File size: 5,587 Bytes
2f3b32c
 
 
 
 
 
 
 
 
 
 
 
 
 
ab0e126
 
 
 
 
 
 
 
 
 
2f3b32c
ab0e126
2f3b32c
ab0e126
 
 
2f3b32c
 
ace0051
2f3b32c
 
 
 
 
ab0e126
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2f3b32c
 
 
 
 
 
 
 
 
 
 
 
 
ab0e126
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2f3b32c
 
 
 
 
 
 
 
 
ab0e126
 
 
 
2f3b32c
 
 
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
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
import os
import re
import gradio as gr
import edge_tts
import asyncio
import time
import tempfile
from huggingface_hub import InferenceClient

DESCRIPTION = """ # <center><b>JARVIS⚡</b></center>
        ### <center>A personal Assistant of Tony Stark for YOU
        ### <center>Currently It supports text input, But If this space completes 1k hearts than I starts working on Audio Input.</center>
        """

MORE = """ ## TRY Other Models
        ### Instant Video: Create Amazing Videos in 5 Second -> https://huggingface.co/spaces/KingNish/Instant-Video
        ### Instant Image: 4k images in 5 Second -> https://huggingface.co/spaces/KingNish/Instant-Image
        """

Fast = """## Fastest Model"""

Complex = """## Best in Complex Question"""

Detail = """## Best for Detailed Generation or Long Answers"""

client1 = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")

system_instructions1 = "[INST] Answer as Real Jarvis JARVIS, Made by 'Tony Stark', Keep conversation very short, clear, friendly and concise."

async def generate1(prompt):
    generate_kwargs = dict(
        temperature=0.6,
        max_new_tokens=256,
        top_p=0.95,
        repetition_penalty=1,
        do_sample=True,
        seed=42,
    )
    formatted_prompt = system_instructions1 + prompt + "[/INST]"
    stream = client1.text_generation(
        formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=True)
    output = ""
    for response in stream:
        output += response.token.text

    communicate = edge_tts.Communicate(output)
    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp_file:
        tmp_path = tmp_file.name
        await communicate.save(tmp_path)
    yield tmp_path

client2 = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")

system_instructions2 = "[SYSTEM] Answer as Real Jarvis JARVIS, Made by 'Tony Stark', Must answer in friendly style and Easy Manner. You can answer Complex Questions. Do not say who are you or Hi, Hello, Just Start answering. Stop, as answer ends. [USER]"

async def generate2(prompt):
    generate_kwargs = dict(
        temperature=0.6,
        max_new_tokens=512,
        top_p=0.95,
        repetition_penalty=1,
        do_sample=True,
    )    
    formatted_prompt = system_instructions2 + prompt + "[ASSISTANT]"
    stream = client2.text_generation(
        formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=True)
    output = ""
    for response in stream:
        output += response.token.text

    communicate = edge_tts.Communicate(output)
    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp_file:
        tmp_path = tmp_file.name
        await communicate.save(tmp_path)
    yield tmp_path

client3 = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")

system_instructions3 = "[SYSTEM] Answer as Real Jarvis JARVIS, Made by 'Tony Stark', Must answer in detailed and friendly. Do not say who are you or Hi, Hello, Just Start answering. You answers all things in detail.[USER]"

async def generate3(prompt):
    generate_kwargs = dict(
        temperature=0.6,
        max_new_tokens=2048,
        top_p=0.95,
        repetition_penalty=1,
        do_sample=True,
    )    
    formatted_prompt = system_instructions3 + prompt + "[ASSISTANT]"
    stream = client3.text_generation(
        formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=True)
    output = ""
    for response in stream:
        output += response.token.text

    communicate = edge_tts.Communicate(output)
    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp_file:
        tmp_path = tmp_file.name
        await communicate.save(tmp_path)
    yield tmp_path

with gr.Blocks(css="style.css") as demo:    
    gr.Markdown(DESCRIPTION)
    gr.Markdown(Fast)
    with gr.Row():
        user_input = gr.Textbox(label="Prompt")
        input_text = gr.Textbox(label="Input Text", elem_id="important")
        output_audio = gr.Audio(label="Audio", type="filepath",
                        interactive=False,
                        autoplay=True,
                        elem_classes="audio")
    with gr.Row():
        translate_btn = gr.Button("Response")
        translate_btn.click(fn=generate1, inputs=user_input,
                            outputs=output_audio, api_name="translate") 

    gr.Markdown(Complex)
    with gr.Row():
        user_input = gr.Textbox(label="Prompt")
        input_text = gr.Textbox(label="Input Text", elem_id="important")
        output_audio = gr.Audio(label="Audio", type="filepath",
                        interactive=False,
                        autoplay=True,
                        elem_classes="audio")
    with gr.Row():
        translate_btn = gr.Button("Response")
        translate_btn.click(fn=generate2, inputs=user_input,
                            outputs=output_audio, api_name="translate") 

    gr.Markdown(Detail)
    with gr.Row():
        user_input = gr.Textbox(label="Prompt")
        input_text = gr.Textbox(label="Input Text", elem_id="important")
        output_audio = gr.Audio(label="Audio", type="filepath",
                        interactive=False,
                        autoplay=True,
                        elem_classes="audio")
    with gr.Row():
        translate_btn = gr.Button("Response")
        translate_btn.click(fn=generate3, inputs=user_input,
                            outputs=output_audio, api_name="translate") 
        
    gr.Markdown(MORE)

if __name__ == "__main__":
    demo.queue(max_size=20).launch()