From b59f4af5d8b4b7f3b0b70e6fbb9d27ad87502ea5 Mon Sep 17 00:00:00 2001
From: AznamirWoW <101997116+AznamirWoW@users.noreply.github.com>
Date: Fri, 3 Jan 2025 18:55:39 -0500
Subject: [PATCH] fixing memory leak ~7MB/s during training

---
 rvc/train/train.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/rvc/train/train.py b/rvc/train/train.py
index 7a748b2d..a407a1f1 100644
--- a/rvc/train/train.py
+++ b/rvc/train/train.py
@@ -654,7 +654,7 @@ def train_and_evaluate(
                     #    loss_disc, _, _ = discriminator_loss_scaled(y_d_hat_r, y_d_hat_g)
                     loss_disc, _, _ = discriminator_loss(y_d_hat_r, y_d_hat_g)
             # Discriminator backward and update
-            epoch_disc_sum += loss_disc
+            epoch_disc_sum += loss_disc.item()
             optim_d.zero_grad()
             scaler.scale(loss_disc).backward()
             scaler.unscale_(optim_d)
@@ -689,7 +689,7 @@ def train_and_evaluate(
                             "value": loss_gen_all,
                             "epoch": epoch,
                         }
-            epoch_gen_sum += loss_gen_all
+            epoch_gen_sum += loss_gen_all.item()
             optim_g.zero_grad()
             scaler.scale(loss_gen_all).backward()
             scaler.unscale_(optim_g)