videomae_kinetics_wlasl_2000_20ep_coR

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

• Loss: 2.5193
• Accuracy: 0.4382

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: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• 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: 35720
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
30.4461	0.05	1786	7.5731	0.0020
28.8085	1.0500	3572	6.7152	0.0240
24.6341	2.0500	5358	5.7700	0.0827
20.2204	3.0500	7145	4.9529	0.1581
16.0475	4.05	8931	4.2523	0.2296
12.2515	5.0500	10717	3.6646	0.3044
9.0329	6.0500	12503	3.2212	0.3488
6.4841	7.0500	14290	2.9584	0.3739
4.5533	8.05	16076	2.7372	0.3996
3.1616	9.0500	17862	2.6446	0.4119
2.219	10.0500	19648	2.5457	0.4254
1.5975	11.0500	21435	2.5348	0.4201
1.2009	12.05	23221	2.5232	0.4270
0.9648	13.0500	25007	2.5305	0.4277
0.7877	14.0500	26793	2.5389	0.4231
0.6812	15.0500	28580	2.5198	0.4349
0.5795	16.05	30366	2.5196	0.4331
0.5066	17.0500	32152	2.5130	0.4336
0.453	18.0500	33938	2.5175	0.4385
0.3928	19.0499	35720	2.5193	0.4382

Framework versions

• Transformers 4.46.1
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.20.1