File size: 10,183 Bytes
e2d8169 5511ee4 e1925db e2d8169 5511ee4 3339344 e2d8169 5511ee4 9de903b 7b8f4ee e2d8169 5511ee4 e2d8169 5511ee4 e2d8169 c3de96c e2d8169 5511ee4 e2d8169 e1925db 13e5f34 8f194c4 e2d8169 950cada e2d8169 e1925db 5511ee4 c3de96c 9cad072 0353e7c 5511ee4 e2d8169 5511ee4 99d0cca 13e5f34 b3d62ee 5511ee4 8f194c4 b3d62ee 7b8f4ee b3d62ee 5511ee4 b3d62ee c3de96c b3d62ee 13e5f34 b3d62ee c3de96c b3d62ee dfe7f52 5511ee4 e2d8169 dfe7f52 e2d8169 e1925db 771ff84 3339344 8f194c4 3339344 f2d7c4e e2d8169 009081e 3339344 e2d8169 9cad072 e2d8169 e35a4fa e2d8169 5511ee4 7b8f4ee 97b2304 e1925db 8f194c4 0840021 97b2304 97edf64 97b2304 5511ee4 e2d8169 8f194c4 a26c262 e2d8169 e1925db 8a2a9a9 e2d8169 01963f4 e2d8169 5511ee4 e2d8169 9b3281c e2d8169 5336f00 e2d8169 9b3281c e2d8169 5336f00 e2d8169 5511ee4 dfe7f52 7b8f4ee dfe7f52 7b8f4ee dfe7f52 7b8f4ee dfe7f52 5511ee4 e35a4fa |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 |
import os
import time
from typing import List, Tuple, Optional
import google.generativeai as genai
import gradio as gr
# Get the Google API key from environment variables
GOOGLE_API_KEY = os.environ.get("GOOGLE_API_KEY")
# Define the title and subtitle for the Gradio interface
TITLE = """<h1 align="center">🏋️ AI Personal Trainer Playground 💪</h1>"""
SUBTITLE = """<h3 align="center">Upload your workout video and let the AI analyze your form 🖇️</h3>"""
# Define the prompt for the AI model
Prompt = """
You are the world's best fitness expert. Your goal is to analyze in detail how people perform their exercises and sports movements. Watch the provided video carefully and give them constructive feedback in at least 10 sentences. Focus on the following aspects:
Form and Technique: Identify any issues with the form and technique of the exercises being performed. Provide specific suggestions for improvement.
Repetitions and Sets: Count the number of repetitions and sets for each exercise. Ensure they match the intended workout plan.
Pacing and Timing: Evaluate the pacing and timing of the exercises. Suggest any adjustments needed to optimize performance.
Overall Performance: Give an overall assessment of the workout, highlighting strengths and areas for improvement.
Remember to be encouraging and supportive in your feedback. Your goal is to help them improve and stay motivated. Thank you!
"""
# Function to preprocess stop sequences
def preprocess_stop_sequences(stop_sequences: str) -> Optional[List[str]]:
if not stop_sequences:
return None
return [sequence.strip() for sequence in stop_sequences.split(",")]
# Function to handle user input
def user(text_prompt: str, chatbot: List[Tuple[str, str]]):
return "", chatbot + [[text_prompt, None]]
# Function to handle bot response
def bot(
google_key: str,
model_name: str,
video_prompt,
temperature: float,
max_output_tokens: int,
stop_sequences: str,
top_k: int,
top_p: float,
text_prompt_component: str,
chatbot: List[Tuple[str, str]]
):
# Use the provided Google API key or the one from environment variables
google_key = google_key if google_key else GOOGLE_API_KEY
if not google_key:
raise ValueError(
"GOOGLE_API_KEY is not set. "
"Please follow the instructions in the README to set it up.")
# Combine the user input with the predefined prompt
user_input = chatbot[-1][0]
combined_prompt = Prompt + "\n" + user_input
# Configure the generative AI model
genai.configure(api_key=google_key)
generation_config = genai.types.GenerationConfig(
temperature=temperature,
max_output_tokens=max_output_tokens,
stop_sequences=preprocess_stop_sequences(stop_sequences=stop_sequences),
top_k=top_k,
top_p=top_p)
# Handle video prompt if provided
if video_prompt is not None:
model = genai.GenerativeModel(model_name)
# Upload the video file
video_file = genai.upload_file(path=video_prompt)
while video_file.state.name == "PROCESSING":
print('.', end='')
time.sleep(10)
video_file = genai.get_file(video_file.name)
if video_file.state.name == "FAILED":
raise ValueError(video_file.state.name)
# Generate content based on the video and prompt
response = model.generate_content(
contents=[video_file, combined_prompt],
stream=True,
generation_config=generation_config,
request_options={"timeout": 600})
response.resolve()
else:
model = genai.GenerativeModel(model_name)
response = model.generate_content(
combined_prompt,
stream=True,
generation_config=generation_config)
response.resolve()
# Streaming effect for chatbot response
chatbot[-1][1] = ""
for chunk in response:
for i in range(0, len(chunk.text), 10):
section = chunk.text[i:i + 10]
chatbot[-1][1] += section
time.sleep(0.01)
yield chatbot
# Define Gradio components
google_key_component = gr.Textbox(
label="GOOGLE API KEY",
value="",
type="password",
placeholder="...",
info="You have to provide your own GOOGLE_API_KEY for this app to function properly",
visible=GOOGLE_API_KEY is None
)
video_prompt_component = gr.Video(label="Video", autoplay=True)
model_selection = gr.Dropdown(["gemini-1.5-flash-latest", "gemini-1.5-pro-latest"], label="Select Gemini Model", value="gemini-1.5-pro-latest")
chatbot_component = gr.Chatbot(
label='Gemini',
bubble_full_width=False,
scale=3, height=500
)
text_prompt_component = gr.Textbox(
placeholder="Hi there!",
label="Ask me anything and press Enter"
)
run_button_component = gr.Button()
temperature_component = gr.Slider(
minimum=0,
maximum=1.0,
value=0.6,
step=0.05,
label="Temperature",
info=(
"Temperature controls the degree of randomness in token selection. Lower "
"temperatures are good for prompts that expect a true or correct response, "
"while higher temperatures can lead to more diverse or unexpected results. "
))
max_output_tokens_component = gr.Slider(
minimum=1,
maximum=2048,
value=1024,
step=1,
label="Token limit",
info=(
"Token limit determines the maximum amount of text output from one prompt. A "
"token is approximately four characters. The default value is 2048."
))
stop_sequences_component = gr.Textbox(
label="Add stop sequence",
value="",
type="text",
placeholder="STOP, END",
info=(
"A stop sequence is a series of characters (including spaces) that stops "
"response generation if the model encounters it. The sequence is not included "
"as part of the response. You can add up to five stop sequences."
))
top_k_component = gr.Slider(
minimum=1,
maximum=40,
value=32,
step=1,
label="Top-K",
info=(
"Top-k changes how the model selects tokens for output. A top-k of 1 means the "
"selected token is the most probable among all tokens in the model's "
"vocabulary (also called greedy decoding), while a top-k of 3 means that the "
"next token is selected from among the 3 most probable tokens (using "
"temperature)."
))
top_p_component = gr.Slider(
minimum=0,
maximum=1,
value=1,
step=0.01,
label="Top-P",
info=(
"Top-p changes how the model selects tokens for output. Tokens are selected "
"from most probable to least until the sum of their probabilities equals the "
"top-p value. For example, if tokens A, B, and C have a probability of .3, .2, "
"and .1 and the top-p value is .5, then the model will select either A or B as "
"the next token (using temperature). "
))
# Define user and bot inputs
user_inputs = [text_prompt_component, chatbot_component]
bot_inputs = [
google_key_component,
model_selection,
video_prompt_component,
temperature_component,
max_output_tokens_component,
stop_sequences_component,
top_k_component,
top_p_component,
text_prompt_component,
chatbot_component
]
# Create the Gradio interface
with gr.Blocks() as demo:
gr.HTML(TITLE)
gr.HTML(SUBTITLE)
with gr.Column():
google_key_component.render()
with gr.Row():
video_prompt_component.render()
chatbot_component.render()
text_prompt_component.render()
run_button_component.render()
with gr.Accordion("Parameters", open=False):
model_selection.render()
temperature_component.render()
max_output_tokens_component.render()
stop_sequences_component.render()
with gr.Accordion("Advanced", open=False):
top_k_component.render()
top_p_component.render()
# Define the interaction between user input and bot response
run_button_component.click(
fn=user,
inputs=user_inputs,
outputs=[text_prompt_component, chatbot_component],
queue=False
).then(
fn=bot, inputs=bot_inputs, outputs=[chatbot_component]
)
text_prompt_component.submit(
fn=user,
inputs=user_inputs,
outputs=[text_prompt_component, chatbot_component],
queue=False
).then(
fn=bot, inputs=bot_inputs, outputs=[chatbot_component]
)
# Define example inputs for the Gradio interface
gr.Examples(
fn=bot,
inputs=bot_inputs,
outputs=[chatbot_component],
examples=[
[
"",
"gemini-1.5-pro-latest",
"./example1.mp4",
.7,
1024,
"",
32,
1,
"Give me some tips to improve my deadlift.",
[("", "")]
],
[
"",
"gemini-1.5-pro-latest",
"./example2.mp4",
.7,
1024,
"",
32,
1,
"How is my form?",
[("", "")]
],
[
"",
"gemini-1.5-pro-latest",
"./example3.mp4",
.7,
1024,
"",
32,
1,
"What improvements can I make?",
[("", "")]
],
[
"",
"gemini-1.5-pro-latest",
"./example4.mp4",
.7,
1024,
"",
32,
1,
"I just started working out. I'm not sure I'm doing it right. Can you check?",
[("", "")]
]
],
#cache_examples="lazy",
)
# Launch the Gradio interface
demo.queue(max_size=99).launch(debug=False, show_error=True) |