Merge branch 'master' into jomayeri/aio-file-offset

deepspeed/runtime/zero/parameter_offload.py

@@ -133,7 +133,6 @@ class DeepSpeedZeRoOffload(object):
         self.persistent_parameters = self.mark_persistent_parameters(self.param_numel_persistence_threshold,
                                                                      self.model_persistence_threshold)
 
-        self.param_coordinators = {}
         self._prefetch_bucket_sz = int(prefetch_bucket_size)
         self._max_reuse_distance_in_numel = int(max_reuse_distance)
         self._max_available_parameters_in_numel = int(max_live_parameters)
@@ -141,12 +140,21 @@ class DeepSpeedZeRoOffload(object):
         ) if overlap_comm else get_accelerator().default_stream()
 
         if not hasattr(module, "ds_inflight_param_registry"):
-            module.ds_inflight_param_registry = dict()
+            module.ds_inflight_param_registry = InflightParamRegistry()
-            # we need two registries, one for training and one for eval. They will be used when creating PartitionedParameterCoordinator
-            module.ds_inflight_param_registry[True] = InflightParamRegistry()
-            module.ds_inflight_param_registry[False] = InflightParamRegistry()
         self.__inflight_param_registry = module.ds_inflight_param_registry
 
+        self.param_coordinator = PartitionedParameterCoordinator(
+            prefetch_bucket_sz=self._prefetch_bucket_sz,
+            max_reuse_distance_in_numel=self._max_reuse_distance_in_numel,
+            max_available_parameters_in_numel=self._max_available_parameters_in_numel,
+            allgather_stream=self.__allgather_stream,
+            inflight_param_registry=self.__inflight_param_registry,
+            prefetch_nvme=self.offload_device == OffloadDeviceEnum.nvme,
+            timers=self.timers,
+            zero_quantized_weights=self.zero_quantized_weights,
+            zero_quantized_nontrainable_weights=self.zero_quantized_nontrainable_weights,
+        )
+
         self.forward_hooks = []
         self.backward_hooks = []
         self.setup_zero_stage3_hooks()
@@ -161,26 +169,13 @@ class DeepSpeedZeRoOffload(object):
         """Partitioning Parameters that were not partitioned usually if parameters
         of modules whose input parameters do not require grad computation do not
         trigger post call and will therefore will remain unpartitioned"""
-        self.get_param_coordinator(training=self.module.training).release_and_reset_all(self.module)
+        self.get_param_coordinator().release_and_reset_all(self.module)
         for param in iter_params(self.module, recurse=True):
             if param.ds_status != ZeroParamStatus.NOT_AVAILABLE:
                 raise RuntimeError(f"{param.ds_summary()} expected to be released")
 
-    def get_param_coordinator(self, training):
+    def get_param_coordinator(self):
-        if not training in self.param_coordinators:
+        return self.param_coordinator
-            self.param_coordinators[training] = PartitionedParameterCoordinator(
-                prefetch_bucket_sz=self._prefetch_bucket_sz,
-                max_reuse_distance_in_numel=self._max_reuse_distance_in_numel,
-                max_available_parameters_in_numel=self._max_available_parameters_in_numel,
-                allgather_stream=self.__allgather_stream,
-                inflight_param_registry=self.__inflight_param_registry[training],
-                prefetch_nvme=self.offload_device == OffloadDeviceEnum.nvme,
-                timers=self.timers,
-                zero_quantized_weights=self.zero_quantized_weights,
-                zero_quantized_nontrainable_weights=self.zero_quantized_nontrainable_weights,
-            )
-
-        return self.param_coordinators[training]
 
     def empty_partition_cache(self):
         self.partition_all_parameters()
@@ -228,14 +223,14 @@ class DeepSpeedZeRoOffload(object):
 
         #reset step if in inference mode
         @instrument_w_nvtx
-        def _end_of_forward_hook(module, *args):
+        def _start_of_forward_hook(module, *args):
 
-            if not torch._C.is_grad_enabled():
+            self.get_param_coordinator().reset_step()
-                self.get_param_coordinator(training=False).reset_step()
+
+        self.module.register_forward_pre_hook(_start_of_forward_hook)
 
         #likely one of them should be enough but just to be safe
         self._register_hooks_recursively(self.module)
-        self.module.register_forward_hook(_end_of_forward_hook)
 
         # Add top module to stack trace
         global FWD_MODULE_STACK
@@ -447,7 +442,7 @@ class DeepSpeedZeRoOffload(object):
         global FWD_MODULE_STACK
         FWD_MODULE_STACK.append(sub_module)
 
-        param_coordinator = self.get_param_coordinator(training=sub_module.training)
+        param_coordinator = self.get_param_coordinator()
         param_coordinator.trace_prologue(sub_module)
         if param_coordinator.is_record_trace():
             param_coordinator.record_module(sub_module)
@@ -460,7 +455,7 @@ class DeepSpeedZeRoOffload(object):
         see_memory_usage(f"After sub module function {sub_module.__class__.__name__} {sub_module.id} before release",
                          force=False)
 
-        param_coordinator = self.get_param_coordinator(training=sub_module.training)
+        param_coordinator = self.get_param_coordinator()
         param_coordinator.release_sub_module(sub_module)
 
         see_memory_usage(f"After sub module function {sub_module.__class__.__name__}  {sub_module.id} after release",
@@ -468,8 +463,8 @@ class DeepSpeedZeRoOffload(object):
 
     @torch.no_grad()
     def pre_sub_module_backward_function(self, sub_module):
-        assert sub_module.training, "backward pass is invalid for module in evaluation mode"
+        # assert sub_module.training, "backward pass is invalid for module in evaluation mode"
-        param_coordinator = self.get_param_coordinator(training=True)
+        param_coordinator = self.get_param_coordinator()
         param_coordinator.trace_prologue(sub_module)
         if param_coordinator.is_record_trace():
             param_coordinator.record_module(sub_module)
@@ -477,12 +472,12 @@ class DeepSpeedZeRoOffload(object):
 
     @torch.no_grad()
     def post_sub_module_backward_function(self, sub_module):
-        assert sub_module.training, "backward pass is invalid for module in evaluation mode"
+        # assert sub_module.training, "backward pass is invalid for module in evaluation mode"
         see_memory_usage(
             f"After sub module backward function {sub_module.__class__.__name__} {sub_module.id} before release",
             force=False)
 
-        self.get_param_coordinator(training=True).release_sub_module(sub_module)
+        self.get_param_coordinator().release_sub_module(sub_module)
 
         see_memory_usage(
             f"After sub module backward function {sub_module.__class__.__name__} {sub_module.id} after release",

deepspeed/runtime/zero/partitioned_param_coordinator.py

@@ -18,6 +18,7 @@ from deepspeed.runtime.swap_tensor.partitioned_param_swapper import PartitionedP
 from deepspeed.utils.debug import debug_module2name_id, debug_param2name_id
 from deepspeed.accelerator import get_accelerator
 import deepspeed.runtime.compiler as compiler
+from deepspeed.runtime.compiler import is_compiling
 
 import logging
 
@@ -92,7 +93,7 @@ class PartitionedParameterCoordinator:
         # keeps track of the number of submodules invoked so far.
         self.__step_id: int = 0
         # network tracing mode
-        self.__trace_mode: ZeRoTraceMode = ZeRoTraceMode.RECORD
+        self.__trace_mode: ZeRoTraceMode = ZeRoTraceMode.INVALID
         # sequence of submodules/parameters in forward pass + backward pass
         self.__submodule_order: Iterable[Module] = []
         self.__param_order: Iterable[__class__.__ParamInTrace] = []
@@ -188,6 +189,9 @@ class PartitionedParameterCoordinator:
     @compiler.disable
     def record_module(self, sub_module: Module) -> None:
         """adds sub module to trace"""
+        if is_compiling():
+            return
+
         if not self.is_record_trace():
             raise RuntimeError(f"attempted to record trace when status = {self.__trace_mode}")
 
@@ -195,6 +199,8 @@ class PartitionedParameterCoordinator:
         self.__step_id_module_fetched_for[sub_module.id].append(self.__step_id)
 
     def record_parameters(self, sub_module: Module) -> None:
+        if is_compiling():
+            return
         """adds sub module to trace"""
         if not self.is_record_trace():
             raise RuntimeError(f"attempted to record trace when status = {self.__trace_mode}")
@@ -209,8 +215,12 @@ class PartitionedParameterCoordinator:
         for sub_module in self.__submodule_order:
             self.record_parameters(sub_module)
 
+    @compiler.disable
     def reset_step(self) -> None:
         """indicate that we have completed one fwd+bwd for the model"""
+        if is_compiling():
+            return
+
         self._clean_inflight_param_registry()
 
         if not self.is_complete_trace():  # not self.trace_complete:

deepspeed/runtime/zero/stage3.py

@@ -593,8 +593,8 @@ class DeepSpeedZeroOptimizer_Stage3(ZeROOptimizer):
 
         return device_buffer
 
-    def _get_param_coordinator(self, training):
+    def _get_param_coordinator(self):
-        return self.parameter_offload.get_param_coordinator(training)
+        return self.parameter_offload.get_param_coordinator()
 
     def _configure_offloading(self, offload_optimizer_config, offload_param_config):
         ###################### offload optimizer setup ##################################
@@ -1874,7 +1874,7 @@ class DeepSpeedZeroOptimizer_Stage3(ZeROOptimizer):
         see_memory_usage(f"In step before checking overflow", force=False)
 
         print_rank_0("Finished Tracing at Beginning of Step")
-        self._get_param_coordinator(training=True).hierarchy = 0
+        self._get_param_coordinator().hierarchy = 0
 
         print_rank_0("Finished Tracing at Beginning of Step")
 
@@ -2258,8 +2258,6 @@ class DeepSpeedZeroOptimizer_Stage3(ZeROOptimizer):
         else:
             self.loss_scaler.backward(loss.float(), retain_graph=retain_graph)
 
-        self._get_param_coordinator(training=True).reset_step()
-
         if self.swap_optimizer:
             self.optimizer_swapper.post_backward()
 

tests/unit/runtime/zero/test_zero.py

@@ -1628,3 +1628,48 @@ class TestEmptyParameterGroup(DistributedTest):
             optimizer=optimizer,
             config=config_dict,
         )
+
+
+class TestZero3SwitchModes(DistributedTest):
+    world_size = 2
+
+    @pytest.mark.parametrize("prefetch_ratio", [0.0, 0.5, 1.0])
+    def test(self, prefetch_ratio, zero_stage=3):
+
+        hidden_dim = 10
+        model = SimpleModel(hidden_dim)
+
+        prefetch_bucket_size = int(sum([p.numel() for p in model.parameters(recurse=True)]) * prefetch_ratio)
+        config_dict = {
+            "train_micro_batch_size_per_gpu": 2,
+            "gradient_accumulation_steps": 2,
+            "zero_optimization": {
+                "stage": zero_stage,
+                "stage3_prefetch_bucket_size": prefetch_bucket_size
+            },
+            "optimizer": {
+                "type": "Adam",
+                "params": {
+                    "lr": 1e-3
+                }
+            },
+            "fp16": {
+                "enabled": True,
+                "initial_scale_power": 8
+            }
+        }
+
+        model, _, _, _ = deepspeed.initialize(config=config_dict, model=model, model_parameters=model.parameters())
+        data_loader = random_dataloader(model=model, total_samples=16, hidden_dim=hidden_dim, device=model.device)
+
+        for _ in range(3):
+            model.train()
+            for batch in data_loader:
+                loss = model(batch[0], batch[1])
+                model.backward(loss)
+                model.step()
+
+            model.eval()
+            with torch.no_grad():
+                for batch in data_loader:
+                    loss = model(batch[0], batch[1])