Created
February 13, 2024 20:44
-
-
Save CoffeeVampir3/7a7c5b78347aaee263c660ac8461b7b3 to your computer and use it in GitHub Desktop.
moondream tagger
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
from transformers import AutoModelForCausalLM, CodeGenTokenizerFast as Tokenizer | |
from moondream import Moondream | |
from PIL import Image | |
from collections import Counter | |
import glob | |
import os | |
model = Moondream.from_pretrained("/home/blackroot/Desktop/moond/moondream1").to("cuda") | |
tokenizer = Tokenizer.from_pretrained("/home/blackroot/Desktop/moond/moondream1/tokenizer") | |
characters = [ | |
"modeus", | |
"azazel", | |
"malina", | |
"lucifer", | |
"beelzebub", | |
"pandemonica", | |
"cerberus", | |
"millicent" | |
] | |
directory_path = '/home/blackroot/Desktop/moond/moondream1/assets' | |
pattern = f"{directory_path}/*.png" | |
for filepath in glob.glob(pattern): | |
#print(filepath) | |
image = Image.open(filepath) | |
enc_image = model.encode_image(image) | |
# Construct the text file name by replacing .png with .txt | |
txt_filename = os.path.splitext(filepath)[0] + '.txt' | |
txt_filepath = os.path.join(directory_path, txt_filename) | |
words = filepath.split('_') | |
matched_character = "" | |
for word in words: | |
word_cleaned = os.path.splitext(word)[0].lower() | |
if word_cleaned in characters: | |
matched_character = word | |
break | |
answer = model.answer_question(enc_image, f"Describe the character:", tokenizer) | |
answer = f"{matched_character} {answer}" | |
print(answer) | |
with open(txt_filepath, 'w') as txt_file: | |
txt_file.write(answer) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment