videomae-base- finetuned-engine-subset-20230310

Hiperparámetros de entrenamiento:

• learning_rate: 5e-05
• train_batch_size: 6
• eval_batch_size: 6
• seed: 42
• optimizer: Adam con betas=(0.9,0.999) y epsilon=1e-08
• lr_scheduler_type: lineal
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 600

# Ejemplo de uso del modelo
from transformers import VideoMaeForVideoClassification, VideoMAEFeatureExtractor

model = VideoMaeForVideoClassification.from_pretrained('steveice/videomae-base-finetuned-engine-subset-20230310')
extractor = VideoMAEFeatureExtractor.from_pretrained('MCG-NJU/videomae-base')

# Preprocesamiento y clasificación de video
inputs = extractor(frame_sequences, return_tensors='pt')
outputs = model(**inputs)
logits = outputs.logits
# Obtener la etiqueta predicha
predicted_class_idx = logits.argmax(-1)