simonsays1980

Using IAM Role with MFA for secure access

6

I am trying now for hours to access my Athena database (which works charmingly via PyAthena in Python). My setup looks as follows: - I use `AWR.Athena` version `2.0.7.0` -...

Dear Frank, _first of all:_ thanks for providing your package to the community! `Hmisc` has spared me a lot of time transferring numbers out of `R` output into LateX tables....

Rnd with extra value heads

1

## Why are these changes needed? Exploration algorithms become more important as we move to more complex environments that have difficult reward functions (e.g. sparse ones). Some exploration algorithms have...

Motivated by a larger error in my project and inspired by the `VecCheckNan` class in _stable-baselines_ I wrote an environment wrapper for the `BaseEnv` in RLlib that can check numerics...

### Description Following the discussion in #30471 (specifically [this comment](https://github.com/ray-project/ray/issues/30471#issuecomment-1361775334)) and much checking over the last years, I would like to see some tools that simplify tracing an error back...

## Why are these changes needed? I ran into problems when using the `Offline API` with the older `JsonWriter` setup when trying to use the `output_config` with a custom output...

[RLlib] Add total metrics to a response for `Commands.REPORT_SAMPLES`

1

### Description At the moment the `PolicyServerInput` sends *no response* when the `PolicyClient` sends the samples back (see the [code](https://github.com/ray-project/ray/blob/87f22e106575e7eaf9c8004a5427cf7c2232b2c1/rllib/env/policy_server_input.py#L254)). The problem is, if multiple clients are connected, no client...

## Why are these changes needed? This is the third part for moving the `DQN Rainbow` algorithm to our new stack and towards using `EnvRunner API`. See for the other...

## Why are these changes needed? Random n-step sampling introduces an additional source of variablity into experience replay and should therefore increase robustness. ## Related issue number Closes #43299 ##...

## Why are these changes needed? `synchronous_parallel_sample` offers failure tolerant sampling, but was so far not usable when sampling episodes. This PR closes this gap. ## Related issue number ##...