videomae-base-finetuned-good-gestureUnitV12

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:

• Accuracy: 0.8966
• Loss: 0.2937
• Accuracy Gunit: 0.8333
• Accuracy Nothing: 0.9556

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• 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: 160

Training results

Training Loss	Epoch	Step	Accuracy	Validation Loss	Accuracy Gunit	Accuracy Nothing
0.8549	0.1062	17	0.5714	0.6774	0.68	0.4839
0.6437	1.1062	34	0.4643	0.7254	1.0	0.0323
0.6226	2.1063	51	0.6071	0.6527	0.96	0.3226
0.5883	3.1063	68	0.5714	0.6389	1.0	0.2258
0.5136	4.1063	85	0.6964	0.5330	0.84	0.5806
0.4284	5.1063	102	0.8214	0.4506	0.84	0.8065
0.3474	6.1063	119	0.8214	0.3974	0.76	0.8710
0.2859	7.1063	136	0.8214	0.3822	0.64	0.9677
0.3059	8.1062	153	0.8393	0.3763	0.68	0.9677
0.2582	9.0437	160	0.8393	0.3738	0.68	0.9677

Framework versions

• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1