rlpyt

rlpyt copied to clipboard

Bias in EvalCollector

4

`EvalCollector` (e.g. `GpuEvalCollector`) seems to bias towards short trajectories, leading to inaccurate evaluation. The problem occurs when `eval_n_envs > 1`. In [GpuEvalCollector.collect_evaluation()](https://github.com/astooke/rlpyt/blob/master/rlpyt/samplers/parallel/gpu/collectors.py#L125), the list of `TrajInfoCls` is created as a...

kaixin96

DistributedDataParallelCPU is deprecated

1

Hi, in many *_agent.py files you use following import: > from torch.nn.parallel import DistributedDataParallelCPU as DDPC and in PyTorch 1.2 we have the following code: > def DistributedDataParallelCPU(*args, **kwargs): >...

codelast

dear author: I have met this problem how can I slove it? ``` rlpyt.projects.qcpo.qcpo_model.py | normalize_observation : True rlpyt.projects.qcpo.qcpo_model.py | normalize_observation : True rlpyt.projects.qcpo.qcpo_model.py | normalize_observation : True D:\Anaconda\envs\qcpo1\lib\site-packages\rlpyt\samplers\buffer.py:3: UserWarning:...

lalala215

I am trying to load a replaybuffer.pkl file for the launch_dmc_ac_pretrain_1.py script in rlpyt/ul/experiments_for_rl/scripts/dmcontrol/launch_ul/. However, I found that there is no existing file, so I created an empty pkl-file. But...

Gechenseu