Roman Vaxenburg

Results 34 comments of Roman Vaxenburg

Not a real solution, but the (hacky) workaround I'm using to restore checkpoints is this: ```python agent._learner._checkpointer._checkpoint.restore(ckpt) ``` where `agent` is an instance of, say, `acme.agents.tf.dmpo.DistributionalMPO` and `ckpt` looks like:...

I might be wrong, but can you try `ckpt_path = "ckpt-37"` instead (i.e., drop the `.index` part)?

This happens when the number of detected contacts is larger than the allocated contact buffer (as the error message says). It is explained, e.g., [here](https://mujoco.readthedocs.io/en/latest/XMLreference.html#size). One way to increase the...

Unlike the original single-site version, looks like the multiple site generalization doesn't support simultaneous position and orientation fitting, does it?

Thanks so much @saran-t. Unfortunately none of these helped. The only other hint I have is that the error usually appears later in the job, roughly after some 100 million...

No updates on my end. I'm still having this problem..

Thanks for responding @kevinzakka! I'm not using Docker, just a conda environment. I think the error occurs on every GPU I've tried so far: A100 SXM4, RTX 2080Ti, T4 PCIe.

Is there a task in the suite that uses vision (i.e., requires rendering)?

This one, on all machines: `| NVIDIA-SMI 525.60.13 Driver Version: 525.60.13 CUDA Version: 12.0 |`