flowMC-0.4.5

What's Changed

• Update Doc by @kazewong in #228
• Add final_log_prob return value to AdamOptimization.optimize method by @thomasckng in #229
• Batch take step by @kazewong in #232

Full Changelog: flowMC-0.4.4...flowMC-0.4.5

flowMC-0.4.4

1. This release introduces the parallel tempering variant of the RQSpline_MALA bundle
2. Relax the strict dependency version pinning mentioned in #218
3. Add state resource and lambda strategy #210
4. Update the float64 support to remove warnings raised by jax deprecation warning.
5. flowMC-0.4.3 was tanked from Pypi due to a bug in parallel tempering. Skip to version 0.4.4

What's Changed

• Sync by @kazewong in #219
• 216 request to relax strict dependency version pinning by @kazewong in #218
• 210 add states resource by @kazewong in #220
• Make pt resource strategy bundle by @kazewong in #222
• Jim sync by @kazewong in #224
• flowMC 0.4.3 update by @kazewong in #225
• reverse tempering direction by @kazewong in #227

Full Changelog: flowMC-0.4.2...flowMC-0.4.4

flowMC-0.4.2

A minor release that allows float64 sampling

What's Changed

• docs: add HajimeKawahara as a contributor for bug by @allcontributors in #211
• Add more flexible casting instead of fixing float32 by @kazewong in #209
• docs: add matt-graham as a contributor for bug, test, and 2 more by @allcontributors in #213
• docs: add daniel-dodd as a contributor for doc, review, and 2 more by @allcontributors in #212
• Contributors by @kazewong in #214
• update feat: update contributor profiles and contributions in .all-co… by @kazewong in #215

New Contributors

• @allcontributors made their first contribution in #211

Full Changelog: flowMC-0.4.1...flowMC-0.4.2

flowMC-0.4.1

This PR provides a couple of enhancements:

1. LogPDF is now a resource, meaning in the more advance case, the users can specify multiple logPDFs. This helps use cases such as discrete-continuous sampling as well as parallel tempering.
2. Add parallel tempering strategy
3. Strategies within the sampler is now promoted to be a dictionary instead of a list, making strategies to be reusable without recompilation

What's Changed

• Sync by @kazewong in #203
• 204 making logpdf a resource by @kazewong in #205
• 202 add parallel tempering strategy by @kazewong in #208
• 206 make strategies reusable without recompiling by @kazewong in #207

Full Changelog: flowMC-0.4.0...flowMC-0.4.1

flowMC-0.4.0

This is a major release of flowMC. The main changes includes the addition of a new intermediate API centers around the resource-strategy design, where users can customize the sampling algorithm to employ more complex algorithm. There are some breaking changes in the top level API but it should be relatively straightforward to migrate from the older flowMC version to this version.

What's Changed

• 170 lower precision training by @kazewong in #171
• minor fix of the example using a non-jax likelihood by @marylou-gabrie in #176
• New script to generate useful plots by @Qazalbash in #179
• Update Sampler and NFModel classes by @kazewong in #182
• Update optax requirement and remove testing print statement by @thomasckng in #180
• 188 updating project infrastructures by @kazewong in #189
• 190 updating documentation to align with the latest version of flowmc by @kazewong in #196
• 193 improve coverage of flowmc by @kazewong in #197
• 198 generalize buffer by @kazewong in #199

New Contributors

• @Qazalbash made their first contribution in #179
• @thomasckng made their first contribution in #180

Full Changelog: flowMC-0.3.4...flowMC-0.4.0

flowMC-0.3.4

What's Changed

• update example and add print summary to Sampler by @ThibeauWouters in #169
• The previous load model function does not correctly load the model by @kazewong in #172

Full Changelog: flowMC-0.3.3...flowMC-0.3.4

flowMC-0.3.3

What's Changed

• 165 add probability floor to normalizing flow model -> updating Adam optimizer to optimize with constraint by @kazewong in #168

Full Changelog: flowMC-0.3.2...flowMC-0.3.3

flowMC-0.3.2

Optimization strategy is included in this version.

Also fixed a bug in the flow model that prevent training

What's Changed

• 163 implement optimization strategy by @kazewong in #164

Full Changelog: flowMC-0.3.1...flowMC-0.3.2

flowMC-0.3.1

Fixing training bugs and sampling bugs. Previous version incorrectly also start at initial position instead of actually stepping during sampling. This is fixed and tested with community example now.

Full Changelog: flowMC-0.3.0...flowMC-0.3.1

flowMC-0.3.0

The version release focuses on cleaning up the internal API, which should makes the study of auto-tuning way easier now. The top level API remains relatively unchanged. On the performance side, most noticeably the compilation time of the normalizing flow has been reduced significantly.

For the next version of flowMC (0.4.0), the main focus will be some sort of auto-tune and adding more potential strategies in the mix, including 1. using importance sampling with normalizing flow to further improve the sampling results 2. Start the normalizing flow with some variational inference scheme instead of from scratch.

What's Changed

• Bayeux example by @kazewong in #146
• Added plotting utilities and refactored hyperparams by @ThibeauWouters in #143
• small bug fix in keys for stepping optimizer by @tedwards2412 in #148
• Sync branch 134 by @kazewong in #150
• 134 add precommit script and fix formatting by @kazewong in #151
• 144 get rid of random key set by @kazewong in #152
• 153 use scan to reduce nf compilation time by @kazewong in #154
• Sync branch 142 by @kazewong in #155
• 142 sampling from arrays by @kazewong in #156
• Update README.md by @kazewong in #159
• 158 put training loop into nf class by @kazewong in #160
• 157 making sampler composable by @kazewong in #161

New Contributors

• @ThibeauWouters made their first contribution in #143
• @tedwards2412 made their first contribution in #148

Full Changelog: flowMC-0.2.4...flowMC-0.3.0