Fix seed for FSDP wrap

composer/core/state.py

@@ -27,7 +27,8 @@
 from composer.core.serializable import Serializable
 from composer.core.time import Time, Timestamp, TimeUnit
 from composer.devices import Device
-from composer.utils import batch_get, batch_set, dist, ensure_tuple, get_composer_env_dict, is_model_deepspeed
+from composer.utils import (batch_get, batch_set, dist, ensure_tuple, get_composer_env_dict, is_model_deepspeed,
+                            reproducibility)
 from composer.utils.misc import using_torch_2
 
 if TYPE_CHECKING:
@@ -1264,8 +1265,9 @@ def load_model_and_optimizer_state(
             assert self.fsdp_config is not None
             log.info('Wrapping model with FSDP after loading model_state.')
             from composer.trainer.dist_strategy import prepare_fsdp_module
-            prepare_fsdp_module(self.model, self.optimizers, self.fsdp_config, self.precision, self.device,
-                                self.auto_microbatching)
+            with reproducibility.seed_context(self.rank_zero_seed):
+                prepare_fsdp_module(self.model, self.optimizers, self.fsdp_config, self.precision, self.device,
+                                    self.auto_microbatching)
             log.debug('Finished wrapping model with FSDP.')
 
         # Legacy optimizer state load must happen after FSDP monolith

composer/trainer/dist_strategy.py

@@ -243,11 +243,6 @@ def prepare_fsdp_module(
                              'gpu and some ranks are on meta. Either keep all ranks on the same '
                              "device or set fsdp_config['sync_module_states'] = True. Otherwise, "
                              'some weights may be randomly initialized when loading a checkpoint.')
-        # Comment out while we debug deadlock
-        # if fsdp_config['sharding_strategy'] in ('HYBRID_SHARD', '_HYBRID_SHARD_ZERO2'):
-        #     raise ValueError('HSDP (HYBRID_SHARD or _HYBRID_SHARD_ZERO2) requires '
-        #                      'fsdp_config["sync_module_states"] = True or different replicas will '
-        #                      'have different weights.')
 
     # Check if other ranks OOMed after forward/backward pass when using auto microbatching. This
     # may happen when close to memory limit or with uneven memory usage across ranks. Since we

composer/trainer/trainer.py

@@ -1293,7 +1293,8 @@ def __init__(
 
         # FSDP wrap if not using monolith checkpoint on rank 0 only
         if self.state.fsdp_config is not None and fsdp_auto_wrap and not self.state.load_fsdp_monolith_rank0_only:
-            prepare_fsdp_module(model, optimizers, self.state.fsdp_config, precision, device, auto_microbatching)
+            with reproducibility.seed_context(self.state.rank_zero_seed):
+                prepare_fsdp_module(model, optimizers, self.state.fsdp_config, precision, device, auto_microbatching)
 
         # Configure Deepspeed
         if self.state.deepspeed_config is not None:
@@ -1440,7 +1441,8 @@ def __init__(
         # FSDP wrap if model is not yet wrapped and FSDP is enabled. This can happen if
         # load_fsdp_monolith_rank0_only=True but no checkpoint was loaded.
         if not self.state.fsdp_enabled and self.state.fsdp_config is not None and self.state.fsdp_auto_wrap and self.state.load_fsdp_monolith_rank0_only:
-            prepare_fsdp_module(model, optimizers, self.state.fsdp_config, precision, device, auto_microbatching)
+            with reproducibility.seed_context(self.state.rank_zero_seed):
+                prepare_fsdp_module(model, optimizers, self.state.fsdp_config, precision, device, auto_microbatching)
 
         self.engine.run_event(Event.AFTER_LOAD)
 

composer/utils/reproducibility.py

@@ -53,6 +53,7 @@
 import textwrap
 import time
 import warnings
+from contextlib import contextmanager
 from typing import Any, Dict, List
 
 import numpy as np
@@ -62,6 +63,7 @@
 from composer.utils import dist
 
 __all__ = [
+    'seed_context',
     'configure_deterministic_mode',
     'get_random_seed',
     'seed_all',
@@ -76,6 +78,15 @@
 MAX_SEED = 2**32 - 1
 
 
+@contextmanager
+def seed_context(seed: int):
+    """Context manager to store rng_state and reseed for duration of context."""
+    rng_state = get_rng_state()
+    seed_all(seed)
+    yield
+    load_rng_state(rng_state)
+
+
 def configure_deterministic_mode():
     """Configure PyTorch deterministic mode.