Issue with Zero Success Rate and Automatic Interruption during Training for 'microwave' Grasping Task

[yixi1266]

hi, I am encountering two issues:

Zero Success Rate: Despite running the training for a considerable amount of time, the success rate for the task remains consistently zero.

Automatic Interruption: The training process is being automatically interrupted before completion, even though the model is still running.

[fitz0401]

It seems that I encountered the same problem (psl_kitchen.py)

[xcvbn123456]

hi, I am encountering two issues:

Zero Success Rate: Despite running the training for a considerable amount of time, the success rate for the task remains consistently zero.

Automatic Interruption: The training process is being automatically interrupted before completion, even though the model is still running.
Did you solve this problem?

[xcvbn123456]

It seems that I encountered the same problem (psl_kitchen.py)

Did you solve this problem?

[xcvbn123456]

It seems that I encountered the same problem (psl_kitchen.py)

您好，请问后来可以训练起来吗？我遇到了类似的0成功率问题

[Flxiang1]

It seems that I encountered the same problem (psl_kitchen.py)

Hello, I would like to know how you generated images for train and eval, and why my images only have GPU and CPU

[xcvbn123456]

It seems that I encountered the same problem (psl_kitchen.py)

Hello, I would like to know how you generated images for train and eval, and why my images only have GPU and CPU

Maybe your code has not been trained to generate logs. It took me about 4000 rounds to see the curve.Or you can check whether the rlkit-related packages and functions in the train.py file are imported normally.

[xcvbn123456]

It seems that I encountered the same problem (psl_kitchen.py)

Hello, I would like to know how you generated images for train and eval, and why my images only have GPU and CPU

Maybe your code has not been trained to generate logs. It took me about 4000 rounds to see the curve.Or you can check whether the rlkit-related packages and functions in the train.py file are imported normally.

[Flxiang1]

It seems that I encountered the same problem (psl_kitchen.py)

Hello, I would like to know how you generated images for train and eval, and why my images only have GPU and CPU

Maybe your code has not been trained to generate logs. It took me about 4000 rounds to see the curve.Or you can check whether the rlkit-related packages and functions in the train.py file are imported normally.

Okay, thank you very much. I have already formed a curve.

[xcvbn123456]

It seems that I encountered the same problem (psl_kitchen.py)

Hello, I would like to know how you generated images for train and eval, and why my images only have GPU and CPU

Maybe your code has not been trained to generate logs. It took me about 4000 rounds to see the curve.Or you can check whether the rlkit-related packages and functions in the train.py file are imported normally.

Okay, thank you very much. I have already formed a curve.

Have you ever encountered the problem that the code interrupts and jumps to the next task when training the third task (light-v0)?

[Flxiang1]

It seems that I encountered the same problem (psl_kitchen.py)

Hello, I would like to know how you generated images for train and eval, and why my images only have GPU and CPU

Maybe your code has not been trained to generate logs. It took me about 4000 rounds to see the curve.Or you can check whether the rlkit-related packages and functions in the train.py file are imported normally.

Okay, thank you very much. I have already formed a curve.

Have you ever encountered the problem that the code interrupts and jumps to the next task when training the third task (light-v0)?

Sorry, I haven't reached the third task yet. Is there any error message output when the task is interrupted?

[xcvbn123456]

It seems that I encountered the same problem (psl_kitchen.py)

Hello, I would like to know how you generated images for train and eval, and why my images only have GPU and CPU

Maybe your code has not been trained to generate logs. It took me about 4000 rounds to see the curve.Or you can check whether the rlkit-related packages and functions in the train.py file are imported normally.

Okay, thank you very much. I have already formed a curve.

Have you ever encountered the problem that the code interrupts and jumps to the next task when training the third task (light-v0)?

抱歉，我还没有完成第三个任务。任务中断时是否有错误消息输出？

When my code runs the third task, the code runs to a certain round and outputs this error ：

ValueError: Caught ValueError in DataLoader worker process 3.
Original Traceback (most recent call last):
File "/home/cqy/anaconda3/envs/psl/lib/python3.8/site-packages/torch/utils/data/_utils/worker.py", line 302, in _worker_loop
data = fetcher.fetch(index)
File "/home/cqy/anaconda3/envs/psl/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 32, in fetch
data.append(next(self.dataset_iter))
File "/home/cqy/planseqlearn/planseqlearn/replay_buffer.py", line 235, in iter
yield self._sample()
File "/home/cqy/planseqlearn/planseqlearn/replay_buffer.py", line 212, in _sample
idx = np.random.randint(0, episode_len(episode) - self._nstep + 1) + 1
File "mtrand.pyx", line 748, in numpy.random.mtrand.RandomState.randint
File "_bounded_integers.pyx", line 1247, in numpy.random._bounded_integers._rand_int64
ValueError: high <= 0

And then The code jumps out of the third task to the fourth task on its own.
When you run the third task, please check to see if you get the same error.

[xcvbn123456]

It seems that I encountered the same problem (psl_kitchen.py)

Hello, I would like to know how you generated images for train and eval, and why my images only have GPU and CPU

Maybe your code has not been trained to generate logs. It took me about 4000 rounds to see the curve.Or you can check whether the rlkit-related packages and functions in the train.py file are imported normally.

Okay, thank you very much. I have already formed a curve.

Have you ever encountered the problem that the code interrupts and jumps to the next task when training the third task (light-v0)?

Sorry, I haven't reached the third task yet. Is there any error message output when the task is interrupted?

when training the third task (light-v0)，Have you faced some problems?

[LeMaoLeMao]

It seems that I encountered the same problem (psl_kitchen.py)

您好，请问后来可以训练起来吗？我遇到了类似的0成功率问题

This is sparse reward environment, the only thing the policy did is to take random actions near the object and accidentally complete the task. Therefore, it is possible that a run has 0 success rate, the solution is to run across multiple seeds and report average success rate curve