adv-optm 0.1.7__py3-none-any.whl → 0.1.9__py3-none-any.whl

adv_optm/__init__.py

@@ -16,4 +16,4 @@ __all__ = [
     "Lion_Prodigy_adv",
 ]
 
-__version__ = "0.1.7"
+__version__ = "0.1.9"

adv_optm/optim/AdamW_adv.py

@@ -55,7 +55,7 @@ class AdamW_adv(torch.optim.Optimizer):
             the warmup, `alpha` ramps from 0 to its target value. If `None`,
             the scheduler is disabled. (default: None)
         factored (bool): whether to use the factorization or disable it to use
-            the uncompressed optimizer. (default: True)
+            the uncompressed optimizer. (default: False)
     """
 
     def __init__(
@@ -76,7 +76,7 @@ class AdamW_adv(torch.optim.Optimizer):
         beta3_ema: float = 0.9999,
         alpha: float = 5.0,
         t_alpha: int | None = None,
-        factored: bool = True,
+        factored: bool = False,
     ):
         if not (lr >= 0.0):
             raise ValueError(f"Learning-rate should be >= 0.0. Got {lr}")
@@ -86,6 +86,9 @@ class AdamW_adv(torch.optim.Optimizer):
             raise ValueError(f"Epsilon should be >= 0.0. Got {eps}")
         if not (weight_decay >= 0.0):
             raise ValueError(f"Weight-decay should be >= 0.0. Got {weight_decay}")
+        if use_cautious and use_grams:
+            print("Warning: use_cautious is incompatible with use_grams, Disabling use_cautious.")
+            use_cautious = False
 
         defaults = {
             "lr": lr, "betas": betas, "eps": eps, "weight_decay": weight_decay,
@@ -216,7 +219,10 @@ class AdamW_adv(torch.optim.Optimizer):
                 del unpacked_sign_slow
 
                 mt_slow.mul_(beta3_ema).add_(grad_reshaped, alpha=1.0 - beta3_ema)
-                update = mt + (alpha_t * mt_slow) if beta1 > 0 else grad_reshaped + (alpha_t * mt_slow)
+                if beta1 > 0:
+                    update = torch.add(mt, mt_slow, alpha=alpha_t)
+                else:
+                    update = torch.add(grad_reshaped, mt_slow, alpha=alpha_t)
             else:
                 update = mt.clone() if beta1 > 0 else grad_reshaped.clone()
             del grad_reshaped
@@ -262,7 +268,10 @@ class AdamW_adv(torch.optim.Optimizer):
             if self.use_AdEMAMix:
                 exp_avg_slow = state['exp_avg_slow']
                 exp_avg_slow.mul_(beta3_ema).add_(grad, alpha=1 - beta3_ema)
-                update = exp_avg + (alpha_t * exp_avg_slow) if beta1 > 0 else grad + (alpha_t * exp_avg_slow)
+                if beta1 > 0:
+                    update = torch.add(exp_avg, exp_avg_slow, alpha=alpha_t)
+                else:
+                    update = torch.add(grad, exp_avg_slow, alpha=alpha_t)
             else:
                 update = exp_avg.clone() if beta1 > 0 else grad.clone()
 

adv_optm/optim/Adopt_adv.py

@@ -62,8 +62,18 @@ class Adopt_adv(torch.optim.Optimizer):
             the warmup, `alpha` ramps from 0 to its target value. If `None`,
             the scheduler is disabled and the full `alpha` value is used from
             the start. (default: None)
+        Simplified_AdEMAMix (bool): whether to use the Simplified AdEMAMix update rule.
+            This changes the EMA to accumulator and the update numerator to `alpha_grad * grad + mt`, which can be
+            more responsive, especially for small batch sizes. Enabling this will
+            automatically disable `use_AdEMAMix`, `use_cautious`, `use_grams`,
+            and `use_atan2`. (default: False)
+        alpha_grad (float): Mixing coefficient for the Simplified AdEMAMix update rule
+            (only used when `Simplified_AdEMAMix` is `True`). Controls the weight of the
+            current gradient. For small batch sizes, use high values (e.g., 10-100) to be
+            more responsive. For large batch sizes, use low values (e.g., 0-1) for
+            stability. (default: 100.0)
         factored (bool): whether to use the factorization or disable it to use
-            the uncompressed optimizer. (default: True)
+            the uncompressed optimizer. (default: False)
     """
 
     def __init__(
@@ -77,14 +87,16 @@ class Adopt_adv(torch.optim.Optimizer):
         vector_reshape: bool = True,
         stochastic_rounding: bool = True,
         use_atan2: bool = False,
-        use_cautious: bool = True,
+        use_cautious: bool = False,
         use_grams: bool = False,
         use_orthograd: bool = False,
         use_AdEMAMix: bool = False,
         beta3_ema: float = 0.9999,
         alpha: float = 5.0,
         t_alpha: int | None = None,
-        factored: bool = True,
+        Simplified_AdEMAMix: bool = False,
+        alpha_grad: float = 100.0,
+        factored: bool = False,
     ):
         if not (lr >= 0.0):
             raise ValueError(f"Learning-rate should be >= 0.0. Got {lr}")
@@ -94,19 +106,34 @@ class Adopt_adv(torch.optim.Optimizer):
             raise ValueError(f"Epsilon should be >= 0.0. Got {eps}")
         if not (weight_decay >= 0.0):
             raise ValueError(f"Weight-decay should be >= 0.0. Got {weight_decay}")
+        if use_cautious and use_grams:
+            print("Warning: use_cautious is incompatible with use_grams, Disabling use_cautious.")
+            use_cautious = False
+        if betas[0] == 0.0 and Simplified_AdEMAMix:
+            raise ValueError(f"Beta1 cannot be 0.0 when using Simplified_AdEMAMix. Got {betas[0]}")
+        if use_AdEMAMix and Simplified_AdEMAMix:
+            print("Warning: use_AdEMAMix is incompatible with Simplified_AdEMAMix, Disabling use_AdEMAMix.")
+        if use_grams and Simplified_AdEMAMix:
+            print("Warning: use_grams is incompatible with Simplified_AdEMAMix, Disabling use_grams.")
+        if use_cautious and Simplified_AdEMAMix:
+            print("Warning: use_cautious is incompatible with Simplified_AdEMAMix, Disabling use_cautious.")
+        if use_atan2 and Simplified_AdEMAMix:
+            print("Warning: use_atan2 is incompatible with Simplified_AdEMAMix. Disabling use_atan2.")
+            use_atan2 = False
 
         defaults = {
             "lr": lr, "betas": betas, "eps": eps, "weight_decay": weight_decay,
             "vector_reshape": vector_reshape, "beta3_ema": beta3_ema, "alpha": alpha,
-            "t_alpha": t_alpha,
+            "t_alpha": t_alpha, "alpha_grad": alpha_grad, 
         }
         self.clip_lambda = clip_lambda
         self.stochastic_rounding = stochastic_rounding
-        self.use_atan2 = use_atan2
-        self.use_cautious = use_cautious
-        self.use_grams = use_grams
+        self.use_atan2 = use_atan2 and not Simplified_AdEMAMix
+        self.use_cautious = use_cautious and not Simplified_AdEMAMix
+        self.use_grams = use_grams and not Simplified_AdEMAMix
         self.use_orthograd = use_orthograd
-        self.use_AdEMAMix = use_AdEMAMix
+        self.use_AdEMAMix = use_AdEMAMix and not Simplified_AdEMAMix
+        self.Simplified_AdEMAMix = Simplified_AdEMAMix
         self.factored = factored
         super().__init__(params, defaults)
 
@@ -185,6 +212,8 @@ class Adopt_adv(torch.optim.Optimizer):
             alpha_t = alpha
             if t_alpha is not None and t_alpha > 0 and current_step < t_alpha:
                 alpha_t = min(current_step * alpha / t_alpha, alpha)
+        if self.Simplified_AdEMAMix:
+            alpha_grad = group["alpha_grad"]
 
         if state['factored']:
             d1, d2 = state['effective_shape']
@@ -224,7 +253,10 @@ class Adopt_adv(torch.optim.Optimizer):
             del denom
 
             # ADOPT Step B: Update momentum m_t using normalized gradient
-            mt.mul_(beta1).add_(normalized_grad, alpha=1.0 - beta1)
+            if self.Simplified_AdEMAMix:
+                mt.mul_(beta1).add_(normalized_grad, alpha=1.0)
+            else:
+                mt.mul_(beta1).add_(normalized_grad, alpha=1.0 - beta1)
             if self.use_grams:
                 mt = grad_reshaped.sign() * mt.abs()
             elif self.use_cautious:
@@ -235,8 +267,10 @@ class Adopt_adv(torch.optim.Optimizer):
 
             if self.use_AdEMAMix:
                 mt_slow.mul_(beta3_ema).add_(normalized_grad, alpha=1.0 - beta3_ema)
-                update = mt + (alpha_t * mt_slow)
+                update = torch.add(mt, m_slow, alpha=alpha_t)
                 update = update.view(p.shape)
+            elif self.Simplified_AdEMAMix:
+                update = torch.add(mt, grad_reshaped, alpha=alpha_grad)
             else:
                 update = mt.view(p.shape)
 
@@ -283,7 +317,10 @@ class Adopt_adv(torch.optim.Optimizer):
             del denom
 
             # ADOPT Step B: Update momentum m_t
-            m.mul_(beta1).add_(normalized_grad, alpha=1.0 - beta1)
+            if self.Simplified_AdEMAMix:
+                m.mul_(beta1).add_(normalized_grad, alpha=1.0)
+            else:
+                m.mul_(beta1).add_(normalized_grad, alpha=1.0 - beta1)
 
             if self.use_grams:
                 m = grad.sign() * m.abs()
@@ -295,9 +332,11 @@ class Adopt_adv(torch.optim.Optimizer):
 
             if self.use_AdEMAMix:
                 m_slow.mul_(beta3_ema).add_(normalized_grad, alpha=1.0 - beta3_ema)
-                update = m + (alpha_t * m_slow)
+                update = torch.add(m, m_slow, alpha=alpha_t)
+            elif self.Simplified_AdEMAMix:
+                update = torch.add(m, grad, alpha=alpha_grad)
             else:
-                update = m
+                update = m.clone()
 
             if self.use_atan2:
                 update.mul_(group['lr'] * 1.2732395447351628)

adv_optm/optim/Lion_Prodigy_adv.py

@@ -33,8 +33,6 @@ class Lion_Prodigy_adv(torch.optim.Optimizer):
             (default: 0.0).
         factored (bool): whether to use the factorization or use the
             uncompressed optimizer. (default: True)
-        variance_reduction (bool): whether to use the variance reduction technique
-            from "Convergence Analysis of the Lion Optimizer" (arXiv:2508.12327v1). (default: False).
         d0 (float):
             Initial D estimate for D-adaptation (default 1e-6). Rarely needs changing.
         d_coef (float):
@@ -66,7 +64,6 @@ class Lion_Prodigy_adv(torch.optim.Optimizer):
         use_cautious: bool = False,
         clip_threshold: float = 0.0,
         factored: bool = True,
-        variance_reduction: bool = False,
         # prodigy parameters
         beta3: float = None,
         d0: float = 1e-6,
@@ -97,7 +94,6 @@ class Lion_Prodigy_adv(torch.optim.Optimizer):
         self.stochastic_rounding = stochastic_rounding
         self.use_cautious = use_cautious
         self.factored = factored
-        self.variance_reduction = variance_reduction
         self.fsdp_in_use = fsdp_in_use
         super().__init__(params, defaults)
         # Global state for accumulating metrics across parameter updates within a single step.
@@ -183,12 +179,8 @@ class Lion_Prodigy_adv(torch.optim.Optimizer):
                 state['mv_m_nmf'] = torch.zeros(d2, device=p.device, dtype=dtype)
                 packed_d2 = (d2 + 7) // 8
                 state['sign'] = torch.zeros((d1, packed_d2), dtype=torch.uint8, device=p.device)
-                if self.variance_reduction:
-                    state['prev_grad'] = torch.zeros((d1, d2), device=p.device, dtype=dtype)
             else: # Fallback to standard Lion
                 state['exp_avg'] = torch.zeros_like(p, device=p.device, dtype=dtype)
-                if self.variance_reduction:
-                    state['prev_grad'] = torch.zeros_like(p, device=p.device, dtype=dtype)
 
         if state['factored']:
             # Factored Path
@@ -215,20 +207,7 @@ class Lion_Prodigy_adv(torch.optim.Optimizer):
             update_for_param = signed_update.view(p.shape).mul(self.dlr)
 
             # Update momentum m_t = β2*m_{t-1} + (1-β2)*lr*g_t
-            if self.variance_reduction:
-                if state['step'] == 1:
-                    exp_avg.copy_(grad_reshaped)
-                else:
-                    # Heuristic Prodigy-STORM update
-                    correction = exp_avg.sub(state['prev_grad'])
-                    grad_alpha = self.d * (1 - self.beta2) + self.beta2
-                    exp_avg.copy_(grad_reshaped).mul_(grad_alpha).add_(correction, alpha=self.beta2)
-                    del correction, grad_alpha
-                state['prev_grad'].copy_(grad_reshaped)
-            else:
-                # Standard Prodigy-Lion
-                alpha = self.d * (1 - self.beta2)
-                exp_avg.mul_(self.beta2).add_(grad_reshaped, alpha=alpha)
+            exp_avg.mul_(self.beta2).add_(grad_reshaped, alpha=self.d * (1 - self.beta2))
             del grad_reshaped
 
             # Compress new momentum m_t and store factors
@@ -254,20 +233,7 @@ class Lion_Prodigy_adv(torch.optim.Optimizer):
             update_for_param = signed_update.mul(self.dlr)
 
             # Update momentum 
-            if self.variance_reduction:
-                if state['step'] == 1:
-                    exp_avg.copy_(grad)
-                else:
-                    # Heuristic Prodigy-STORM update
-                    correction = exp_avg.sub(state['prev_grad'])
-                    grad_alpha = self.d * (1 - self.beta2) + self.beta2
-                    exp_avg.copy_(grad).mul_(grad_alpha).add_(correction, alpha=self.beta2)
-                    del grad_alpha, correction
-                state['prev_grad'].copy_(grad)
-            else:
-                # Standard Prodigy-Lion
-                alpha = self.d * (1 - self.beta2)
-                exp_avg.mul_(self.beta2).add_(grad, alpha=alpha)
+            exp_avg.mul_(self.beta2).add_(grad, alpha=self.d * (1 - self.beta2))
 
         # --- Accumulate Prodigy stats ---
         d0, safeguard_warmup, slice_p = group['d0'], group['safeguard_warmup'], group['slice_p']
@@ -298,7 +264,7 @@ class Lion_Prodigy_adv(torch.optim.Optimizer):
         else:
             p.data.add_(-update_for_param)
 
-            del update_for_param
+        del update_for_param
 
     @torch.no_grad()
     def step(self, closure: Optional[callable] = None):

adv_optm/optim/Lion_adv.py

@@ -33,8 +33,6 @@ class Lion_adv(torch.optim.Optimizer):
             (default: 0.0).
         factored (bool): whether to use the factorization or use the
             uncompressed optimizer. (default: True)
-        variance_reduction (bool): whether to use the variance reduction technique
-            from "Convergence Analysis of the Lion Optimizer" (arXiv:2508.12327v1). (default: False).
     """
 
     def __init__(
@@ -49,7 +47,6 @@ class Lion_adv(torch.optim.Optimizer):
         use_cautious: bool = False,
         clip_threshold: float = 0.0,
         factored: bool = True,
-        variance_reduction: bool = False,
     ):
         if not lr > 0.0:
             raise ValueError(f"Learning rate must be > 0.0, but got {lr}")
@@ -69,7 +66,6 @@ class Lion_adv(torch.optim.Optimizer):
         self.stochastic_rounding = stochastic_rounding
         self.use_cautious = use_cautious
         self.factored = factored
-        self.variance_reduction = variance_reduction
         super().__init__(params, defaults)
 
     @property
@@ -122,12 +118,8 @@ class Lion_adv(torch.optim.Optimizer):
                 state['mv_m_nmf'] = torch.zeros(d2, device=p.device, dtype=dtype)
                 packed_d2 = (d2 + 7) // 8
                 state['sign'] = torch.zeros((d1, packed_d2), dtype=torch.uint8, device=p.device)
-                if self.variance_reduction:
-                    state['prev_grad'] = torch.zeros((d1, d2), device=p.device, dtype=dtype)
             else: # Fallback to standard Lion
                 state['exp_avg'] = torch.zeros_like(p, device=p.device, dtype=dtype)
-                if self.variance_reduction:
-                    state['prev_grad'] = torch.zeros_like(p, device=p.device, dtype=dtype)
 
         state['step'] += 1
         beta1, beta2 = group["betas"]
@@ -157,21 +149,9 @@ class Lion_adv(torch.optim.Optimizer):
             # Parameter update
             update_for_param = signed_update.view(p.shape).mul_(lr)
 
-            # Update momentum
-            if self.variance_reduction:
-                if state['step'] == 1:
-                    exp_avg.copy_(grad_reshaped)
-                else:
-                    # Use the simplified STORM update: m_t = g_t + β₂ * (m_{t-1} - g_{t-1})
-                    correction = exp_avg.sub(state['prev_grad'])
-                    # Calculate the new momentum and store it back into exp_avg
-                    exp_avg.copy_(grad_reshaped).add_(correction, alpha=beta2)
-                    del correction
-                # Update prev_grad for the next iteration
-                state['prev_grad'].copy_(grad_reshaped)
-            else:
-                # Standard Lion momentum update
-                exp_avg.mul_(beta2).add_(grad_reshaped, alpha=1-beta2)
+            # Standard Lion momentum update
+            exp_avg.mul_(beta2).add_(grad_reshaped, alpha=1-beta2)
+            del grad_reshaped
 
             # Compress new momentum m_t and store factors
             state['sign'] = _pack_bools(exp_avg > 0)
@@ -195,21 +175,8 @@ class Lion_adv(torch.optim.Optimizer):
 
             update_for_param = signed_update.mul_(lr)
 
-            # Update momentum 
-            if self.variance_reduction:
-                if state['step'] == 1:
-                    exp_avg.copy_(grad)
-                else:
-                    # Use the simplified STORM update: m_t = g_t + β₂ * (m_{t-1} - g_{t-1})
-                    correction = exp_avg.sub(state['prev_grad'])
-                    # Calculate the new momentum and store it back into exp_avg
-                    exp_avg.copy_(grad).add_(correction, alpha=beta2)
-                    del correction
-                # Update prev_grad for the next iteration
-                state['prev_grad'].copy_(grad)
-            else:
-                # Standard Lion momentum update
-                exp_avg.mul_(beta2).add_(grad, alpha=1-beta2)
+            # Standard Lion momentum update
+            exp_avg.mul_(beta2).add_(grad, alpha=1-beta2)
 
         if group["weight_decay"] != 0:
             if p.dtype == torch.bfloat16 and self.stochastic_rounding:
@@ -225,7 +192,7 @@ class Lion_adv(torch.optim.Optimizer):
         else:
             p.data.add_(-update_for_param)
 
-            del update_for_param
+        del update_for_param
 
     @torch.no_grad()
     def step(self, closure: Optional[callable] = None):

adv_optm/optim/Prodigy_adv.py

@@ -64,7 +64,7 @@ class Prodigy_adv(torch.optim.Optimizer):
             more responsive. For large batch sizes, use low values (e.g., 0-1) for
             stability. (default: 100.0)
         factored (bool): whether to use the factorization or disable it to use
-            the uncompressed optimizer. (default: True)
+            the uncompressed optimizer. (default: False)
         d0 (float):
             Initial D estimate for D-adaptation (default 1e-6). Rarely needs changing.
         d_coef (float):
@@ -82,6 +82,9 @@ class Prodigy_adv(torch.optim.Optimizer):
         slice_p (int): Reduce memory usage by calculating LR adaptation statistics on only every 
             pth entry of each tensor. For values greater than 1 this an an approximation to standard 
             Prodigy. Values ~11 are reasonable (default 11).
+        prodigy_steps (int): If greater than zero, disable Prodigy's stepsize adjustments
+            after the specified optimiser step and release all state memory required by Prodigy
+            (default: 0).
     """
 
     def __init__(
@@ -103,7 +106,7 @@ class Prodigy_adv(torch.optim.Optimizer):
         t_alpha: int | None = None,
         Simplified_AdEMAMix: bool = False,
         alpha_grad: float = 100.0,
-        factored: bool = True,
+        factored: bool = False,
         # prodigy parameters
         beta3: float = None,
         d0: float = 1e-6,
@@ -112,6 +115,7 @@ class Prodigy_adv(torch.optim.Optimizer):
         safeguard_warmup: bool = False,
         fsdp_in_use: bool = False,
         slice_p: int = 11,
+        prodigy_steps: int = 0,
     ):
         if not (lr >= 0.0):
             raise ValueError(f"Learning-rate should be >= 0.0. Got {lr}")
@@ -121,8 +125,13 @@ class Prodigy_adv(torch.optim.Optimizer):
             raise ValueError(f"Epsilon should be >= 0.0. Got {eps}")
         if not (weight_decay >= 0.0):
             raise ValueError(f"Weight-decay should be >= 0.0. Got {weight_decay}")
+        if not (prodigy_steps >= 0):
+            raise ValueError(f"prodigy_steps should be >= 0. Got {prodigy_steps}")
+        if use_cautious and use_grams:
+            print("Warning: use_cautious is incompatible with use_grams, Disabling use_cautious.")
+            use_cautious = False
         if betas[0] == 0.0 and Simplified_AdEMAMix:
-            raise ValueError(f"Beta 1 cannot be 0.0 when using Simplified_AdEMAMix. Got {betas[0]}")
+            raise ValueError(f"Beta1 cannot be 0.0 when using Simplified_AdEMAMix. Got {betas[0]}")
         if use_AdEMAMix and Simplified_AdEMAMix:
             print("Warning: use_AdEMAMix is incompatible with Simplified_AdEMAMix, Disabling use_AdEMAMix.")
         if use_grams and Simplified_AdEMAMix:
@@ -140,7 +149,7 @@ class Prodigy_adv(torch.optim.Optimizer):
             "beta3_ema": beta3_ema, "alpha": alpha, "t_alpha": t_alpha,
             "beta3": beta3, "d": d0, "d0": d0, "d_max": d0, "d_numerator": 0.0, "d_coef": d_coef,
             "growth_rate": growth_rate, "safeguard_warmup": safeguard_warmup, "k": 0, "slice_p": slice_p,
-            "fsdp_in_use": fsdp_in_use,
+            "fsdp_in_use": fsdp_in_use, "prodigy_steps": prodigy_steps,
             "alpha_grad": alpha_grad, 
         }
         self.stochastic_rounding = stochastic_rounding
@@ -293,7 +302,10 @@ class Prodigy_adv(torch.optim.Optimizer):
                 torch.where(unpacked_sign_slow, mt_slow, -mt_slow, out=mt_slow)
                 del unpacked_sign_slow
                 mt_slow.mul_(beta3_ema).add_(grad_reshaped, alpha=self.d * (1.0 - beta3_ema))
-                update = mt + (alpha_t * mt_slow) if self.beta1 > 0 else grad_reshaped + (alpha_t * mt_slow)
+                if self.beta1 > 0:
+                    update = torch.add(mt, mt_slow, alpha=alpha_t)
+                else:
+                    update = torch.add(grad_reshaped, mt_slow, alpha=alpha_t)
             elif self.Simplified_AdEMAMix:
                 update = torch.add(mt, grad_reshaped, alpha=alpha_grad * self.d)
             else:
@@ -344,7 +356,10 @@ class Prodigy_adv(torch.optim.Optimizer):
             if self.use_AdEMAMix:
                 exp_avg_slow = state['exp_avg_slow']
                 exp_avg_slow.mul_(beta3_ema).add_(grad, alpha=self.d * (1.0 - beta3_ema))
-                update = exp_avg + (alpha_t * exp_avg_slow) if self.beta1 > 0 else grad + (alpha_t * exp_avg_slow)
+                if self.beta1 > 0:
+                    update = torch.add(exp_avg, exp_avg_slow, alpha=alpha_t)
+                else:
+                    update = torch.add(grad, exp_avg_slow, alpha=alpha_t)
             elif self.Simplified_AdEMAMix:
                 update = torch.add(exp_avg, grad, alpha=alpha_grad * self.d)
             else:
@@ -364,19 +379,27 @@ class Prodigy_adv(torch.optim.Optimizer):
             update.mul_(self.dlr)
 
         # --- Accumulate Prodigy stats ---
-        d0, safeguard_warmup, slice_p = group['d0'], group['safeguard_warmup'], group['slice_p']
-        s, p0 = state['s'], state['p0']
-        grad_flat = grad.flatten().float()
-        p_flat = p.data.flatten().float()
-        p0 = p0.float()
+        prodigy_steps = group['prodigy_steps']
+        if prodigy_steps <= 0 or group['k'] < prodigy_steps:
+            d0, safeguard_warmup, slice_p = group['d0'], group['safeguard_warmup'], group['slice_p']
+            s, p0 = state['s'], state['p0']
+            grad_flat = grad.flatten().float()
+            p_flat = p.data.flatten().float()
+            p0 = p0.float()
 
-        self.d_numerator += (self.d / d0) * self.dlr * torch.dot(grad_flat[::slice_p], p0.data - p_flat[::slice_p]).item()
+            self.d_numerator += (self.d / d0) * self.dlr * torch.dot(grad_flat[::slice_p], p0.data - p_flat[::slice_p]).item()
 
-        alpha = ((self.d / d0) * self.d) if safeguard_warmup else ((self.d / d0) * self.dlr)
-        s.mul_(self.beta3).add_(grad_flat[::slice_p], alpha=alpha)
-        self.d_denom += s.abs().sum().item()
+            alpha = ((self.d / d0) * self.d) if safeguard_warmup else ((self.d / d0) * self.dlr)
+            s.mul_(self.beta3).add_(grad_flat[::slice_p], alpha=alpha)
+            self.d_denom += s.abs().sum().item()
 
-        del s, p0, grad_flat, p_flat, alpha
+            del s, p0, grad_flat, p_flat, alpha
+        else:
+            # Free memory if prodigy_steps is reached
+            if 's' in state:
+                del state['s']
+            if 'p0' in state:
+                del state['p0']
 
         # Decoupled weight decay
         if group["weight_decay"] != 0:
@@ -413,29 +436,43 @@ class Prodigy_adv(torch.optim.Optimizer):
     def calculate_d(self):
         """Calculates the new `d` based on the accumulated stats."""
         g_group = self.param_groups[0]
-        d_max, d_coef, growth_rate = g_group['d_max'], g_group['d_coef'], g_group['growth_rate']
         
-        if self.fsdp_in_use and dist.is_available() and dist.is_initialized():
-            # Use the device of the first parameter to avoid hardcoding '.cuda()'
-            device = self.param_groups[0]['params'][0].device
-            dist_tensor = torch.tensor([self.d_numerator, self.d_denom], device=device)
-            dist.all_reduce(dist_tensor, op=dist.ReduceOp.SUM)
-            global_d_numerator = dist_tensor[0].item()
-            global_d_denom = dist_tensor[1].item()
-        else:
-            global_d_numerator = self.d_numerator
-            global_d_denom = self.d_denom
-
-        d_hat = self.d
-        if global_d_denom > 0:
-            d_hat = d_coef * global_d_numerator / global_d_denom
-            if self.d == g_group['d0']:
-                self.d = max(self.d, d_hat)
-            d_max = max(d_max, d_hat)
-            self.d = min(d_max, self.d * growth_rate)
-
+        # Only perform d-adaptation if prodigy_steps has not been reached
+        prodigy_active = not (g_group.get('prodigy_steps', 0) > 0 and g_group['k'] >= g_group['prodigy_steps'])
+
+        if prodigy_active:
+            d_max, d_coef, growth_rate = g_group['d_max'], g_group['d_coef'], g_group['growth_rate']
+            
+            if self.fsdp_in_use and dist.is_available() and dist.is_initialized():
+                # Use the device of the first parameter to avoid hardcoding '.cuda()'
+                device = self.param_groups[0]['params'][0].device
+                dist_tensor = torch.tensor([self.d_numerator, self.d_denom], device=device)
+                dist.all_reduce(dist_tensor, op=dist.ReduceOp.SUM)
+                global_d_numerator = dist_tensor[0].item()
+                global_d_denom = dist_tensor[1].item()
+            else:
+                global_d_numerator = self.d_numerator
+                global_d_denom = self.d_denom
+
+            d_hat = self.d
+            if global_d_denom > 0:
+                if self.Simplified_AdEMAMix and g_group['alpha_grad'] > 0:
+                    # A simple and effective hack to make prodigy compatible with Simplified_AdEMAMix large step sizes
+                    # by diving by alpha_grad we make sure that d_numerator that was influenced by (alpha_grad * grad)
+                    # are now normalized by /alpha_grad. this is a heuristic way since the update is also influenced by
+                    # the increasing and decaying accumulator but it's effective and it worked for me (for Lora/Finetune).
+                    global_d_numerator /= g_group['alpha_grad']
+                d_hat = d_coef * global_d_numerator / global_d_denom
+                if self.d == g_group['d0']:
+                    self.d = max(self.d, d_hat)
+                d_max = max(d_max, d_hat)
+                self.d = min(d_max, self.d * growth_rate)
+
+            for group in self.param_groups:
+                group['d_numerator'] = global_d_numerator
+                group['d'] = self.d
+                group['d_max'] = d_max
+        
+        # Increment step counter for all groups, regardless of whether d was updated
         for group in self.param_groups:
-            group['d_numerator'] = global_d_numerator
-            group['d'] = self.d
-            group['d_max'] = d_max
             group['k'] += 1

adv_optm-0.1.9.dist-info/METADATA

@@ -0,0 +1,174 @@
+Metadata-Version: 2.4
+Name: adv_optm
+Version: 0.1.9
+Summary: A family of highly efficient, lightweight yet powerful optimizers.
+Home-page: https://github.com/Koratahiu/Advanced_Optimizers
+Author: Koratahiu
+Author-email: hiuhonor@gmail.com
+License: Apache 2.0
+Keywords: llm,fine-tuning,memory-efficient,low-rank,compression,pytorch,optimizer,adam
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Operating System :: OS Independent
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: torch>=2.0
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: keywords
+Dynamic: license
+Dynamic: license-file
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
+
+# Advanced Optimizers (AIO)
+
+A comprehensive, all-in-one collection of optimization algorithms for deep learning, designed for maximum efficiency, minimal memory footprint, and superior performance across diverse model architectures and training scenarios.
+
+[![PyPI](https://img.shields.io/pypi/v/adv_optm)](https://pypi.org/project/adv_optm/)
+
+---
+
+## 📦 Installation
+
+```bash
+pip install adv_optm
+```
+
+---
+
+## 🧠 Core Innovations
+
+This library integrates multiple state-of-the-art optimization techniques validated through extensive research and practical training, with 1-bit compression for optimizer states:
+
+### **Memory-Efficient Optimization (SMMF-inspired)**
+- **Paper**: [SMMF: Square-Matricized Momentum Factorization](https://arxiv.org/abs/2412.08894)
+- **Approach**: Uses rank-1 non-negative matrix factorization with reconstruction cycle (factor → reconstruct → update → factor)
+- **Innovation**: 
+  - First moment split into **1-bit sign + absolute value**
+  - Final storage: **four factored vectors + one 1-bit sign state**
+  - Preserves Adam-like update quality with drastically reduced memory
+
+---
+
+## ⚡ Performance Characteristics
+
+### Memory Efficiency (SDXL Model - 6.5GB)
+| Optimizer | Memory Usage | Description |
+|-----------|--------------|-------------|
+| `Adopt_Factored` | 328 MB | 4 small vectors + 1-bit state |
+| `Adopt_Factored + AdEMAMix` | 625 MB | 6 small vectors + two 1-bit states |
+| `Simplified_AdEMAMix` | 328 MB | Same as standard factored (no extra state) |
+
+### Speed Comparison (SDXL, Batch Size 4)
+| Optimizer | Speed | Notes |
+|-----------|-------|-------|
+| `Adafactor` | ~8.5s/it | Baseline |
+| `Adopt_Factored` | ~10s/it | +18% overhead from compression |
+| `Adopt_Factored + AdEMAMix` | ~12s/it | +41% overhead (3 factored states) |
+
+---
+
+## 🧪 Available Optimizers
+
+### Standard Optimizers (All support `factored=True/False`)
+| Optimizer | Description | Best For |
+|-----------|-------------|----------|
+| `Adam_Adv` | Advanced Adam implementation | General purpose |
+| `Adopt_Adv` | Adam-variant with independent beta2 | Stable training for small batch size regimes |
+| `Prodigy_Adv` | Prodigy with D-Adaptation | Adam with automatic LR tuning |
+| `Simplified_AdEMAMix` | Adam variant with accumulator momentum | Small/large batch training when tuned correctly |
+| `Lion_Adv` | Advanced Lion implementation | Memory-constrained environments |
+| `Prodigy_Lion_Adv` | Prodigy + Lion combination | Lion with automatic LR tuning |
+
+### Feature Matrix
+| Feature | Adam_Adv | Adopt_Adv | Prodigy_Adv | Simplified_AdEMAMix | Lion_Adv |
+|---------|----------|-----------|-------------|---------------------|----------|
+| Factored | ✓ | ✓ | ✓ | ✓ | ✓ |
+| AdEMAMix | ✓ | ✓ | ✓ | ✗ | ✗ |
+| Simplified_AdEMAMix | ✗ | ✗ | ✓ | ✓ | ✗ |
+| OrthoGrad | ✓ | ✓ | ✓ | ✓ | ✓ |
+| Grams | ✓ | ✓ | ✓ | ✗ | ✗ |
+| Cautious | ✓ | ✓ | ✓ | ✗ | ✓ |
+| atan2 | ✓ | ✓ | ✓ | ✗ | ✗ |
+| Stochastic Rounding | ✓ | ✓ | ✓ | ✓ | ✓ |
+| Fused Backward Pass | ✓ | ✓ | ✓ | ✓ | ✓ |
+
+---
+
+## ⚙️ Key Features & Parameters
+
+### Comprehensive Feature Guide
+
+| Feature | Description | Recommended Usage | Performance Impact | Theoretical Basis | Compatibility |
+|---------|-------------|-------------------|--------------------|-------------------|--------------|
+| **Factored** | Memory-efficient optimization using rank-1 factorization | Enable for large models (>1B params) or limited VRAM | +12-41% time overhead, 1-bit memory usage | [SMMF](https://arxiv.org/abs/2412.08894) | All optimizers |
+| **AdEMAMix** | Dual EMA system for momentum | Use for long training runs (10k+ steps) | +1 state memory. | [AdEMAMix](https://arxiv.org/abs/2409.03137) | Adam/Adopt/Prodigy |
+| **Simplified_AdEMAMix** | Accumulator-based momentum | Small batch training (≤32) | Same memory as standard, no extra overhead | [Schedule-Free Connections](https://arxiv.org/abs/2502.02431) | Adam/Prodigy |
+| **OrthoGrad** | Removes gradient component parallel to weights | Full finetuning without weight decay | +33% time overhead, no memory impact | [Grokking at Edge](https://github.com/LucasPrietoAl/grokking-at-the-edge-of-numerical-stability) | All optimizers |
+| **Stochastic Rounding** | Improves precision for BF16 training | BF16 training | Minimal overhead (<5%) | [Revisiting BFloat16 Training](https://arxiv.org/abs/2010.06192) | All optimizers |
+| **atan2** | Robust eps replacement + built-in clipping | Use with Adopt or unstable training | No overhead | [Adam-atan2](https://github.com/lucidrains/adam-atan2-pytorch) | Adam/Adopt/prodigy |
+| **Cautious** | Update only when the direction align with the gradients | should faster the convergence | No overhead | [C-Optim](https://github.com/kyleliang919/C-Optim) | Adam/Adopt/prodigy |
+| **Grams** | Update direction from the gradients | should have a stronger effect than cautious | No overhead | [Grams](https://github.com/Gunale0926/Grams) | Adam/Adopt/prodigy |
+
+---
+
+## Simplified_AdEMAMix Parameters
+Simplified_AdEMAMix replaces standard momentum with an accumulator for better small-large batch performance.
+
+| Parameter | Recommended Values | Description |
+|-----------|---------------------|-------------|
+| `beta1` | 0.9 (large BS), 0.99-0.9999 (small BS) | Determines memory length of accumulator |
+| `alpha` | 100-10 (small BS), 1-0 (large BS) | Gradient smoothing factor |
+
+**Alpha Tuning Guide**:
+| Batch Size | Recommended α | Rationale |
+|------------|---------------|-----------|
+| Small (≤32) | 100, 50, 20, 10 | Emphasizes recent gradients for quick adaptation |
+| Medium (32-512) | 10, 5, 2, 1 | Balanced approach |
+| Large (≥512) | 1, 0.5, 0 | Emphasizes historical gradients for stability |
+
+⚠️ **Important**: Use **~100x smaller learning rate** with Simplified_AdEMAMix compared to AdamW (e.g., 1e-6 instead of 1e-4)
+
+### 📊 Performance Validation
+Small Batch Training (SDXL, BS=2, 1.8K steps)
+![Training Comparison](https://github.com/user-attachments/assets/7eff0671-cc59-47fc-8b63-d5205456d649)
+
+- **🟢 Prodigy_adv** (beta1=0.9, d0=1e-5): Final LR=2.9e-4
+- **🔵 Prodigy_adv + Simplified_AdEMAMix** (beta1=0.99, α=100, d0=1e-7): Final LR=5.8e-6
+
+**Results**:
+- Simplified_AdEMAMix shows faster convergence and better final performance
+- D-Adaptation automatically handles aggressive updates (50x smaller LR)
+- Generated samples show significantly better quality with Simplified_AdEMAMix
+
+---
+
+## ⚠️ Known Limitations
+
+### 1. Prodigy_Adv Sensitivity
+- Highly sensitive to gradient modifications (Adopt normalization, low-rank factorization)
+- May fail to increase learning rate in some LoRA scenarios
+- **Fix**: Disable factorization or set beta1=0
+
+### 2. Aggressive Learning Rates
+- Can destabilize factored first moment
+- **Recommendation**: Check Prodigy learning rate as reference for safe LR threshold
+
+---
+
+## 📚 References
+
+1. [SMMF: Square-Matricized Momentum Factorization](https://arxiv.org/abs/2412.08894)
+2. [The AdEMAMix Optimizer: Better, Faster, Older](https://arxiv.org/abs/2409.03137)
+3. [Connections between Schedule-Free Optimizers, AdEMAMix, and Accelerated SGD Variants](https://arxiv.org/abs/2502.02431)
+
+---

adv_optm-0.1.9.dist-info/RECORD

@@ -0,0 +1,19 @@
+adv_optm/__init__.py,sha256=hHL2QwlnQMvIggC9ejOxGOKq65DnnYaHC1ScPQMuIIw,306
+adv_optm/optim/AdamW_adv.py,sha256=Pu0TB14dOhcq9kwXclMIeKCI6ef_P0emwzxPu6xuBM0,14252
+adv_optm/optim/Adopt_adv.py,sha256=71o9BHV3XFefJX21G37PKG96D09x-PSU0eW3Q7WkAjs,17427
+adv_optm/optim/Lion_Prodigy_adv.py,sha256=kIAGXoMbDNRg5reKXtUC_vQQ2gyM-NXPB-Pv9zSpiE8,12787
+adv_optm/optim/Lion_adv.py,sha256=05j_j6LIzHW5b79DVwMIf1FZHVNB8xnStNVjlOdVkCE,8256
+adv_optm/optim/Prodigy_adv.py,sha256=NykG5gcAHjmhlMutknOjAoYKI-K6e5lA3Q9J9vkqnz0,22357
+adv_optm/optim/Simplified_AdEMAMix.py,sha256=opIZjnGJ03-DDAIHTZyJBMReVfgusGDb8FZSWMU3-UM,9774
+adv_optm/optim/__init__.py,sha256=pcP865H2j1tut2VfTUhzQh7V8TF_tzPjqFnjMfFed2k,382
+adv_optm/util/BF16_Stochastic_Rounding.py,sha256=Q5H0BcogmE4atP65dLoI21HKSf50lRdsBDfeF6v9Tbg,1548
+adv_optm/util/Effective_Shape.py,sha256=TBvIk1V8IuTbbBsxuekJA4e_v8JlR5Nujtut8RTWAm4,318
+adv_optm/util/NNMF.py,sha256=yRf5IP5Sjq0Uf0DxN0Q8NxEGSdD-f1ULziLVDOjY8K4,639
+adv_optm/util/One_Bit_Boolean.py,sha256=Wat49esdwohuN-OHOFMW8D0aOQgV9cP5Rl8z6yfmpos,1068
+adv_optm/util/OrthoGrad.py,sha256=NzInuBQGy_Ja__M1R9XbvqVaQ0fhGbtGgFE9YON7B3I,707
+adv_optm/util/__init__.py,sha256=qoyIF0jcLjs_vSEcsv36clw5LFNBEbifyXrrVxMH-G4,349
+adv_optm-0.1.9.dist-info/licenses/LICENSE,sha256=HrhfyXIkWY2tGFK11kg7vPCqhgh5DcxleloqdhrpyMY,11558
+adv_optm-0.1.9.dist-info/METADATA,sha256=IvocLvlwTsZ5WPmO6ZsVffmybwZRf3tr_ALojuwL6dw,8422
+adv_optm-0.1.9.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+adv_optm-0.1.9.dist-info/top_level.txt,sha256=iNfBIIzu-lPrQ7jyC56WBCcbkRwitM2nJ15-MRQ_6fg,9
+adv_optm-0.1.9.dist-info/RECORD,,

adv_optm-0.1.7.dist-info/METADATA

@@ -1,130 +0,0 @@
-Metadata-Version: 2.4
-Name: adv_optm
-Version: 0.1.7
-Summary: A family of highly efficient, lightweight yet powerful optimizers.
-Home-page: https://github.com/Koratahiu/Advanced_Optimizers
-Author: Koratahiu
-Author-email: hiuhonor@gmail.com
-License: Apache 2.0
-Keywords: llm,fine-tuning,memory-efficient,low-rank,compression,pytorch,optimizer,adam
-Classifier: Programming Language :: Python :: 3
-Classifier: License :: OSI Approved :: Apache Software License
-Classifier: Operating System :: OS Independent
-Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
-Classifier: Topic :: Software Development :: Libraries :: Python Modules
-Requires-Python: >=3.8
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: torch>=2.0
-Dynamic: author
-Dynamic: author-email
-Dynamic: classifier
-Dynamic: description
-Dynamic: description-content-type
-Dynamic: home-page
-Dynamic: keywords
-Dynamic: license
-Dynamic: license-file
-Dynamic: requires-dist
-Dynamic: requires-python
-Dynamic: summary
-
-# Advanced Optimizers
-
-This repo introduces a new family of highly efficient, lightweight yet powerful optimizers, born from extensive research into recent academic literature and validated through practical training runs across diverse models.
-
----
-
-### Install
-
-`pip install adv_optm`
-
----
-
-### Theory (Inspired by SMMF)
-
-Based primarily on:  
-**[SMMF: Square-Matricized Momentum Factorization for Memory-Efficient Optimization](https://arxiv.org/abs/2412.08894)**
-
-The core innovation:
-- Uses fast, non-negative matrix factorization (NNMF - rank 1), but **reconstructs the full state before each update** to preserve momentum accuracy, then re-factors afterward (factor → reconstruct → update → factor cycle).
-- For the *signed first moment*, we split into **sign + absolute value**:
-  - Sign is stored as **1-bit state** via bitwise ops (SMMF originally used 8-bit with 7 bits wasted).
-  - Absolute value goes through the factor/reconstruct cycle using two factored vectors + the signed state.
-- Final storage: **four factored vectors + one 1-bit sign**.
-- Updates behave like full-state Adam but with drastically reduced memory.
-
-> ✅ **TL;DR**: Lightweight, strong, memory-efficient optimizer.
-
----
-
-### Memory Cost
-
-- **Adopt_Factored** for full SDXL finetune: **328 MB** (4 small vectors + 1-bit state)
-- **Adopt_Factored with AdEMAMix** for full SDXL finetune: **625 MB** (6 small vectors + two 1-bit states)
-> SDXL is 6.5GB model.
-
----
-
-### ⏱️ Speed (my tests in SDXL - BS 4)
-
-- **Adopt_Factored**: ~10s/it
-- **Adopt_Factored with AdEMAMix**: ~12s/it
-- **Adafactor**: ~8.5s/it  
-→ Overhead from compression/reconstruction cycles.
-→ It's faster than [MLorc](https://arxiv.org/abs/2506.01897) (~12s/it), which uses RSVD compression, and should be the fastest momentum compression (AFAIK).
-
----
-
-### 📈 Performance
-
-- **Better than Adafactor, and CAME factorzation methods**
-- **Comparable or identical to Adam** (see SMMF paper results)
-
----
-
-### Available Optimizers (all support `Factored` toggle)
-
-Set `Factored=False` to disable factorization and run as a full uncompressed optimizer (like vanilla Adam).
-
-1. **Adam**
-2. **Prodigy**
-3. **Adopt**
-
----
-
-### Bonus Features (Built-in)
-
-- **Fused Backward Pass**
-
-- **Stochastic Rounding (SR)**: Improves quality and convergence for **BF16 training**.
-
-- **[AdEMAMix](https://arxiv.org/abs/2409.03137)**  
-  → This adds a second, slow-moving EMA, which is combined with the primary momentum to stabilize updates, especially during long runs of full finetuning.
-  → A higher value of beta3 (e.g., 0.9999) gives the EMA a longer memory, making it more stable but slower to adapt. A lower value (e.g., 0.999) is often better for shorter training runs (2k-4k steps).
-  → When `factored` is true, it compresses the new momentum in the same way as the first moment (1-bit state + 2 vectors). However, this introduces noticeable overhead as we are compressing/reconstructing a third state each step.
-
-  ⚠️ **Note**: AdEMAMix updates are more aggressive than normal Adam/Adopt, so use a x2-x5 smaller LR than usual (or use Prodigy).
-
-- **[`atan2` smoothing & scaling](https://github.com/lucidrains/adam-atan2-pytorch)**  
-  → Robust `eps` replacement (no tuning!) + built-in gradient clipping  
-  → *Ideal for ADOPT* (which normally needs higher `eps` and clipping), so `use_atan2` is all-in-one for it.
-
-- **[OrthoGrad](https://github.com/LucasPrietoAl/grokking-at-the-edge-of-numerical-stability)**  
-  → Removes gradient component parallel to weights → prevents "naïve loss minimization" (NLM) → reduces natural overfitting  
-  → Perfect for fine-tuning the direction of existing features (e.g., full finetune or training a trained LoRA) without weight decay erasing prior knowledge.
-
-  ⚠️ **Note**: OrthoGrad introduces **~33% time overhead**, so take this into account.
-
-- **[Grams: Gradient Descent with Adaptive Momentum Scaling](https://github.com/Gunale0926/Grams)**  
-  → Eliminates the need for 1-bit momentum sign storage by using the **sign of gradients** for the first moment.
-
-  ⚠️ **Not recommended for small batch sizes**: gradients are too noisy, which can destabilize momentum (tested for Prodigy and it made the optimizer slower to find the LR or converge in BS 4).
-
-### Other Notes
-
-- **Adopt** skips the first step (only initializes the states) and has built-in clipping (sticking to the original optimizer), but we skip both of these when you enable `use_atan2`; as the optimizer becomes scale-invariant and the values of the states won't cause any issues or instability.
-
-- When `use_atan2` is True, `eps` will be ignored and you should also disable any gradient clipping.
-
----

adv_optm-0.1.7.dist-info/RECORD

@@ -1,19 +0,0 @@
-adv_optm/__init__.py,sha256=CZ_tjWWk5d5D8q_R0rcr8vvwlZyY_44zyAcIAmN_SDY,306
-adv_optm/optim/AdamW_adv.py,sha256=ZeNzk2tWbyd2QDI5hp4InwG3iuHHfqLrlhr_VmcQfRM,13884
-adv_optm/optim/Adopt_adv.py,sha256=rzBWfFOPrMuC6vwETsw7QPKmVXcv4IJRDCTj-6eU1Qk,14798
-adv_optm/optim/Lion_Prodigy_adv.py,sha256=JMss9X8lRpIU4E34PfFpWMMal_XNvZ8Yuqc6i7R5wIQ,14588
-adv_optm/optim/Lion_adv.py,sha256=BA4bSEhJiQ7BhGLDRn9nuMlBrLVh-OMscbmSTeGgRmI,10137
-adv_optm/optim/Prodigy_adv.py,sha256=gJL2r32R3xGD62jMR55ZyKxRv0yL70XHxj4FzEJbFc4,20196
-adv_optm/optim/Simplified_AdEMAMix.py,sha256=opIZjnGJ03-DDAIHTZyJBMReVfgusGDb8FZSWMU3-UM,9774
-adv_optm/optim/__init__.py,sha256=pcP865H2j1tut2VfTUhzQh7V8TF_tzPjqFnjMfFed2k,382
-adv_optm/util/BF16_Stochastic_Rounding.py,sha256=Q5H0BcogmE4atP65dLoI21HKSf50lRdsBDfeF6v9Tbg,1548
-adv_optm/util/Effective_Shape.py,sha256=TBvIk1V8IuTbbBsxuekJA4e_v8JlR5Nujtut8RTWAm4,318
-adv_optm/util/NNMF.py,sha256=yRf5IP5Sjq0Uf0DxN0Q8NxEGSdD-f1ULziLVDOjY8K4,639
-adv_optm/util/One_Bit_Boolean.py,sha256=Wat49esdwohuN-OHOFMW8D0aOQgV9cP5Rl8z6yfmpos,1068
-adv_optm/util/OrthoGrad.py,sha256=NzInuBQGy_Ja__M1R9XbvqVaQ0fhGbtGgFE9YON7B3I,707
-adv_optm/util/__init__.py,sha256=qoyIF0jcLjs_vSEcsv36clw5LFNBEbifyXrrVxMH-G4,349
-adv_optm-0.1.7.dist-info/licenses/LICENSE,sha256=HrhfyXIkWY2tGFK11kg7vPCqhgh5DcxleloqdhrpyMY,11558
-adv_optm-0.1.7.dist-info/METADATA,sha256=BEKyVG9zVdb9WThOw9YtgWZ_zqDmErumpY5Fr-AkbX0,5846
-adv_optm-0.1.7.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-adv_optm-0.1.7.dist-info/top_level.txt,sha256=iNfBIIzu-lPrQ7jyC56WBCcbkRwitM2nJ15-MRQ_6fg,9
-adv_optm-0.1.7.dist-info/RECORD,,