22

app.py

@@ -25,12 +25,6 @@ HF_ENABLE = False
 LOCAL_DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
 
 
-# def gpu_decorator(fn):
-#     if IS_HF_SPACE and HF_ENABLE and spaces is not None:
-#         return spaces.GPU(fn)
-#     return fn
-
-
 def local_move2gpu(x):
     """Move models to GPU on local environment. No-op on HuggingFace Spaces (ZeroGPU handles it)."""
     if IS_HF_SPACE:
@@ -45,37 +39,27 @@ _model = None
 _separator = None
 
 
-@spaces.GPU
-def get_model():
-    """加载 YingMusicSinger 模型 / Load YingMusicSinger model."""
+def _load_model_impl():
+    """Internal: load YingMusicSinger (no GPU decorator, called inside GPU context)."""
     download_files(task="infer")
     global _model
     if _model is None:
         from src.YingMusicSinger.infer.YingMusicSinger import YingMusicSinger
-
         _model = YingMusicSinger.from_pretrained("ASLP-lab/YingMusic-Singer")
-    _model = local_move2gpu(_model)
-    _model.eval()
+        _model = local_move2gpu(_model)
+        _model.eval()
     return _model
 
 
-@spaces.GPU
-def get_separator():
-    """
-    加载 MelBandRoformer 分离模型 / Load MelBandRoformer separator.
-    Returns a Separator instance ready for inference.
-    """
+def _load_separator_impl():
+    """Internal: load MelBandRoformer separator (no GPU decorator, called inside GPU context)."""
     download_files(task="infer")
     global _separator
     if _separator is None:
-        from src.third_party.MusicSourceSeparationTraining.inference_api import (
-            Separator,
-        )
-
+        from src.third_party.MusicSourceSeparationTraining.inference_api import Separator
         _separator = Separator(
             config_path="ckpts/config_vocals_mel_band_roformer_kj.yaml",
             checkpoint_path="ckpts/MelBandRoformer.ckpt",
-            # device=device,
         )
     return _separator
 
@@ -83,19 +67,12 @@ def get_separator():
 # ---------------------------------------------------------------------------
 # Vocal separation utilities / 人声分离工具
 # ---------------------------------------------------------------------------
-@spaces.GPU
-def separate_vocals(
-    audio_path: str,
-    device: str = "cuda:0",
-) -> tuple:
+def _separate_vocals_impl(audio_path: str) -> tuple:
     """
-    使用 MelBandRoformer 将音频分离为人声和伴奏。
     Separate audio into vocals and accompaniment using MelBandRoformer.
-
-    Returns:
-        (vocals_path, accompaniment_path)
+    Must be called within an active GPU context.
     """
-    separator = get_separator()
+    separator = _load_separator_impl()
 
     wav, sr = torchaudio.load(audio_path)
     vocal_wav, inst_wav, out_sr = separator.separate(wav, sr)
@@ -121,24 +98,20 @@ def mix_vocal_and_accompaniment(
     vocal_wav, vocal_sr = torchaudio.load(vocal_path)
     accomp_wav, accomp_sr = torchaudio.load(accomp_path)
 
-    # 统一采样率至人声采样率 / Resample accompaniment to match vocal sample rate
     if accomp_sr != vocal_sr:
         accomp_wav = torchaudio.functional.resample(accomp_wav, accomp_sr, vocal_sr)
 
-    # 统一声道数 / Match channel count
     if vocal_wav.shape[0] != accomp_wav.shape[0]:
         if vocal_wav.shape[0] == 1:
             vocal_wav = vocal_wav.expand(accomp_wav.shape[0], -1)
         else:
             accomp_wav = accomp_wav.expand(vocal_wav.shape[0], -1)
 
-    # 对齐长度（以较短者为准）/ Align to shorter length
     min_len = min(vocal_wav.shape[1], accomp_wav.shape[1])
     vocal_wav = vocal_wav[:, :min_len]
     accomp_wav = accomp_wav[:, :min_len]
 
     mixed = vocal_wav * vocal_gain + accomp_wav
-    # 防止 clipping / Prevent clipping
     peak = mixed.abs().max()
     if peak > 1.0:
         mixed = mixed / peak
@@ -150,6 +123,8 @@ def mix_vocal_and_accompaniment(
 
 # ---------------------------------------------------------------------------
 # Inference wrapper / 推理入口
+# Single @spaces.GPU scope covers ALL heavy work (separation + synthesis)
+# so models stay resident in GPU memory across steps within one call.
 # ---------------------------------------------------------------------------
 @spaces.GPU
 def synthesize(
@@ -169,15 +144,12 @@ def synthesize(
     主合成流程 / Main synthesis pipeline.
 
     1. (可选) 用 MelBandRoformer 分离参考音频和旋律音频的人声与伴奏
-       (Optional) Separate vocals & accompaniment from both ref and melody audio via MelBandRoformer
     2. 送入 YingMusicSinger 合成
-       Run YingMusicSinger inference
     3. (可选) 将合成人声与旋律音频的伴奏混合
-       (Optional) Mix synthesised vocals with melody accompaniment
     """
     import random
 
-    # ---- 输入校验 / Input validation -----------------------------------------
+    # ---- 输入校验 / Input validation ----------------------------------------
     if ref_audio is None:
         raise gr.Error("请上传参考音频 / Please upload Reference Audio")
     if melody_audio is None:
@@ -192,31 +164,24 @@ def synthesize(
         melody_audio if isinstance(melody_audio, str) else melody_audio[0]
     )
 
-    # seed = -1 means random / seed 为 -1 时随机生成
     actual_seed = int(seed)
     if actual_seed < 0:
         actual_seed = random.randint(0, 2**31 - 1)
 
-    device = "cuda:0" if torch.cuda.is_available() else "cpu"
-
-    # ---- Step 1: 人声分离（参考音频 + 旋律音频）/ Vocal separation for both (optional) ----
+    # ---- Step 1: 人声分离（合并在同一 GPU 上下文中）/ Vocal separation (same GPU context) ----
     melody_accomp_path = None
     actual_ref_path = ref_audio_path
     actual_melody_path = melody_audio_path
 
     if separate_vocals_flag:
-        # 分离参考音频 / Separate reference audio
-        ref_vocals_path, _ = separate_vocals(ref_audio_path, device=device)
+        ref_vocals_path, _ = _separate_vocals_impl(ref_audio_path)
         actual_ref_path = ref_vocals_path
 
-        # 分离旋律音频 / Separate melody audio
-        melody_vocals_path, melody_accomp_path = separate_vocals(
-            melody_audio_path, device=device
-        )
+        melody_vocals_path, melody_accomp_path = _separate_vocals_impl(melody_audio_path)
         actual_melody_path = melody_vocals_path
 
-    # ---- Step 2: 模型推理 / Model inference ----------------------------------
-    model = get_model()
+    # ---- Step 2: 模型推理 / Model inference (same GPU context) ---------------
+    model = _load_model_impl()
 
     audio_tensor, sr = model(
         ref_audio_path=actual_ref_path,
@@ -231,7 +196,6 @@ def synthesize(
         seed=actual_seed,
     )
 
-    # 先保存纯人声合成结果 / Save raw vocal synthesis result
     vocal_out_path = os.path.join(tempfile.mkdtemp(), "vocal_output.wav")
     torchaudio.save(vocal_out_path, audio_tensor.to("cpu"), sample_rate=sr)
 
@@ -247,6 +211,45 @@ def synthesize(
         return vocal_out_path
 
 
+# ---------------------------------------------------------------------------
+# Example presets / 预设示例
+# ---------------------------------------------------------------------------
+EXAMPLES_MELODY_CONTROL = [
+    # [ref_audio, melody_audio, ref_text, target_text, sep, mix, sil, t_shift, nfe, cfg, seed]
+    [
+        "examples/melody_control/ref_01.wav",
+        "examples/melody_control/melody_01.wav",
+        "该体谅的不执着|如果那天我",
+        "好多天|看不完你",
+        True, False, 0.5, 0.5, 32, 3.0, -1,
+    ],
+    [
+        "examples/melody_control/ref_02.wav",
+        "examples/melody_control/melody_02.wav",
+        "月光下的身影|渐渐模糊",
+        "星光照亮前路|指引方向",
+        True, False, 0.5, 0.5, 32, 3.0, -1,
+    ],
+]
+
+EXAMPLES_LYRIC_EDIT = [
+    [
+        "examples/lyric_edit/ref_01.wav",
+        "examples/lyric_edit/melody_01.wav",
+        "该体谅的不执着|如果那天我",
+        "忘不掉的笑容|留在心里面",
+        True, False, 0.5, 0.5, 32, 3.0, -1,
+    ],
+    [
+        "examples/lyric_edit/ref_02.wav",
+        "examples/lyric_edit/melody_02.wav",
+        "夜深了还不睡|想着你的脸",
+        "春风又吹过来|带走我思念",
+        True, False, 0.5, 0.5, 32, 3.0, -1,
+    ],
+]
+
+
 # ---------------------------------------------------------------------------
 # Custom CSS / 自定义样式
 # ---------------------------------------------------------------------------
@@ -274,23 +277,50 @@ CUSTOM_CSS = """
 /* ---------- Header / 头部 ---------- */
 #app-header {
     text-align: center;
-    padding: 2.5rem 1rem 1.5rem;
+    padding: 1.8rem 1rem 0.5rem;
 }
 #app-header h1 {
-    font-family: 'Playfair Display', serif !important;
-    font-size: 2.6rem !important;
-    font-weight: 800 !important;
-    background: linear-gradient(135deg, #f48c06, #e85d04, #dc2f02);
-    -webkit-background-clip: text;
-    -webkit-text-fill-color: transparent;
-    margin-bottom: 0.3rem !important;
-    letter-spacing: -0.02em;
+    font-size: 1.45rem !important;
+    font-weight: 700 !important;
+    line-height: 1.4;
+    margin-bottom: 0.6rem !important;
+}
+#app-header .badges img {
+    display: inline-block;
+    margin: 3px 2px;
+    vertical-align: middle;
 }
-#app-header p {
+#app-header .authors {
     color: var(--text-muted);
-    font-size: 1.05rem;
-    margin-top: 0;
+    font-size: 0.92rem;
+    margin: 0.5rem 0 0.2rem;
+    line-height: 1.7;
 }
+#app-header .affiliations {
+    color: var(--text-muted);
+    font-size: 0.85rem;
+    margin-bottom: 0.5rem;
+}
+#app-header .lang-links a {
+    color: var(--primary-light);
+    text-decoration: none;
+    margin: 0 4px;
+    font-size: 0.9rem;
+}
+#app-header .lang-links a:hover { text-decoration: underline; }
+
+/* ---------- Disclaimer ---------- */
+#disclaimer {
+    border-left: 4px solid var(--primary);
+    background: var(--accent-glow);
+    border-radius: 8px;
+    padding: 12px 16px;
+    margin: 8px 0 16px;
+    font-size: 0.88rem;
+    color: var(--text-muted);
+    line-height: 1.6;
+}
+#disclaimer strong { color: var(--primary-light); }
 
 /* ---------- Section labels / 分区标题 ---------- */
 .section-title {
@@ -305,6 +335,12 @@ CUSTOM_CSS = """
     margin-bottom: 12px !important;
 }
 
+/* ---------- Example tabs ---------- */
+.example-tab-label {
+    font-weight: 600 !important;
+    font-size: 0.95rem !important;
+}
+
 /* ---------- Run button / 合成按钮 ---------- */
 #run-btn {
     background: linear-gradient(135deg, #e85d04, #dc2f02) !important;
@@ -331,23 +367,93 @@ CUSTOM_CSS = """
 }
 """
 
+# ---------------------------------------------------------------------------
+# Header HTML / 头部 HTML
+# ---------------------------------------------------------------------------
+HEADER_HTML = """
+<div id="app-header" align="center">
+  <h1>
+    🏆 LyricEditBench: The first benchmark for melody-preserving lyric modification evaluation<br>
+    🚀 <a href="https://github.com/ASLP-lab/YingMusic-Singer" target="_blank">YingMusic-Singer</a>
+  </h1>
+
+  <div class="lang-links">
+    <a href="">English</a> ｜ <a href="README_ZH.md">中文</a>
+  </div>
+
+  <div class="badges" style="margin: 10px 0;">
+    <img src="https://img.shields.io/badge/Python-3.10-3776AB?logo=python&logoColor=white" alt="Python">
+    <img src="https://img.shields.io/badge/License-CC%20BY--NC--SA%204.0-lightgrey" alt="License">
+    <a href="https://arxiv.org/abs/0.0" target="_blank">
+      <img src="https://img.shields.io/badge/arXiv-0.0-b31b1b?logo=arxiv&logoColor=white" alt="arXiv">
+    </a>
+    <a href="https://github.com/ASLP-lab/YingMusic-Singer" target="_blank">
+      <img src="https://img.shields.io/badge/GitHub-YingMusic--Singer-181717?logo=github&logoColor=white" alt="GitHub">
+    </a>
+    <a href="https://huggingface.co/spaces/ASLP-lab/YingMusic-Singer" target="_blank">
+      <img src="https://img.shields.io/badge/🤗%20HuggingFace-Space-FFD21E" alt="HuggingFace Space">
+    </a>
+    <a href="https://huggingface.co/ASLP-lab/YingMusic-Singer" target="_blank">
+      <img src="https://img.shields.io/badge/🤗%20HuggingFace-Model-FF9D00" alt="HuggingFace Model">
+    </a>
+    <a href="https://huggingface.co/datasets/ASLP-lab/LyricEditBench" target="_blank">
+      <img src="https://img.shields.io/badge/🤗%20HuggingFace-LyricEditBench-FF6F00" alt="LyricEditBench">
+    </a>
+    <a href="https://discord.gg/RXghgWyvrn" target="_blank">
+      <img src="https://img.shields.io/badge/Discord-Join%20Us-5865F2?logo=discord&logoColor=white" alt="Discord">
+    </a>
+    <a href="https://github.com/ASLP-lab/YingMusic-Singer/blob/main/assets/wechat_qr.png" target="_blank">
+      <img src="https://img.shields.io/badge/WeChat-Group-07C160?logo=wechat&logoColor=white" alt="WeChat">
+    </a>
+    <a href="http://www.npu-aslp.org/" target="_blank">
+      <img src="https://img.shields.io/badge/🏫%20ASLP-Lab-4A90D9" alt="ASLP Lab">
+    </a>
+  </div>
+
+  <p class="authors">
+    <a href="https://orcid.org/0009-0005-5957-8936" target="_blank"><b>Chunbo Hao</b></a>¹² &nbsp;·&nbsp;
+    <a href="https://orcid.org/0009-0003-2602-2910" target="_blank"><b>Junjie Zheng</b></a>² &nbsp;·&nbsp;
+    <a href="https://orcid.org/0009-0001-6706-0572" target="_blank"><b>Guobin Ma</b></a>¹ &nbsp;·&nbsp;
+    <b>Yuepeng Jiang</b>¹ &nbsp;·&nbsp;
+    <b>Huakang Chen</b>¹ &nbsp;·&nbsp;
+    <b>Wenjie Tian</b>¹ &nbsp;·&nbsp;
+    <a href="https://orcid.org/0009-0003-9258-4006" target="_blank"><b>Gongyu Chen</b></a>² &nbsp;·&nbsp;
+    <a href="https://orcid.org/0009-0005-5413-6725" target="_blank"><b>Zihao Chen</b></a>² &nbsp;·&nbsp;
+    <b>Lei Xie</b>¹
+  </p>
+  <p class="affiliations">
+    <sup>1</sup> Northwestern Polytechnical University &nbsp;·&nbsp; <sup>2</sup> Giant Network
+  </p>
+</div>
+"""
+
+DISCLAIMER_HTML = """
+<div id="disclaimer">
+  <strong>⚠️ Disclaimer / 免责声明：</strong>
+  YingMusic-Singer enables the creation of singing voices with modified lyrics, supporting applications
+  in artistic creation and entertainment. Potential risks include unauthorized voice cloning and copyright
+  infringement. To ensure responsible deployment, users should obtain consent for voice usage, disclose
+  AI involvement, and verify musical originality.
+</div>
+"""
+
 
 # ---------------------------------------------------------------------------
 # Build the Gradio UI / 构建界面
 # ---------------------------------------------------------------------------
 def build_ui():
+    # Shared input components referenced by both Examples tabs and main form
+    ALL_INPUTS_ORDER = None  # defined after components are created
+
     with gr.Blocks(
         css=CUSTOM_CSS, title="YingMusic Singer", theme=gr.themes.Base()
     ) as demo:
+
         # ---- Header / 头部 ----
-        gr.HTML(
-            """
-            <div id="app-header">
-                <h1>♫ YingMusic Singer</h1>
-                <p>基于参考音色与旋律音频的歌声合成系统 &nbsp;·&nbsp; Singing Voice Synthesis</p>
-            </div>
-            """
-        )
+        gr.HTML(HEADER_HTML)
+        gr.HTML(DISCLAIMER_HTML)
+
+        gr.HTML("<hr style='border-color:#30363d; margin: 8px 0 18px;'>")
 
         # ================================================================
         # ROW 1 – 音频输入 / Audio Inputs + 歌词 / Lyrics (side by side)
@@ -459,42 +565,61 @@ def build_ui():
             elem_id="output-audio",
         )
 
-        # ---- 联动：未开启分离时，禁用伴奏混合 ----
-        # ---- Disable mix checkbox when separation is off ----
+        # ================================================================
+        # ROW 5 – 示例 / Examples (two tabs)
+        # ================================================================
+        gr.HTML("<hr style='border-color:#30363d; margin: 20px 0 12px;'>")
+        gr.Markdown("#### 🎵 预设示例 / Example Presets", elem_classes="section-title")
+        gr.Markdown(
+            "<small style='color:#8b949e;'>点击任意行自动填入输入区域 / Click any row to auto-fill the inputs above</small>"
+        )
+
+        # All inputs in order expected by synthesize()
+        _all_inputs = [
+            ref_audio, melody_audio, ref_text, target_text,
+            separate_vocals_flag, mix_accompaniment_flag,
+            sil_len_to_end, t_shift, nfe_step, cfg_strength, seed,
+        ]
+
+        with gr.Tabs():
+            with gr.Tab("🎼 Melody Control"):
+                gr.Examples(
+                    examples=EXAMPLES_MELODY_CONTROL,
+                    inputs=_all_inputs,
+                    label="Melody Control Examples",
+                    examples_per_page=5,
+                )
+            with gr.Tab("✏️ Lyric Edit"):
+                gr.Examples(
+                    examples=EXAMPLES_LYRIC_EDIT,
+                    inputs=_all_inputs,
+                    label="Lyric Edit Examples",
+                    examples_per_page=5,
+                )
+
+        # ================================================================
+        # Event wiring / 事件绑定
+        # ================================================================
+
+        # 联动：未开启分离时，禁用伴奏混合
         separate_vocals_flag.change(
-            fn=lambda sep: gr.update(
-                interactive=sep, value=False if not sep else False
-            ),
+            fn=lambda sep: gr.update(interactive=sep, value=False if not sep else False),
             inputs=[separate_vocals_flag],
             outputs=[mix_accompaniment_flag],
         )
 
-        # ---- 绑定事件 / Wire up ----
         run_btn.click(
             fn=synthesize,
-            inputs=[
-                ref_audio,
-                melody_audio,
-                ref_text,
-                target_text,
-                separate_vocals_flag,
-                mix_accompaniment_flag,
-                sil_len_to_end,
-                t_shift,
-                nfe_step,
-                cfg_strength,
-                seed,
-            ],
+            inputs=_all_inputs,
             outputs=output_audio,
         )
 
         # ---- 页脚 / Footer ----
         gr.Markdown(
             """
-            ---
-            <center style="color:#8b949e; font-size:0.85rem;">
-            YingMusic Singer &nbsp;·&nbsp; 基于 Flow Matching + VAE / Powered by Flow Matching + VAE &nbsp;·&nbsp;
-            用 <code>|</code> 分隔歌词中的乐句 / Use <code>|</code> to separate phrases in lyrics
+            <center style="color:#8b949e; font-size:0.85rem; margin-top: 16px;">
+            Use <code>|</code> to separate lyric phrases &nbsp;·&nbsp;
+            用 <code>|</code> 分隔歌词乐句
             </center>
             """,
         )
@@ -508,4 +633,4 @@ def build_ui():
 if __name__ == "__main__":
     demo = build_ui()
     demo.queue()
-    demo.launch(server_name="0.0.0.0", server_port=7860, share=False)
+    demo.launch(server_name="0.0.0.0", server_port=7860, share=False)