Merge pull request #233 from kozistr/feature/lots-of-stuffs

[Feature] Lots of stuffs

Author: kozistr

README.md

@@ -13,6 +13,12 @@ Major version is updated! (`v2.12.0` -> `v3.0.0`) (#164)
 * Implement `GaLore` optimizer. (#224, #228)
   * [Memory-Efficient LLM Training by Gradient Low-Rank Projection](https://arxiv.org/abs/2403.03507)
 * Implement `Adalite` optimizer. (#225, #229)
+* Implement `bSAM` optimizer. (#212, #233)
+  * [SAM as an Optimal Relaxation of Bayes](https://arxiv.org/abs/2210.01620)
+* Implement `Schedule-Free` optimizer. (#230, #233)
+  * [Schedule-Free optimizers](https://github.com/facebookresearch/schedule_free)
+* Implement `EMCMC`. (#231, #233)
+  * [Entropy-MCMC: Sampling from flat basins with ease](https://www.semanticscholar.org/paper/Entropy-MCMC%3A-Sampling-from-Flat-Basins-with-Ease-Li-Zhang/fd95de3f24fc4f955a6fe5719d38d1d06136e0cd) 
 
 ### Fix
 
@@ -35,4 +41,5 @@ thanks to @sdbds, @i404788
 
 ## Diff
 
-[2.12.0...3.0.0](https://github.com/kozistr/pytorch_optimizer/compare/v2.12.0...v3.0.0)
+* from the previous major version : [2.0.0...3.0.0](https://github.com/kozistr/pytorch_optimizer/compare/v2.0.0...v3.0.0)
+* from the previous version: [2.12.0...3.0.0](https://github.com/kozistr/pytorch_optimizer/compare/v2.12.0...v3.0.0)

docs/changelogs/v3.0.0.md

@@ -96,6 +96,10 @@
     :docstring:
     :members:
 
+::: pytorch_optimizer.BSAM
+    :docstring:
+    :members:
+
 ::: pytorch_optimizer.CAME
     :docstring:
     :members:
@@ -236,6 +240,14 @@
     :docstring:
     :members:
 
+::: pytorch_optimizer.ScheduleFreeSGD
+    :docstring:
+    :members:
+
+::: pytorch_optimizer.ScheduleFreeAdamW
+    :docstring:
+    :members:
+
 ::: pytorch_optimizer.AccSGD
     :docstring:
     :members:

docs/index.md

@@ -84,6 +84,10 @@
     :docstring:
     :members:
 
+::: pytorch_optimizer.optimizer.utils.reg_noise
+    :docstring:
+    :members:
+
 ## Newton methods
 
 ::: pytorch_optimizer.optimizer.shampoo_utils.power_iteration

docs/optimizer.md

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "pytorch_optimizer"
-version = "2.12.0"
+version = "3.0.0"
 description = "optimizer & lr scheduler & objective function collections in PyTorch"
 license = "Apache-2.0"
 authors = ["kozistr <[email protected]>"]
@@ -12,13 +12,13 @@ documentation = "https://pytorch-optimizers.readthedocs.io/en/latest"
 keywords = [
     "pytorch", "deep-learning", "optimizer", "lr scheduler", "A2Grad", "ASGD", "AccSGD", "AdaBelief", "AdaBound",
     "AdaDelta", "AdaFactor", "AdaMax", "AdaMod", "AdaNorm", "AdaPNM", "AdaSmooth", "AdaHessian", "Adai", "Adalite",
-    "AdamP", "AdamS", "Adan", "AggMo", "Aida", "AliG", "Amos", "Apollo", "AvaGrad", "CAME", "DAdaptAdaGrad",
+    "AdamP", "AdamS", "Adan", "AggMo", "Aida", "AliG", "Amos", "Apollo", "AvaGrad", "bSAM", "CAME", "DAdaptAdaGrad",
     "DAdaptAdam", "DAdaptAdan", "DAdaptSGD", "DAdaptLion", "DiffGrad", "Fromage", "GaLore", "Gravity", "GSAM", "LARS",
     "Lamb", "Lion", "LOMO", "Lookahead", "MADGRAD", "MSVAG", "Nero", "NovoGrad", "PAdam", "PCGrad", "PID", "PNM",
-    "Prodigy", "QHAdam", "QHM", "RAdam", "Ranger", "Ranger21", "RotoGrad", "SAM", "SGDP", "Shampoo", "ScalableShampoo",
-    "SGDW", "SignSGD", "SM3", "SopihaH", "SRMM", "SWATS", "Tiger", "WSAM", "Yogi", "BCE", "BCEFocal", "Focal",
-    "FocalCosine", "SoftF1", "Dice", "LDAM", "Jaccard", "Bi-Tempered", "Tversky", "FocalTversky", "LovaszHinge",
-    "bitsandbytes",
+    "Prodigy", "QHAdam", "QHM", "RAdam", "Ranger", "Ranger21", "RotoGrad", "SAM", "ScheduleFreeSGD",
+    "ScheduleFreeAdamW", "SGDP", "Shampoo", "ScalableShampoo", "SGDW", "SignSGD", "SM3", "SopihaH", "SRMM", "SWATS",
+    "Tiger", "WSAM", "Yogi", "BCE", "BCEFocal", "Focal", "FocalCosine", "SoftF1", "Dice", "LDAM", "Jaccard",
+    "Bi-Tempered", "Tversky", "FocalTversky", "LovaszHinge", "bitsandbytes",
 ]
 classifiers = [
     "License :: OSI Approved :: Apache Software License",
@@ -50,7 +50,7 @@ bitsandbytes = { version = "^0.43", optional = true }
 
 [tool.poetry.dev-dependencies]
 isort = { version = "^5", python = ">=3.8" }
-black = { version = "^24", python = ">=3.8"}
+black = { version = "^24", python = ">=3.8" }
 ruff = "*"
 pytest = "*"
 pytest-cov = "*"

docs/util.md

@@ -79,7 +79,8 @@
 from pytorch_optimizer.optimizer.ranger import Ranger
 from pytorch_optimizer.optimizer.ranger21 import Ranger21
 from pytorch_optimizer.optimizer.rotograd import RotoGrad
-from pytorch_optimizer.optimizer.sam import GSAM, SAM, WSAM
+from pytorch_optimizer.optimizer.sam import BSAM, GSAM, SAM, WSAM
+from pytorch_optimizer.optimizer.schedulefree import ScheduleFreeAdamW, ScheduleFreeSGD
 from pytorch_optimizer.optimizer.sgd import ASGD, SGDW, AccSGD, SignSGD
 from pytorch_optimizer.optimizer.sgdp import SGDP
 from pytorch_optimizer.optimizer.shampoo import ScalableShampoo, Shampoo
@@ -186,6 +187,9 @@
     Aida,
     GaLore,
     Adalite,
+    BSAM,
+    ScheduleFreeSGD,
+    ScheduleFreeAdamW,
 ]
 OPTIMIZERS: Dict[str, OPTIMIZER] = {str(optimizer.__name__).lower(): optimizer for optimizer in OPTIMIZER_LIST}