This model was trained on 209 random screenshots from the TV-show Scavengers Reign (2023) on MAX. I've used GPT-4o for captioning. I've I've kept most of the default settings of the 24gb LoRA config except for the steps which I've set to 4.000 steps.Workflow<ol><li>I used a short Python script to grab a 1.000 random images from a MP4 file</li><li>Then I used czkawka (<a target="_blank" rel="ugc" href="https://github.com/qarmin/czkawka">github</a>) to get rid of any duplicate or similar images</li><li>I've made a list of all charactes appearances, removing those that appeared the most often to avoid biases within the model</li><li>After that, I checked all the images manually and picked the 209 most aesthetic</li><li>I used a custom GPT (<a target="_blank" rel="ugc" href="https://chatgpt.com/g/g-jigGyrMzl-scavengers-reign-captioner">scavengers reign GPT</a>) for captioning</li><li>Finally I've trained the model with ostris ai-toolkit (<a target="_blank" rel="ugc" href="https://github.com/ostris/ai-toolkit">github</a>).</li></ol>Code:<pre><code>import cv2
import random

mp4_directory = ''
output_directory = ''
frames_to_extract = 120
base_name = "Random_screenshot"
list_of_random_frames = []
frame_distance = 100
first_frame = 0 

count = 0

vidcap = cv2.VideoCapture(mp4_directory)
totalFrames = vidcap.get(cv2.CAP_PROP_FRAME_COUNT)
while count &lt; frames_to_extract:
 count += 1
 count_str = str(count)
 frames_skipped = -1
 while True:
 randomFrameNumber = random.randint(0, totalFrames)
 frames_skipped +=1
 if frames_skipped &gt; 0:
 print(f"Frame Skipped {frames_skipped}")
 if all(abs(randomFrameNumber - frame) &gt; frame_distance and randomFrameNumber&gt; first_frame for frame in list_of_random_frames):
 break
 list_of_random_frames.append(randomFrameNumber)
 photo_output = output_directory + base_name + count_str + ".png"
 vidcap.set(cv2.CAP_PROP_POS_FRAMES,randomFrameNumber)
 success, image = vidcap.read()
 if success:
 cv2.imwrite(photo_output, image)
 print(f"Saving image to: {photo_output}")</code></pre>PS: If you want the dataset please contact me. I just don't want to get CivitAI in copyright trouble.

Files with SHA256: c912bc7b619e4c2a9026b4ca29b77820f08f2ad684c8ef04146d570071e47244