Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -5,53 +5,37 @@ import librosa
|
|
5 |
import numpy as np
|
6 |
import torch
|
7 |
|
8 |
-
from transformers import SpeechT5Processor, SpeechT5ForTextToSpeech, SpeechT5HifiGan
|
9 |
from transformers import AutoProcessor, AutoModelForCausalLM
|
10 |
|
11 |
|
12 |
-
checkpoint = "microsoft/speecht5_tts"
|
13 |
-
tts_processor = SpeechT5Processor.from_pretrained(checkpoint)
|
14 |
-
tts_model = SpeechT5ForTextToSpeech.from_pretrained(checkpoint)
|
15 |
-
vocoder = SpeechT5HifiGan.from_pretrained("microsoft/speecht5_hifigan")
|
16 |
|
17 |
|
18 |
vqa_processor = AutoProcessor.from_pretrained("ronniet/git-large-vqa-env")
|
19 |
vqa_model = AutoModelForCausalLM.from_pretrained("ronniet/git-large-vqa-env")
|
20 |
|
21 |
-
def tts(text):
|
22 |
-
|
23 |
-
|
24 |
|
25 |
-
|
26 |
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
|
31 |
-
|
32 |
-
# # load one of the provided speaker embeddings at random
|
33 |
-
# idx = np.random.randint(len(speaker_embeddings))
|
34 |
-
# key = list(speaker_embeddings.keys())[idx]
|
35 |
-
# speaker_embedding = np.load(speaker_embeddings[key])
|
36 |
|
37 |
-
|
38 |
-
# np.random.shuffle(speaker_embedding)
|
39 |
|
40 |
-
|
41 |
-
# x = (np.random.rand(512) >= 0.5) * 1.0
|
42 |
-
# x[x == 0] = -1.0
|
43 |
-
# speaker_embedding *= x
|
44 |
|
45 |
-
|
46 |
-
|
47 |
-
speaker_embedding = np.load("cmu_us_bdl_arctic-wav-arctic_a0009.npy")
|
48 |
-
|
49 |
-
speaker_embedding = torch.tensor(speaker_embedding).unsqueeze(0)
|
50 |
-
|
51 |
-
speech = tts_model.generate_speech(input_ids, speaker_embedding, vocoder=vocoder)
|
52 |
-
|
53 |
-
speech = (speech.numpy() * 32767).astype(np.int16)
|
54 |
-
return (16000, speech)
|
55 |
|
56 |
|
57 |
# captioner = pipeline(model="microsoft/git-base")
|
@@ -70,15 +54,15 @@ def predict(image, prompt):
|
|
70 |
text_ids = vqa_model.generate(pixel_values=pixel_values, input_ids=prompt_ids, max_length=50)
|
71 |
text = vqa_processor.batch_decode(text_ids, skip_special_tokens=True)[0][len(prompt):]
|
72 |
|
73 |
-
audio = tts(text)
|
74 |
|
75 |
-
return text
|
76 |
|
77 |
|
78 |
demo = gr.Interface(
|
79 |
fn=predict,
|
80 |
inputs=[gr.Image(type="pil",label="Environment"), gr.Textbox(label="Prompt", value="What is in the scene?")],
|
81 |
-
outputs=
|
82 |
css=".gradio-container {background-color: #002A5B}",
|
83 |
theme=gr.themes.Soft() #.set(
|
84 |
# button_primary_background_fill="#AAAAAA",
|
|
|
5 |
import numpy as np
|
6 |
import torch
|
7 |
|
8 |
+
# from transformers import SpeechT5Processor, SpeechT5ForTextToSpeech, SpeechT5HifiGan
|
9 |
from transformers import AutoProcessor, AutoModelForCausalLM
|
10 |
|
11 |
|
12 |
+
# checkpoint = "microsoft/speecht5_tts"
|
13 |
+
# tts_processor = SpeechT5Processor.from_pretrained(checkpoint)
|
14 |
+
# tts_model = SpeechT5ForTextToSpeech.from_pretrained(checkpoint)
|
15 |
+
# vocoder = SpeechT5HifiGan.from_pretrained("microsoft/speecht5_hifigan")
|
16 |
|
17 |
|
18 |
vqa_processor = AutoProcessor.from_pretrained("ronniet/git-large-vqa-env")
|
19 |
vqa_model = AutoModelForCausalLM.from_pretrained("ronniet/git-large-vqa-env")
|
20 |
|
21 |
+
# def tts(text):
|
22 |
+
# if len(text.strip()) == 0:
|
23 |
+
# return (16000, np.zeros(0).astype(np.int16))
|
24 |
|
25 |
+
# inputs = tts_processor(text=text, return_tensors="pt")
|
26 |
|
27 |
+
# # limit input length
|
28 |
+
# input_ids = inputs["input_ids"]
|
29 |
+
# input_ids = input_ids[..., :tts_model.config.max_text_positions]
|
30 |
|
31 |
+
# speaker_embedding = np.load("cmu_us_bdl_arctic-wav-arctic_a0009.npy")
|
|
|
|
|
|
|
|
|
32 |
|
33 |
+
# speaker_embedding = torch.tensor(speaker_embedding).unsqueeze(0)
|
|
|
34 |
|
35 |
+
# speech = tts_model.generate_speech(input_ids, speaker_embedding, vocoder=vocoder)
|
|
|
|
|
|
|
36 |
|
37 |
+
# speech = (speech.numpy() * 32767).astype(np.int16)
|
38 |
+
# return (16000, speech)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
39 |
|
40 |
|
41 |
# captioner = pipeline(model="microsoft/git-base")
|
|
|
54 |
text_ids = vqa_model.generate(pixel_values=pixel_values, input_ids=prompt_ids, max_length=50)
|
55 |
text = vqa_processor.batch_decode(text_ids, skip_special_tokens=True)[0][len(prompt):]
|
56 |
|
57 |
+
# audio = tts(text)
|
58 |
|
59 |
+
return text
|
60 |
|
61 |
|
62 |
demo = gr.Interface(
|
63 |
fn=predict,
|
64 |
inputs=[gr.Image(type="pil",label="Environment"), gr.Textbox(label="Prompt", value="What is in the scene?")],
|
65 |
+
outputs=gr.Textbox(label="Caption"),
|
66 |
css=".gradio-container {background-color: #002A5B}",
|
67 |
theme=gr.themes.Soft() #.set(
|
68 |
# button_primary_background_fill="#AAAAAA",
|