finetuned-Accident-MultipleLabels-Video-subset-v2-new

This model is a fine-tuned version of MCG-NJU/videomae-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.7905
• Accuracy: 0.3333

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.2	4	1.9753	0.1406
No log	1.2	8	1.9765	0.0938
1.7405	2.2	12	2.0965	0.0938
1.7405	3.2	16	2.1152	0.0469
1.4691	4.2	20	2.1130	0.0469

Framework versions

• Transformers 4.35.2
• Pytorch 2.2.0.dev20231202+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0