How should i get correspond tensorboard graph ? about forex-dqn

[noootown]

Hi can you tell me more about your question?
Feel free to write in chinese

[yuanzhanghu]

谢谢！我怎么才能通过程序运行， 得到你贴出来的那些图？

[yuanzhanghu]

Also i didn't see any DQN related code in your RLBrain file. Could you introduce a bit about your algorithm ? Thanks!

[noootown]

You should be able to found some information about DQN on the internet. I don't think my code is too complicate too understand and please remember, Google is your best friend.

[noootown]

About the first question, please runpython3 train/index.py -h and you'll see help info, thanks

[yuanzhanghu]

From your RLBrain file, i didn't see you were using CNN layer. That's why i said i didn't see DQN code. maybe i missed something in your code ?
For first question, i got this from 'python3 train/index.py -h':
yhu@yhu-Aspire-M3920:~/workspace/deeplearning/Forex-DQN$ python3 train/index.py -h │
python3 train/index.py │
-d

: assign the output directory to data/ │
-h --help: help │
-l : load ckpt file from /history/ │
-o: command line output result │
-t: testing mode, default: training mode

Which one can i use to generate tensorboard graph ? Sorry maybe i have too many questions ?

[yuanzhanghu]

Sorry, i am still new to AI area, just figured out how to use tensorboard. Just ignore my first question please. Still trying to understand your algo... 👍

[noootown]

Sorry, I was too busy and didn't look into your first question, to generate tensorboard, you must run tensorboard --logdir=XXX, please checkout tensorboard API.
And as for DQN, it is not a CNN algorithm, I think you should find more informations about DQN and understand it.

If you are a beginner and trying to look into AI or deep learning, I think my project is not a good starter kit since it contains advance algorithm like DQN. You can checkout 莫煩's video.

加油 :)