search

cvlab-stonybrook / LearningToCountEverything

MIT License
357 stars 72 forks source link

Experiments on CARPK #32

Open zyweven opened 1 year ago

zyweven commented 1 year ago

Hello I use the pre-trained model provided by you to conduct shot-3 experiments on CARPK, and the result is MAE: 43.02, RMSE: 58.12. Is this normal? How many examples are used in the CARPK experiment in the paper?

Viresh-R commented 1 year ago

Hey, the CARPK result mentioned in the paper uses 12 exemplars(as mentioned in the paper). We randomly sample 12 exemplars from different training images, and use this fixed set of exemplars for the evaluation. Also, we do not do the test time adaption for the pertained model. I don't recall the exact 3-shot results, but it was definitely worse than the 12-shot result.

donggoing commented 9 months ago

@Viresh-R Could you please release the train and inference code for CARPK?