fix all-gather overlap the model_checkpoint is 0

configs/7B_sft.py

@@ -1,7 +1,7 @@
 JOB_NAME = "7b_train"
 DO_ALERT = False
 
-SEQ_LEN = 4096
+SEQ_LEN = 2048
 HIDDEN_SIZE = 4096
 NUM_ATTENTION_HEAD = 32
 MLP_RATIO = 8 / 3
@@ -49,9 +49,9 @@ VALID_FOLDER = "/path/to/dataset"
 data = dict(
     seq_len=SEQ_LEN,
     # micro_num means the number of micro_batch contained in one gradient update
-    micro_num=1,
+    micro_num=4,
     # packed_length = micro_bsz * SEQ_LEN
-    micro_bsz=1,
+    micro_bsz=2,
     # defaults to the value of micro_num
     valid_micro_num=4,
     # defaults to 0, means disable evaluate
@@ -163,7 +163,7 @@ pipeline parallel (dict):
 """
 parallel = dict(
     zero1=dict(size=-1, fsdp=False),
-    tensor=dict(size=8, sp="intern", intern_overlap=True, reduce_scatter_overlap=True),
+    tensor=dict(size=4, sp="intern", intern_overlap=True, reduce_scatter_overlap=True),
     pipeline=dict(size=1, interleaved_overlap=True),
 )
 

internlm/model/overlap_handler.py

@@ -315,7 +315,7 @@ class FSTPOverlapHandler:
         # 1. register post_backward_hook @head module to prefetch for the last block's last module
         # 2. register pre_backward_hook @fstp_module to wait handle for current module and to prefetch for next module
         # 3. register post_backward_hook @fstp_module to release resource
-        if self.model_checkpoint is False:
+        if not self.model_checkpoint:
             for head in self.head:
                 head.register_full_backward_hook(_post_backward_hook_for_head)
 

internlm/train/training_internlm.py

@@ -407,6 +407,7 @@ def initialize_llm_profile(profiling: bool = False, start_time: str = None):
 tgs_list = []
 tflops_list = []
 tflops_list_2 = []
+loss_list = []
 
 
 @llm_timeout(func_name="record_current_batch_training_metrics")
@@ -599,11 +600,12 @@ def record_current_batch_training_metrics(
             step_count=batch_count,
             cur_step_loss=loss.item(),
         )
-
+        loss_list.append(loss.item())
         if batch_count >= 5:
             tgs_list.append(tgs_origin)
             tflops_list.append(tflops)
             tflops_list_2.append(tflops_2)
+            
         if batch_count == gpc.config.data.total_steps - 1:
             print(tgs_list, flush=True)
             avg_tgs = sum(tgs_list) / len(tgs_list)
@@ -625,3 +627,5 @@ def record_current_batch_training_metrics(
                 if abs(tf - avg_tflops_2) > 10:
                     tflops_list_2.remove(tf)
             print(f"avg_tflops_2: {sum(tflops_list_2)/len(tflops_list_2)}", flush=True)
+            
+            print("loss: ", loss_list, flush=True)