videomae-surf-analytics-runpod

This model is a fine-tuned version of facebook/timesformer-base-finetuned-k400 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4027
• Accuracy: 0.8838
• F1: 0.8838

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: 12
• eval_batch_size: 12
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 610

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.6712	0.1016	62	0.8671	0.6680	0.6623
0.3119	1.1016	124	0.5911	0.7884	0.7887
0.2505	2.1016	186	0.5297	0.8008	0.8002
0.207	3.1016	248	0.5970	0.7801	0.7787
0.1743	4.1016	310	0.5612	0.8050	0.7984
0.1005	5.1016	372	0.4027	0.8838	0.8838
0.0147	6.1016	434	0.4360	0.8589	0.8573
0.0573	7.1016	496	0.4451	0.8714	0.8697
0.0143	8.1016	558	0.4099	0.8672	0.8666
0.1311	9.0852	610	0.4056	0.8755	0.8752

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.1+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1