Merge pull request #318 from kozistr/feature/grams-optimizer

[Feature] Implement `Grams` optimizer

README.md

@@ -10,7 +10,7 @@
 
 ## The reasons why you use `pytorch-optimizer`.
 
-* Wide range of supported optimizers. Currently, **86 optimizers (+ `bitsandbytes`, `qgalore`, `torchao`)**, **16 lr schedulers**, and **13 loss functions** are supported!
+* Wide range of supported optimizers. Currently, **87 optimizers (+ `bitsandbytes`, `qgalore`, `torchao`)**, **16 lr schedulers**, and **13 loss functions** are supported!
 * Including many variants such as `Cautious`, `AdamD`, `Gradient Centrailiaztion`
 * Easy to use, clean, and tested codes
 * Active maintenance
@@ -195,6 +195,7 @@ get_supported_optimizers(['adam*', 'ranger*'])
 | APOLLO        | *SGD-like Memory, AdamW-level Performance*                                                        | [github](https://github.com/zhuhanqing/APOLLO)                                                                 | <https://arxiv.org/abs/2412.05270>                                                          | [cite](https://github.com/zhuhanqing/APOLLO?tab=readme-ov-file#-citation)                                                           |
 | MARS          | *Unleashing the Power of Variance Reduction for Training Large Models*                            | [github](https://github.com/AGI-Arena/MARS)                                                                    | <https://arxiv.org/abs/2411.10438>                                                          | [cite](https://github.com/AGI-Arena/MARS/tree/main?tab=readme-ov-file#citation)                                                     |
 | SGDSaI        | *No More Adam: Learning Rate Scaling at Initialization is All You Need*                           | [github](https://github.com/AnonymousAlethiometer/SGD_SaI)                                                     | <https://arxiv.org/abs/2411.10438>                                                          | [cite](https://github.com/AnonymousAlethiometer/SGD_SaI?tab=readme-ov-file#citation)                                                |
+| Grams         | *Grams: Gradient Descent with Adaptive Momentum Scaling*                                          |                                                                                                                | <https://arxiv.org/abs/2412.17107>                                                          | [cite](https://ui.adsabs.harvard.edu/abs/2024arXiv241217107C/exportcitation)                                                        |
 
 ## Supported LR Scheduler
 

docs/changelogs/v3.3.3.md

@@ -0,0 +1,6 @@
+### Change Log
+
+### Feature
+
+* Implement `Grams` optimizer. (#317, #318)
+    * [Grams: Gradient Descent with Adaptive Momentum Scaling](https://arxiv.org/abs/2412.17107) 

docs/index.md

@@ -10,7 +10,7 @@
 
 ## The reasons why you use `pytorch-optimizer`.
 
-* Wide range of supported optimizers. Currently, **86 optimizers (+ `bitsandbytes`, `qgalore`, `torchao`)**, **16 lr schedulers**, and **13 loss functions** are supported!
+* Wide range of supported optimizers. Currently, **87 optimizers (+ `bitsandbytes`, `qgalore`, `torchao`)**, **16 lr schedulers**, and **13 loss functions** are supported!
 * Including many variants such as `Cautious`, `AdamD`, `Gradient Centrailiaztion`
 * Easy to use, clean, and tested codes
 * Active maintenance
@@ -195,6 +195,7 @@ get_supported_optimizers(['adam*', 'ranger*'])
 | APOLLO        | *SGD-like Memory, AdamW-level Performance*                                                        | [github](https://github.com/zhuhanqing/APOLLO)                                                                 | <https://arxiv.org/abs/2412.05270>                                                          | [cite](https://github.com/zhuhanqing/APOLLO?tab=readme-ov-file#-citation)                                                           |
 | MARS          | *Unleashing the Power of Variance Reduction for Training Large Models*                            | [github](https://github.com/AGI-Arena/MARS)                                                                    | <https://arxiv.org/abs/2411.10438>                                                          | [cite](https://github.com/AGI-Arena/MARS/tree/main?tab=readme-ov-file#citation)                                                     |
 | SGDSaI        | *No More Adam: Learning Rate Scaling at Initialization is All You Need*                           | [github](https://github.com/AnonymousAlethiometer/SGD_SaI)                                                     | <https://arxiv.org/abs/2411.10438>                                                          | [cite](https://github.com/AnonymousAlethiometer/SGD_SaI?tab=readme-ov-file#citation)                                                |
+| Grams         | *Grams: Gradient Descent with Adaptive Momentum Scaling*                                          |                                                                                                                | <https://arxiv.org/abs/2412.17107>                                                          | [cite](https://ui.adsabs.harvard.edu/abs/2024arXiv241217107C/exportcitation)                                                        |
 
 ## Supported LR Scheduler
 

docs/optimizer.md

@@ -188,6 +188,10 @@
     :docstring:
     :members:
 
+::: pytorch_optimizer.Grams
+    :docstring:
+    :members:
+
 ::: pytorch_optimizer.Gravity
     :docstring:
     :members:

docs/visualization.md

@@ -162,6 +162,10 @@
 
 ![image](https://raw.githubusercontent.com/kozistr/pytorch_optimizer/main/docs/visualizations/rastrigin_GaLore.png)
 
+### Grams
+
+![image](https://raw.githubusercontent.com/kozistr/pytorch_optimizer/main/docs/visualizations/rastrigin_Grams.png)
+
 ### Gravity
 
 ![image](https://raw.githubusercontent.com/kozistr/pytorch_optimizer/main/docs/visualizations/rastrigin_Gravity.png)
@@ -484,6 +488,10 @@
 
 ![image](https://raw.githubusercontent.com/kozistr/pytorch_optimizer/main/docs/visualizations/rosenbrock_GaLore.png)
 
+### Grams
+
+![image](https://raw.githubusercontent.com/kozistr/pytorch_optimizer/main/docs/visualizations/rosenbrock_Grams.png)
+
 ### Gravity
 
 ![image](https://raw.githubusercontent.com/kozistr/pytorch_optimizer/main/docs/visualizations/rosenbrock_Gravity.png)

pyproject.toml

@@ -14,10 +14,10 @@ keywords = [
     "AdaDelta", "AdaFactor", "AdaMax", "AdamG", "AdaMod", "AdaNorm", "AdaPNM", "AdaSmooth", "AdEMAMix", "ADOPT",
     "AdaHessian", "Adai", "Adalite", "AdaLomo", "AdamMini", "AdamP", "AdamS", "Adan", "AggMo", "Aida", "AliG", "Amos",
     "Apollo", "APOLLO", "AvaGrad", "bSAM", "CAME", "DAdaptAdaGrad", "DAdaptAdam", "DAdaptAdan", "DAdaptSGD",
-    "DAdaptLion", "DeMo", "DiffGrad", "FAdam", "Fromage", "FTRL", "GaLore", "Gravity", "GrokFast", "GSAM", "Kate",
-    "Lamb", "LaProp", "LARS", "Lion", "LOMO", "Lookahead", "MADGRAD", "MARS", "MSVAG", "Muno", "Nero", "NovoGrad",
-    "PAdam", "PCGrad", "PID", "PNM", "Prodigy", "QHAdam", "QHM", "RAdam", "Ranger", "Ranger21", "RotoGrad", "SAM",
-    "ScheduleFreeSGD", "ScheduleFreeAdamW", "ScheduleFreeRAdam", "SGDP", "Shampoo", "ScalableShampoo", "SGDW",
+    "DAdaptLion", "DeMo", "DiffGrad", "FAdam", "Fromage", "FTRL", "GaLore", "Grams", "Gravity", "GrokFast", "GSAM",
+    "Kate", "Lamb", "LaProp", "LARS", "Lion", "LOMO", "Lookahead", "MADGRAD", "MARS", "MSVAG", "Muno", "Nero",
+    "NovoGrad", "PAdam", "PCGrad", "PID", "PNM", "Prodigy", "QHAdam", "QHM", "RAdam", "Ranger", "Ranger21", "RotoGrad",
+    "SAM", "ScheduleFreeSGD", "ScheduleFreeAdamW", "ScheduleFreeRAdam", "SGDP", "Shampoo", "ScalableShampoo", "SGDW",
     "SignSGD", "SM3", "SOAP", "SopihaH", "SRMM", "StableAdamW", "SWATS", "Tiger", "TRAC", "WSAM", "Yogi", "BCE",
     "BCEFocal", "Focal", "FocalCosine", "SoftF1", "Dice", "LDAM", "Jaccard", "Bi-Tempered", "Tversky", "FocalTversky",
     "LovaszHinge", "bitsandbytes", "WSD", "QGaLore",

pytorch_optimizer/__init__.py

@@ -105,6 +105,7 @@
     FAdam,
     Fromage,
     GaLore,
+    Grams,
     Gravity,
     GrokFastAdamW,
     Kate,

pytorch_optimizer/optimizer/__init__.py

@@ -46,6 +46,7 @@
 from pytorch_optimizer.optimizer.ftrl import FTRL
 from pytorch_optimizer.optimizer.galore import GaLore
 from pytorch_optimizer.optimizer.gc import centralize_gradient
+from pytorch_optimizer.optimizer.grams import Grams
 from pytorch_optimizer.optimizer.gravity import Gravity
 from pytorch_optimizer.optimizer.grokfast import GrokFastAdamW
 from pytorch_optimizer.optimizer.kate import Kate
@@ -282,6 +283,7 @@ def load_optimizer(optimizer: str) -> OPTIMIZER:
     LaProp,
     MARS,
     SGDSaI,
+    Grams,
 ]
 OPTIMIZERS: Dict[str, OPTIMIZER] = {str(optimizer.__name__).lower(): optimizer for optimizer in OPTIMIZER_LIST}
 

pytorch_optimizer/optimizer/grams.py

@@ -0,0 +1,108 @@
+import math
+
+import torch
+
+from pytorch_optimizer.base.exception import NoSparseGradientError
+from pytorch_optimizer.base.optimizer import BaseOptimizer
+from pytorch_optimizer.base.types import BETAS, CLOSURE, DEFAULTS, LOSS, PARAMETERS
+
+
+class Grams(BaseOptimizer):
+    r"""Gradient Descent with Adaptive Momentum Scaling.
+
+    :param params: PARAMETERS. iterable of parameters to optimize or dicts defining parameter groups.
+    :param lr: float. learning rate.
+    :param betas: BETAS. coefficients used for computing running averages of gradient and the squared hessian trace.
+    :param weight_decay: float. weight decay (L2 penalty).
+    :param weight_decouple: bool. decoupled weight decay.
+    :param eps: float. term added to the denominator to improve numerical stability.
+    """
+
+    def __init__(
+        self,
+        params: PARAMETERS,
+        lr: float = 1e-3,
+        betas: BETAS = (0.9, 0.999),
+        weight_decay: float = 0.0,
+        weight_decouple: bool = True,
+        eps: float = 1e-6,
+        **kwargs,
+    ):
+        self.validate_learning_rate(lr)
+        self.validate_betas(betas)
+        self.validate_non_negative(weight_decay, 'weight_decay')
+        self.validate_non_negative(eps, 'eps')
+
+        defaults: DEFAULTS = {
+            'lr': lr,
+            'betas': betas,
+            'weight_decay': weight_decay,
+            'weight_decouple': weight_decouple,
+            'eps': eps,
+        }
+
+        super().__init__(params, defaults)
+
+    def __str__(self) -> str:
+        return 'Grams'
+
+    @torch.no_grad()
+    def reset(self):
+        for group in self.param_groups:
+            group['step'] = 0
+            for p in group['params']:
+                state = self.state[p]
+
+                state['exp_avg'] = torch.zeros_like(p)
+                state['exp_avg_sq'] = torch.zeros_like(p)
+
+    @torch.no_grad()
+    def step(self, closure: CLOSURE = None) -> LOSS:
+        loss: LOSS = None
+        if closure is not None:
+            with torch.enable_grad():
+                loss = closure()
+
+        for group in self.param_groups:
+            if 'step' in group:
+                group['step'] += 1
+            else:
+                group['step'] = 1
+
+            beta1, beta2 = group['betas']
+
+            bias_correction1: float = self.debias(beta1, group['step'])
+            bias_correction2_sq: float = math.sqrt(self.debias(beta2, group['step']))
+
+            for p in group['params']:
+                if p.grad is None:
+                    continue
+
+                grad = p.grad
+                if grad.is_sparse:
+                    raise NoSparseGradientError(str(self))
+
+                state = self.state[p]
+                if len(state) == 0:
+                    state['exp_avg'] = torch.zeros_like(p)
+                    state['exp_avg_sq'] = torch.zeros_like(p)
+
+                exp_avg, exp_avg_sq = state['exp_avg'], state['exp_avg_sq']
+                exp_avg.lerp_(grad, weight=beta1)
+                exp_avg_sq.mul_(beta2).addcmul_(grad, grad, value=1.0 - beta2)
+
+                update = (exp_avg / bias_correction1) / (exp_avg_sq / bias_correction2_sq).sqrt_().add_(group['eps'])
+                update.abs_().mul_(grad.sign())
+
+                self.apply_weight_decay(
+                    p,
+                    grad,
+                    lr=group['lr'],
+                    weight_decay=group['weight_decay'],
+                    weight_decouple=group['weight_decouple'],
+                    fixed_decay=False,
+                )
+
+                p.add_(update, alpha=-group['lr'])
+
+        return loss