tusimple / rl-multishot-reid Goto Github PK

In the second step of how to run the code, it is written about preparing data files and no such description is mentioned, could you please explain upon the same.

Hi Dear authors,
could you please upload your pre-trained model and the test demo?
Then we can avoid train this model repeatedly.
Thank you

Instead, mxnet/tools/im2rec.py is there, please explain how to modify the pre-processing file to run it successfully. I installed mxnet using pip install mxnet==0.11.0 and also tried version 1.0.0 but none of them have a bin folder.
Maybe you could help me by telling what mxnet version had you used?

when I run preprocess_prid_image.py,I got this problem.The folder recs shows Permission denied

I have tried training the RL phase model with the baseline training samples(recs) , but it seems that the training sample preparing is different between RL training and baseline, could you share the data preprocessing scripts for RL training thank you

For preparation of Mars data, following error is seen:
Traceback (most recent call last):
File "preprocess_mars_image.py", line 100, in
gen_train()
File "preprocess_mars_image.py", line 85, in gen_train
save_train(pool)
File "preprocess_mars_image.py", line 46, in save_train
ts = random.sample(pool, 96)
File "/usr/lib/python2.7/random.py", line 323, in sample
raise ValueError("sample larger than population")
ValueError: sample larger than population

I have tried by reducing the sample size (ts = random.sample(pool, 96)) but the issue pertains. Do i need to change pool size? Any suggestion please?
Regards,

Hi，I want to re-implement this code to a pytorch version do you have some suggestions?
about the state data, the memory and the agent？
Thank you so much,