint_path(hps.model_dir, "D_*.pth"), net_d, optim_d
         )
@@ -477,10 +478,6 @@ def train_and_evaluate(rank, epoch, hps, nets, optims, scaler, loaders, writers,
                 if loss_kl > 9:
                     loss_kl = 9
 
-                # print([global_step, lr])
-                print(
-                    f"[loss_disc={loss_disc:.3f} | loss_gen={loss_gen:.3f} | loss_fm={loss_fm:.3f} | loss_mel={loss_mel:.3f} | loss_kl={loss_kl:.3f}]"
-                )
                 scalar_dict = {
                     "loss/g/total": loss_gen_all,
                     "loss/d/total": loss_disc,
@@ -559,15 +556,15 @@ def train_and_evaluate(rank, epoch, hps, nets, optims, scaler, loaders, writers,
             else:
                 ckpt = net_g.state_dict()
             print(
-                "Saving small model... %s_e%s:%s"
+                "saving ckpt %s_e%s:%s"
                 % (
                     hps.name,
                     epoch,
                     save_final(
                         ckpt,
                         hps.sample_rate,
                         hps.if_f0,
-                        hps.name + f"_e{epoch}_s{global_step}",
+                        hps.name + "_e%s_s%s" % (epoch, global_step),
                         epoch,
                         hps.version,
                         hps,
@@ -576,16 +573,20 @@ def train_and_evaluate(rank, epoch, hps, nets, optims, scaler, loaders, writers,
             )
 
     if rank == 0:
-        print(f"Epoch {epoch}: {epoch_recorder.record()}")
+        print(
+            f"{hps.name} | epoch={epoch} | step={global_step} | {epoch_recorder.record()} | loss_disc={loss_disc:.3f} | loss_gen={loss_gen:.3f} | loss_fm={loss_fm:.3f} | loss_mel={loss_mel:.3f} | loss_kl={loss_kl:.3f}"
+        )
     if epoch >= hps.total_epoch and rank == 0:
-        print("The training is over, finalizing the program...")
+        print(
+            f"Training has been successfully completed with {epoch} epoch and {global_step} steps."
+        )
 
         if hasattr(net_g, "module"):
             ckpt = net_g.module.state_dict()
         else:
             ckpt = net_g.state_dict()
         print(
-            "Saving final model... %s"
+            "Saving final checkpoint: %s"
             % (
                 save_final(
                     ckpt, hps.sample_rate, hps.if_f0, hps.name, epoch, hps.version, hps

rvc/train/utils.py

@@ -83,12 +83,12 @@ def load_checkpoint(checkpoint_path, model, optimizer=None, load_opt=1):
     learning_rate = checkpoint_dict["learning_rate"]
     if optimizer is not None and load_opt == 1:
         optimizer.load_state_dict(checkpoint_dict["optimizer"])
-    print(f"Loaded checkpoint '{checkpoint_path}' (epochs {iteration})")
+    print(f"Loaded checkpoint '{checkpoint_path}' (epoch {iteration})")
     return model, optimizer, learning_rate, iteration
 
 
 def save_checkpoint(model, optimizer, learning_rate, iteration, checkpoint_path):
-    print(f"Saving model '{checkpoint_path}' (epochs {iteration})")
+    print(f"Saving model '{checkpoint_path}' (epoch {iteration})")
     if hasattr(model, "module"):
         state_dict = model.module.state_dict()
     else:
@@ -218,34 +218,6 @@ def get_hparams():
         required=True,
         help="if caching the dataset in GPU memory, 1 or 0",
     )
-    parser.add_argument(
-        "-sof",
-        "--stop_on_fit",
-        type=int,
-        required=False,
-        help="if retraining mode collapses, 1 or 0",
-    )
-    parser.add_argument(
-        "-sm",
-        "--smoothness",
-        type=float,
-        required=False,
-        help="smoothness for --stop_on_fit",
-    )
-    parser.add_argument(
-        "-rc",
-        "--retrain_collapse",
-        type=int,
-        required=False,
-        help="if retraining mode collapses, 1 or 0",
-    )
-    parser.add_argument(
-        "-ct",
-        "--collapse_threshold",
-        type=int,
-        required=False,
-        help="if retraining mode collapses, 1 or 0",
-    )
     args = parser.parse_args()
     name = args.experiment_dir
     experiment_dir = os.path.join("./logs", args.experiment_dir)
@@ -267,11 +239,6 @@ def get_hparams():
     hparams.if_latest = args.if_latest
     hparams.save_every_weights = args.save_every_weights
     hparams.if_cache_data_in_gpu = args.if_cache_data_in_gpu
-    hparams.if_stop_on_fit = args.stop_on_fit
-    hparams.smoothness = args.smoothness
-    hparams.if_retrain_collapse = args.retrain_collapse
-    if args.collapse_threshold != None:
-        hparams.collapse_threshold = args.collapse_threshold * 0.01
     hparams.data.training_files = f"{experiment_dir}/filelist.txt"
     return hparams
 

tabs/inference/inference.py

@@ -1,6 +1,6 @@
 import os, sys
 import gradio as gr
-import re
+import regex as re
 import shutil
 import datetime
 import random
@@ -39,10 +39,7 @@
     os.path.join(root, file)
     for root, _, files in os.walk(model_root, topdown=False)
     for file in files
-    if (
-        file.endswith((".pth", ".onnx"))
-        and not (file.startswith("G_") or file.startswith("D_"))
-    )
+    if file.endswith((".pth", ".onnx"))
 ]
 
 indexes_list = [
@@ -76,10 +73,7 @@ def change_choices():
         os.path.join(root, file)
         for root, _, files in os.walk(model_root, topdown=False)
         for file in files
-        if (
-            file.endswith((".pth", ".onnx"))
-            and not (file.startswith("G_") or file.startswith("D_"))
-        )
+        if file.endswith((".pth", ".onnx"))
     ]
 
     indexes_list = [
@@ -268,14 +262,7 @@ def inference_tab():
                     value=False,
                     interactive=True,
                 )
-                pitch = gr.Slider(
-                    minimum=-12,
-                    maximum=12,
-                    step=1,
-                    label=i18n("Pitch"),
-                    value=0,
-                    interactive=True,
-                )
+                pitch = gr.Slider(-12, 12, 0, label=i18n("Pitch"))
                 filter_radius = gr.Slider(
                     minimum=0,
                     maximum=7,

tabs/tts/tts.py

@@ -39,10 +39,7 @@
     os.path.join(root, file)
     for root, _, files in os.walk(model_root, topdown=False)
     for file in files
-    if (
-        file.endswith((".pth", ".onnx"))
-        and not (file.startswith("G_") or file.startswith("D_"))
-    )
+    if file.endswith((".pth", ".onnx"))
 ]
 
 indexes_list = [
@@ -67,10 +64,7 @@ def change_choices():
         os.path.join(root, file)
         for root, _, files in os.walk(model_root, topdown=False)
         for file in files
-        if (
-            file.endswith((".pth", ".onnx"))
-            and not (file.startswith("G_") or file.startswith("D_"))
-        )
+        if file.endswith((".pth", ".onnx"))
     ]
 
     indexes_list = [
@@ -257,14 +251,7 @@ def tts_tab():
                 interactive=True,
             )
 
-            pitch = gr.Slider(
-                minimum=-12,
-                maximum=12,
-                step=1,
-                label=i18n("Pitch"),
-                value=0,
-                interactive=True,
-            )
+            pitch = gr.Slider(-12, 12, 0, label=i18n("Pitch"))
             filter_radius = gr.Slider(
                 minimum=0,
                 maximum=7,