Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -2,31 +2,34 @@ import gradio as gr
|
|
2 |
import torch
|
3 |
import torchaudio
|
4 |
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
|
5 |
-
# from huggingface_hub import InferenceClient # Removed
|
6 |
from ttsmms import download, TTS
|
7 |
from langdetect import detect
|
8 |
-
from gradio_client import Client
|
9 |
|
|
|
10 |
# Load ASR Model
|
|
|
11 |
asr_model_name = "Futuresony/Future-sw_ASR-24-02-2025"
|
12 |
processor = Wav2Vec2Processor.from_pretrained(asr_model_name)
|
13 |
asr_model = Wav2Vec2ForCTC.from_pretrained(asr_model_name)
|
14 |
|
15 |
-
#
|
16 |
-
#
|
17 |
-
|
18 |
-
|
19 |
-
# def format_prompt(user_input): # Removed
|
20 |
-
# return f"{user_input}" # Removed
|
21 |
|
|
|
22 |
# Load TTS Models
|
|
|
23 |
swahili_dir = download("swh", "./data/swahili")
|
24 |
english_dir = download("eng", "./data/english")
|
25 |
|
26 |
swahili_tts = TTS(swahili_dir)
|
27 |
english_tts = TTS(english_dir)
|
28 |
|
|
|
29 |
# ASR Function
|
|
|
30 |
def transcribe(audio_file):
|
31 |
speech_array, sample_rate = torchaudio.load(audio_file)
|
32 |
resampler = torchaudio.transforms.Resample(orig_freq=sample_rate, new_freq=16000)
|
@@ -38,39 +41,62 @@ def transcribe(audio_file):
|
|
38 |
transcription = processor.batch_decode(predicted_ids)[0]
|
39 |
return transcription
|
40 |
|
41 |
-
#
|
|
|
|
|
42 |
def generate_text(prompt):
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
result
|
47 |
-
|
48 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
49 |
|
|
|
50 |
# TTS Function
|
|
|
51 |
def text_to_speech(text):
|
52 |
-
print(f"Converting text to speech (type: {type(text)})
|
53 |
lang = detect(text)
|
54 |
wav_path = "./output.wav"
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
-
|
59 |
-
|
|
|
|
|
|
|
60 |
return wav_path
|
61 |
|
|
|
62 |
# Combined Processing Function
|
|
|
63 |
def process_audio(audio):
|
64 |
-
print(f"Processing audio
|
|
|
65 |
transcription = transcribe(audio)
|
66 |
-
print(f"Transcription
|
|
|
67 |
generated_text = generate_text(transcription)
|
68 |
-
print(f"Generated
|
69 |
-
|
70 |
-
|
71 |
-
|
|
|
|
|
72 |
|
|
|
73 |
# Gradio Interface
|
|
|
74 |
with gr.Blocks() as demo:
|
75 |
gr.Markdown("<p align='center' style='font-size: 20px;'>End-to-End ASR, Text Generation, and TTS</p>")
|
76 |
gr.HTML("<center>Upload or record audio. The model will transcribe, generate a response, and read it out.</center>")
|
@@ -88,4 +114,4 @@ with gr.Blocks() as demo:
|
|
88 |
)
|
89 |
|
90 |
if __name__ == "__main__":
|
91 |
-
demo.launch()
|
|
|
2 |
import torch
|
3 |
import torchaudio
|
4 |
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
|
|
|
5 |
from ttsmms import download, TTS
|
6 |
from langdetect import detect
|
7 |
+
from gradio_client import Client
|
8 |
|
9 |
+
# =========================
|
10 |
# Load ASR Model
|
11 |
+
# =========================
|
12 |
asr_model_name = "Futuresony/Future-sw_ASR-24-02-2025"
|
13 |
processor = Wav2Vec2Processor.from_pretrained(asr_model_name)
|
14 |
asr_model = Wav2Vec2ForCTC.from_pretrained(asr_model_name)
|
15 |
|
16 |
+
# =========================
|
17 |
+
# Load Text Generation Model via Gradio Client
|
18 |
+
# =========================
|
19 |
+
llm_client = Client("Futuresony/Mr.Events")
|
|
|
|
|
20 |
|
21 |
+
# =========================
|
22 |
# Load TTS Models
|
23 |
+
# =========================
|
24 |
swahili_dir = download("swh", "./data/swahili")
|
25 |
english_dir = download("eng", "./data/english")
|
26 |
|
27 |
swahili_tts = TTS(swahili_dir)
|
28 |
english_tts = TTS(english_dir)
|
29 |
|
30 |
+
# =========================
|
31 |
# ASR Function
|
32 |
+
# =========================
|
33 |
def transcribe(audio_file):
|
34 |
speech_array, sample_rate = torchaudio.load(audio_file)
|
35 |
resampler = torchaudio.transforms.Resample(orig_freq=sample_rate, new_freq=16000)
|
|
|
41 |
transcription = processor.batch_decode(predicted_ids)[0]
|
42 |
return transcription
|
43 |
|
44 |
+
# =========================
|
45 |
+
# Text Generation Function (Safe)
|
46 |
+
# =========================
|
47 |
def generate_text(prompt):
|
48 |
+
print(f"[DEBUG] Generating text for prompt: {prompt} (type: {type(prompt)})")
|
49 |
+
|
50 |
+
result = llm_client.predict(query=prompt, api_name="/chat")
|
51 |
+
print(f"[DEBUG] /chat returned: {result} (type: {type(result)})")
|
52 |
+
|
53 |
+
# Ensure result is always a string
|
54 |
+
if not isinstance(result, str):
|
55 |
+
try:
|
56 |
+
result = " ".join(map(str, result)) if isinstance(result, (list, tuple)) else str(result)
|
57 |
+
except Exception as e:
|
58 |
+
print(f"[ERROR] Failed to convert result to string: {e}")
|
59 |
+
result = "Error: Unable to generate text."
|
60 |
+
|
61 |
+
return result.strip()
|
62 |
|
63 |
+
# =========================
|
64 |
# TTS Function
|
65 |
+
# =========================
|
66 |
def text_to_speech(text):
|
67 |
+
print(f"[DEBUG] Converting text to speech: {text} (type: {type(text)})")
|
68 |
lang = detect(text)
|
69 |
wav_path = "./output.wav"
|
70 |
+
try:
|
71 |
+
if lang == "sw":
|
72 |
+
swahili_tts.synthesis(text, wav_path=wav_path)
|
73 |
+
else:
|
74 |
+
english_tts.synthesis(text, wav_path=wav_path)
|
75 |
+
except Exception as e:
|
76 |
+
print(f"[ERROR] TTS synthesis failed: {e}")
|
77 |
+
return None
|
78 |
return wav_path
|
79 |
|
80 |
+
# =========================
|
81 |
# Combined Processing Function
|
82 |
+
# =========================
|
83 |
def process_audio(audio):
|
84 |
+
print(f"[DEBUG] Processing audio: {audio} (type: {type(audio)})")
|
85 |
+
|
86 |
transcription = transcribe(audio)
|
87 |
+
print(f"[DEBUG] Transcription: {transcription}")
|
88 |
+
|
89 |
generated_text = generate_text(transcription)
|
90 |
+
print(f"[DEBUG] Generated Text: {generated_text}")
|
91 |
+
|
92 |
+
speech_path = text_to_speech(generated_text)
|
93 |
+
print(f"[DEBUG] Speech Path: {speech_path}")
|
94 |
+
|
95 |
+
return transcription, generated_text, speech_path
|
96 |
|
97 |
+
# =========================
|
98 |
# Gradio Interface
|
99 |
+
# =========================
|
100 |
with gr.Blocks() as demo:
|
101 |
gr.Markdown("<p align='center' style='font-size: 20px;'>End-to-End ASR, Text Generation, and TTS</p>")
|
102 |
gr.HTML("<center>Upload or record audio. The model will transcribe, generate a response, and read it out.</center>")
|
|
|
114 |
)
|
115 |
|
116 |
if __name__ == "__main__":
|
117 |
+
demo.launch()
|