Roger Creus
Roger Creus
Experiencing the same when using LoRa requests...
How can we create a Slack account from rlltedevelo-msx2832.slack.com?
> Hi @yuyecreus, could you provide the Android phone you've used for the profiling, and more details about the steps you've done to get the profiling results in Android Studio?...
Coming back to this after quite a long time. Can you confirm there still exists this API mismatch? I think the Gym API > EnvPool API and I realized that...
@yooceii
> Roughly speaking, you should do > > ``` > eval_env.obs_rms = env.obs_rms > eval_env.return_rms = env.return_rms > ``` For everyone else, `env.envs[0].obs_rms` will access the Reward RMS wrapper because...
Hey! This looks pretty impressive! Just curious, what is the state of this PR?
Nice! Looking forward to the results
I'm curious about how it performs in other environments (e.g. atari?)
Still seeing this, only when using LoRa. I am currently using LLama3-8b, `tensor_parallel_size=8` and `max_model_len=1250`. The same run without using LoRa works flawlessly. This might be related: https://stackoverflow.com/questions/68106457/pytorch-cuda-error-an-illegal-memory-access-was-encountered The root...