Closed ghost closed 4 years ago
@deepseek
How many GPUs did you use?
It seems that your results are trained on the custom dataset.
We can guess what the problem is.
dataset
training environment
It is recommended that you should search for your optimal training setting by referring to the above mentions.
@youngwanLEE
train.py
Centermask2 still needs further improvement
Did you find a solution? I'm also using 1 GPU.
@youngwanLEE
centermask_V_39_eSE_FPN_ms_3x
andcentermask_lite_V_19_slim_dw_eSE_FPN_ms_4x
total_loss
doesn't stabilize below 0.6How to reproduce:
python train.py
inference.py
Both v19 and v39 get bad results, and won't decline further below ~
total_loss: 0.6
Samples