Upload 4 files

README.md

@@ -4,7 +4,7 @@ emoji: ⚡
 colorFrom: gray
 colorTo: indigo
 sdk: gradio
-sdk_version: 5.34.2
+sdk_version: 5.43.1
 app_file: app.py
 license: mit
 pinned: true

app.py

@@ -6,7 +6,7 @@ import random
 import logging
 import time
 import soundfile as sf
-from infer_rvc_python.main import download_manager
+from infer_rvc_python.main import download_manager, load_hu_bert, Config
 import zipfile
 import edge_tts
 import asyncio
@@ -21,14 +21,53 @@ import numpy as np
 import urllib.request
 import shutil
 import threading
+import argparse
+import sys
+
+parser = argparse.ArgumentParser(description="Run the app with optional sharing")
+parser.add_argument(
+    '--share',
+    action='store_true',
+    help='Enable sharing mode'
+)
+parser.add_argument(
+    '--theme',
+    type=str,
+    default="aliabid94/new-theme",
+    help='Set the theme (default: aliabid94/new-theme)'
+)
+args = parser.parse_args()
+
+IS_COLAB = True if ('google.colab' in sys.modules or args.share) else False
+IS_ZERO_GPU = os.getenv("SPACES_ZERO_GPU")
 
 logging.getLogger("infer_rvc_python").setLevel(logging.ERROR)
 
 converter = BaseLoader(only_cpu=False, hubert_path=None, rmvpe_path=None)
+converter.hu_bert_model = load_hu_bert(Config(only_cpu=False), converter.hubert_path)
+
+test_model = "https://huggingface.co/sail-rvc/aAldeano_Minecraft__RVC_V2_-_500_Epochs_/resolve/main/model.pth?download=true, https://huggingface.co/sail-rvc/Aldeano_Minecraft__RVC_V2_-_500_Epochs_/resolve/main/model.index?download=true"
+test_names = ["model.pth", "model.index"]
+
+for url, filename in zip(test_model.split(", "), test_names):
+    try:
+        download_manager(
+            url=url,
+            path=".",
+            extension="",
+            overwrite=False,
+            progress=True,
+        )
+        if not os.path.isfile(filename):
+            raise FileNotFoundError
+    except Exception:
+        with open(filename, "wb") as f:
+            pass
 
 title = "<center><strong><font size='7'>RVC⚡ZERO</font></strong></center>"
-description = "This demo is provided for educational and research purposes only. The authors and contributors of this project do not endorse or encourage any misuse or unethical use of this software. Any use of this software for purposes other than those intended is solely at the user's own risk. The authors and contributors shall not be held responsible for any damages or liabilities arising from the use of this demo inappropriately."
-theme = "aliabid94/new-theme"
+description = "This demo is provided for educational and research purposes only. The authors and contributors of this project do not endorse or encourage any misuse or unethical use of this software. Any use of this software for purposes other than those intended is solely at the user's own risk. The authors and contributors shall not be held responsible for any damages or liabilities arising from the use of this demo inappropriately." if IS_ZERO_GPU else ""
+theme = args.theme
+delete_cache_time = (3200, 3200) if IS_ZERO_GPU else (86400, 86400)
 
 PITCH_ALGO_OPT = [
     "pm",
@@ -39,6 +78,26 @@ PITCH_ALGO_OPT = [
 ]
 
 
+async def get_voices_list(proxy=None):
+    """Print all available voices."""
+    from edge_tts import list_voices
+    voices = await list_voices(proxy=proxy)
+    voices = sorted(voices, key=lambda voice: voice["ShortName"])
+
+    table = [
+        {
+            "ShortName": voice["ShortName"],
+            "Gender": voice["Gender"],
+            "ContentCategories": ", ".join(voice["VoiceTag"]["ContentCategories"]),
+            "VoicePersonalities": ", ".join(voice["VoiceTag"]["VoicePersonalities"]),
+            "FriendlyName": voice["FriendlyName"],
+        }
+        for voice in voices
+    ]
+
+    return table
+
+
 def find_files(directory):
     file_paths = []
     for filename in os.listdir(directory):
@@ -106,19 +165,23 @@ def find_my_model(a_, b_):
     return model, index
 
 
-def get_file_size(url):
-
+def ensure_valid_file(url):
     if "huggingface" not in url:
         raise ValueError("Only downloads from Hugging Face are allowed")
 
     try:
-        with urllib.request.urlopen(url) as response:
-            info = response.info()
-            content_length = info.get("Content-Length")
+        request = urllib.request.Request(url, method="HEAD")
+        with urllib.request.urlopen(request) as response:
+            content_length = response.headers.get("Content-Length")
+
+        if content_length is None:
+            raise ValueError("No Content-Length header found")
 
         file_size = int(content_length)
-        if file_size > 500000000:
-            raise ValueError("The file is too large. You can only download files up to 500 MB in size.")
+        if file_size > 900000000 and IS_ZERO_GPU:
+            raise ValueError("The file is too large. Max allowed is 900 MB.")
+
+        return file_size
 
     except Exception as e:
         raise e
@@ -130,7 +193,7 @@ def clear_files(directory):
     shutil.rmtree(directory)
 
 
-def get_my_model(url_data):
+def get_my_model(url_data, progress=gr.Progress(track_tqdm=True)):
 
     if not url_data:
         return None, None
@@ -147,12 +210,9 @@ def get_my_model(url_data):
     os.makedirs(directory, exist_ok=True)
 
     try:
-        get_file_size(a_)
-        if b_:
-            get_file_size(b_)
-
         valid_url = [a_] if not b_ else [a_, b_]
         for link in valid_url:
+            ensure_valid_file(link)
             download_manager(
                 url=link,
                 path=directory,
@@ -194,13 +254,13 @@ def get_my_model(url_data):
         t.start()
 
 
-def add_audio_effects(audio_list):
+def add_audio_effects(audio_list, type_output):
     print("Audio effects")
 
     result = []
     for audio_path in audio_list:
         try:
-            output_path = f'{os.path.splitext(audio_path)[0]}_effects.wav'
+            output_path = f'{os.path.splitext(audio_path)[0]}_effects.{type_output}'
 
             # Initialize audio effects plugins
             board = Pedalboard(
@@ -211,13 +271,23 @@ def add_audio_effects(audio_list):
                  ]
             )
 
+            # Temporary WAV to hold processed data before exporting
+            temp_wav = f'{os.path.splitext(audio_path)[0]}_temp.wav'
+
             with AudioFile(audio_path) as f:
-                with AudioFile(output_path, 'w', f.samplerate, f.num_channels) as o:
-                    # Read one second of audio at a time, until the file is empty:
+                with AudioFile(temp_wav, 'w', f.samplerate, f.num_channels) as o:
                     while f.tell() < f.frames:
                         chunk = f.read(int(f.samplerate))
                         effected = board(chunk, f.samplerate, reset=False)
                         o.write(effected)
+
+            # Convert with pydub to desired output type
+            audio_seg = AudioSegment.from_file(temp_wav, format=type_output)
+            audio_seg.export(output_path, format=type_output, bitrate=("320k" if type_output == "mp3" else None))
+
+            # Clean up temp file
+            os.remove(temp_wav)
+
             result.append(output_path)
         except Exception as e:
             traceback.print_exc()
@@ -227,13 +297,13 @@ def add_audio_effects(audio_list):
     return result
 
 
-def apply_noisereduce(audio_list):
+def apply_noisereduce(audio_list, type_output):
     # https://github.com/sa-if/Audio-Denoiser
     print("Noice reduce")
 
     result = []
     for audio_path in audio_list:
-        out_path = f'{os.path.splitext(audio_path)[0]}_noisereduce.wav'
+        out_path = f"{os.path.splitext(audio_path)[0]}_noisereduce.{type_output}"
 
         try:
             # Load audio file
@@ -254,7 +324,7 @@ def apply_noisereduce(audio_list):
             )
 
             # Save reduced audio to file
-            reduced_audio.export(out_path, format="wav")
+            reduced_audio.export(out_path, format=type_output, bitrate=("320k" if type_output == "mp3" else None))
             result.append(out_path)
 
         except Exception as e:
@@ -266,13 +336,17 @@ def apply_noisereduce(audio_list):
 
 
 @spaces.GPU()
-def convert_now(audio_files, random_tag, converter):
-    return converter(
-        audio_files,
-        random_tag,
-        overwrite=False,
-        parallel_workers=8
-    )
+def convert_now(audio_files, random_tag, converter, type_output, steps):
+    for step in range(steps):
+        audio_files = converter(
+            audio_files,
+            random_tag,
+            overwrite=False,
+            parallel_workers=(2 if IS_COLAB else 8),
+            type_output=type_output,
+        )
+
+    return audio_files
 
 
 def run(
@@ -287,6 +361,8 @@ def run(
     c_b_p,
     active_noise_reduce,
     audio_effects,
+    type_output,
+    steps,
 ):
     if not audio_files:
         raise ValueError("The audio pls")
@@ -316,17 +392,17 @@ def run(
         respiration_median_filtering=r_m_f,
         envelope_ratio=e_r,
         consonant_breath_protection=c_b_p,
-        resample_sr=44100 if audio_files[0].endswith('.mp3') else 0, 
+        resample_sr=0,
     )
     time.sleep(0.1)
 
-    result = convert_now(audio_files, random_tag, converter)
+    result = convert_now(audio_files, random_tag, converter, type_output, steps)
 
     if active_noise_reduce:
-        result = apply_noisereduce(result)
+        result = apply_noisereduce(result, type_output)
 
     if audio_effects:
-        result = add_audio_effects(result)
+        result = add_audio_effects(result, type_output)
 
     return result
 
@@ -485,10 +561,30 @@ def sound_gui():
         type="filepath",
         # format="mp3",
         autoplay=True,
-        visible=False,
+        visible=True,
+        interactive=False,
+        elem_id="audio_tts",
+    )
+
+
+def steps_conf():
+    return gr.Slider(
+        minimum=1,
+        maximum=3,
+        label="Steps",
+        value=1,
+        step=1,
+        interactive=True,
     )
 
 
+def format_output_gui():
+    return gr.Dropdown(
+        label="Format output:",
+        choices=["wav", "mp3", "flac"],
+        value="wav",
+    )
+
 def denoise_conf():
     return gr.Checkbox(
         False,
@@ -571,9 +667,18 @@ def show_components_down(value_active):
         visible=value_active
     )
 
+CSS = """
+#audio_tts {
+  visibility: hidden;   /* invisible but still takes space */
+  height: 0px;
+  width: 0px;
+  max-width: 0px;
+  max-height: 0px;
+}
+"""
 
 def get_gui(theme):
-    with gr.Blocks(theme=theme, fill_width=True, fill_height=True, delete_cache=(3200, 3200)) as app:
+    with gr.Blocks(theme=theme, css=CSS, fill_width=True, fill_height=False, delete_cache=delete_cache_time) as app:
         gr.Markdown(title)
         gr.Markdown(description)
 
@@ -608,7 +713,7 @@ def get_gui(theme):
 
         down_active_gui = down_active_conf()
         down_info = gr.Markdown(
-            "Provide a link to a zip file, like this one: `https://huggingface.co/mrmocciai/Models/resolve/main/Genshin%20Impact/ayaka-v2.zip?download=true`, or separate links with a comma for the .pth and .index files, like this: `https://huggingface.co/sail-rvc/ayaka-jp/resolve/main/model.pth?download=true, https://huggingface.co/sail-rvc/ayaka-jp/resolve/main/model.index?download=true`",
+            f"Provide a link to a zip file, like this one: `https://huggingface.co/MrDawg/ToothBrushing/resolve/main/ToothBrushing.zip?download=true`, or separate links with a comma for the .pth and .index files, like this: `{test_model}`",
             visible=False
         )
         with gr.Row():
@@ -634,17 +739,20 @@ def get_gui(theme):
             [model, indx]
         )
 
-        algo = pitch_algo_conf()
-        algo_lvl = pitch_lvl_conf()
-        indx_inf = index_inf_conf()
-        res_fc = respiration_filter_conf()
-        envel_r = envelope_ratio_conf()
-        const = consonant_protec_conf()
-        with gr.Row():
-            with gr.Column():
-                with gr.Row():
-                    denoise_gui = denoise_conf()
-                    effects_gui = effects_conf()
+        with gr.Accordion(label="Advanced settings", open=False):
+            algo = pitch_algo_conf()
+            algo_lvl = pitch_lvl_conf()
+            indx_inf = index_inf_conf()
+            res_fc = respiration_filter_conf()
+            envel_r = envelope_ratio_conf()
+            const = consonant_protec_conf()
+            steps_gui = steps_conf()
+            format_out = format_output_gui()
+            with gr.Row():
+                with gr.Column():
+                    with gr.Row():
+                        denoise_gui = denoise_conf()
+                        effects_gui = effects_conf()
         button_base = button_conf()
         output_base = output_conf()
 
@@ -662,6 +770,8 @@ def get_gui(theme):
                 const,
                 denoise_gui,
                 effects_gui,
+                format_out,
+                steps_gui,
             ],
             outputs=[output_base],
         )
@@ -723,9 +833,11 @@ def get_gui(theme):
 
 
 if __name__ == "__main__":
-
-    tts_voice_list = asyncio.new_event_loop().run_until_complete(edge_tts.list_voices())
-    voices = sorted([f"{v['ShortName']}-{v['Gender']}" for v in tts_voice_list])
+    tts_voice_list = asyncio.new_event_loop().run_until_complete(get_voices_list(proxy=None))
+    voices = sorted([
+        (" - ".join(reversed(v["FriendlyName"].split("-"))).replace("Microsoft ", "").replace("Online (Natural)", f"({v['Gender']})").strip(), f"{v['ShortName']}-{v['Gender']}")
+        for v in tts_voice_list
+    ])
 
     app = get_gui(theme)
 
@@ -733,10 +845,9 @@ if __name__ == "__main__":
 
     app.launch(
         max_threads=40,
-        share=False,
+        share=IS_COLAB,
         show_error=True,
         quiet=False,
-        debug=False,
+        debug=IS_COLAB,
         ssr_mode=False,
-        allowed_paths=["./downloads/"],
     )

pre-requirements.txt

@@ -1 +1,2 @@
-pip==24.0
+pip==23.0.1
+Setuptools<=80.6.0

requirements.txt

@@ -1,8 +1,11 @@
-torch==2.2.0
-infer-rvc-python==1.1.0
+torch==2.5.1
+infer-rvc-python
 edge-tts
 pedalboard
 noisereduce
 numpy==1.23.5
 transformers<=4.48.3
 pydantic==2.10.6
+gradio==5.43.1
+spaces
+matplotlib-inline