ViViT_WLASL_250_epochs

This model is a fine-tuned version of google/vivit-b-16x2-kinetics400 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 4.0544
• Top 1 Accuracy: 0.2617
• Top 5 Accuracy: 0.5577
• Top 10 Accuracy: 0.6670
• Accuracy: 0.2617
• Precision: 0.2325
• Recall: 0.2617
• F1: 0.2253

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: 893000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Top 1 Accuracy	Top 5 Accuracy	Top 10 Accuracy	Accuracy	Precision	Recall	F1
30.5598	0.004	3572	7.6528	0.0010	0.0038	0.0064	0.0010	0.0008	0.0010	0.0004
29.9841	1.0040	7144	7.5548	0.0046	0.0120	0.0176	0.0046	0.0006	0.0046	0.0009
28.2597	2.0040	10716	7.2959	0.0125	0.0337	0.0495	0.0125	0.0053	0.0125	0.0048
26.1127	3.0040	14289	6.9165	0.0304	0.0748	0.1223	0.0301	0.0108	0.0301	0.0120
23.7044	4.004	17861	6.4996	0.0447	0.1407	0.2102	0.0447	0.0182	0.0447	0.0196
20.6604	5.0040	21433	6.0328	0.0822	0.2288	0.3121	0.0822	0.0421	0.0822	0.0434
17.6287	6.0040	25005	5.5622	0.1210	0.3041	0.4213	0.1210	0.0714	0.1210	0.0742
14.3215	7.0040	28578	5.0794	0.1576	0.3797	0.4951	0.1573	0.0998	0.1573	0.1038
10.5032	8.004	32150	4.6439	0.1915	0.4494	0.5695	0.1915	0.1353	0.1915	0.1386
7.2387	9.0040	35722	4.2461	0.2247	0.5123	0.6297	0.2255	0.1676	0.2255	0.1721
3.9708	10.0040	39294	3.9632	0.2485	0.5587	0.6701	0.2487	0.2034	0.2487	0.2046
2.1244	11.0040	42867	3.7748	0.2587	0.5753	0.6872	0.2587	0.2258	0.2587	0.2220
1.3992	12.004	46439	3.6907	0.2543	0.5794	0.6885	0.2543	0.2279	0.2543	0.2210
1.0175	13.0040	50011	3.7060	0.2503	0.5738	0.6874	0.2503	0.2176	0.2503	0.2142
0.914	14.0040	53583	3.6819	0.2648	0.5804	0.6915	0.2648	0.2380	0.2648	0.2311
0.7522	15.0040	57156	3.7360	0.2561	0.5758	0.6969	0.2564	0.2325	0.2564	0.2235
1.045	16.004	60728	3.7846	0.2638	0.5723	0.6877	0.2635	0.2470	0.2635	0.2327
0.8234	17.0040	64300	3.8910	0.2574	0.5692	0.6724	0.2572	0.2386	0.2572	0.2261
0.7311	18.0040	67872	4.0142	0.2561	0.5585	0.6680	0.2561	0.2402	0.2561	0.2262
1.0981	19.0040	71445	4.0544	0.2617	0.5577	0.6670	0.2617	0.2325	0.2617	0.2253

Framework versions

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