up

README.md

@@ -1,8 +1,8 @@
 ---
-title: Asv
-emoji: 📉
+title: Unispeech Speaker Verification
+emoji: 💻
 colorFrom: blue
-colorTo: blue
+colorTo: purple
 sdk: gradio
 app_file: app.py
 pinned: false
@@ -12,7 +12,7 @@ pinned: false
 
 `title`: _string_  
 Display title for the Space
-
+ 
 `emoji`: _string_  
 Space emoji (emoji-only character allowed)
 

app.py

@@ -0,0 +1,131 @@
+import torch
+import gradio as gr
+from torchaudio.sox_effects import apply_effects_file
+from transformers import AutoFeatureExtractor, AutoModelForAudioXVector
+
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+
+STYLE = """
+<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/bootstrap@5.1.3/dist/css/bootstrap.min.css" integrity="sha256-YvdLHPgkqJ8DVUxjjnGVlMMJtNimJ6dYkowFFvp4kKs=" crossorigin="anonymous">
+"""
+OUTPUT_OK = STYLE + """
+    <div class="container">
+        <div class="row"><h1 style="text-align: center">The speakers are</h1></div>
+        <div class="row"><h1 class="display-1 text-success" style="text-align: center">{:.1f}%</h1></div>
+        <div class="row"><h1 style="text-align: center">similar</h1></div>
+        <div class="row"><h1 class="text-success" style="text-align: center">Welcome, human!</h1></div>
+        <div class="row"><small style="text-align: center">(You must get at least 85% to be considered the same person)</small><div class="row">
+    </div>
+"""
+OUTPUT_FAIL = STYLE + """
+    <div class="container">
+        <div class="row"><h1 style="text-align: center">The speakers are</h1></div>
+        <div class="row"><h1 class="display-1 text-danger" style="text-align: center">{:.1f}%</h1></div>
+        <div class="row"><h1 style="text-align: center">similar</h1></div>
+        <div class="row"><h1 class="text-danger" style="text-align: center">You shall not pass!</h1></div>
+        <div class="row"><small style="text-align: center">(You must get at least 85% to be considered the same person)</small><div class="row">
+    </div>
+"""
+
+EFFECTS = [
+    ['remix', '-'],
+    ["channels", "1"],
+    ["rate", "16000"],
+    ["gain", "-1.0"],
+    ["silence", "1", "0.1", "0.1%", "-1", "0.1", "0.1%"],
+    ['trim', '0', '10'],
+]
+
+THRESHOLD = 0.85
+
+model_name = "microsoft/unispeech-sat-base-plus-sv"
+feature_extractor = AutoFeatureExtractor.from_pretrained(model_name)
+model = AutoModelForAudioXVector.from_pretrained(model_name).to(device)
+cosine_sim = torch.nn.CosineSimilarity(dim=-1)
+
+
+def similarity_fn(path1, path2):
+    if not (path1 and path2):
+        return '<b style="color:red">ERROR: Please record audio for *both* speakers!</b>'
+
+    wav1, _ = apply_effects_file(path1, EFFECTS)
+    wav2, _ = apply_effects_file(path2, EFFECTS)
+    print(wav1.shape, wav2.shape)
+
+    input1 = feature_extractor(wav1.squeeze(0), return_tensors="pt", sampling_rate=16000).input_values.to(device)
+    input2 = feature_extractor(wav2.squeeze(0), return_tensors="pt", sampling_rate=16000).input_values.to(device)
+
+    with torch.no_grad():
+        emb1 = model(input1).embeddings
+        emb2 = model(input2).embeddings
+    emb1 = torch.nn.functional.normalize(emb1, dim=-1).cpu()
+    emb2 = torch.nn.functional.normalize(emb2, dim=-1).cpu()
+    similarity = cosine_sim(emb1, emb2).numpy()[0]
+
+    if similarity >= THRESHOLD:
+        output = OUTPUT_OK.format(similarity * 100)
+    else:
+        output = OUTPUT_FAIL.format(similarity * 100)
+
+    return output
+
+
+inputs = [
+    gr.inputs.Audio(source="microphone", type="filepath", optional=True, label="Speaker #1"),
+    gr.inputs.Audio(source="microphone", type="filepath", optional=True, label="Speaker #2"),
+]
+output = gr.outputs.HTML(label="")
+
+
+description = (
+    "This demo will compare two speech samples and determine if they are from the same speaker. "
+    "Try it with your own voice!"
+)
+article = (
+    "<p style='text-align: center'>"
+    "<a href='https://huggingface.co/microsoft/unispeech-sat-large-sv' target='_blank'>🎙️ Learn more about UniSpeech-SAT</a> | "
+    "<a href='https://arxiv.org/abs/2110.05752' target='_blank'>📚 UniSpeech-SAT paper</a> | "
+    "<a href='https://www.danielpovey.com/files/2018_icassp_xvectors.pdf' target='_blank'>📚 X-Vector paper</a>"
+    "</p>"
+)
+
+interface = gr.Interface(
+    fn=similarity_fn,
+    inputs=inputs,
+    outputs=output,
+    title="Voice Authentication with UniSpeech-SAT + X-Vectors",
+    description=description,
+    article=article,
+    layout="horizontal",
+    theme="huggingface",
+    allow_flagging=False,
+    live=False,
+    examples=[
+        ["samples/cate_blanch.wav", "samples/cate_blanch_2.wav"],
+        ["samples/cate_blanch.wav", "samples/cate_blanch_3.wav"],
+        ["samples/cate_blanch_2.wav", "samples/cate_blanch_3.wav"],
+        ["samples/naomi_watts.wav", "samples/cate_blanch_3.wav"],
+        ["samples/naomi_watts.wav", "samples/denzel_wash.wav"],
+        ["samples/arnold_schwarz.wav", "samples/denzel_wash.wav"],
+        ["samples/arnold_schwarz.wav", "samples/cate_blanch.wav"],
+        ["samples/arnold_schwarz.wav", "samples/cate_blanch_2.wav"],
+        ["samples/arnold_schwarz.wav", "samples/heath_ledger_2.wav"],
+        ["samples/dicaprio_2.wav", "samples/crowe.wav"],
+        ["samples/KirstenDunst.wav", "samples/crowe.wav"],
+        ["samples/cate_blanch_3.wav", "samples/KirstenDunst.wav"],
+        ["samples/Zendaya.wav", "samples/cate_blanch_3.wav"],
+        ["samples/dicaprio.wav", "samples/denzel_wash.wav"],
+        ["samples/cate_blanch_2.wav", "samples/cate_blanch_3.wav"],
+        ["samples/cate_blanch_2.wav", "samples/cate_blanch_3.wav"],
+        ["samples/crowe.wav", "samples/crowe_2.wav"],
+        ["samples/crowe.wav", "samples/crowe_3.wav"],
+        ["samples/dicaprio.wav", "samples/dicaprio_2.wav"],
+        ["samples/heath_ledger.wav", "samples/heath_ledger_2.wav"],
+        ["samples/heath_ledger.wav", "samples/heath_ledger_3.wav"],
+        ["samples/heath_ledger.wav", "samples/heath_ledger_4.wav"],
+        ["samples/heath_ledger_3.wav", "samples/heath_ledger_2.wav"],
+        ["samples/heath_ledger_4.wav", "samples/heath_ledger_2.wav"],
+        ["samples/heath_ledger_3.wav", "samples/heath_ledger_4.wav"],
+    ]
+)
+interface.launch(enable_queue=True)

requirements.txt

@@ -0,0 +1,2 @@
+git+https://github.com/huggingface/transformers
+torchaudio