Gemini vision

This commit is contained in:
Taradepan R 2023-12-14 17:15:13 +00:00
parent 1174275b5b
commit b44e7a1269
3 changed files with 60 additions and 15 deletions

View File

@ -23,6 +23,7 @@ Make a [Replicate](https://replicate.com), [OpenAI](https://beta.openai.com/), a
``` ```
export OPENAI_API_KEY=<token> export OPENAI_API_KEY=<token>
export ELEVENLABS_API_KEY=<eleven-token> export ELEVENLABS_API_KEY=<eleven-token>
export GEMINI_API_KEY=<Gemini-api-key>
``` ```
Make a new voice in Eleven and get the voice id of that voice using their [get voices](https://elevenlabs.io/docs/api-reference/voices) API, or by clicking the flask icon next to the voice in the VoiceLab tab. Make a new voice in Eleven and get the voice id of that voice using their [get voices](https://elevenlabs.io/docs/api-reference/voices) API, or by clicking the flask icon next to the voice in the VoiceLab tab.
@ -42,4 +43,5 @@ In another terminal, run the narrator:
```bash ```bash
python narrator.py python narrator.py
``` ```
choose the model by selecting between 1 or 2.

View File

@ -6,8 +6,13 @@ import time
import simpleaudio as sa import simpleaudio as sa
import errno import errno
from elevenlabs import generate, play, set_api_key, voices from elevenlabs import generate, play, set_api_key, voices
import google.generativeai as genai
import PIL.Image
client = OpenAI() client = OpenAI()
genai.configure(api_key = os.environ.get("GEMINI_API_KEY"))
set_api_key(os.environ.get("ELEVENLABS_API_KEY")) set_api_key(os.environ.get("ELEVENLABS_API_KEY"))
@ -76,26 +81,63 @@ def analyze_image(base64_image, script):
def main(): def main():
script = [] script = []
while True: model = input("Which model would you like to use? 1. GPT-4 Vision 2. Gemini Pro Vision \n")
# path to your image
image_path = os.path.join(os.getcwd(), "./frames/frame.jpg")
# getting the base64 encoding if model == "1":
base64_image = encode_image(image_path) print("using GPT-4 Vision")
# analyze posture
print("👀 David is watching...") print("👀 David is watching...")
analysis = analyze_image(base64_image, script=script)
print("🎙️ David says:") while True:
print(analysis) # path to your image
image_path = os.path.join(os.getcwd(), "./frames/frame.jpg")
play_audio(analysis)
script = script + [{"role": "assistant", "content": analysis}] # getting the base64 encoding
base64_image = encode_image(image_path)
analysis = analyze_image(base64_image, script=script)
# wait for 5 seconds
time.sleep(5) print("🎙️ David says:")
print(analysis)
play_audio(analysis)
script = script + [{"role": "assistant", "content": analysis}]
# wait for 5 seconds
time.sleep(5)
elif model == "2":
print("using Gemini Pro Vision")
print("👀 David is watching...")
while True:
# path to your image
image_path = os.path.join(os.getcwd(), "./frames/frame.jpg")
img = PIL.Image.open(image_path)
model = genai.GenerativeModel('gemini-pro-vision')
response = model.generate_content(["""
You are Sir David Attenborough. Narrate the picture of the human as if it is a nature documentary.
Make it snarky and funny. Don't repeat yourself. Make it short. If I do anything remotely interesting, make a big deal about it!
"""+"refer to these previous narrations".join(script), img])
response_text = response.text
print("🎙️ David says:")
print(response_text)
play_audio(response_text)
script = script + [response_text]
# wait for 5 seconds
time.sleep(5)
else:
print("Please enter a valid model number")
if __name__ == "__main__": if __name__ == "__main__":

View File

@ -39,3 +39,4 @@ typing_extensions==4.8.0
urllib3==2.0.7 urllib3==2.0.7
wcwidth==0.2.10 wcwidth==0.2.10
websockets==12.0 websockets==12.0
google-generativeai==0.3.1