Spaces:
Runtime error
Runtime error
File size: 2,578 Bytes
9f5785e 5bcaefa a5f7b84 5bcaefa 4b144cc 9f5785e 5bcaefa 9f5785e 0b90662 9f5785e a5f7b84 9f5785e 0b90662 5bcaefa 0b90662 5bcaefa 5ddeb31 4b144cc 5ddeb31 4b144cc 5ddeb31 a5f7b84 5ddeb31 0b90662 a5f7b84 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 |
import requests
import torch
import numpy as np
from transformers import AutoTokenizer
from parler_tts import ParlerTTSForConditionalGeneration
from pydub import AudioSegment
import simpleaudio as sa
# Hugging Face API URL for Roberta model
API_URL_ROBERTA = "https://api-inference.huggingface.co/models/deepset/roberta-base-squad2"
device = "cpu"
if torch.cuda.is_available():
device = "cuda:0"
if torch.backends.mps.is_available():
device = "mps"
if torch.xpu.is_available():
device = "xpu"
torch_dtype = torch.float16 if device != "cpu" else torch.float32
model = ParlerTTSForConditionalGeneration.from_pretrained("parler-tts/parler_tts_mini_v0.1").to(device, dtype=torch_dtype)
tokenizer = AutoTokenizer.from_pretrained("parler-tts/parler_tts_mini_v0.1")
def query_roberta(api_token, prompt, context):
payload = {
"inputs": {
"question": prompt,
"context": context
}
}
headers = {"Authorization": f"Bearer {api_token}"}
response = requests.post(API_URL_ROBERTA, headers=headers, json=payload)
try:
response_json = response.json()
if 'error' in response_json:
raise ValueError(response_json['error'])
return response_json
except ValueError as e:
print(f"ValueError: {e}")
return {"error": str(e)}
except Exception as e:
print(f"Exception: {e}")
return {"error": "An unexpected error occurred"}
def generate_speech(answer):
input_ids = tokenizer(answer, return_tensors="pt").input_ids.to(device)
generation = model.generate(input_ids=input_ids).to(torch.float32)
audio_arr = generation.cpu().numpy().squeeze()
# Convert numpy array to audio segment
audio_segment = AudioSegment(
audio_arr.tobytes(),
frame_rate=model.config.sampling_rate,
sample_width=audio_arr.dtype.itemsize,
channels=1
)
# Play the audio using simpleaudio
try:
play_obj = sa.play_buffer(
audio_segment.raw_data,
num_channels=1,
bytes_per_sample=audio_segment.sample_width,
sample_rate=audio_segment.frame_rate
)
play_obj.wait_done() # Wait until the audio is done playing
except Exception as e:
print(f"Error playing audio: {e}")
def gradio_interface(api_token, prompt, context):
answer = query_roberta(api_token, prompt, context)
if 'error' in answer:
return answer['error'], None
generate_speech(answer.get('answer', ''))
return answer.get('answer', 'No answer found'), None |