Duplicate from maxmax20160403/sovits5.0

Co-authored-by: maxmax <maxmax20160403@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,34 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,14 @@
+---
+title: Sovits5.0
+emoji: 🏃
+colorFrom: pink
+colorTo: blue
+sdk: gradio
+sdk_version: 3.28.1
+app_file: app.py
+pinned: false
+license: mit
+duplicated_from: maxmax20160403/sovits5.0
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,200 @@
+from vits.models import SynthesizerInfer
+from omegaconf import OmegaConf
+import torchcrepe
+import torch
+import io
+import os
+import gradio as gr
+import librosa
+import numpy as np
+import soundfile
+
+import logging
+
+logging.getLogger('numba').setLevel(logging.WARNING)
+logging.getLogger('markdown_it').setLevel(logging.WARNING)
+logging.getLogger('urllib3').setLevel(logging.WARNING)
+logging.getLogger('matplotlib').setLevel(logging.WARNING)
+
+
+def load_svc_model(checkpoint_path, model):
+    assert os.path.isfile(checkpoint_path)
+    checkpoint_dict = torch.load(checkpoint_path, map_location="cpu")
+    saved_state_dict = checkpoint_dict["model_g"]
+    state_dict = model.state_dict()
+    new_state_dict = {}
+    for k, v in state_dict.items():
+        new_state_dict[k] = saved_state_dict[k]
+    model.load_state_dict(new_state_dict)
+    return model
+
+
+def compute_f0_nn(filename, device):
+    audio, sr = librosa.load(filename, sr=16000)
+    assert sr == 16000
+    # Load audio
+    audio = torch.tensor(np.copy(audio))[None]
+    # Here we'll use a 20 millisecond hop length
+    hop_length = 320
+    # Provide a sensible frequency range for your domain (upper limit is 2006 Hz)
+    # This would be a reasonable range for speech
+    fmin = 50
+    fmax = 1000
+    # Select a model capacity--one of "tiny" or "full"
+    model = "full"
+    # Pick a batch size that doesn't cause memory errors on your gpu
+    batch_size = 512
+    # Compute pitch using first gpu
+    pitch, periodicity = torchcrepe.predict(
+        audio,
+        sr,
+        hop_length,
+        fmin,
+        fmax,
+        model,
+        batch_size=batch_size,
+        device=device,
+        return_periodicity=True,
+    )
+    pitch = np.repeat(pitch, 2, -1)  # 320 -> 160 * 2
+    periodicity = np.repeat(periodicity, 2, -1)  # 320 -> 160 * 2
+    # CREPE was not trained on silent audio. some error on silent need filter.
+    periodicity = torchcrepe.filter.median(periodicity, 9)
+    pitch = torchcrepe.filter.mean(pitch, 9)
+    pitch[periodicity < 0.1] = 0
+    pitch = pitch.squeeze(0)
+    return pitch
+
+
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+hp = OmegaConf.load("configs/base.yaml")
+model = SynthesizerInfer(
+    hp.data.filter_length // 2 + 1,
+    hp.data.segment_size // hp.data.hop_length,
+    hp)
+load_svc_model("vits_pretrain/sovits5.0-48k-debug.pth", model)
+model.eval()
+model.to(device)
+
+
+def svc_change(argswave, argsspk):
+
+    argsppg = "svc_tmp.ppg.npy"
+    os.system(f"python whisper/inference.py -w {argswave} -p {argsppg}")
+
+    spk = np.load(argsspk)
+    spk = torch.FloatTensor(spk)
+
+    ppg = np.load(argsppg)
+    ppg = np.repeat(ppg, 2, 0)  # 320 PPG -> 160 * 2
+    ppg = torch.FloatTensor(ppg)
+
+    pit = compute_f0_nn(argswave, device)
+    pit = torch.FloatTensor(pit)
+
+    len_pit = pit.size()[0]
+    len_ppg = ppg.size()[0]
+    len_min = min(len_pit, len_ppg)
+    pit = pit[:len_min]
+    ppg = ppg[:len_min, :]
+
+    with torch.no_grad():
+
+        spk = spk.unsqueeze(0).to(device)
+        source = pit.unsqueeze(0).to(device)
+        source = model.pitch2source(source)
+
+        hop_size = hp.data.hop_length
+        all_frame = len_min
+        hop_frame = 10
+        out_chunk = 2500  # 25 S
+        out_index = 0
+        out_audio = []
+        has_audio = False
+
+        while (out_index + out_chunk < all_frame):
+            has_audio = True
+            if (out_index == 0):  # start frame
+                cut_s = out_index
+                cut_s_48k = 0
+            else:
+                cut_s = out_index - hop_frame
+                cut_s_48k = hop_frame * hop_size
+
+            if (out_index + out_chunk + hop_frame > all_frame):  # end frame
+                cut_e = out_index + out_chunk
+                cut_e_48k = 0
+            else:
+                cut_e = out_index + out_chunk + hop_frame
+                cut_e_48k = -1 * hop_frame * hop_size
+
+            sub_ppg = ppg[cut_s:cut_e, :].unsqueeze(0).to(device)
+            sub_pit = pit[cut_s:cut_e].unsqueeze(0).to(device)
+            sub_len = torch.LongTensor([cut_e - cut_s]).to(device)
+            sub_har = source[:, :, cut_s *
+                             hop_size:cut_e * hop_size].to(device)
+            sub_out = model.inference(sub_ppg, sub_pit, spk, sub_len, sub_har)
+            sub_out = sub_out[0, 0].data.cpu().detach().numpy()
+
+            sub_out = sub_out[cut_s_48k:cut_e_48k]
+            out_audio.extend(sub_out)
+            out_index = out_index + out_chunk
+
+        if (out_index < all_frame):
+            if (has_audio):
+                cut_s = out_index - hop_frame
+                cut_s_48k = hop_frame * hop_size
+            else:
+                cut_s = 0
+                cut_s_48k = 0
+            sub_ppg = ppg[cut_s:, :].unsqueeze(0).to(device)
+            sub_pit = pit[cut_s:].unsqueeze(0).to(device)
+            sub_len = torch.LongTensor([all_frame - cut_s]).to(device)
+            sub_har = source[:, :, cut_s * hop_size:].to(device)
+            sub_out = model.inference(sub_ppg, sub_pit, spk, sub_len, sub_har)
+            sub_out = sub_out[0, 0].data.cpu().detach().numpy()
+
+            sub_out = sub_out[cut_s_48k:]
+            out_audio.extend(sub_out)
+        out_audio = np.asarray(out_audio)
+
+    return out_audio
+
+
+def svc_main(sid, input_audio):
+    if input_audio is None:
+        return "You need to upload an audio", None
+    sampling_rate, audio = input_audio
+    audio = (audio / np.iinfo(audio.dtype).max).astype(np.float32)
+    if len(audio.shape) > 1:
+        audio = librosa.to_mono(audio.transpose(1, 0))
+    if sampling_rate != 16000:
+        audio = librosa.resample(audio, orig_sr=sampling_rate, target_sr=16000)
+    if (len(audio) > 16000*100):
+        audio = audio[:16000*100]
+    wav_path = "temp.wav"
+    soundfile.write(wav_path, audio, 16000, format="wav")
+    out_audio = svc_change(wav_path, f"configs/singers/singer00{sid}.npy")
+    return "Success", (48000, out_audio)
+
+
+app = gr.Blocks()
+with app:
+    with gr.Tabs():
+        with gr.TabItem("sovits 5.0"):
+            gr.Markdown(value="""
+                基于开源数据:Multi-Singer
+                
+                https://github.com/Multi-Singer/Multi-Singer.github.io
+
+                [轻度伴奏可以无需去伴奏]就能直接进行歌声转换的SVC库
+                """)
+            sid = gr.Dropdown(label="音色", choices=[
+                              "22", "33", "47", "51"], value="47")
+            vc_input3 = gr.Audio(label="上传音频")
+            vc_submit = gr.Button("转换", variant="primary")
+            vc_output1 = gr.Textbox(label="状态信息")
+            vc_output2 = gr.Audio(label="转换音频")
+        vc_submit.click(svc_main, [sid, vc_input3], [vc_output1, vc_output2])
+
+    app.launch()

configs/base.yaml

@@ -0,0 +1,69 @@
+train:
+  model: "sovits"
+  seed: 1234
+  epochs: 10000
+  learning_rate: 2e-4
+  betas: [0.8, 0.99]
+  lr_decay: 0.999875
+  eps: 1e-9
+  batch_size: 8
+  c_stft: 5
+  c_mel: 2.5
+  c_kl: 1.0
+  port: 8001
+  pretrain: ""
+#############################
+data: 
+  training_files: "files/train.txt"
+  validation_files: "files/valid.txt"
+  segment_size: 12000  # WARNING: base on hop_length
+  max_wav_value: 32768.0
+  sampling_rate: 48000
+  filter_length: 2048
+  hop_length: 480
+  win_length: 2048
+  mel_channels: 80
+  mel_fmin: 0.0
+  mel_fmax: 24000.0
+#############################
+vits:
+  ppg_dim: 1024
+  spk_dim: 256
+  gin_channels: 256
+  inter_channels: 192
+  hidden_channels: 192
+  filter_channels: 512
+#############################
+gen:
+  upsample_input: 192
+  upsample_rates: [6,5,4,2,2]
+  upsample_kernel_sizes: [20,15,8,4,4]
+  upsample_initial_channel: 256
+  resblock_kernel_sizes: [3,7,11]
+  resblock_dilation_sizes: [[1,3,5], [1,3,5], [1,3,5]]
+#############################
+mpd:
+  periods: [2,3,5,7,11]
+  kernel_size: 5
+  stride: 3
+  use_spectral_norm: False
+  lReLU_slope: 0.2
+#############################
+mrd:
+  resolutions: "[(1024, 120, 600), (2048, 240, 1200), (512, 50, 240)]" # (filter_length, hop_length, win_length)
+  use_spectral_norm: False
+  lReLU_slope: 0.2
+#############################
+log:
+  info_interval: 100
+  eval_interval: 5
+  save_interval: 5
+  num_audio: 6
+  pth_dir: 'chkpt'
+  log_dir: 'logs'
+#############################
+dist_config:
+  dist_backend: "nccl"
+  dist_url: "tcp://localhost:54321"
+  world_size: 1
+

configs/singers/singer0001.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e2879921d43bdbf11fc5d6ac91f434f905a2c5e59d75368bfbf3c6bdbddcb3cf
+size 1152

configs/singers/singer0002.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fbe5c7925c2fdb514e2c5b450de1d2737ec7f86f1c65eeb488c1888c0b9a7069
+size 1152

configs/singers/singer0003.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5665126aeb6c6fab89c79b90debf2ce2e64b321076dcb414089eff8848eac8b4
+size 1152

configs/singers/singer0004.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:79f0fe5993e9adcaeae25b0fa68265d40c9c1b5539ca12d6e438477de2177819
+size 1152

configs/singers/singer0005.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1158fb447929cf9400a31675cf9992fd3ed7558e061562189d9e6bf56d83fb2a
+size 1152

configs/singers/singer0006.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06c1fd3a9afaa7944e4b81b7ca787e667b0dae8c7e90c6d24177245449f4e940
+size 1152

configs/singers/singer0007.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36611b9e57545332b9fb97fd35a356fbe8d60258f2f5e2232168481bb6dfab5b
+size 1152

configs/singers/singer0008.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8584ad6f3569a1307082cd410085d9a562807e962274b89b72487c7bc79124d4
+size 1152

configs/singers/singer0009.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b069db4e3e5ca389ffba974c74eab46caf4c60545773e5f7e5e253310619073e
+size 1152

configs/singers/singer0010.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7d4d92735e4bac1618e89198d113013db09061b6c1f74ba0c500b70b097cd407
+size 1152

configs/singers/singer0011.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:942388b4276dc06ee365f59c324ce1642e4bf810dcc99992739787e3b9ad135d
+size 1152

configs/singers/singer0012.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3411efcf4ee4f534cea2b742c2eca166ae971efbceab21fb41b77b8923a1ba3a
+size 1152

configs/singers/singer0013.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e8e30cd1bce61405db194278dd7bf207d16abf656dd22f9a20f29e3657674f3
+size 1152

configs/singers/singer0014.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9cc8200753b4ba7605c9a13bf454b100025965135c5d816f7440ec53a2e6dd4
+size 1152

configs/singers/singer0015.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dcb58688e51dbdeb22e5dd85d27ff3904c4594c78420b8e9c9ab481adbecc5fe
+size 1152

configs/singers/singer0016.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:66a3c6162b8c937e9e8bbdc806b873866afce4b110664831642f7b41922bbf39
+size 1152

configs/singers/singer0017.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84782c98c930bd980f350837f4b3e8e193c49ef46aef9f92471c6136659975a9
+size 1152

configs/singers/singer0018.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:731ebafda06aecedfd79941978149a0f87595f04e24eab7ed5300defe9070fc0
+size 1152

configs/singers/singer0019.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d88e620994e4413c4c58ffb9239ef46ded60ff3eab0715c7af96cbe4092198f
+size 1152

configs/singers/singer0020.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e5abaabe5457a20161351dcf5f8737d63a2a92fb1de1842ea9e92e47b9ca6fe
+size 1152

configs/singers/singer0021.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d7f99c92c89a44c1f2dd0688f033f0593c8c88b0537b092928bfbaa63a8d3e9
+size 1152

configs/singers/singer0022.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33becb1da48b12ba4957a0ef0b25bbd51e100d5762ebc4c7d381f6b957e682a2
+size 1152

configs/singers/singer0023.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f49cbaf3f7653f48f80854a513a334f31dca719a09cca66e257995ce4a741a9
+size 1152

configs/singers/singer0024.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:92ed584994d56473c8bab0799d213e927c5a2928facef2b93a2f95f764d868b4
+size 1152

configs/singers/singer0025.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14b7e1f55393d5beaa2f3bbd0ef7f2be7e108993c680acb265ff24df19f7062b
+size 1152

configs/singers/singer0026.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:92ecc9aa68f136960c00e98aaca16e92c38960bc7eb9687aee90190972974726
+size 1152

configs/singers/singer0027.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5a8a1c2a445179d38664fb55c84ee9a36350beee50efa9f850d29b394447bfa
+size 1152

configs/singers/singer0028.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b79b8266c8d368dc99f49a347b2631e1e5cfb44056b5a9ab4470b42f9851ee35
+size 1152

configs/singers/singer0029.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60fa5fd9e8ba14d7f6d67304842f16382f7d2e739969bde9551222ff8c282775
+size 1152

configs/singers/singer0030.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f5070e4196c91fa713aed20aedb2a570a7b2ad8301ee61f59821dafaea3c6a7
+size 1152

configs/singers/singer0031.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47f4f8c065be1c5448c1b80e5c99087e7357cf1f8a8a55f2d844ccf1ca4931e6
+size 1152

configs/singers/singer0032.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:019f40cf49cb7ccb44fb9c6a9f6345e84f837185a1642623144b4e2969c8738b
+size 1152

configs/singers/singer0033.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e05e212c93fc9e7b13174dd76721ee891bb4ea8bb1638a4c43523ed65d30f67
+size 1152

configs/singers/singer0034.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:715a089dd9b3e5cbf021b0f41055f59208911e49cccf375ecf8b82544f325c3d
+size 1152

configs/singers/singer0035.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9af8cd05182ec53ff573bce53dad049759bea1de5656915f414910eaf47f61ed
+size 1152

configs/singers/singer0036.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3cec474244d86acfd24d6abf7e033b24b40b838cba2fcd3b4d0e5611313d67ef
+size 1152

configs/singers/singer0037.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:316e3435d373e352fe95fcb2ec0ab1c8afdeb270ce9f13c940ba91187eecdcf3
+size 1152

configs/singers/singer0038.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e6458e251512dab86abce504490de6762f9c2de66ddbc853c24c3d05eb39c96
+size 1152

configs/singers/singer0039.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c2e484ae33eef7ac92dd784e9e3b9bca7e6c0838d50b43c674da47620f281f20
+size 1152

configs/singers/singer0040.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b3a104163ad4cf87caff70b845b2c3e70190ce430a8f21247d350ef102071dc
+size 1152

configs/singers/singer0041.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:962ba35045f952562bbf68239c8abfda4e1888118fae7ef19814282abee2d28e
+size 1152

configs/singers/singer0042.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0cf0871ba7e939c90f2f027862f80e11151d8b1a21b6624ee05f184d024b35a3
+size 1152

configs/singers/singer0043.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9494ed20a9b095d19cce17619a6372ba3371f980c643078ffda8649a30ac2f8b
+size 1152

configs/singers/singer0044.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c12949caa6176fbe5f323cf643d29eef14af9a3ee03be27c938d8bb6fc2922f1
+size 1152

configs/singers/singer0045.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:222adf210792d1b2745ef98b717f57e0d309d8176e9b59ff56063c1e2001728d
+size 1152

configs/singers/singer0046.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6070f66f028928114493a363e4117636afefb1a094c54ffc01f89ef261ad1882
+size 1152