Update & Supplement with new custom optimizer

scripts/run_finetune_with_custom_optim.sh

@@ -18,10 +18,8 @@ optim=dummy
 # Select an optimizer from the following options:
 # - 'adamw_torch'
 # - 'adafactor'
+# - 'lamb'
 # - 'sgd'
-# - 'lion_8bit'
-# - 'lion_32bit'
-# - 'rmsprop'
 # Additional optimizers are shown below
 learning_rate=1e-5
 lr_schedule=cosine
@@ -187,6 +185,64 @@ elif [ "${optim}" == "lamb" ]; then
   optim_suffix_args+=" --optim_beta1 ${beta1}"
   optim_suffix_args+=" --optim_beta2 ${beta2}"
   optim_suffix_args+=" --optim_weight_decay ${weight_decay}"
+elif [ "${optim}" == "lion" ]; then
+  optim_suffix_args="--use_customized_optim 1"
+  optim_suffix_args+=" --customized_optim ${optim}"
+  optim_suffix_args+=" --optim_beta1 ${beta1}"
+  optim_suffix_args+=" --optim_beta2 ${beta2}"
+  optim_suffix_args+=" --optim_weight_decay ${weight_decay}"
+elif [ "${optim}" == "adamw" ]; then
+  optim_suffix_args="--use_customized_optim 1"
+  optim_suffix_args+=" --customized_optim ${optim}"
+  optim_suffix_args+=" --optim_beta1 ${beta1}"
+  optim_suffix_args+=" --optim_beta2 ${beta2}"
+  optim_suffix_args+=" --optim_weight_decay ${weight_decay}"
+elif [ "${optim}" == "adafactor" ]; then
+  optim_suffix_args="--use_customized_optim 1"
+  optim_suffix_args+=" --customized_optim ${optim}"
+  optim_suffix_args+=" --optim_beta1 ${beta1}"
+  optim_suffix_args+=" --optim_beta2 ${beta2}"
+  optim_suffix_args+=" --optim_weight_decay ${weight_decay}"
+elif [ "${optim}" == "came" ]; then
+  optim_suffix_args="--use_customized_optim 1"
+  optim_suffix_args+=" --customized_optim ${optim}"
+  optim_suffix_args+=" --optim_beta1 ${beta1}"
+  optim_suffix_args+=" --optim_beta2 ${beta2}"
+  optim_suffix_args+=" --optim_beta3 ${beta3}"
+  optim_suffix_args+=" --optim_weight_decay ${weight_decay}"
+elif [ "${optim}" == "qhadam" ]; then
+  optim_suffix_args="--use_customized_optim 1"
+  optim_suffix_args+=" --customized_optim ${optim}"
+  optim_suffix_args+=" --optim_beta1 ${beta1}"
+  optim_suffix_args+=" --optim_beta2 ${beta2}"
+  optim_suffix_args+=" --optim_weight_decay ${weight_decay}"
+elif [ "${optim}" == "adahessian" ]; then
+  optim_suffix_args="--use_customized_optim 1"
+  optim_suffix_args+=" --customized_optim ${optim}"
+  optim_suffix_args+=" --optim_beta1 ${beta1}"
+  optim_suffix_args+=" --optim_beta2 ${beta2}"
+  optim_suffix_args+=" --optim_weight_decay ${weight_decay}"
+elif [ "${optim}" == "swats" ]; then
+  optim_suffix_args="--use_customized_optim 1"
+  optim_suffix_args+=" --customized_optim ${optim}"
+  optim_suffix_args+=" --optim_beta1 ${beta1}"
+  optim_suffix_args+=" --optim_beta2 ${beta2}"
+  optim_suffix_args+=" --optim_weight_decay ${weight_decay}"
+elif [ "${optim}" == "qhm" ]; then
+  optim_suffix_args="--use_customized_optim 1"
+  optim_suffix_args+=" --customized_optim ${optim}"
+  optim_suffix_args+=" --optim_momentum ${momentum}"
+  optim_suffix_args+=" --optim_weight_decay ${weight_decay}"
+elif [ "${optim}" == "sgd" ]; then
+  optim_suffix_args="--use_customized_optim 1"
+  optim_suffix_args+=" --customized_optim ${optim}"
+  optim_suffix_args+=" --optim_momentum ${momentum}"
+  optim_suffix_args+=" --optim_weight_decay ${weight_decay}"
+elif [ "${optim}" == "rmsprop" ]; then
+  optim_suffix_args="--use_customized_optim 1"
+  optim_suffix_args+=" --customized_optim ${optim}"
+  optim_suffix_args+=" --optim_momentum ${momentum}"
+  optim_suffix_args+=" --optim_weight_decay ${weight_decay}"
 elif [ "${optim}" == "adamax" ]; then
   optim_suffix_args="--use_customized_optim 1"
   optim_suffix_args+=" --customized_optim ${optim}"
@@ -270,7 +326,7 @@ else
 fi
 
 # Finetune
-exp_id=alpaca_${optim}_lr-${learning_rate}_beta1-${beta1}_beta2-${beta2}_lr-sched-${lr_schedule}_model-$(basename ${model_name_or_path})_batch-size-${batch_size}x${gradient_accumulation_steps}_seed-${seed}
+exp_id=alpaca_${optim}_lr-${learning_rate}_beta1-${beta1}_beta2-${beta2}_beta3-${beta3}_momentum-${momentum}_lr-sched-${lr_schedule}_model-$(basename ${model_name_or_path})_batch-size-${batch_size}x${gradient_accumulation_steps}_seed-${seed}
 echo "$(date): ${exp_id}..."
 
 tmp_dir=tmp

src/lmflow/args.py

@@ -35,7 +35,9 @@ class OptimizerNames():
     ADABOUND = "adabound"
     LARS = "lars"
     LAMB = "lamb"
+    LION = "lion"
     ADAMAX = "adamax"
+    ADAFACTOR = "adafactor"
     NADAM = "nadam"
     RADAM = "radam"
     ADAMP = "adamp"
@@ -44,6 +46,14 @@ class OptimizerNames():
     SOPHIA = "sophia"
     ADAN = "adan"
     ADAM = "adam"
+    ADAMW = "adamw"
+    ADAHESSIAN = "adahessian"
+    CAME = "came"
+    QHADAM = "qhadam"
+    QHM = "qhm"
+    SWATS = "swats"
+    SGD = "sgd"
+    RMSPROP = "rmsprop"
     NOVOGRAD = "novograd"
     ADADELTA = "adadelta"
     ADAGRAD = "adagrad"

src/lmflow/optim/adafactor.py

@@ -0,0 +1,158 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+
+import torch
+import math
+from torch.optim.optimizer import Optimizer
+class Adafactor(Optimizer):
+    """Implements Adafactor algorithm.
+    This implementation is based on: `Adafactor: Adaptive Learning Rates with Sublinear Memory Cost`
+    (see https://arxiv.org/abs/1804.04235)
+
+    Note that this optimizer internally adjusts the learning rate depending on the
+    *scale_parameter*, *relative_step* and *warmup_init* options.
+
+    To use a manual (external) learning rate schedule you should set `scale_parameter=False` and
+    `relative_step=False`.
+
+    Arguments:
+        params (iterable): iterable of parameters to optimize or dicts defining parameter groups
+        lr (float, optional): external learning rate (default: None)
+        eps (tuple[float, float]): regularization constants for square gradient
+            and parameter scale respectively (default: (1e-30, 1e-3))
+        clip_threshold (float): threshold of root mean square of final gradient update (default: 1.0)
+        decay_rate (float): coefficient used to compute running averages of square gradient (default: -0.8)
+        beta1 (float): coefficient used for computing running averages of gradient (default: None)
+        weight_decay (float, optional): weight decay (L2 penalty) (default: 0)
+        scale_parameter (bool): if True, learning rate is scaled by root mean square of parameter (default: True)
+        warmup_init (bool): time-dependent learning rate computation depends on
+            whether warm-up initialization is being used (default: False)
+    """
+
+    def __init__(self, params, lr=None, eps=1e-30, eps_scale=1e-3, clip_threshold=1.0,
+                 decay_rate=-0.8, betas=None, weight_decay=0.0, scale_parameter=True, warmup_init=False):
+        relative_step = not lr
+        if warmup_init and not relative_step:
+            raise ValueError('warmup_init requires relative_step=True')
+
+        beta1 = None if betas is None else betas[0]   # make it compat with standard betas arg
+        defaults = dict(lr=lr, eps=eps, eps_scale=eps_scale, clip_threshold=clip_threshold, decay_rate=decay_rate,
+                        beta1=beta1, weight_decay=weight_decay, scale_parameter=scale_parameter,
+                        relative_step=relative_step, warmup_init=warmup_init)
+        super(Adafactor, self).__init__(params, defaults)
+
+    @staticmethod
+    def _get_lr(param_group, param_state):
+        if param_group['relative_step']:
+            min_step = 1e-6 * param_state['step'] if param_group['warmup_init'] else 1e-2
+            lr_t = min(min_step, 1.0 / math.sqrt(param_state['step']))
+            param_scale = 1.0
+            if param_group['scale_parameter']:
+                param_scale = max(param_group['eps_scale'], param_state['RMS'])
+            param_group['lr'] = lr_t * param_scale
+        return param_group['lr']
+
+    @staticmethod
+    def _get_options(param_group, param_shape):
+        factored = len(param_shape) >= 2
+        use_first_moment = param_group['beta1'] is not None
+        return factored, use_first_moment
+
+    @staticmethod
+    def _rms(tensor):
+        return tensor.norm(2) / (tensor.numel() ** 0.5)
+
+    def _approx_sq_grad(self, exp_avg_sq_row, exp_avg_sq_col):
+        r_factor = (exp_avg_sq_row / exp_avg_sq_row.mean(dim=-1, keepdim=True)).rsqrt_().unsqueeze(-1)
+        c_factor = exp_avg_sq_col.unsqueeze(-2).rsqrt()
+        return torch.mul(r_factor, c_factor)
+
+    @torch.no_grad()
+    def step(self, closure=None):
+        """Performs a single optimization step.
+        Arguments:
+            closure (callable, optional): A closure that reevaluates the model and returns the loss.
+        """
+        loss = None
+        if closure is not None:
+            with torch.enable_grad():
+                loss = closure()
+
+        for group in self.param_groups:
+            for p in group['params']:
+                if p.grad is None:
+                    continue
+                grad = p.grad
+                if grad.dtype in {torch.float16, torch.bfloat16}:
+                    grad = grad.float()
+                if grad.is_sparse:
+                    raise RuntimeError('Adafactor does not support sparse gradients.')
+
+                state = self.state[p]
+
+                factored, use_first_moment = self._get_options(group, grad.shape)
+                # State Initialization
+                if len(state) == 0:
+                    state['step'] = 0
+
+                    if use_first_moment:
+                        # Exponential moving average of gradient values
+                        state['exp_avg'] = torch.zeros_like(grad)
+                    if factored:
+                        state['exp_avg_sq_row'] = torch.zeros(grad.shape[:-1]).to(grad)
+                        state['exp_avg_sq_col'] = torch.zeros(grad.shape[:-2] + grad.shape[-1:]).to(grad)
+                    else:
+                        state['exp_avg_sq'] = torch.zeros_like(grad)
+
+                    state['RMS'] = 0
+                else:
+                    if use_first_moment:
+                        state['exp_avg'] = state['exp_avg'].to(grad)
+                    if factored:
+                        state['exp_avg_sq_row'] = state['exp_avg_sq_row'].to(grad)
+                        state['exp_avg_sq_col'] = state['exp_avg_sq_col'].to(grad)
+                    else:
+                        state['exp_avg_sq'] = state['exp_avg_sq'].to(grad)
+
+                p_fp32 = p
+                if p.dtype in {torch.float16, torch.bfloat16}:
+                    p_fp32 = p_fp32.float()
+
+                state['step'] += 1
+                state['RMS'] = self._rms(p_fp32)
+                lr_t = self._get_lr(group, state)
+
+                beta2t = 1.0 - math.pow(state['step'], group['decay_rate'])
+                update = grad ** 2 + group['eps']
+                if factored:
+                    exp_avg_sq_row = state['exp_avg_sq_row']
+                    exp_avg_sq_col = state['exp_avg_sq_col']
+
+                    exp_avg_sq_row.mul_(beta2t).add_(update.mean(dim=-1), alpha=1.0 - beta2t)
+                    exp_avg_sq_col.mul_(beta2t).add_(update.mean(dim=-2), alpha=1.0 - beta2t)
+
+                    # Approximation of exponential moving average of square of gradient
+                    update = self._approx_sq_grad(exp_avg_sq_row, exp_avg_sq_col)
+                    update.mul_(grad)
+                else:
+                    exp_avg_sq = state['exp_avg_sq']
+
+                    exp_avg_sq.mul_(beta2t).add_(update, alpha=1.0 - beta2t)
+                    update = exp_avg_sq.rsqrt().mul_(grad)
+
+                update.div_((self._rms(update) / group['clip_threshold']).clamp_(min=1.0))
+                update.mul_(lr_t)
+
+                if use_first_moment:
+                    exp_avg = state['exp_avg']
+                    exp_avg.mul_(group['beta1']).add_(update, alpha=1 - group['beta1'])
+                    update = exp_avg
+
+                if group['weight_decay'] != 0:
+                    p_fp32.add_(p_fp32, alpha=-group['weight_decay'] * lr_t)
+
+                p_fp32.add_(-update)
+                if p.dtype in {torch.float16, torch.bfloat16}:
+                    p.copy_(p_fp32)
+
+        return loss