Fail to run the example #5
Description
Hi @felix0901, i follow your tutorial and try to ./run_ConfX.sh in my conda virtual environment, but it seems that there are some bugs in your given example. Could you help me to check that and give me some help?
Max constraint: 31782000000000.0 Min constraint: 726440000000.0 Set constraint: 16254220000000.0 Epoch 0: New best reward: -1.756339e+07 Episode 1: succeed Episode 2: succeed Epoch 2: New best reward: -1.042791e+07 Episode 3: succeed Epoch 3: New best reward: -8.091650e+06 Episode 4: succeed Episode 5: succeed Episode 6: succeed Episode 7: succeed Episode 8: succeed Episode 9: succeed Episode 10: succeed Episode 11: finding ++++++++++++++++++++ 96 8 96 ++++++++++++++++++++ /.../confuciux/src/ConfX/rl_confx.py:144: RuntimeWarning: invalid value encountered in subtract rewards = (rewards - rewards.mean()) / (rewards.std() + EPISIOLON) /.../.conda/envs/confxEnv/lib/python3.6/site-packages/numpy/core/_methods.py:193: RuntimeWarning: invalid value encountered in subtract x = asanyarray(arr - arrmean) Episode 12: finding Traceback (most recent call last): File "main.py", line 241, in <module> scores = policy_graident(n_episodes=epoch_rl, outfile=outfile, chkpt_file=chkpt_file, eps=0.0, temperature=1) File "main.py", line 45, in policy_graident action, log_prob = agent.act(state, infos, eps,temperature) File "/.../confuciux/src/ConfX/rl_confx.py", line 88, in act action =m.sample() File "/.../.conda/envs/confxEnv/lib/python3.6/site-packages/torch/distributions/categorical.py", line 106, in sample samples_2d = torch.multinomial(probs_2d, sample_shape.numel(), True).T RuntimeError: invalid multinomial distribution (encountering probability entry < 0)
Thank you!