Edit model card

vivit-surf-analytics-runpod

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: 0.7609
  • Accuracy: 0.9163
  • F1: 0.9154

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: 1
  • eval_batch_size: 1
  • 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: 22230

Training results

Training Loss Epoch Step Validation Loss Accuracy F1
0.0 0.0333 741 0.9082 0.9070 0.9068
0.5976 1.0333 1482 2.5471 0.7302 0.7286
0.1188 2.0333 2223 1.0145 0.8698 0.8695
0.0001 3.0333 2964 1.1956 0.8465 0.8384
0.5026 4.0333 3705 0.8190 0.8651 0.8608
0.0 5.0333 4446 1.1466 0.8372 0.8377
0.0 6.0333 5187 1.0804 0.8419 0.8358
0.0 7.0333 5928 0.8535 0.8930 0.8909
0.0 8.0333 6669 0.6512 0.9070 0.9070
0.0001 9.0333 7410 0.8475 0.8884 0.8887
0.0001 10.0333 8151 0.7335 0.8977 0.8972
0.0 11.0333 8892 0.7774 0.9070 0.9054
0.0 12.0333 9633 0.7346 0.9116 0.9107
0.0 13.0333 10374 0.7609 0.9163 0.9154
0.0 14.0333 11115 0.7560 0.9070 0.9074
0.0 15.0333 11856 0.8036 0.9163 0.9151
0.0 16.0333 12597 0.7962 0.9163 0.9151
0.0 17.0333 13338 0.7821 0.9163 0.9147
0.0 18.0333 14079 0.7898 0.9163 0.9149
0.0 19.0333 14820 1.0166 0.8791 0.8748
0.0 20.0333 15561 0.8697 0.8977 0.8968
0.0 21.0333 16302 0.7670 0.9023 0.9017
0.0 22.0333 17043 0.7399 0.9116 0.9107
0.0 23.0333 17784 0.7458 0.9116 0.9107
0.0 24.0333 18525 0.7701 0.8977 0.8969
0.0 25.0333 19266 0.7924 0.9023 0.9014
0.0 26.0333 20007 0.7955 0.9023 0.9014
0.0 27.0333 20748 0.8675 0.8977 0.8969
0.0 28.0333 21489 0.8671 0.8977 0.8969
0.0 29.0333 22230 0.8665 0.8977 0.8969

Framework versions

  • Transformers 4.41.2
  • Pytorch 2.3.1+cu121
  • Datasets 2.19.2
  • Tokenizers 0.19.1
Downloads last month
1
Safetensors
Model size
88.6M params
Tensor type
F32
·
This model can be loaded on Inference API (serverless).

Finetuned from