Which results in paper correspond to the finetune command? about frozen-in-time HOT 5 CLOSED

This is finetuning from imagenet initialisation, to finetune from the pretrained model download it
https://www.robots.ox.ac.uk/~maxbain/frozen-in-time/models/cc-webvid2m-4f_stformer_b_16_224.pth.tar
Then add
"load_checkpoint": PATH_TO_DOWNLOADED_MODEL
to configs/msrvtt_4f_i21k.json

from frozen-in-time.

Thanks! I will try to implement the pretaining with this checkpoint.

BTW, I have finetuned the checkpoint WebVid2M+CC3M+COCO, 4-frames, base_patch_16_224. I achieve:

    epoch          : 27
    loss_0         : 0.5593496548643809
    val_loss_0     : 1.0900163816081152
    val_0_t2v_metrics_R1: 28.4
    val_0_t2v_metrics_R5: 55.8
    val_0_t2v_metrics_R10: 67.5
    val_0_t2v_metrics_R50: 88.6
    val_0_t2v_metrics_MedR: 4.0
    val_0_t2v_metrics_MeanR: 29.498
    val_0_t2v_metrics_geometric_mean_R1-R5-R10: 47.46994959864713
    val_0_v2t_metrics_R1: 28.6
    val_0_v2t_metrics_R5: 56.2
    val_0_v2t_metrics_R10: 68.4
    val_0_v2t_metrics_R50: 88.9
    val_0_v2t_metrics_MedR: 4.0
    val_0_v2t_metrics_MeanR: 26.8585
    val_0_v2t_metrics_geometric_mean_R1-R5-R10: 47.90558524371724
Validation performance didn't improve for 10 epochs. Training stops.

Are these results correct? The R1 is slightly lower than reported values (31.0). Is this regular fluctuation?

from frozen-in-time.

I got the similar results as in paper with batch_size = 64.

from frozen-in-time.

run the test script
test.py --resume PATH_TO_FINETUNED_CHECKPOINT --sliding_window_stride 12

the sliding window stride argument adds temporal averaging over multiple frame samples :)

from frozen-in-time.

Exactly! I got further improvements with this testing script. R1 comes to 33.7.

from frozen-in-time.