Mava icon indicating copy to clipboard operation
Mava copied to clipboard

🦁 A research-friendly codebase for fast experimentation of multi-agent reinforcement learning in JAX

Results 160 Mava issues
Sort by recently updated
recently updated
newest added

### What do you want to test? Jax MAPPO full integration test ### Outline of test structure * Unit tests (if possible) * Test component interactions * Ensure not crashing...

test

**Please describe the purpose of the feature. Is it related to a problem? ** Implement a reverb dataset component in the Mava 2.0 (redesign) codebase. Describe the solution you'd like...

enhancement

### What do you want to test? Sorting utils (in `mava/utils/sort_utils`). ### Outline of test structure * Unit tests ### Definition of done Passing checks, cover edge cases. ### Mandatory...

test

### Describe the bug Even when the mode rgb_array is passed into rendering for the debug env it still displays a window, this can be a nuisance when running locally....

bug

### Describe the bug When we split the prng keys and generate new keys (`key, subkey = jax.random.split(key)`), we don't have a consistent way of using these new keys. We...

refactor

### Feature Custom checkpointers for `jax` systems ### Proposal @RuanJohn found a nice easy way to do this. ### Testing Unique tests will need to be created. Checkpointers should also...

enhancement

### Please describe what needs to be maintained? This is really messy and is all over the show in environment loops and loop wrappers, so it would be really nice...

maintenance

### Please describe what needs to be maintained? Investigate whether the use of enums is still relevant after Tensorflow support is deprecated. ### Describe the outcome you'd like When using...

Maintenance

## What? New components were added to use JAX Environments - these components are the Jax environment loop and the Jax executor environment loop. Additionally, a seed value can now...

size/L

Hi, I found you use reverb to store data. For distributed training, is reverb fast? For example, how much does time cost to get a batch from the remote replay...