timesformer-new-finetuned-yt-sample_rate16

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: 0.6420
• Precision: 0.8440
• Recall: 0.7083
• F1-score: 0.7702
• Accuracy: 0.7820

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 7000

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1-score	Accuracy
0.5503	0.05	350	0.4886	0.7819	0.8214	0.8012	0.7897
0.4942	1.05	700	0.4466	0.8	0.8214	0.8106	0.8020
0.4497	2.05	1050	0.4485	0.8075	0.8051	0.8063	0.8005
0.3557	3.05	1400	0.5576	0.8791	0.6384	0.7397	0.7682
0.2741	4.05	1750	0.5533	0.8611	0.6920	0.7673	0.7836
0.2634	5.05	2100	0.7618	0.9271	0.5298	0.6742	0.7360
0.1744	6.05	2450	0.6420	0.8440	0.7083	0.7702	0.7820

Framework versions

• Transformers 4.46.3
• Pytorch 2.0.0+cu117
• Datasets 3.1.0
• Tokenizers 0.20.3