videomae-base-finetuned-ucf101-subset-face

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:

• Loss: 2.5315
• Accuracy: 0.6389

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: 3
• eval_batch_size: 3
• 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: 2000

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.8253	9.005	100	1.7977	0.1667
1.6427	19.005	200	2.1211	0.1667
1.0603	29.005	300	2.5370	0.1944
0.6361	39.005	400	1.9683	0.4444
0.7149	49.005	500	2.8125	0.3889
0.3396	59.005	600	2.2497	0.5556
0.3026	69.005	700	1.7178	0.6389
0.3043	79.005	800	2.5029	0.6111
0.1636	89.005	900	2.7748	0.6111
0.1292	99.005	1000	2.1868	0.6389
0.5229	109.005	1100	2.4543	0.6111
0.0016	119.005	1200	1.7452	0.75
0.0013	129.005	1300	2.5026	0.6111
0.0011	139.005	1400	2.3153	0.6389
0.0011	149.005	1500	1.7536	0.75
0.0028	159.005	1600	2.5384	0.6389
0.0605	169.005	1700	2.6368	0.6111
0.2064	179.005	1800	2.3678	0.6667
0.0013	189.005	1900	2.4561	0.6389
0.0009	199.005	2000	2.5315	0.6389

Framework versions

• Transformers 4.45.0
• Pytorch 2.4.1+cu118
• Datasets 3.0.0
• Tokenizers 0.20.0