Flashbax

Latest version: v0.1.3

Safety actively analyzes 722491 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

0.1.3

Summary
- Fixed unexpected behavior when sampling from trajectory buffer and prioritised trajectory buffer.
- Added ability to sample from multiple buffers (mixed experience replay).
- Fixed some checks and test cases.

What's Changed
* Fix Jumanji environment spec access in examples by rk1a in https://github.com/instadeepai/flashbax/pull/24
* Update README.md by EdanToledo in https://github.com/instadeepai/flashbax/pull/25
* fix: linter for readme by EdanToledo in https://github.com/instadeepai/flashbax/pull/28
* fix: be explicit about squeeze dim in prioritised sampling to avoid flattening (1,1) arrays by callumtilbury in https://github.com/instadeepai/flashbax/pull/27
* feat: add a prioritised item buffer by EdanToledo in https://github.com/instadeepai/flashbax/pull/19
* chore: Type ignore on max_length_time_axis by callumtilbury in https://github.com/instadeepai/flashbax/pull/35
* feat: simple utility to get timestep count by callumtilbury in https://github.com/instadeepai/flashbax/pull/34
* fix: error when vault's rel_dir is a parent directory by callumtilbury in https://github.com/instadeepai/flashbax/pull/32
* jax.tree_map -> jax.tree.map by garymm in https://github.com/instadeepai/flashbax/pull/38
* feat: Mixed Experience Replay 🤝 by callumtilbury in https://github.com/instadeepai/flashbax/pull/30
* prioritised_trajectory_buffer: allow adding max_length_time_axis by garymm in https://github.com/instadeepai/flashbax/pull/40
* Treat warnings as errors by mickvangelderen in https://github.com/instadeepai/flashbax/pull/41
* Pass max_length_time_axis instead of max_size by mickvangelderen in https://github.com/instadeepai/flashbax/pull/43
* Address `sample_sequence_length` greater than `min_length_time_axis` by mickvangelderen in https://github.com/instadeepai/flashbax/pull/45
* Fix: Overflow error in PER by EdanToledo in https://github.com/instadeepai/flashbax/pull/48
* Fix/core trajectory buffer by EdanToledo in https://github.com/instadeepai/flashbax/pull/50
* fix:prioritised trajectory buffer by EdanToledo in https://github.com/instadeepai/flashbax/pull/51
* fix: update release.yml by SimonDuToit in https://github.com/instadeepai/flashbax/pull/54
* chore: bump version by SimonDuToit in https://github.com/instadeepai/flashbax/pull/55

New Contributors
* rk1a made their first contribution in https://github.com/instadeepai/flashbax/pull/24
* garymm made their first contribution in https://github.com/instadeepai/flashbax/pull/38
* mickvangelderen made their first contribution in https://github.com/instadeepai/flashbax/pull/41
* SimonDuToit made their first contribution in https://github.com/instadeepai/flashbax/pull/54

**Full Changelog**: https://github.com/instadeepai/flashbax/compare/v0.1.2...v0.1.3

0.1.2

What's Changed
* fix: improve the compression api of vaults by callumtilbury in https://github.com/instadeepai/flashbax/pull/22


**Full Changelog**: https://github.com/instadeepai/flashbax/compare/v0.1.1...v0.1.2

0.1.1

What's Changed
* fix: mkdocs render issue of asterisk by callumtilbury in https://github.com/instadeepai/flashbax/pull/12
* feat: Vault Compression by callumtilbury in https://github.com/instadeepai/flashbax/pull/14
* add: IQL Example by JemmaLDaniel in https://github.com/instadeepai/flashbax/pull/13
* fix: incorrect current index after a Vault read. by callumtilbury in https://github.com/instadeepai/flashbax/pull/20
* chore: bump version to 0.1.1 by callumtilbury in https://github.com/instadeepai/flashbax/pull/21

New Contributors
* JemmaLDaniel made their first contribution in https://github.com/instadeepai/flashbax/pull/13

**Full Changelog**: https://github.com/instadeepai/flashbax/compare/v0.1.0...v0.1.1

0.1.0

What's Changed
* New Buffer: Item Buffer - Simple and easy to use. Dedicated for independent data items.
* Vault: Vault is an efficient mechanism for saving Flashbax buffers to persistent data storage, e.g. for use in offline reinforcement learning.

**Full Changelog**: https://github.com/instadeepai/flashbax/compare/v0.0.1...v0.1.0

0.0.1

**Full Changelog**: https://github.com/instadeepai/flashbax/commits/v0.0.1

Links

Releases

© 2025 Safety CLI Cybersecurity Inc. All Rights Reserved.