fix bugs for pipeline

internlm/core/scheduler/pipeline_scheduler.py

@@ -154,13 +154,7 @@ class PipelineScheduler(BaseScheduler):
         self._tensor_shape = tensor_shape
 
     def pre_processing(self, engine):
-        types = set()
-
-        for param in engine.model.parameters():
-            types.add(param.dtype)
-        assert len(types) == 1, f"Mixed types of parameter detected, {types}"
-
-        self.dtype = types.pop()
+        self.dtype = gpc.config.model.get("dtype", torch.half)
 
     @staticmethod
     def _call_engine(engine, data):  # pylint: disable=W0237
@@ -430,6 +424,7 @@ class PipelineScheduler(BaseScheduler):
                     comm.send_obj_meta(output_obj)
                     need_forward_meta = False  # send only once.
                 # Send the forward computation output to the next stage
+                assert output_obj.dtype == self.dtype
                 comm.send_forward(output_obj, scatter_gather_tensors=self.scatter_gather_tensors)
 
         output, label = pack_return_tensors(return_tensors) if len(return_tensors) > 0 else (None, None)
@@ -537,6 +532,7 @@ class PipelineScheduler(BaseScheduler):
             # Send the output of forward computation of this pipeline stage to the next pipeline stage as input for
             # forward computation
             if not gpc.is_last_rank(ParallelMode.PIPELINE):
+                assert output_obj.dtype == self.dtype
                 comm.send_forward(output_obj, scatter_gather_tensors=self.scatter_gather_tensors)
 
             input_objs.append(input_obj)
@@ -572,6 +568,7 @@ class PipelineScheduler(BaseScheduler):
             if gpc.is_last_rank(ParallelMode.PIPELINE):
                 output_obj_grad = None
             else:
+                assert output_obj.dtype == self.dtype
                 output_obj_grad = comm.send_forward_recv_backward(
                     output_obj,
                     backward_recv_shapes,
@@ -984,6 +981,7 @@ class InterleavedPipelineScheduler(PipelineScheduler):
             # in this iteration; receive tensors for next iteration).
             if k != (num_warmup_microsteps - 1) or not receive_extra_backward:
                 # Normal warm-up communication process, or no need to prepare backward input for the 1F1B stage
+                assert output_obj.dtype == self.dtype
                 input_obj = comm.send_forward_recv_forward(
                     output_obj,
                     input_shape,
@@ -995,6 +993,7 @@ class InterleavedPipelineScheduler(PipelineScheduler):
                 if self._communication_overlap:
                     # In this case, we should handle forward and backward communication separately, consistent with the
                     # overlap version of the 1F1B stage
+                    assert output_obj.dtype == self.dtype
                     input_obj = comm.send_forward_recv_forward(
                         output_obj,
                         input_shape,
@@ -1011,6 +1010,7 @@ class InterleavedPipelineScheduler(PipelineScheduler):
                 else:
                     # In this case, we should handle forward and backward communication together, consistent with the
                     # non-overlap version of the 1F1B stage
+                    assert output_obj.dtype == self.dtype
                     input_obj, output_obj_grad = comm.send_forward_backward_recv_forward_backward(
                         output_obj,
                         None,  # no backward grad to send
@@ -1203,6 +1203,7 @@ class InterleavedPipelineScheduler(PipelineScheduler):
             output_shape = self._output_obj_shapes[next_backward_chunk_id] if recv_next else None
 
             # Communicate objs.
+            assert output_obj.dtype == self.dtype
             input_obj, output_obj_grad = comm.send_forward_backward_recv_forward_backward(
                 output_obj,
                 input_obj_grad,

internlm/train/utils.py

@@ -1,7 +1,9 @@
 from typing import Dict, Tuple
 
 import torch
+import torch.distributed as dist
 
+from internlm.core.context.parallel_context import ParallelMode
 from internlm.core.context.parallel_context import global_context as gpc
 from internlm.model.utils import is_gate_param, is_moe_param, is_norm_param
 
@@ -74,7 +76,10 @@ def split_params_into_different_groups_for_optimizer(param_groups: Tuple[Dict])
         pgroup["params"] = origin_params
 
     for _, g in new_groups.items():
-        if g["params"]:
+        # remove empty group, especially for fp32 group
+        is_empty = torch.tensor(bool(g["params"]), device=torch.cuda.current_device())
+        dist.all_reduce(is_empty, group=gpc.get_group(ParallelMode.MODEL))
+        if is_empty:
             param_groups.append(g)
 
     return tuple(param_groups)