beit-large-patch16-512: types of film shots

Model description

This model is a fine-tuned version of microsoft/beit-large-patch16-512 on the szymonrucinski/types-of-film-shots dataset. It achieves the following results on the evaluation set:

• Loss: 1.2335
• Accuracy: 0.6763

usage

from transformers import pipeline
from PIL import Image
import requests

pipe = pipeline(
    "image-classification",
    model="pszemraj/beit-large-patch16-512-film-shot-classifier",
)

url = "https://cdn-uploads.huggingface.co/production/uploads/60bccec062080d33f875cd0c/9YqYvv188ZccCMSzuv0KW.png"
image = Image.open(requests.get(url, stream=True).raw)
result = pipe(image)[0]
print(result)

try some of these:

class labels

The dataset contains the following labels:

"id2label": {
    "0": "ambiguous",
    "1": "closeUp",
    "2": "detail",
    "3": "extremeLongShot",
    "4": "fullShot",
    "5": "longShot",
    "6": "mediumCloseUp",
    "7": "mediumShot"
  },

as plaintext:

ambiguous, close up, detail, extreme long shot, full shot, long shot, medium close up, medium shot

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 2
• eval_batch_size: 4
• seed: 24414
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 6.0
• mixed_precision_training: Native AMP

Training results

Framework versions

• Transformers 4.38.0.dev0
• Pytorch 2.2.0+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2