videomae-base-finetuned-bleeding-exp_0

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: 1.4958
• Accuracy: 0.5

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: 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: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.04	2	0.6967	0.5
No log	1.04	4	0.6799	0.75
No log	2.04	6	0.6721	0.75
No log	3.04	8	0.6742	0.75
0.6424	4.04	10	0.6927	0.25
0.6424	5.04	12	0.7295	0.5
0.6424	6.04	14	0.8047	0.5
0.6424	7.04	16	0.8589	0.5
0.6424	8.04	18	0.8842	0.5
0.6123	9.04	20	0.9349	0.5
0.6123	10.04	22	0.9543	0.5
0.6123	11.04	24	0.9924	0.5
0.6123	12.04	26	1.0729	0.5
0.6123	13.04	28	1.2268	0.5
0.3641	14.04	30	1.3759	0.5
0.3641	15.04	32	1.4344	0.5
0.3641	16.04	34	1.4563	0.5
0.3641	17.04	36	1.4365	0.5
0.3641	18.04	38	1.4343	0.5
0.4378	19.04	40	1.4375	0.5
0.4378	20.04	42	1.4530	0.5
0.4378	21.04	44	1.4732	0.5
0.4378	22.04	46	1.4877	0.5
0.4378	23.04	48	1.4919	0.5
0.222	24.04	50	1.4958	0.5

Framework versions

• Transformers 4.40.2
• Pytorch 1.12.0
• Datasets 2.19.1
• Tokenizers 0.19.1