fix(train): support ddp in Windows (gloo backend) (#394)

src/so_vits_svc_fork/train.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import os
 import warnings
 from logging import getLogger
 from multiprocessing import cpu_count
@@ -10,6 +11,7 @@
 import torch
 from lightning.pytorch.accelerators import MPSAccelerator, TPUAccelerator
 from lightning.pytorch.loggers import TensorBoardLogger
+from lightning.pytorch.strategies.ddp import DDPStrategy
 from lightning.pytorch.tuner import Tuner
 from torch.cuda.amp import autocast
 from torch.nn import functional as F
@@ -75,7 +77,13 @@ def train(
 
     datamodule = VCDataModule(hparams)
     strategy = (
-        "ddp_find_unused_parameters_true" if torch.cuda.device_count() > 1 else "auto"
+        (
+            "ddp_find_unused_parameters_true"
+            if os.name != "nt"
+            else DDPStrategy(find_unused_parameters=True, process_group_backend="gloo")
+        )
+        if torch.cuda.device_count() > 1
+        else "auto"
     )
     LOG.info(f"Using strategy: {strategy}")
     trainer = pl.Trainer(