Grokfast support (#1917)

src/axolotl/core/trainer_builder.py

@@ -1287,6 +1287,18 @@ def get_post_trainer_create_callbacks(self, trainer):
 
         if self.cfg.lisa_step_interval and self.cfg.lisa_n_layers:
             callbacks.append(lisa_callback_factory(trainer))
+
+        if self.cfg.plugins:
+            plugin_manager = PluginManager.get_instance()
+            callbacks.extend(
+                [
+                    cb
+                    for cb in plugin_manager.add_callbacks_post_trainer(
+                        self.cfg, trainer
+                    )
+                    if cb
+                ]
+            )
         return callbacks
 
     def _get_trainer_cls(self):

src/axolotl/integrations/base.py

@@ -140,7 +140,7 @@ def create_lr_scheduler(
 
     def add_callbacks_pre_trainer(self, cfg, model):  # pylint: disable=unused-argument
         """
-        Adds callbacks to the trainer before training.
+        setup callbacks before creating the trainer.
 
         Parameters:
         cfg (dict): The configuration for the plugin.
@@ -155,14 +155,15 @@ def add_callbacks_post_trainer(
         self, cfg, trainer
     ):  # pylint: disable=unused-argument
         """
-        Adds callbacks to the trainer after training.
+        Adds callbacks to the trainer after creating the trainer.
+        This is useful for callbacks that require access to the model or trainer.
 
         Parameters:
         cfg (dict): The configuration for the plugin.
         trainer (object): The trainer object for training.
 
         Returns:
-        List[callable]: A list of callback functions to be added to the TrainingArgs
+        List[callable]: A list of callback functions to be added
         """
         return []
 
@@ -393,7 +394,9 @@ def add_callbacks_pre_trainer(self, cfg, model):
         """
         callbacks = []
         for plugin in self.plugins.values():
-            callbacks.extend(plugin.add_callbacks_pre_trainer(cfg, model))
+            plugin_callbacks = plugin.add_callbacks_pre_trainer(cfg, model)
+            if plugin_callbacks:  # if the plugin returned a list of callbacks
+                callbacks.extend(plugin_callbacks)
         return callbacks
 
     def add_callbacks_post_trainer(self, cfg, trainer):
@@ -409,7 +412,9 @@ def add_callbacks_post_trainer(self, cfg, trainer):
         """
         callbacks = []
         for plugin in self.plugins.values():
-            callbacks.extend(plugin.add_callbacks_post_trainer(cfg, trainer))
+            plugin_callbacks = plugin.add_callbacks_post_trainer(cfg, trainer)
+            if plugin_callbacks:
+                callbacks.extend(plugin_callbacks)
         return callbacks
 
     def post_train_unload(self, cfg):

src/axolotl/integrations/grokfast/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2024 Jaerin Lee, Bong Gyun Kang, Kihoon Kim, Kyoung Mu Lee
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

src/axolotl/integrations/grokfast/README.md

@@ -0,0 +1,13 @@
+# Grokfast Optimizer
+
+See https://github.com/ironjr/grokfast
+
+### Usage
+
+```yaml
+plugins:
+  - axolotl.integrations.grokfast.GrokfastPlugin
+
+grokfast_alpha: 2.0
+grokfast_lamb: 0.98
+```

src/axolotl/integrations/grokfast/__init__.py

@@ -0,0 +1,50 @@
+"""
+Grokfast plugin for Axolotl
+"""
+import logging
+
+from transformers.trainer_callback import TrainerCallback
+
+from ..base import BasePlugin
+from .args import GrokfastArgs  # pylint: disable=unused-import. # noqa: F401
+from .optimizer import gradfilter_ema
+
+LOG = logging.getLogger("axolotl.integrations.grokfast")
+
+
+class GrokfastCallbackHandler(TrainerCallback):
+    """
+    Transformer trainer callbacks for Grokfast
+    """
+
+    def __init__(self, *args_, alpha=0.98, lamb=2.0, **kwargs):
+        super().__init__(*args_, **kwargs)
+        self.grads = None
+        self.alpha = alpha
+        self.lamb = lamb
+
+    def on_train_begin(self, *args_, **kwargs):  # pylint: disable=unused-argument
+        self.grads = None
+
+    def on_pre_optimizer_step(
+        self, args_, state, control, **kwargs
+    ):  # pylint: disable=unused-argument
+        model = kwargs.pop("model")
+        self.grads = gradfilter_ema(model, self.grads, alpha=self.alpha, lamb=self.lamb)
+        return control
+
+
+class GrokfastPlugin(BasePlugin):
+    """
+    Plugin for Grokfast optimizer integraton with Axolotl.
+    """
+
+    def get_input_args(self):
+        return "axolotl.integrations.grokfast.GrokfastArgs"
+
+    def add_callbacks_post_trainer(self, cfg, trainer):
+        LOG.info("Adding Grokfast callback to the trainer")
+        callback = GrokfastCallbackHandler(
+            alpha=cfg.grokfast_alpha, lamb=cfg.grokfast_lamb
+        )
+        return [callback]

src/axolotl/integrations/grokfast/args.py

@@ -0,0 +1,15 @@
+"""
+config args for grokfast plugin
+"""
+from typing import Optional
+
+from pydantic import BaseModel
+
+
+class GrokfastArgs(BaseModel):
+    """
+    Input args for Grokfast optimizer.
+    """
+
+    grokfast_alpha: Optional[float] = 0.98
+    grokfast_lamb: Optional[float] = 2.0

src/axolotl/integrations/grokfast/optimizer.py

@@ -0,0 +1,63 @@
+# Copyright: MIT License (c) 2024 Jaerin Lee, Bong Gyun Kang, Kihoon Kim, Kyoung Mu Lee
+# Reference: https://github.com/ironjr/grokfast
+
+# pylint: skip-file
+from collections import deque
+from typing import Dict, Literal, Optional
+
+import torch
+import torch.nn as nn
+
+
+def gradfilter_ma(
+    m: nn.Module,
+    grads: Optional[Dict[str, deque]] = None,
+    window_size: int = 100,
+    lamb: float = 5.0,
+    filter_type: Literal["mean", "sum"] = "mean",
+    warmup: bool = True,
+    trigger: bool = False,  # For ablation study.
+) -> Dict[str, deque]:
+    if grads is None:
+        grads = {
+            n: deque(maxlen=window_size)
+            for n, p in m.named_parameters()
+            if p.requires_grad and p.grad is not None
+        }
+
+    for n, p in m.named_parameters():
+        if p.requires_grad and p.grad is not None:
+            grads[n].append(p.grad.data.detach())  # .cpu())
+
+            # Modify the gradients.
+            if not warmup or len(grads[n]) == window_size and not trigger:
+                if filter_type == "mean":
+                    avg = sum(grads[n]) / len(grads[n])
+                elif filter_type == "sum":
+                    avg = sum(grads[n])
+                else:
+                    raise ValueError(f"Unrecognized filter_type {filter_type}")
+                p.grad.data = p.grad.data + avg * lamb
+
+    return grads
+
+
+def gradfilter_ema(
+    m: nn.Module,
+    grads: Optional[Dict[str, torch.Tensor]] = None,
+    alpha: float = 0.98,
+    lamb: float = 2.0,
+) -> Dict[str, torch.Tensor]:
+    if grads is None:
+        grads = {
+            n: p.grad.data.detach()
+            for n, p in m.named_parameters()
+            if p.requires_grad and p.grad is not None
+        }
+
+    for n, p in m.named_parameters():
+        if p.requires_grad and p.grad is not None:
+            grads[n] = grads[n] * alpha + p.grad.data.detach() * (1 - alpha)
+            p.grad.data = p.grad.data + grads[n] * lamb
+
+    return grads

src/axolotl/utils/config/models/input/v0_4_1/__init__.py

@@ -783,6 +783,8 @@ class Config:
     is_mistral_derived_model: Optional[bool] = Field(default=None)
     is_qwen_derived_model: Optional[bool] = Field(default=None)
 
+    plugins: Optional[List[str]] = Field(default=None)
+
     @field_validator("datasets", mode="before")
     @classmethod
     def deprecate_sharegpt_datasets(cls, datasets):