Luckeciano Melo

icon indicating a website link luckeciano.github.io icon indicating an email [email protected]

icon company University of Oxford icon location Investigating Intelligence. Primary focus on Deep Reinforcement Learning. PhD Student at University of Oxford.

Results 5 comments of Luckeciano Melo

Baseline results on v2 benchmarks

Hello @avnishn, Thanks for providing us your PR. Quick question: is it possible to also share the "progress.csv" files of the baseline runs? It would save a lot of time...

Is this reward function good for competition evaluation?

Hey @smsong, Actually, I commited a mistake. The v_tgt is not null at initial position (I saw a point in the map, but there is an arrow as well). I'm...

How to install mujoco-py in a server without root privileges?

Thank you so much @ezhang7423! Really appreciate your help. After installing these conda packages, I also had an issue to find lGL (`cannot find -lGL`). Mujoco-py README says to create...

[BUG]clip gradient not working

I am having the same issue with accelerate+deepspeed zero2. My config is: pytorch 2.5.1 accelerate 1.4.0 transformers 4.49.0 deepspeed 0.15.4 (also tested with 0.16.6, but the issue persists) I am...

[BUG]clip gradient not working

Ok, after some investigation I found out that, at least in my case, gradient clipping is happening, but transformers trainer is logging the wrong thing. In transformers' trainer, the grad...