From 58431c4ceff05bd070ada370b237ad344a86acea Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E6=BA=90=E6=96=87=E9=9B=A8?=
 <41315874+fumiama@users.noreply.github.com>
Date: Sun, 3 Sep 2023 16:48:38 +0800
Subject: [PATCH 1/3] feat(webui): add mps support

---
 webui.py | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/webui.py b/webui.py
index 2dfe081..58cea49 100644
--- a/webui.py
+++ b/webui.py
@@ -8,6 +8,10 @@ from text import cleaned_text_to_sequence, get_bert
 from text.cleaner import clean_text
 import gradio as gr
 import webbrowser
+import sys, os
+
+if sys.platform == "darwin":
+    os.environ["PYTORCH_ENABLE_MPS_FALLBACK"] = "1"
 
 def get_text(text, language_str, hps):
     norm_text, phone, tone, word2ph = clean_text(text, language_str)
@@ -58,7 +62,15 @@ if __name__ == "__main__":
     args = parser.parse_args()
     hps = utils.get_hparams_from_file(args.config)
 
-    device = "cuda:0" if torch.cuda.is_available() else "cpu"
+    device = (
+        "cuda:0"
+        if torch.cuda.is_available()
+        else (
+            "mps"
+            if sys.platform == "darwin" and torch.backends.mps.is_available()
+            else "cpu"
+        )
+    )
     net_g = SynthesizerTrn(
         len(symbols),
         hps.data.filter_length // 2 + 1,

From bce9eb02512e02757d0b2f9d99369087c0752c74 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E6=BA=90=E6=96=87=E9=9B=A8?=
 <41315874+fumiama@users.noreply.github.com>
Date: Sun, 3 Sep 2023 17:04:42 +0800
Subject: [PATCH 2/3] fix: mps

---
 webui.py | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/webui.py b/webui.py
index 58cea49..9a7d5af 100644
--- a/webui.py
+++ b/webui.py
@@ -1,3 +1,8 @@
+import sys, os
+
+if sys.platform == "darwin":
+    os.environ["PYTORCH_ENABLE_MPS_FALLBACK"] = "1"
+
 import torch
 import argparse
 import commons
@@ -8,10 +13,7 @@ from text import cleaned_text_to_sequence, get_bert
 from text.cleaner import clean_text
 import gradio as gr
 import webbrowser
-import sys, os
 
-if sys.platform == "darwin":
-    os.environ["PYTORCH_ENABLE_MPS_FALLBACK"] = "1"
 
 def get_text(text, language_str, hps):
     norm_text, phone, tone, word2ph = clean_text(text, language_str)

From 5a6f8245372d9c7217cdc2faef31ad7e46eaedc3 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E6=BA=90=E6=96=87=E9=9B=A8?=
 <41315874+fumiama@users.noreply.github.com>
Date: Sun, 3 Sep 2023 17:13:30 +0800
Subject: [PATCH 3/3] optimize(webui): memory usage

---
 webui.py | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/webui.py b/webui.py
index 9a7d5af..7a58eea 100644
--- a/webui.py
+++ b/webui.py
@@ -15,6 +15,9 @@ import gradio as gr
 import webbrowser
 
 
+net_g = None
+
+
 def get_text(text, language_str, hps):
     norm_text, phone, tone, word2ph = clean_text(text, language_str)
     phone, tone, language = cleaned_text_to_sequence(phone, tone, language_str)
@@ -27,6 +30,7 @@ def get_text(text, language_str, hps):
             word2ph[i] = word2ph[i] * 2
         word2ph[0] += 1
     bert = get_bert(norm_text, word2ph, language_str)
+    del word2ph
 
     assert bert.shape[-1] == len(phone)
 
@@ -37,16 +41,19 @@ def get_text(text, language_str, hps):
     return bert, phone, tone, language
 
 def infer(text, sdp_ratio, noise_scale, noise_scale_w, length_scale, sid):
-    bert, phones, tones, lang_ids = get_text(text, "ZH", hps,)
+    global net_g
+    bert, phones, tones, lang_ids = get_text(text, "ZH", hps)
     with torch.no_grad():
         x_tst=phones.to(device).unsqueeze(0)
         tones=tones.to(device).unsqueeze(0)
         lang_ids=lang_ids.to(device).unsqueeze(0)
         bert = bert.to(device).unsqueeze(0)
         x_tst_lengths = torch.LongTensor([phones.size(0)]).to(device)
+        del phones
         speakers = torch.LongTensor([hps.data.spk2id[sid]]).to(device)
-        audio = net_g.infer(x_tst, x_tst_lengths, speakers, tones, lang_ids,bert, sdp_ratio=sdp_ratio
+        audio = net_g.infer(x_tst, x_tst_lengths, speakers, tones, lang_ids, bert, sdp_ratio=sdp_ratio
                            , noise_scale=noise_scale, noise_scale_w=noise_scale_w, length_scale=length_scale)[0][0,0].data.cpu().float().numpy()
+        del x_tst, tones, lang_ids, bert, x_tst_lengths, speakers
         return audio
 
 def tts_fn(text, speaker, sdp_ratio, noise_scale, noise_scale_w, length_scale):