Spaces:
Runtime error
Runtime error
Upload with huggingface_hub
Browse files- DESCRIPTION.md +1 -0
- README.md +1 -1
- app.py +1 -9
DESCRIPTION.md
ADDED
@@ -0,0 +1 @@
|
|
|
1 |
+
This demo identifies if two speakers are the same person using Gradio's Audio and HTML components.
|
README.md
CHANGED
@@ -1,7 +1,7 @@
|
|
1 |
|
2 |
---
|
3 |
title: same-person-or-different
|
4 |
-
emoji:
|
5 |
colorFrom: indigo
|
6 |
colorTo: indigo
|
7 |
sdk: gradio
|
1 |
|
2 |
---
|
3 |
title: same-person-or-different
|
4 |
+
emoji: 🔥
|
5 |
colorFrom: indigo
|
6 |
colorTo: indigo
|
7 |
sdk: gradio
|
app.py
CHANGED
@@ -1,13 +1,9 @@
|
|
1 |
-
# URL: https://huggingface.co/spaces/gradio/same-person-or-different
|
2 |
-
# DESCRIPTION: This demo identifies if two speakers are the same person using Gradio's Audio and HTML components.
|
3 |
-
# imports
|
4 |
import gradio as gr
|
5 |
import torch
|
6 |
from torchaudio.sox_effects import apply_effects_file
|
7 |
from transformers import AutoFeatureExtractor, AutoModelForAudioXVector
|
8 |
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
9 |
|
10 |
-
# define outputs for HTML component
|
11 |
OUTPUT_OK = (
|
12 |
"""
|
13 |
<div class="container">
|
@@ -31,8 +27,6 @@ OUTPUT_FAIL = (
|
|
31 |
"""
|
32 |
)
|
33 |
|
34 |
-
# load model and define constants
|
35 |
-
|
36 |
EFFECTS = [
|
37 |
["remix", "-"],
|
38 |
["channels", "1"],
|
@@ -50,7 +44,6 @@ model = AutoModelForAudioXVector.from_pretrained(model_name).to(device)
|
|
50 |
cosine_sim = torch.nn.CosineSimilarity(dim=-1)
|
51 |
|
52 |
|
53 |
-
# define core fn
|
54 |
def similarity_fn(path1, path2):
|
55 |
if not (path1 and path2):
|
56 |
return '<b style="color:red">ERROR: Please record audio for *both* speakers!</b>'
|
@@ -76,7 +69,6 @@ def similarity_fn(path1, path2):
|
|
76 |
|
77 |
return output
|
78 |
|
79 |
-
#define inputs, outputs, description, article and examples
|
80 |
inputs = [
|
81 |
gr.inputs.Audio(source="microphone", type="filepath", optional=True, label="Speaker #1"),
|
82 |
gr.inputs.Audio(source="microphone", type="filepath", optional=True, label="Speaker #2"),
|
@@ -99,7 +91,7 @@ examples = [
|
|
99 |
["samples/cate_blanch.mp3", "samples/cate_blanch_2.mp3"],
|
100 |
["samples/cate_blanch.mp3", "samples/heath_ledger.mp3"],
|
101 |
]
|
102 |
-
|
103 |
interface = gr.Interface(
|
104 |
fn=similarity_fn,
|
105 |
inputs=inputs,
|
|
|
|
|
|
|
1 |
import gradio as gr
|
2 |
import torch
|
3 |
from torchaudio.sox_effects import apply_effects_file
|
4 |
from transformers import AutoFeatureExtractor, AutoModelForAudioXVector
|
5 |
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
6 |
|
|
|
7 |
OUTPUT_OK = (
|
8 |
"""
|
9 |
<div class="container">
|
27 |
"""
|
28 |
)
|
29 |
|
|
|
|
|
30 |
EFFECTS = [
|
31 |
["remix", "-"],
|
32 |
["channels", "1"],
|
44 |
cosine_sim = torch.nn.CosineSimilarity(dim=-1)
|
45 |
|
46 |
|
|
|
47 |
def similarity_fn(path1, path2):
|
48 |
if not (path1 and path2):
|
49 |
return '<b style="color:red">ERROR: Please record audio for *both* speakers!</b>'
|
69 |
|
70 |
return output
|
71 |
|
|
|
72 |
inputs = [
|
73 |
gr.inputs.Audio(source="microphone", type="filepath", optional=True, label="Speaker #1"),
|
74 |
gr.inputs.Audio(source="microphone", type="filepath", optional=True, label="Speaker #2"),
|
91 |
["samples/cate_blanch.mp3", "samples/cate_blanch_2.mp3"],
|
92 |
["samples/cate_blanch.mp3", "samples/heath_ledger.mp3"],
|
93 |
]
|
94 |
+
|
95 |
interface = gr.Interface(
|
96 |
fn=similarity_fn,
|
97 |
inputs=inputs,
|