diff --git a/infer-web.py b/infer-web.py
index 47596d539..9e1765bd6 100644
--- a/infer-web.py
+++ b/infer-web.py
@@ -16,7 +16,7 @@
from i18n.i18n import I18nAuto
from configs.config import Config
from sklearn.cluster import MiniBatchKMeans
-import torch, platform
+import torch
import numpy as np
import gradio as gr
import faiss
@@ -32,9 +32,14 @@
import shutil
import logging
+# เพิ่มส่วนตรวจสอบ DirectML
+try:
+ import torch_directml
+ has_dml = True
+except ImportError:
+ has_dml = False
logging.getLogger("numba").setLevel(logging.WARNING)
-logging.getLogger("httpx").setLevel(logging.WARNING)
logger = logging.getLogger(__name__)
@@ -52,73 +57,63 @@
config = Config()
vc = VC(config)
-
-
+
if config.dml == True:
-
def forward_dml(ctx, x, scale):
ctx.scale = scale
res = x.clone().detach()
return res
fairseq.modules.grad_multiply.GradMultiply.forward = forward_dml
+
i18n = I18nAuto()
logger.info(i18n)
-# 判断是否有能用来训练和加速推理的N卡
+
+# ==========================================
+# ส่วนตรวจสอบ GPU (NVIDIA & AMD/DirectML)
+# ==========================================
ngpu = torch.cuda.device_count()
gpu_infos = []
mem = []
if_gpu_ok = False
+# 1. ตรวจสอบการ์ดจอ NVIDIA ก่อน
if torch.cuda.is_available() or ngpu != 0:
for i in range(ngpu):
gpu_name = torch.cuda.get_device_name(i)
if any(
value in gpu_name.upper()
for value in [
- "10",
- "16",
- "20",
- "30",
- "40",
- "A2",
- "A3",
- "A4",
- "P4",
- "A50",
- "500",
- "A60",
- "70",
- "80",
- "90",
- "M4",
- "T4",
- "TITAN",
- "4060",
- "L",
- "6000",
+ "10", "16", "20", "30", "40", "A2", "A3", "A4", "P4", "A50", "500", "A60", "70", "80", "90", "M4", "T4", "TITAN",
]
):
- # A10#A100#V100#A40#P40#M40#K80#A4500
- if_gpu_ok = True # 至少有一张能用的N卡
+ if_gpu_ok = True
gpu_infos.append("%s\t%s" % (i, gpu_name))
mem.append(
int(
- torch.cuda.get_device_properties(i).total_memory
- / 1024
- / 1024
- / 1024
- + 0.4
+ torch.cuda.get_device_properties(i).total_memory / 1024 / 1024 / 1024 + 0.4
)
)
+
+# 2. ถ้าไม่เจอ NVIDIA ให้ตรวจสอบ AMD ผ่าน DirectML (ถ้ามีติดตั้งไว้)
+if not if_gpu_ok and has_dml and torch_directml.is_available():
+ if_gpu_ok = True
+ dml_device_count = torch_directml.device_count()
+ for i in range(dml_device_count):
+ gpu_name = torch_directml.device_name(i)
+ gpu_infos.append("%s\t%s (DirectML)" % (i, gpu_name))
+ # DirectML ไม่ได้คืนค่า memory ง่ายๆ เหมือน CUDA จึงกำหนดค่ากลางไว้เผื่อคำนวณ batch_size
+ mem.append(8)
+
if if_gpu_ok and len(gpu_infos) > 0:
gpu_info = "\n".join(gpu_infos)
- default_batch_size = min(mem) // 2
+ default_batch_size = min(mem) // 2 if mem else 4 # ป้องกัน error ถ้า memory ว่าง
else:
- gpu_info = i18n("很遗憾您这没有能用的显卡来支持您训练")
+ gpu_info = "Unfortunately, there is no compatible GPU available for training. Falling back to CPU."
default_batch_size = 1
-gpus = "-".join([i[0] for i in gpu_infos])
+gpus = "-".join([i[0].split('\t')[0] for i in gpu_infos]) if gpu_infos else ""
+# ==========================================
class ToolButton(gr.Button, gr.components.FormComponent):
"""Small button with single emoji as text, fits inside gradio forms"""
@@ -129,35 +124,24 @@ def __init__(self, **kwargs):
def get_block_name(self):
return "button"
-
weight_root = os.getenv("weight_root")
weight_uvr5_root = os.getenv("weight_uvr5_root")
index_root = os.getenv("index_root")
-outside_index_root = os.getenv("outside_index_root")
names = []
for name in os.listdir(weight_root):
if name.endswith(".pth"):
names.append(name)
index_paths = []
-
-
-def lookup_indices(index_root):
- global index_paths
- for root, dirs, files in os.walk(index_root, topdown=False):
- for name in files:
- if name.endswith(".index") and "trained" not in name:
- index_paths.append("%s/%s" % (root, name))
-
-
-lookup_indices(index_root)
-lookup_indices(outside_index_root)
+for root, dirs, files in os.walk(index_root, topdown=False):
+ for name in files:
+ if name.endswith(".index") and "trained" not in name:
+ index_paths.append("%s/%s" % (root, name))
uvr5_names = []
for name in os.listdir(weight_uvr5_root):
if name.endswith(".pth") or "onnx" in name:
uvr5_names.append(name.replace(".pth", ""))
-
def change_choices():
names = []
for name in os.listdir(weight_root):
@@ -173,24 +157,20 @@ def change_choices():
"__type__": "update",
}
-
def clean():
return {"value": "", "__type__": "update"}
-
def export_onnx(ModelPath, ExportedPath):
from infer.modules.onnx.export import export_onnx as eo
eo(ModelPath, ExportedPath)
-
sr_dict = {
"32k": 32000,
"40k": 40000,
"48k": 48000,
}
-
def if_done(done, p):
while 1:
if p.poll() is None:
@@ -202,8 +182,6 @@ def if_done(done, p):
def if_done_multi(done, ps):
while 1:
- # poll==None代表进程未结束
- # 只要有一个进程未结束都不停
flag = 1
for p in ps:
if p.poll() is None:
@@ -220,6 +198,7 @@ def preprocess_dataset(trainset_dir, exp_dir, sr, n_p):
os.makedirs("%s/logs/%s" % (now_dir, exp_dir), exist_ok=True)
f = open("%s/logs/%s/preprocess.log" % (now_dir, exp_dir), "w")
f.close()
+ per = 3.0 if config.is_half else 3.7
cmd = '"%s" infer/modules/train/preprocess.py "%s" %s %s "%s/logs/%s" %s %.1f' % (
config.python_cmd,
trainset_dir,
@@ -228,12 +207,10 @@ def preprocess_dataset(trainset_dir, exp_dir, sr, n_p):
now_dir,
exp_dir,
config.noparallel,
- config.preprocess_per,
+ per,
)
- logger.info("Execute: " + cmd)
- # , stdin=PIPE, stdout=PIPE,stderr=PIPE,cwd=now_dir
+ logger.info(cmd)
p = Popen(cmd, shell=True)
- # 煞笔gr, popen read都非得全跑完了再一次性读取, 不用gr就正常读一句输出一句;只能额外弄出一个文本流定时读
done = [False]
threading.Thread(
target=if_done,
@@ -254,7 +231,6 @@ def preprocess_dataset(trainset_dir, exp_dir, sr, n_p):
yield log
-# but2.click(extract_f0,[gpus6,np7,f0method8,if_f0_3,trainset_dir4],[info2])
def extract_f0_feature(gpus, n_p, f0method, if_f0, exp_dir, version19, gpus_rmvpe):
gpus = gpus.split("-")
os.makedirs("%s/logs/%s" % (now_dir, exp_dir), exist_ok=True)
@@ -272,11 +248,10 @@ def extract_f0_feature(gpus, n_p, f0method, if_f0, exp_dir, version19, gpus_rmvp
f0method,
)
)
- logger.info("Execute: " + cmd)
+ logger.info(cmd)
p = Popen(
cmd, shell=True, cwd=now_dir
- ) # , stdin=PIPE, stdout=PIPE,stderr=PIPE
- # 煞笔gr, popen read都非得全跑完了再一次性读取, 不用gr就正常读一句输出一句;只能额外弄出一个文本流定时读
+ )
done = [False]
threading.Thread(
target=if_done,
@@ -303,15 +278,14 @@ def extract_f0_feature(gpus, n_p, f0method, if_f0, exp_dir, version19, gpus_rmvp
config.is_half,
)
)
- logger.info("Execute: " + cmd)
+ logger.info(cmd)
p = Popen(
cmd, shell=True, cwd=now_dir
- ) # , shell=True, stdin=PIPE, stdout=PIPE, stderr=PIPE, cwd=now_dir
+ )
ps.append(p)
- # 煞笔gr, popen read都非得全跑完了再一次性读取, 不用gr就正常读一句输出一句;只能额外弄出一个文本流定时读
done = [False]
threading.Thread(
- target=if_done_multi, #
+ target=if_done_multi,
args=(
done,
ps,
@@ -326,10 +300,10 @@ def extract_f0_feature(gpus, n_p, f0method, if_f0, exp_dir, version19, gpus_rmvp
exp_dir,
)
)
- logger.info("Execute: " + cmd)
+ logger.info(cmd)
p = Popen(
cmd, shell=True, cwd=now_dir
- ) # , shell=True, stdin=PIPE, stdout=PIPE, stderr=PIPE, cwd=now_dir
+ )
p.wait()
done = [True]
while 1:
@@ -344,19 +318,12 @@ def extract_f0_feature(gpus, n_p, f0method, if_f0, exp_dir, version19, gpus_rmvp
log = f.read()
logger.info(log)
yield log
- # 对不同part分别开多进程
- """
- n_part=int(sys.argv[1])
- i_part=int(sys.argv[2])
- i_gpu=sys.argv[3]
- exp_dir=sys.argv[4]
- os.environ["CUDA_VISIBLE_DEVICES"]=str(i_gpu)
- """
+
leng = len(gpus)
ps = []
for idx, n_g in enumerate(gpus):
cmd = (
- '"%s" infer/modules/train/extract_feature_print.py %s %s %s %s "%s/logs/%s" %s %s'
+ '"%s" infer/modules/train/extract_feature_print.py %s %s %s %s "%s/logs/%s" %s'
% (
config.python_cmd,
config.device,
@@ -366,15 +333,13 @@ def extract_f0_feature(gpus, n_p, f0method, if_f0, exp_dir, version19, gpus_rmvp
now_dir,
exp_dir,
version19,
- config.is_half,
)
)
- logger.info("Execute: " + cmd)
+ logger.info(cmd)
p = Popen(
cmd, shell=True, cwd=now_dir
- ) # , shell=True, stdin=PIPE, stdout=PIPE, stderr=PIPE, cwd=now_dir
+ )
ps.append(p)
- # 煞笔gr, popen read都非得全跑完了再一次性读取, 不用gr就正常读一句输出一句;只能额外弄出一个文本流定时读
done = [False]
threading.Thread(
target=if_done_multi,
@@ -417,16 +382,12 @@ def get_pretrained_models(path_str, f0_str, sr2):
sr2,
)
return (
- (
- "assets/pretrained%s/%sG%s.pth" % (path_str, f0_str, sr2)
- if if_pretrained_generator_exist
- else ""
- ),
- (
- "assets/pretrained%s/%sD%s.pth" % (path_str, f0_str, sr2)
- if if_pretrained_discriminator_exist
- else ""
- ),
+ "assets/pretrained%s/%sG%s.pth" % (path_str, f0_str, sr2)
+ if if_pretrained_generator_exist
+ else "",
+ "assets/pretrained%s/%sD%s.pth" % (path_str, f0_str, sr2)
+ if if_pretrained_discriminator_exist
+ else "",
)
@@ -452,7 +413,7 @@ def change_version19(sr2, if_f0_3, version19):
)
-def change_f0(if_f0_3, sr2, version19): # f0method8,pretrained_G14,pretrained_D15
+def change_f0(if_f0_3, sr2, version19):
path_str = "" if version19 == "v1" else "_v2"
return (
{"visible": if_f0_3, "__type__": "update"},
@@ -461,7 +422,6 @@ def change_f0(if_f0_3, sr2, version19): # f0method8,pretrained_G14,pretrained_D
)
-# but3.click(click_train,[exp_dir1,sr2,if_f0_3,save_epoch10,total_epoch11,batch_size12,if_save_latest13,pretrained_G14,pretrained_D15,gpus16])
def click_train(
exp_dir1,
sr2,
@@ -478,7 +438,6 @@ def click_train(
if_save_every_weights18,
version19,
):
- # 生成filelist
exp_dir = "%s/logs/%s" % (now_dir, exp_dir1)
os.makedirs(exp_dir, exist_ok=True)
gt_wavs_dir = "%s/0_gt_wavs" % (exp_dir)
@@ -545,8 +504,6 @@ def click_train(
with open("%s/filelist.txt" % exp_dir, "w") as f:
f.write("\n".join(opt))
logger.debug("Write filelist done")
- # 生成config#无需生成config
- # cmd = python_cmd + " train_nsf_sim_cache_sid_load_pretrain.py -e mi-test -sr 40k -f0 1 -bs 4 -g 0 -te 10 -se 5 -pg pretrained/f0G40k.pth -pd pretrained/f0D40k.pth -l 1 -c 0"
logger.info("Use gpus: %s", str(gpus16))
if pretrained_G14 == "":
logger.info("No pretrained Generator")
@@ -581,9 +538,9 @@ def click_train(
save_epoch10,
"-pg %s" % pretrained_G14 if pretrained_G14 != "" else "",
"-pd %s" % pretrained_D15 if pretrained_D15 != "" else "",
- 1 if if_save_latest13 == i18n("是") else 0,
- 1 if if_cache_gpu17 == i18n("是") else 0,
- 1 if if_save_every_weights18 == i18n("是") else 0,
+ 1 if if_save_latest13 == "Yes" else 0,
+ 1 if if_cache_gpu17 == "Yes" else 0,
+ 1 if if_save_every_weights18 == "Yes" else 0,
version19,
)
)
@@ -600,21 +557,19 @@ def click_train(
save_epoch10,
"-pg %s" % pretrained_G14 if pretrained_G14 != "" else "",
"-pd %s" % pretrained_D15 if pretrained_D15 != "" else "",
- 1 if if_save_latest13 == i18n("是") else 0,
- 1 if if_cache_gpu17 == i18n("是") else 0,
- 1 if if_save_every_weights18 == i18n("是") else 0,
+ 1 if if_save_latest13 == "Yes" else 0,
+ 1 if if_cache_gpu17 == "Yes" else 0,
+ 1 if if_save_every_weights18 == "Yes" else 0,
version19,
)
)
- logger.info("Execute: " + cmd)
+ logger.info(cmd)
p = Popen(cmd, shell=True, cwd=now_dir)
p.wait()
- return "训练结束, 您可查看控制台训练日志或实验文件夹下的train.log"
+ return "Training finished. You can check the console log or train.log in the experiment folder."
-# but4.click(train_index, [exp_dir1], info3)
def train_index(exp_dir1, version19):
- # exp_dir = "%s/logs/%s" % (now_dir, exp_dir1)
exp_dir = "logs/%s" % (exp_dir1)
os.makedirs(exp_dir, exist_ok=True)
feature_dir = (
@@ -623,10 +578,10 @@ def train_index(exp_dir1, version19):
else "%s/3_feature768" % (exp_dir)
)
if not os.path.exists(feature_dir):
- return "请先进行特征提取!"
+ return "Please perform feature extraction first!"
listdir_res = list(os.listdir(feature_dir))
if len(listdir_res) == 0:
- return "请先进行特征提取！"
+ return "Please perform feature extraction first!"
infos = []
npys = []
for name in sorted(listdir_res):
@@ -662,10 +617,9 @@ def train_index(exp_dir1, version19):
infos.append("%s,%s" % (big_npy.shape, n_ivf))
yield "\n".join(infos)
index = faiss.index_factory(256 if version19 == "v1" else 768, "IVF%s,Flat" % n_ivf)
- # index = faiss.index_factory(256if version19=="v1"else 768, "IVF%s,PQ128x4fs,RFlat"%n_ivf)
- infos.append("training")
+ infos.append("Training index...")
yield "\n".join(infos)
- index_ivf = faiss.extract_index_ivf(index) #
+ index_ivf = faiss.extract_index_ivf(index)
index_ivf.nprobe = 1
index.train(big_npy)
faiss.write_index(
@@ -673,7 +627,8 @@ def train_index(exp_dir1, version19):
"%s/trained_IVF%s_Flat_nprobe_%s_%s_%s.index"
% (exp_dir, n_ivf, index_ivf.nprobe, exp_dir1, version19),
)
- infos.append("adding")
+
+ infos.append("Adding index...")
yield "\n".join(infos)
batch_size_add = 8192
for i in range(0, big_npy.shape[0], batch_size_add):
@@ -684,34 +639,12 @@ def train_index(exp_dir1, version19):
% (exp_dir, n_ivf, index_ivf.nprobe, exp_dir1, version19),
)
infos.append(
- "成功构建索引 added_IVF%s_Flat_nprobe_%s_%s_%s.index"
+ "Index built successfully: added_IVF%s_Flat_nprobe_%s_%s_%s.index"
% (n_ivf, index_ivf.nprobe, exp_dir1, version19)
)
- try:
- link = os.link if platform.system() == "Windows" else os.symlink
- link(
- "%s/added_IVF%s_Flat_nprobe_%s_%s_%s.index"
- % (exp_dir, n_ivf, index_ivf.nprobe, exp_dir1, version19),
- "%s/%s_IVF%s_Flat_nprobe_%s_%s_%s.index"
- % (
- outside_index_root,
- exp_dir1,
- n_ivf,
- index_ivf.nprobe,
- exp_dir1,
- version19,
- ),
- )
- infos.append("链接索引到外部-%s" % (outside_index_root))
- except:
- infos.append("链接索引到外部-%s失败" % (outside_index_root))
-
- # faiss.write_index(index, '%s/added_IVF%s_Flat_FastScan_%s.index'%(exp_dir,n_ivf,version19))
- # infos.append("成功构建索引，added_IVF%s_Flat_FastScan_%s.index"%(n_ivf,version19))
yield "\n".join(infos)
-# but5.click(train1key, [exp_dir1, sr2, if_f0_3, trainset_dir4, spk_id5, gpus6, np7, f0method8, save_epoch10, total_epoch11, batch_size12, if_save_latest13, pretrained_G14, pretrained_D15, gpus16, if_cache_gpu17], info3)
def train1key(
exp_dir1,
sr2,
@@ -738,12 +671,10 @@ def get_info_str(strr):
infos.append(strr)
return "\n".join(infos)
- # step1:处理数据
- yield get_info_str(i18n("step1:正在处理数据"))
+ yield get_info_str("Step No. 1 Processing data...")
[get_info_str(_) for _ in preprocess_dataset(trainset_dir4, exp_dir1, sr2, np7)]
- # step2a:提取音高
- yield get_info_str(i18n("step2:正在提取音高&正在提取特征"))
+ yield get_info_str("Step No. 2 Extracting pitch & features...")
[
get_info_str(_)
for _ in extract_f0_feature(
@@ -751,8 +682,7 @@ def get_info_str(strr):
)
]
- # step3a:训练模型
- yield get_info_str(i18n("step3a:正在训练模型"))
+ yield get_info_str("Step No. 3 Training model...")
click_train(
exp_dir1,
sr2,
@@ -769,16 +699,12 @@ def get_info_str(strr):
if_save_every_weights18,
version19,
)
- yield get_info_str(
- i18n("训练结束, 您可查看控制台训练日志或实验文件夹下的train.log")
- )
+ yield get_info_str("The Training For Model is Complete. Please Check to the train.log's File in the experiment log folder.")
- # step3b:训练索引
[get_info_str(_) for _ in train_index(exp_dir1, version19)]
- yield get_info_str(i18n("全流程结束！"))
+ yield get_info_str("All Processes is Completed.!!!")
-# ckpt_path2.change(change_info_,[ckpt_path2],[sr__,if_f0__])
def change_info_(ckpt_path):
if not os.path.exists(ckpt_path.replace(os.path.basename(ckpt_path), "train.log")):
return {"__type__": "update"}, {"__type__": "update"}, {"__type__": "update"}
@@ -806,27 +732,23 @@ def change_f0_method(f0method8):
return {"visible": visible, "__type__": "update"}
-with gr.Blocks(title="RVC WebUI") as app:
- gr.Markdown("## RVC WebUI")
+with gr.Blocks(title="The RVC WebUI And Voice Remover WebUI Editor By DELTA SYNTH.") as app:
+ gr.Markdown("**The Official English RVC Editor WebUI.**")
gr.Markdown(
- value=i18n(
- "本软件以MIT协议开源, 作者不对软件具备任何控制力, 使用软件者、传播软件导出的声音者自负全责.
如不认可该条款, 则不能使用或引用软件包内任何代码和文件. 详见根目录LICENSE."
- )
+ value="***This Software is open source under the MIT License. The author has no control over the software. Users and distributors of the exported voices bear full responsibility.***"
)
with gr.Tabs():
- with gr.TabItem(i18n("模型推理")):
+ with gr.TabItem("Use The Inference for Model to Rendering."):
with gr.Row():
- sid0 = gr.Dropdown(label=i18n("推理音色"), choices=sorted(names))
+ sid0 = gr.Dropdown(label="Choose to Inference the Model's Voice To Use it.", choices=sorted(names))
with gr.Column():
- refresh_button = gr.Button(
- i18n("刷新音色列表和索引路径"), variant="primary"
- )
- clean_button = gr.Button(i18n("卸载音色省显存"), variant="primary")
+ refresh_button = gr.Button("Click to Refresh For All Model Voice List", variant="primary")
+ clean_button = gr.Button("Click to Reset Voice File To Clear VRAM ( Save VRAM )", variant="primary")
spk_item = gr.Slider(
minimum=0,
maximum=2333,
step=1,
- label=i18n("请选择说话人id"),
+ label="Input to the Speaker's ID",
value=0,
visible=False,
interactive=True,
@@ -834,41 +756,35 @@ def change_f0_method(f0method8):
clean_button.click(
fn=clean, inputs=[], outputs=[sid0], api_name="infer_clean"
)
- with gr.TabItem(i18n("单次推理")):
+ with gr.TabItem("1. Inference For 1 File Using."):
with gr.Group():
with gr.Row():
with gr.Column():
vc_transform0 = gr.Number(
- label=i18n("变调(整数, 半音数量, 升八度12降八度-12)"),
- value=0,
+ label="1.1. Input how much to Pitch Shift ( Integer, Semitones, +12 Octave Up is Same the kid Voice , -12 Octave Down is same the Adult guy voice. )", value=0
)
input_audio0 = gr.Textbox(
- label=i18n(
- "输入待处理音频文件路径(默认是正确格式示例)"
- ),
- placeholder="C:\\Users\\Desktop\\audio_example.wav",
+ label="1.2. Input Audio Path (Example of correct format)",
+ placeholder="H:\Export to UTAU\Test.wav",
+ value="H:\Export to UTAU\Test.wav",
)
file_index1 = gr.Textbox(
- label=i18n(
- "特征检索库文件路径,为空则使用下拉的选择结果"
- ),
- placeholder="C:\\Users\\Desktop\\model_example.index",
+ label="1.3. Input Index Path (Leave blank to use dropdown)",
+ placeholder="Logs\A.index",
+ value="Logs\A.index",
interactive=True,
)
file_index2 = gr.Dropdown(
- label=i18n("自动检测index路径,下拉式选择(dropdown)"),
+ label="Auto-detect Index Path (Dropdown)",
+ value="Logs\A.index",
choices=sorted(index_paths),
interactive=True,
)
f0method0 = gr.Radio(
- label=i18n(
- "选择音高提取算法,输入歌声可用pm提速,harvest低音好但巨慢无比,crepe效果好但吃GPU,rmvpe效果最好且微吃GPU"
- ),
- choices=(
- ["pm", "harvest", "crepe", "rmvpe"]
- if config.dml == False
- else ["pm", "harvest", "rmvpe"]
- ),
+ label="1.4. Click to Use the Pitch Extraction Algorithm ( The pm is fast but low quality, The harvest have a the bass upgrade but is so long time, The crepe have more the better Quality of Voice and Save down the GPU using, The RMVPE have the best quality of Voice and the best save down and slight the GPU using. )",
+ choices=["pm", "harvest", "crepe", "rmvpe"]
+ if config.dml == False
+ else ["pm", "harvest", "rmvpe"],
value="rmvpe",
interactive=True,
)
@@ -877,51 +793,43 @@ def change_f0_method(f0method8):
resample_sr0 = gr.Slider(
minimum=0,
maximum=48000,
- label=i18n("后处理重采样至最终采样率，0为不进行重采样"),
- value=0,
+ label="1.5. Input Resample Rate File.",
+ value=48000,
step=1,
interactive=True,
)
rms_mix_rate0 = gr.Slider(
minimum=0,
maximum=1,
- label=i18n(
- "输入源音量包络替换输出音量包络融合比例，越靠近1越使用输出包络"
- ),
- value=0.25,
+ label="1.6. Input for The Volume Envelope Mix Rate (Closer to 1 = Use output envelope)",
+ value=0.3,
interactive=True,
)
protect0 = gr.Slider(
minimum=0,
maximum=0.5,
- label=i18n(
- "保护清辅音和呼吸声，防止电音撕裂等artifact，拉满0.5不开启，调低加大保护力度但可能降低索引效果"
- ),
- value=0.33,
+ label="1.7. Input to the Protect Voiceless Consonants&Breath rate. ( For 0.5 = disabled this, lower = stronger protection but may reduce index accuracy. )",
+ value=0.40,
step=0.01,
interactive=True,
)
filter_radius0 = gr.Slider(
minimum=0,
maximum=7,
- label=i18n(
- ">=3则使用对harvest音高识别的结果使用中值滤波，数值为滤波半径，使用可以削弱哑音"
- ),
- value=3,
+ label="1.8. Input to The Median Filter Radius for Harvest Pitch ( more than 3 is enables, reduces breathiness )",
+ value=1,
step=1,
interactive=True,
)
index_rate1 = gr.Slider(
minimum=0,
maximum=1,
- label=i18n("检索特征占比"),
- value=0.75,
+ label="1.9. Input to the Retrieval Ratio Feature rate ( Index Rate )",
+ value=0,
interactive=True,
)
f0_file = gr.File(
- label=i18n(
- "F0曲线文件, 可选, 一行一个音高, 代替默认F0及升降调"
- ),
+ label="F0 Curve File (Optional, replaces default F0)",
visible=False,
)
@@ -929,21 +837,15 @@ def change_f0_method(f0method8):
fn=change_choices,
inputs=[],
outputs=[sid0, file_index2],
- api_name="infer_refresh",
+ api_name="Refresh The Inference.",
)
- # file_big_npy1 = gr.Textbox(
- # label=i18n("特征文件路径"),
- # value="E:\\codes\py39\\vits_vc_gpu_train\\logs\\mi-test-1key\\total_fea.npy",
- # interactive=True,
- # )
+
with gr.Group():
with gr.Column():
- but0 = gr.Button(i18n("转换"), variant="primary")
+ but0 = gr.Button("Click to Start Convert File.", variant="primary")
with gr.Row():
- vc_output1 = gr.Textbox(label=i18n("输出信息"))
- vc_output2 = gr.Audio(
- label=i18n("输出音频(右下角三个点,点了可以下载)")
- )
+ vc_output1 = gr.Textbox(label="The Output File's Info For Converting.")
+ vc_output2 = gr.Audio(label="The Output Audio's File is Complete, Listening This... ( Click to the ... For Downloading The Voice's File. )")
but0.click(
vc.vc_single,
@@ -955,7 +857,6 @@ def change_f0_method(f0method8):
f0method0,
file_index1,
file_index2,
- # file_big_npy1,
index_rate1,
filter_radius0,
resample_sr0,
@@ -963,47 +864,39 @@ def change_f0_method(f0method8):
protect0,
],
[vc_output1, vc_output2],
- api_name="infer_convert",
+ api_name="The Multiple Infer Convert...",
)
- with gr.TabItem(i18n("批量推理")):
+ with gr.TabItem("2. Inference Voice For Multiple File in One Ordering."):
gr.Markdown(
- value=i18n(
- "批量转换, 输入待转换音频文件夹, 或上传多个音频文件, 在指定文件夹(默认opt)下输出转换的音频. "
- )
+ value="Use for Inference Voice For the Multiple File in One Ordering. Please Input to the Voice of Input File's Address For Importing All File And Input to the Voice of Output File's Address For Export All File To The Correct Address Before Click All The Process First!!!."
)
with gr.Row():
with gr.Column():
vc_transform1 = gr.Number(
- label=i18n("变调(整数, 半音数量, 升八度12降八度-12)"),
- value=0,
- )
- opt_input = gr.Textbox(
- label=i18n("指定输出文件夹"), value="opt"
+ label="2.1. Input the Pitch Shift (Integer, Semitones, +12 Octave Up, -12 Octave Down)", value=0
)
+ opt_input = gr.Textbox(label="2.2. Input the Output Folder", value="H:\F. Render file to UTAU\Singer_NAME\Language...")
file_index3 = gr.Textbox(
- label=i18n("特征检索库文件路径,为空则使用下拉的选择结果"),
- value="",
+ label="2.3. Input the Index Path (Leave blank to use dropdown)",
+ value="Logs\A.index",
interactive=True,
)
file_index4 = gr.Dropdown(
- label=i18n("自动检测index路径,下拉式选择(dropdown)"),
+ label="2.4. Input the Auto-detect Index Path (Dropdown)",
choices=sorted(index_paths),
+ value="Logs\A.index",
interactive=True,
)
f0method1 = gr.Radio(
- label=i18n(
- "选择音高提取算法,输入歌声可用pm提速,harvest低音好但巨慢无比,crepe效果好但吃GPU,rmvpe效果最好且微吃GPU"
- ),
- choices=(
- ["pm", "harvest", "crepe", "rmvpe"]
- if config.dml == False
- else ["pm", "harvest", "rmvpe"]
- ),
+ label="2.5. Click to Use the Pitch Extraction Algorithm ( The pm is fast but low quality, The harvest have a the bass upgrade but is so long time, The crepe have more the better Quality of Voice and Save down the GPU using, The RMVPE have the best quality of Voice and the best save down and slight the GPU using. )",
+ choices=["pm", "harvest", "crepe", "rmvpe"]
+ if config.dml == False
+ else ["pm", "harvest", "rmvpe"],
value="rmvpe",
interactive=True,
)
format1 = gr.Radio(
- label=i18n("导出文件格式"),
+ label="2.6. Input to Export Formatting.",
choices=["wav", "flac", "mp3", "m4a"],
value="wav",
interactive=True,
@@ -1013,75 +906,60 @@ def change_f0_method(f0method8):
fn=lambda: change_choices()[1],
inputs=[],
outputs=file_index4,
- api_name="infer_refresh_batch",
+ api_name="Click To Refresh",
)
- # file_big_npy2 = gr.Textbox(
- # label=i18n("特征文件路径"),
- # value="E:\\codes\\py39\\vits_vc_gpu_train\\logs\\mi-test-1key\\total_fea.npy",
- # interactive=True,
- # )
with gr.Column():
resample_sr1 = gr.Slider(
minimum=0,
maximum=48000,
- label=i18n("后处理重采样至最终采样率，0为不进行重采样"),
- value=0,
+ label="2.7. Input the Sample Rate For Export File. ( 0 Hz for none )",
+ value=48000,
step=1,
interactive=True,
)
rms_mix_rate1 = gr.Slider(
minimum=0,
maximum=1,
- label=i18n(
- "输入源音量包络替换输出音量包络融合比例，越靠近1越使用输出包络"
- ),
- value=1,
+ label="2.8. Input the Volume Envelope Mix Rate. (Closer to 1 = Use output envelope)",
+ value=0.3,
interactive=True,
)
protect1 = gr.Slider(
minimum=0,
maximum=0.5,
- label=i18n(
- "保护清辅音和呼吸声，防止电音撕裂等artifact，拉满0.5不开启，调低加大保护力度但可能降低索引效果"
- ),
- value=0.33,
+ label="2.9. Input the Protect Voiceless Consonants/Breath Rate. (0.5 = disabled, lower = stronger protection but may reduce index accuracy)",
+ value=0.4,
step=0.01,
interactive=True,
)
filter_radius1 = gr.Slider(
minimum=0,
maximum=7,
- label=i18n(
- ">=3则使用对harvest音高识别的结果使用中值滤波，数值为滤波半径，使用可以削弱哑音"
- ),
- value=3,
+ label="2.10. Input the Median Filter Radius for Harvest Pitch Rate (>=3 enables, reduces breathiness)",
+ value=1,
step=1,
interactive=True,
)
index_rate2 = gr.Slider(
minimum=0,
maximum=1,
- label=i18n("检索特征占比"),
- value=1,
+ label="2.11. Input the Retrieval Ratio Feature Rate. (Index Rate)",
+ value=0,
interactive=True,
)
with gr.Row():
- dir_input = gr.Textbox(
- label=i18n(
- "输入待处理音频文件夹路径(去文件管理器地址栏拷就行了)"
- ),
- placeholder="C:\\Users\\Desktop\\input_vocal_dir",
- )
- inputs = gr.File(
- file_count="multiple",
- label=i18n("也可批量输入音频文件, 二选一, 优先读文件夹"),
- )
+ dir_input = gr.Textbox(
+ label="2.12. Input the path of the Import audio folder to be the Processed ( Please Ctrl + C to copy it from the Flie address bar of the file manager And Paste in here.):",
+ value="H:\\Export to UTAU\\EN",
+ )
+ inputs = gr.File(
+ file_count="2.13. Input the Multiple File For Importing", label="Upload Audio Files (Optional, folder path takes priority)"
+ )
with gr.Row():
- but1 = gr.Button(i18n("转换"), variant="primary")
- vc_output3 = gr.Textbox(label=i18n("输出信息"))
-
+ but1 = gr.Button("Click to Starting to Convert", variant="primary")
+ vc_output3 = gr.Textbox(label="About All The Output's File...")
but1.click(
vc.vc_multi,
[
@@ -1093,7 +971,6 @@ def change_f0_method(f0method8):
f0method1,
file_index3,
file_index4,
- # file_big_npy2,
index_rate2,
filter_radius1,
resample_sr1,
@@ -1102,96 +979,102 @@ def change_f0_method(f0method8):
format1,
],
[vc_output3],
- api_name="infer_convert_batch",
+ api_name="The Infer Convert Batch...",
)
+ import threading
+
+# สร้าง Lock ไว้ด้านบนสุดของ Class หรือไฟล์
+processing_lock = threading.Lock()
+
+def vc_multi(...):
+ if processing_lock.locked():
+ print("งานเก่ากำลังทำอยู่ รอคิวนะจ๊ะ")
+ return "Already processing, please wait..."
+
+ with processing_lock:
+ # Code การแปลงเสียงเดิมๆ ของคุณอยู่ตรงนี้
+ ...
sid0.change(
fn=vc.get_vc,
inputs=[sid0, protect0, protect1],
outputs=[spk_item, protect0, protect1, file_index2, file_index4],
- api_name="infer_change_voice",
+ api_name="The Inference Voice To Change...",
)
- with gr.TabItem(i18n("伴奏人声分离&去混响&去回声")):
+ with gr.TabItem("2. All Separation, Reverb/Echo, Instrument And Vocal Remover."):
with gr.Group():
gr.Markdown(
- value=i18n(
- "人声伴奏分离批量处理， 使用UVR5模型。
合格的文件夹路径格式举例： E:\\codes\\py39\\vits_vc_gpu\\白鹭霜华测试样例(去文件管理器地址栏拷就行了)。
模型分为三类：
1、保留人声：不带和声的音频选这个，对主人声保留比HP5更好。内置HP2和HP3两个模型，HP3可能轻微漏伴奏但对主人声保留比HP2稍微好一丁点；
2、仅保留主人声：带和声的音频选这个，对主人声可能有削弱。内置HP5一个模型；
3、去混响、去延迟模型（by FoxJoy）：
(1)MDX-Net(onnx_dereverb):对于双通道混响是最好的选择，不能去除单通道混响；
(234)DeEcho:去除延迟效果。Aggressive比Normal去除得更彻底，DeReverb额外去除混响，可去除单声道混响，但是对高频重的板式混响去不干净。
去混响/去延迟，附：
1、DeEcho-DeReverb模型的耗时是另外2个DeEcho模型的接近2倍；
2、MDX-Net-Dereverb模型挺慢的；
3、个人推荐的最干净的配置是先MDX-Net再DeEcho-Aggressive。"
- )
+ value="Batch vocal/accompaniment separation using UVR5.
1. Keep Vocals: Use HP2/HP3.
2. Main Vocal Only: Use HP5.
3. DeReverb/DeEcho: MDX-Net for dual-channel reverb, DeEcho for delay."
)
with gr.Row():
with gr.Column():
dir_wav_input = gr.Textbox(
- label=i18n("输入待处理音频文件夹路径"),
+ label="Input Audio Folder Path",
placeholder="C:\\Users\\Desktop\\todo-songs",
)
wav_inputs = gr.File(
- file_count="multiple",
- label=i18n("也可批量输入音频文件, 二选一, 优先读文件夹"),
+ file_count="multiple", label="Upload Audio Files (Optional, folder path takes priority)"
)
with gr.Column():
- model_choose = gr.Dropdown(
- label=i18n("模型"), choices=uvr5_names
- )
+ model_choose = gr.Dropdown(label="Model", choices=uvr5_names)
agg = gr.Slider(
minimum=0,
maximum=20,
step=1,
- label="人声提取激进程度",
+ label="Vocal Extraction Aggressiveness",
value=10,
interactive=True,
- visible=False, # 先不开放调整
+ visible=False,
)
opt_vocal_root = gr.Textbox(
- label=i18n("指定输出主人声文件夹"), value="opt"
+ label="Output Folder for Main Vocals", value="opt"
)
opt_ins_root = gr.Textbox(
- label=i18n("指定输出非主人声文件夹"), value="opt"
+ label="Output Folder for Instrumentals/Others", value="opt"
)
format0 = gr.Radio(
- label=i18n("导出文件格式"),
+ label="Export Format",
choices=["wav", "flac", "mp3", "m4a"],
value="flac",
interactive=True,
)
- but2 = gr.Button(i18n("转换"), variant="primary")
- vc_output4 = gr.Textbox(label=i18n("输出信息"))
- but2.click(
- uvr,
- [
- model_choose,
- dir_wav_input,
- opt_vocal_root,
- wav_inputs,
- opt_ins_root,
- agg,
- format0,
- ],
- [vc_output4],
- api_name="uvr_convert",
- )
- with gr.TabItem(i18n("训练")):
- gr.Markdown(
- value=i18n(
- "step1: 填写实验配置. 实验数据放在logs下, 每个实验一个文件夹, 需手工输入实验名路径, 内含实验配置, 日志, 训练得到的模型文件. "
+ but2 = gr.Button("Convert", variant="primary")
+ vc_output4 = gr.Textbox(label="Output Info")
+ but2.click(
+ uvr,
+ [
+ model_choose,
+ dir_wav_input,
+ opt_vocal_root,
+ wav_inputs,
+ opt_ins_root,
+ agg,
+ format0,
+ ],
+ [vc_output4],
+ api_name="uvr_convert",
)
+ with gr.TabItem("Training"):
+ gr.Markdown(
+ value="Step No. 1 : Input For About Model Information And the voice's Rate to Start the Process.."
)
with gr.Row():
- exp_dir1 = gr.Textbox(label=i18n("输入实验名"), value="mi-test")
+ exp_dir1 = gr.Textbox(label="1.1. Input Model Name For Making..", value="NameAI_1")
sr2 = gr.Radio(
- label=i18n("目标采样率"),
+ label="Choose For The Sample Rate using...",
choices=["40k", "48k"],
value="40k",
interactive=True,
)
if_f0_3 = gr.Radio(
- label=i18n("模型是否带音高指导(唱歌一定要, 语音可以不要)"),
+ label="1.2. Input To Confirm to Use The Pitch Guidance ( Click True for Singing or Click False for Speech )",
choices=[True, False],
value=True,
interactive=True,
)
version19 = gr.Radio(
- label=i18n("版本"),
+ label="1.3. Input The Model Version",
choices=["v1", "v2"],
- value="v2",
+ value="v1",
interactive=True,
visible=True,
)
@@ -1199,151 +1082,134 @@ def change_f0_method(f0method8):
minimum=0,
maximum=config.n_cpu,
step=1,
- label=i18n("提取音高和处理数据使用的CPU进程数"),
+ label="1.4. Input to how much for using the CPU's Threads for This Model Making..",
value=int(np.ceil(config.n_cpu / 1.5)),
interactive=True,
)
- with gr.Group(): # 暂时单人的, 后面支持最多4人的#数据处理
+ with gr.Group():
gr.Markdown(
- value=i18n(
- "step2a: 自动遍历训练文件夹下所有可解码成音频的文件并进行切片归一化, 在实验目录下生成2个wav文件夹; 暂时只支持单人训练. "
- )
+ value="Step No. 2 : Input The Voice's File Address to Checking The Voice's File Using... ( The Spece Using For the Data File is no more other's file without .wav File And No More than 16 Bit and 40,000-48,000 HZ rate for This file. )"
)
with gr.Row():
trainset_dir4 = gr.Textbox(
- label=i18n("输入训练文件夹路径"),
- value=i18n("E:\\语音音频+标注\\米津玄师\\src"),
+ label="2.1. Input To the Data File's Address For Start The Project...", value="H:\\Data\\Singer_Name"
)
spk_id5 = gr.Slider(
minimum=0,
maximum=4,
step=1,
- label=i18n("请指定说话人id"),
- value=0,
+ label="2.2. Input the Speaker's ID. ( No. 1-4 ).",
+ value=1,
interactive=True,
)
- but1 = gr.Button(i18n("处理数据"), variant="primary")
- info1 = gr.Textbox(label=i18n("输出信息"), value="")
- but1.click(
- preprocess_dataset,
- [trainset_dir4, exp_dir1, sr2, np7],
- [info1],
- api_name="train_preprocess",
- )
- with gr.Group():
- gr.Markdown(
- value=i18n(
- "step2b: 使用CPU提取音高(如果模型带音高), 使用GPU提取特征(选择卡号)"
- )
+ but1 = gr.Button("Start the check for the Data File", variant="primary")
+ info1 = gr.Textbox(label="Input the Output Info", value="")
+ but1.click(
+ preprocess_dataset,
+ [trainset_dir4, exp_dir1, sr2, np7],
+ [info1],
+ api_name="Status_for_the_Process....",
)
+ with gr.Group():
+ gr.Markdown(value="Step No. 3 : Input the Engine for Extracking Pitch & Features to This Model Makking.")
with gr.Row():
with gr.Column():
gpus6 = gr.Textbox(
- label=i18n(
- "以-分隔输入使用的卡号, 例如 0-1-2 使用卡0和卡1和卡2"
- ),
+ label="3.1. Input About the GPU ID's (dash-separated, e.g., 0-1-2)",
value=gpus,
interactive=True,
visible=F0GPUVisible,
)
gpu_info9 = gr.Textbox(
- label=i18n("显卡信息"), value=gpu_info, visible=F0GPUVisible
+ label="About GPU Info...", value=gpu_info, visible=F0GPUVisible
)
with gr.Column():
f0method8 = gr.Radio(
- label=i18n(
- "选择音高提取算法:输入歌声可用pm提速,高质量语音但CPU差可用dio提速,harvest质量更好但慢,rmvpe效果最好且微吃CPU/GPU"
- ),
+ label="3.2. Click to Choose the Pitch of All Features For Algorithm to This Model Makking.",
choices=["pm", "harvest", "dio", "rmvpe", "rmvpe_gpu"],
- value="rmvpe_gpu",
+ value="rmvpe",
interactive=True,
)
gpus_rmvpe = gr.Textbox(
- label=i18n(
- "rmvpe卡号配置：以-分隔输入使用的不同进程卡号,例如0-0-1使用在卡0上跑2个进程并在卡1上跑1个进程"
- ),
+ label="3.3. Input For RMVPE GPU Configuration's File...",
value="%s-%s" % (gpus, gpus),
interactive=True,
visible=F0GPUVisible,
)
- but2 = gr.Button(i18n("特征提取"), variant="primary")
- info2 = gr.Textbox(label=i18n("输出信息"), value="", max_lines=8)
- f0method8.change(
- fn=change_f0_method,
- inputs=[f0method8],
- outputs=[gpus_rmvpe],
- )
- but2.click(
- extract_f0_feature,
- [
- gpus6,
- np7,
- f0method8,
- if_f0_3,
- exp_dir1,
- version19,
- gpus_rmvpe,
- ],
- [info2],
- api_name="train_extract_f0_feature",
- )
+ but2 = gr.Button("Start For Extract the Features File...", variant="primary")
+ info2 = gr.Textbox(label="Output Info", value="", max_lines=8)
+ f0method8.change(
+ fn=change_f0_method,
+ inputs=[f0method8],
+ outputs=[gpus_rmvpe],
+ )
+ but2.click(
+ extract_f0_feature,
+ [
+ gpus6,
+ np7,
+ f0method8,
+ if_f0_3,
+ exp_dir1,
+ version19,
+ gpus_rmvpe,
+ ],
+ [info2],
+ api_name="Status_for_the_Process....",
+ )
with gr.Group():
- gr.Markdown(value=i18n("step3: 填写训练设置, 开始训练模型和索引"))
+ gr.Markdown(value="Step No. 4 : Settings a lot of the Train in Epoch's time & Start Training.")
with gr.Row():
save_epoch10 = gr.Slider(
minimum=1,
maximum=50,
step=1,
- label=i18n("保存频率save_every_epoch"),
- value=5,
+ label="4.1. Input The Checkpoint For Save The Model to Testing...( It's Save Model File For Every Checkpoint )",
+ value=40,
interactive=True,
)
total_epoch11 = gr.Slider(
minimum=2,
maximum=1000,
step=1,
- label=i18n("总训练轮数total_epoch"),
- value=20,
+ label="4.2. Input For Maximum Epoch For this Making The Model.",
+ value=200,
interactive=True,
)
batch_size12 = gr.Slider(
minimum=1,
maximum=40,
step=1,
- label=i18n("每张显卡的batch_size"),
+ label="4.3. Input for the Batch Size per GPU Using ( Start For 1 )...",
value=default_batch_size,
interactive=True,
)
if_save_latest13 = gr.Radio(
- label=i18n("是否仅保存最新的ckpt文件以节省硬盘空间"),
- choices=[i18n("是"), i18n("否")],
- value=i18n("否"),
+ label="4.4. Use To Save the latest .ckpt File ( To Save The HDD Storage. )",
+ choices=["Yes", "No"],
+ value="Yes",
interactive=True,
)
if_cache_gpu17 = gr.Radio(
- label=i18n(
- "是否缓存所有训练集至显存. 10min以下小数据可缓存以加速训练, 大数据缓存会炸显存也加不了多少速"
- ),
- choices=[i18n("是"), i18n("否")],
- value=i18n("否"),
+ label="4.5. Use to Save the Vram For The Forever Time to Model ( No More Than 10 min. for that data only )",
+ choices=["Yes", "No"],
+ value="Yes",
interactive=True,
)
if_save_every_weights18 = gr.Radio(
- label=i18n(
- "是否在每次保存时间点将最终小模型保存至weights文件夹"
- ),
- choices=[i18n("是"), i18n("否")],
- value=i18n("否"),
+ label="4.6. Use for Save The No. of Model Checkpoit Time...",
+ choices=["Yes", "No"],
+ value="Yes",
interactive=True,
)
with gr.Row():
pretrained_G14 = gr.Textbox(
- label=i18n("加载预训练底模G路径"),
+ label="4.7. Input File Address for Pre-trained Generator (G) Path..",
value="assets/pretrained_v2/f0G40k.pth",
interactive=True,
)
pretrained_D15 = gr.Textbox(
- label=i18n("加载预训练底模D路径"),
+ label="4.8. Input File Address for Pre-trained Discriminator (D) Path..",
value="assets/pretrained_v2/f0D40k.pth",
interactive=True,
)
@@ -1363,115 +1229,106 @@ def change_f0_method(f0method8):
[f0method8, gpus_rmvpe, pretrained_G14, pretrained_D15],
)
gpus16 = gr.Textbox(
- label=i18n(
- "以-分隔输入使用的卡号, 例如 0-1-2 使用卡0和卡1和卡2"
- ),
+ label="4.9. Input No. for GPU IDs ( dash-separated, e.g., 0-1-2 )",
value=gpus,
interactive=True,
)
- but3 = gr.Button(i18n("训练模型"), variant="primary")
- but4 = gr.Button(i18n("训练特征索引"), variant="primary")
- but5 = gr.Button(i18n("一键训练"), variant="primary")
- info3 = gr.Textbox(label=i18n("输出信息"), value="", max_lines=10)
- but3.click(
- click_train,
- [
- exp_dir1,
- sr2,
- if_f0_3,
- spk_id5,
- save_epoch10,
- total_epoch11,
- batch_size12,
- if_save_latest13,
- pretrained_G14,
- pretrained_D15,
- gpus16,
- if_cache_gpu17,
- if_save_every_weights18,
- version19,
- ],
- info3,
- api_name="train_start",
- )
- but4.click(train_index, [exp_dir1, version19], info3)
- but5.click(
- train1key,
- [
- exp_dir1,
- sr2,
- if_f0_3,
- trainset_dir4,
- spk_id5,
- np7,
- f0method8,
- save_epoch10,
- total_epoch11,
- batch_size12,
- if_save_latest13,
- pretrained_G14,
- pretrained_D15,
- gpus16,
- if_cache_gpu17,
- if_save_every_weights18,
- version19,
- gpus_rmvpe,
- ],
- info3,
- api_name="train_start_all",
- )
+ but3 = gr.Button("Click to Start the Training To Make This Model Now...", variant="primary")
+ but4 = gr.Button("Click to Start Build the Model Index's File...", variant="primary")
+ but5 = gr.Button("Click to Start All Building This Model...", variant="primary")
+ info3 = gr.Textbox(label="Input for Output Info No.", value="1", max_lines=10)
+ but3.click(
+ click_train,
+ [
+ exp_dir1,
+ sr2,
+ if_f0_3,
+ spk_id5,
+ save_epoch10,
+ total_epoch11,
+ batch_size12,
+ if_save_latest13,
+ pretrained_G14,
+ pretrained_D15,
+ gpus16,
+ if_cache_gpu17,
+ if_save_every_weights18,
+ version19,
+ ],
+ info3,
+ api_name="Status_for_the_Process....",
+ )
+ but4.click(train_index, [exp_dir1, version19], info3)
+ but5.click(
+ train1key,
+ [
+ exp_dir1,
+ sr2,
+ if_f0_3,
+ trainset_dir4,
+ spk_id5,
+ np7,
+ f0method8,
+ save_epoch10,
+ total_epoch11,
+ batch_size12,
+ if_save_latest13,
+ pretrained_G14,
+ pretrained_D15,
+ gpus16,
+ if_cache_gpu17,
+ if_save_every_weights18,
+ version19,
+ gpus_rmvpe,
+ ],
+ info3,
+ api_name="train_start_all",
+ )
- with gr.TabItem(i18n("ckpt处理")):
+ with gr.TabItem("Use Checkpoint Processing"):
with gr.Group():
- gr.Markdown(value=i18n("模型融合, 可用于测试音色融合"))
+ gr.Markdown(value="Model Fusion (Merge Models)")
with gr.Row():
- ckpt_a = gr.Textbox(
- label=i18n("A模型路径"), value="", interactive=True
- )
- ckpt_b = gr.Textbox(
- label=i18n("B模型路径"), value="", interactive=True
- )
+ ckpt_a = gr.Textbox(label="Model A Path", value="", interactive=True)
+ ckpt_b = gr.Textbox(label="Model B Path", value="", interactive=True)
alpha_a = gr.Slider(
minimum=0,
maximum=1,
- label=i18n("A模型权重"),
+ label="Model A Weight",
value=0.5,
interactive=True,
)
with gr.Row():
sr_ = gr.Radio(
- label=i18n("目标采样率"),
+ label="Target Sample Rate",
choices=["40k", "48k"],
value="40k",
interactive=True,
)
if_f0_ = gr.Radio(
- label=i18n("模型是否带音高指导"),
- choices=[i18n("是"), i18n("否")],
- value=i18n("是"),
+ label="Has Pitch Guidance?",
+ choices=["Yes", "No"],
+ value="Yes",
interactive=True,
)
info__ = gr.Textbox(
- label=i18n("要置入的模型信息"),
- value="",
- max_lines=8,
- interactive=True,
+ label="Model Info to Insert", value="", max_lines=8, interactive=True
)
name_to_save0 = gr.Textbox(
- label=i18n("保存的模型名不带后缀"),
- value="",
+ label="Save Name (no extension)",
+ value="NameAI...",
max_lines=1,
interactive=True,
)
version_2 = gr.Radio(
- label=i18n("模型版本型号"),
+ label="Model Version",
choices=["v1", "v2"],
value="v1",
interactive=True,
)
with gr.Row():
- but6 = gr.Button(i18n("融合"), variant="primary")
- info4 = gr.Textbox(label=i18n("输出信息"), value="", max_lines=8)
+ but6 = gr.Button("Merge", variant="primary")
+ info4 = gr.Textbox(label="Output Info", value="", max_lines=8)
but6.click(
merge,
[
@@ -1486,30 +1343,25 @@ def change_f0_method(f0method8):
],
info4,
api_name="ckpt_merge",
- ) # def merge(path1,path2,alpha1,sr,f0,info):
+ )
with gr.Group():
- gr.Markdown(
- value=i18n("修改模型信息(仅支持weights文件夹下提取的小模型文件)")
- )
+ gr.Markdown(value="Modify Model Info")
with gr.Row():
ckpt_path0 = gr.Textbox(
- label=i18n("模型路径"), value="", interactive=True
+ label="Model Path", value="", interactive=True
)
info_ = gr.Textbox(
- label=i18n("要改的模型信息"),
- value="",
- max_lines=8,
- interactive=True,
+ label="Info to Modify", value="", max_lines=8, interactive=True
)
name_to_save1 = gr.Textbox(
- label=i18n("保存的文件名, 默认空为和源文件同名"),
+ label="Save Filename (blank to overwrite)",
value="",
max_lines=8,
interactive=True,
)
with gr.Row():
- but7 = gr.Button(i18n("修改"), variant="primary")
- info5 = gr.Textbox(label=i18n("输出信息"), value="", max_lines=8)
+ but7 = gr.Button("Modify", variant="primary")
+ info5 = gr.Textbox(label="Output Info", value="", max_lines=8)
but7.click(
change_info,
[ckpt_path0, info_, name_to_save1],
@@ -1517,57 +1369,50 @@ def change_f0_method(f0method8):
api_name="ckpt_modify",
)
with gr.Group():
- gr.Markdown(
- value=i18n("查看模型信息(仅支持weights文件夹下提取的小模型文件)")
- )
+ gr.Markdown(value="View Model Info")
with gr.Row():
ckpt_path1 = gr.Textbox(
- label=i18n("模型路径"), value="", interactive=True
+ label="Model Path", value="", interactive=True
)
- but8 = gr.Button(i18n("查看"), variant="primary")
- info6 = gr.Textbox(label=i18n("输出信息"), value="", max_lines=8)
+ but8 = gr.Button("View", variant="primary")
+ info6 = gr.Textbox(label="Output Info", value="", max_lines=8)
but8.click(show_info, [ckpt_path1], info6, api_name="ckpt_show")
with gr.Group():
gr.Markdown(
- value=i18n(
- "模型提取(输入logs文件夹下大文件模型路径),适用于训一半不想训了模型没有自动提取保存小文件模型,或者想测试中间模型的情况"
- )
+ value="Extract Small Model from Logs"
)
with gr.Row():
ckpt_path2 = gr.Textbox(
- label=i18n("模型路径"),
+ label="Model Path",
value="E:\\codes\\py39\\logs\\mi-test_f0_48k\\G_23333.pth",
interactive=True,
)
save_name = gr.Textbox(
- label=i18n("保存名"), value="", interactive=True
+ label="Save Name", value="", interactive=True
)
sr__ = gr.Radio(
- label=i18n("目标采样率"),
+ label="Target Sample Rate",
choices=["32k", "40k", "48k"],
value="40k",
interactive=True,
)
if_f0__ = gr.Radio(
- label=i18n("模型是否带音高指导,1是0否"),
+ label="Pitch Guidance (1: Yes, 0: No)",
choices=["1", "0"],
value="1",
interactive=True,
)
version_1 = gr.Radio(
- label=i18n("模型版本型号"),
+ label="Model Version",
choices=["v1", "v2"],
value="v2",
interactive=True,
)
info___ = gr.Textbox(
- label=i18n("要置入的模型信息"),
- value="",
- max_lines=8,
- interactive=True,
+ label="Model Info to Insert", value="", max_lines=8, interactive=True
)
- but9 = gr.Button(i18n("提取"), variant="primary")
- info7 = gr.Textbox(label=i18n("输出信息"), value="", max_lines=8)
+ but9 = gr.Button("Extract", variant="primary")
+ info7 = gr.Textbox(label="Output Info", value="", max_lines=8)
ckpt_path2.change(
change_info_, [ckpt_path2], [sr__, if_f0__, version_1]
)
@@ -1578,35 +1423,44 @@ def change_f0_method(f0method8):
api_name="ckpt_extract",
)
- with gr.TabItem(i18n("Onnx导出")):
+ with gr.TabItem("ONNX Export"):
with gr.Row():
- ckpt_dir = gr.Textbox(
- label=i18n("RVC模型路径"), value="", interactive=True
- )
+ ckpt_dir = gr.Textbox(label="RVC Model Path", value="", interactive=True)
with gr.Row():
onnx_dir = gr.Textbox(
- label=i18n("Onnx输出路径"), value="", interactive=True
+ label="ONNX Output Path", value="", interactive=True
)
with gr.Row():
infoOnnx = gr.Label(label="info")
with gr.Row():
- butOnnx = gr.Button(i18n("导出Onnx模型"), variant="primary")
+ butOnnx = gr.Button("Export ONNX Model", variant="primary")
butOnnx.click(
export_onnx, [ckpt_dir, onnx_dir], infoOnnx, api_name="export_onnx"
)
- tab_faq = i18n("常见问题解答")
- with gr.TabItem(tab_faq):
+ with gr.TabItem("FAQ"):
try:
- if tab_faq == "常见问题解答":
- with open("docs/cn/faq.md", "r", encoding="utf8") as f:
- info = f.read()
- else:
- with open("docs/en/faq_en.md", "r", encoding="utf8") as f:
- info = f.read()
+ with open("docs/en/faq_en.md", "r", encoding="utf8") as f:
+ info = f.read()
gr.Markdown(value=info)
except:
- gr.Markdown(traceback.format_exc())
+ gr.Markdown("FAQ file not found. Please ensure docs/en/faq_en.md exists.")
+
+ # ==========================================
+ # ส่วนแก้ไขปัญหา Gradio Timeout (ReadTimeout)
+ # ==========================================
+ import gradio.networking
+
+ # สร้างฟังก์ชันดักจับ (Monkey Patch) เพื่อข้ามการเช็ค URL ที่ทำให้เกิด Timeout
+ original_url_ok = gradio.networking.url_ok
+ def custom_url_ok(url):
+ try:
+ return original_url_ok(url)
+ except Exception:
+ # ถ้าโหลดเกิน 3 วินาทีจน Error ให้บังคับ return True เพื่อให้รันหน้าเว็บต่อได้เลย ไม่ต้องแครช
+ return True
+
+ gradio.networking.url_ok = custom_url_ok
if config.iscolab:
app.queue(concurrency_count=511, max_size=1022).launch(share=True)
@@ -1614,6 +1468,12 @@ def change_f0_method(f0method8):
app.queue(concurrency_count=511, max_size=1022).launch(
server_name="0.0.0.0",
inbrowser=not config.noautoopen,
+
+ # --- วิธีเปลี่ยน Port ---
+ # หากรันแล้วยังมีปัญหาพอร์ตค้างหรือชนกัน ให้ลบเครื่องหมาย # หน้าบรรทัด server_port=9378, ออก
+ # และใส่ # หน้าบรรทัด server_port=config.listen_port, แทนครับ
+ # server_port=9378,
+
server_port=config.listen_port,
quiet=True,
)