AbhishekGhosh1998 commited on
Commit
e9207c6
·
verified ·
1 Parent(s): 06ac8e4

granularity results echo/tierbtrace @ 2026-08-26T20:03:32Z from sp-aiinfra-interactive-aghosh-b5xzj-8xb200-moe-cg-0

Browse files
.gitattributes CHANGED
@@ -48,3 +48,4 @@ echo/tierbprobe/tierbprobe_gbs8/QWEN3_30B__echo__matched__seq4096__gbs8__cap2p00
48
  echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep filter=lfs diff=lfs merge=lfs -text
49
  echo/tierbtrace/tierbtrace_gbs8/QWEN3_30B__echo__matched__seq4096__gbs8__cap1p11328125/nsys/QWEN3_30B__echo__matched__seq4096__gbs8__cap1p11328125.nsys-rep filter=lfs diff=lfs merge=lfs -text
50
  echo/tierbtrace/tierbtrace_gbs8/DSV2_LITE_L24__echo__matched__seq8192__gbs8__cap1p0755208333333333/nsys/DSV2_LITE_L24__echo__matched__seq8192__gbs8__cap1p0755208333333333.nsys-rep filter=lfs diff=lfs merge=lfs -text
 
 
48
  echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep filter=lfs diff=lfs merge=lfs -text
49
  echo/tierbtrace/tierbtrace_gbs8/QWEN3_30B__echo__matched__seq4096__gbs8__cap1p11328125/nsys/QWEN3_30B__echo__matched__seq4096__gbs8__cap1p11328125.nsys-rep filter=lfs diff=lfs merge=lfs -text
50
  echo/tierbtrace/tierbtrace_gbs8/DSV2_LITE_L24__echo__matched__seq8192__gbs8__cap1p0755208333333333/nsys/DSV2_LITE_L24__echo__matched__seq8192__gbs8__cap1p0755208333333333.nsys-rep filter=lfs diff=lfs merge=lfs -text
51
+ echo/tierbtrace/tierbtrace_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__cap1p08984375/nsys/QWEN2_57B__echo__matched__seq4096__gbs8__cap1p08984375.nsys-rep filter=lfs diff=lfs merge=lfs -text
echo/tierbtrace/tierbtrace_gbs8/DSV2_LITE_L24__echo__matched__seq8192__gbs8__cap1p0755208333333333/nsys/train.log CHANGED
@@ -1569,3 +1569,47 @@ Capture range ended in the application.
1569
  Generating '/tmp/nsys-report-bfe7.qdstrm'
1570
 
1571
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1569
  Generating '/tmp/nsys-report-bfe7.qdstrm'
1570
 
1571
 
1572
+ Collection stop is not allowed in this state.
1573
+ W0826 19:55:34.589000 711460 torch/distributed/elastic/agent/server/api.py:719] Received 15 death signal, shutting down workers
1574
+ W0826 19:55:34.590000 711460 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 711575 closing signal SIGTERM
1575
+ Generated:
1576
+ /workspace/tierb_study_20260826_144644/echo/tierbtrace/tierbtrace_gbs8/DSV2_LITE_L24__echo__matched__seq8192__gbs8__cap1p0755208333333333/nsys/DSV2_LITE_L24__echo__matched__seq8192__gbs8__cap1p0755208333333333.nsys-rep
1577
+ W0826 19:55:34.602000 711460 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 711576 closing signal SIGTERM
1578
+ W0826 19:55:34.614000 711460 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 711577 closing signal SIGTERM
1579
+ W0826 19:55:34.628000 711460 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 711578 closing signal SIGTERM
1580
+ W0826 19:55:34.635000 711460 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 711579 closing signal SIGTERM
1581
+ W0826 19:55:34.645000 711460 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 711580 closing signal SIGTERM
1582
+ W0826 19:55:34.665000 711460 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 711581 closing signal SIGTERM
1583
+ W0826 19:55:34.672000 711460 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 711582 closing signal SIGTERM
1584
+ Traceback (most recent call last):
1585
+ File "/usr/local/bin/torchrun", line 8, in <module>
1586
+ sys.exit(main())
1587
+ ^^^^^^
1588
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 355, in wrapper
1589
+ return f(*args, **kwargs)
1590
+ ^^^^^^^^^^^^^^^^^^
1591
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/run.py", line 892, in main
1592
+ run(args)
1593
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/run.py", line 883, in run
1594
+ elastic_launch(
1595
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/launcher/api.py", line 139, in __call__
1596
+ return launch_agent(self._config, self._entrypoint, list(args))
1597
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1598
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/launcher/api.py", line 261, in launch_agent
1599
+ result = agent.run()
1600
+ ^^^^^^^^^^^
1601
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/metrics/api.py", line 138, in wrapper
1602
+ result = f(*args, **kwargs)
1603
+ ^^^^^^^^^^^^^^^^^^
1604
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/agent/server/api.py", line 711, in run
1605
+ result = self._invoke_run(role)
1606
+ ^^^^^^^^^^^^^^^^^^^^^^
1607
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/agent/server/api.py", line 870, in _invoke_run
1608
+ time.sleep(monitor_interval)
1609
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/multiprocessing/api.py", line 84, in _terminate_process_handler
1610
+ raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval)
1611
+ torch.distributed.elastic.multiprocessing.api.SignalException: Process 711460 got signal: 15
1612
+ The target application terminated. One or more process it created re-parented.
1613
+ Waiting for termination of re-parented processes.
1614
+ Use the `--wait` option to modify this behavior.
1615
+ There are no active sessions.
echo/tierbtrace/tierbtrace_gbs8/DSV2_LITE_L24__echo__matched__seq8192__gbs8__cap1p0755208333333333/nsys/verify_trace_ranks.log ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ report : /workspace/tierb_study_20260826_144644/echo/tierbtrace/tierbtrace_gbs8/DSV2_LITE_L24__echo__matched__seq8192__gbs8__cap1p0755208333333333/nsys/DSV2_LITE_L24__echo__matched__seq8192__gbs8__cap1p0755208333333333.nsys-rep
2
+ kernels : 213065
3
+ devices : 8 of 8 -> [0, 1, 2, 3, 4, 5, 6, 7]
4
+ processes: 1
5
+ PASS: all ranks present
echo/tierbtrace/tierbtrace_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__cap1p08984375/nsys/QWEN2_57B__echo__matched__seq4096__gbs8__cap1p08984375.nsys-rep ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:02425365be4575804ac83416a996d4b93a01ae084b61880f965d3d5ffed62bf9
3
+ size 7395326
echo/tierbtrace/tierbtrace_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__cap1p08984375/nsys/run.env ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ backend=echo
2
+ config=QWEN2_57B
3
+ family=tierB
4
+ repo=/workspace/Megatron-MoE-Echo
5
+ branch=bracing-regions-through-kernels
6
+ commit=9a236793
7
+ num_layers=10
8
+ hidden=3584
9
+ ffn_hidden=2560
10
+ moe_ffn_hidden=2560
11
+ num_experts=64
12
+ topk=8
13
+ experts_per_rank=8
14
+ expert_params=27525120
15
+ heads=28
16
+ kv_groups=4
17
+ kv_channels=
18
+ qk_layernorm=0
19
+ shared_ffn=20480
20
+ shared_expert_overlap=0
21
+ mla=0
22
+ dense_first_layers=0
23
+ moe_layers=10
24
+ rotary_base=1000000
25
+ rotary_scaling=
26
+ qkv_bias=1
27
+ router_balancing=aux_loss
28
+ aux_loss_coeff=1e-2
29
+ router_pre_softmax=1
30
+ router_fusion=0
31
+ router_dtype=fp32
32
+ device_initiated_grouped_gemm=1
33
+ norm_epsilon=1e-6
34
+ seq_length=4096
35
+ micro_batch_size=1
36
+ global_batch_size=8
37
+ num_microbatches=1
38
+ assignments_per_rank=32768
39
+ mean_tokens_per_expert=4096
40
+ slot_policy=matched
41
+ slots_per_rank=8
42
+ capacity_mode=static
43
+ capacity=1.08984375
44
+ cuda_graph=1
45
+ probe_mode=0
46
+ async_grad_reduce=0
47
+ branch_role=timing
48
+ memory_dump_mode=none
49
+ memory_snapshot_path=
50
+ cuda_graph_warmup_steps=3
51
+ recompute=none
52
+ force_balance=0
53
+ train_iters=10
54
+ ep_size=8
55
+ echo_dispatch_overlap=1
56
+ echo_dispatch_recompute=0
57
+ ultraep_quota_min_tokens=
58
+ ultraep_weight_sync_join=
59
+ profile_enable=1
60
+ kernel_trace=1
61
+ fp8_recipe=mxfp8
echo/tierbtrace/tierbtrace_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__cap1p08984375/nsys/train.log ADDED
@@ -0,0 +1,1231 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ==============================================================
2
+ QWEN2_57B / echo (slots/rank=8 policy=matched)
3
+ E=64 K=8 H'=2560 m=8
4
+ L=10 seq=4096 gbs=8 (1 microbatches)
5
+ cap_mode=static cap=1.08984375 cg=1 recompute=none
6
+ memory_dump=none snapshots=/workspace/tierb_study_20260826_144644/echo/tierbtrace/tierbtrace_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__cap1p08984375/nsys/memory_snapshots/snapshot.pickle
7
+ repo=/workspace/Megatron-MoE-Echo @ 9a236793
8
+ ==============================================================
9
+ W0826 19:56:00.419000 723080 torch/distributed/run.py:766]
10
+ W0826 19:56:00.419000 723080 torch/distributed/run.py:766] *****************************************
11
+ W0826 19:56:00.419000 723080 torch/distributed/run.py:766] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed.
12
+ W0826 19:56:00.419000 723080 torch/distributed/run.py:766] *****************************************
13
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
14
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
15
+ using world size: 8, data-parallel size: 8, context-parallel size: 1, hierarchical context-parallel sizes: None, tensor-model-parallel size: 1, pipeline-model-parallel size: 1
16
+ WARNING: overriding default arguments for tokenizer_type:GPT2BPETokenizer with tokenizer_type:GPT2BPETokenizer
17
+ Number of virtual stages per pipeline stage: None
18
+ accumulate and all-reduce gradients in fp32 for bfloat16 data type.
19
+ using torch.bfloat16 for parameters ...
20
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
21
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
22
+ ------------------------ arguments ------------------------
23
+ account_for_embedding_in_pipeline_split ......... False
24
+ account_for_loss_in_pipeline_split .............. False
25
+ accumulate_allreduce_grads_in_fp32 .............. True
26
+ adam_beta1 ...................................... 0.9
27
+ adam_beta2 ...................................... 0.999
28
+ adam_eps ........................................ 1e-08
29
+ add_bias_linear ................................. False
30
+ add_position_embedding .......................... False
31
+ add_qkv_bias .................................... True
32
+ adlr_autoresume ................................. False
33
+ adlr_autoresume_interval ........................ 1000
34
+ align_grad_reduce ............................... True
35
+ align_param_gather .............................. False
36
+ app_tag_run_name ................................ None
37
+ app_tag_run_version ............................. 0.0.0
38
+ apply_layernorm_1p .............................. False
39
+ apply_query_key_layer_scaling ................... False
40
+ apply_residual_connection_post_layernorm ........ False
41
+ apply_rope_fusion ............................... True
42
+ async_save ...................................... None
43
+ async_tensor_model_parallel_allreduce ........... True
44
+ attention_backend ............................... AttnBackend.auto
45
+ attention_dropout ............................... 0.0
46
+ attention_softmax_in_fp32 ....................... False
47
+ auto_detect_ckpt_format ......................... False
48
+ barrier_with_L1_time ............................ True
49
+ bert_binary_head ................................ True
50
+ bert_embedder_type .............................. megatron
51
+ bert_load ....................................... None
52
+ bf16 ............................................ True
53
+ bias_dropout_fusion ............................. True
54
+ bias_gelu_fusion ................................ False
55
+ bias_swiglu_fusion .............................. True
56
+ biencoder_projection_dim ........................ 0
57
+ biencoder_shared_query_context_model ............ False
58
+ block_data_path ................................. None
59
+ calc_ft_timeouts ................................ False
60
+ calculate_per_token_loss ........................ False
61
+ check_for_large_grads ........................... False
62
+ check_for_nan_in_loss_and_grad .................. False
63
+ check_for_spiky_loss ............................ False
64
+ check_weight_hash_across_dp_replicas_interval ... None
65
+ ckpt_assume_constant_structure .................. False
66
+ ckpt_convert_format ............................. None
67
+ ckpt_convert_save ............................... None
68
+ ckpt_convert_update_legacy_dist_opt_format ...... False
69
+ ckpt_format ..................................... torch_dist
70
+ ckpt_fully_parallel_load ........................ False
71
+ ckpt_fully_parallel_save ........................ True
72
+ ckpt_fully_parallel_save_deprecated ............. False
73
+ ckpt_step ....................................... None
74
+ classes_fraction ................................ 1.0
75
+ clip_grad ....................................... 1.0
76
+ clone_scatter_output_in_embedding ............... True
77
+ config_logger_dir ...............................
78
+ consumed_train_samples .......................... 0
79
+ consumed_valid_samples .......................... 0
80
+ context_parallel_size ........................... 1
81
+ cp_comm_type .................................... ['p2p']
82
+ create_attention_mask_in_dataloader ............. True
83
+ cross_entropy_fusion_impl ....................... native
84
+ cross_entropy_loss_fusion ....................... False
85
+ cuda_graph_scope ................................ full_iteration
86
+ cuda_graph_warmup_steps ......................... 3
87
+ data_args_path .................................. None
88
+ data_cache_path ................................. None
89
+ data_parallel_random_init ....................... False
90
+ data_parallel_sharding_strategy ................. no_shard
91
+ data_parallel_size .............................. 8
92
+ data_path ....................................... ['/workspace/data/wikitext/wikitext103_gpt2_text_document']
93
+ data_per_class_fraction ......................... 1.0
94
+ data_sharding ................................... True
95
+ dataloader_type ................................. single
96
+ ddp_average_in_collective ....................... False
97
+ ddp_bucket_size ................................. None
98
+ ddp_num_buckets ................................. None
99
+ ddp_pad_buckets_for_high_nccl_busbw ............. False
100
+ decoder_first_pipeline_num_layers ............... None
101
+ decoder_last_pipeline_num_layers ................ None
102
+ decoder_num_layers .............................. None
103
+ decoder_seq_length .............................. None
104
+ decoupled_lr .................................... None
105
+ decoupled_min_lr ................................ None
106
+ decrease_batch_size_if_needed ................... False
107
+ defer_embedding_wgrad_compute ................... False
108
+ delay_wgrad_compute ............................. False
109
+ deprecated_use_mcore_models ..................... True
110
+ deterministic_mode .............................. False
111
+ dino_bottleneck_size ............................ 256
112
+ dino_freeze_last_layer .......................... 1
113
+ dino_head_hidden_size ........................... 2048
114
+ dino_local_crops_number ......................... 10
115
+ dino_local_img_size ............................. 96
116
+ dino_norm_last_layer ............................ False
117
+ dino_teacher_temp ............................... 0.07
118
+ dino_warmup_teacher_temp ........................ 0.04
119
+ dino_warmup_teacher_temp_epochs ................. 30
120
+ disable_bf16_reduced_precision_matmul ........... False
121
+ disable_mamba_mem_eff_path ...................... False
122
+ disable_straggler_on_startup .................... False
123
+ dist_ckpt_format_deprecated ..................... None
124
+ dist_ckpt_strictness ............................ assume_ok_unexpected
125
+ distribute_saved_activations .................... False
126
+ distributed_backend ............................. nccl
127
+ distributed_timeout_minutes ..................... 10
128
+ embedding_init_method_std ....................... None
129
+ embedding_path .................................. None
130
+ empty_unused_memory_level ....................... 0
131
+ enable_cuda_graph ............................... True
132
+ enable_experimental ............................. False
133
+ enable_ft_package ............................... False
134
+ enable_gloo_process_groups ...................... True
135
+ enable_msc ...................................... True
136
+ enable_one_logger ............................... True
137
+ encoder_num_layers .............................. 10
138
+ encoder_seq_length .............................. 4096
139
+ end_weight_decay ................................ 0.1
140
+ eod_mask_loss ................................... False
141
+ error_injection_rate ............................ 0
142
+ error_injection_type ............................ transient_error
143
+ eval_interval ................................... 1000000
144
+ eval_iters ...................................... 0
145
+ evidence_data_path .............................. None
146
+ exit_duration_in_mins ........................... None
147
+ exit_interval ................................... None
148
+ exit_on_missing_checkpoint ...................... False
149
+ exit_signal_handler ............................. False
150
+ exp_avg_dtype ................................... torch.float32
151
+ exp_avg_sq_dtype ................................ torch.float32
152
+ expert_model_parallel_size ...................... 8
153
+ expert_tensor_parallel_size ..................... 1
154
+ export_force_local_attention .................... False
155
+ export_kd_cfg ................................... None
156
+ export_kd_teacher_ckpt_format ................... None
157
+ export_kd_teacher_load .......................... None
158
+ export_kv_cache_quant ........................... False
159
+ export_legacy_megatron .......................... False
160
+ export_model_type ............................... GPTModel
161
+ export_moe_apply_probs_on_input ................. False
162
+ export_qk_l2_norm ............................... False
163
+ export_quant_cfg ................................ None
164
+ export_real_quant_cfg ........................... None
165
+ export_te_mcore_model ........................... False
166
+ external_cuda_graph ............................. False
167
+ ffn_hidden_size ................................. 2560
168
+ finetune ........................................ False
169
+ finetune_data_split ............................. train
170
+ finetune_hf_dataset ............................. None
171
+ first_last_layers_bf16 .......................... False
172
+ flash_decode .................................... False
173
+ fp16 ............................................ False
174
+ fp16_lm_cross_entropy ........................... False
175
+ fp32_residual_connection ........................ False
176
+ fp8 ............................................. e4m3
177
+ fp8_amax_compute_algo ........................... most_recent
178
+ fp8_amax_history_len ............................ 1
179
+ fp8_interval .................................... 1
180
+ fp8_margin ...................................... 0
181
+ fp8_param_gather ................................ False
182
+ fp8_recipe ...................................... mxfp8
183
+ fp8_wgrad ....................................... True
184
+ fsdp_double_buffer .............................. False
185
+ global_batch_size ............................... 8
186
+ grad_reduce_in_bf16 ............................. False
187
+ gradient_accumulation_fusion .................... True
188
+ gradient_reduce_div_fusion ...................... True
189
+ group_query_attention ........................... True
190
+ head_lr_mult .................................... 1.0
191
+ heterogeneous_layers_config_encoded_json ........ None
192
+ heterogeneous_layers_config_path ................ None
193
+ hidden_dropout .................................. 0.0
194
+ hidden_size ..................................... 3584
195
+ hierarchical_context_parallel_sizes ............. None
196
+ high_priority_stream_groups ..................... []
197
+ hybrid_attention_ratio .......................... 0.0
198
+ hybrid_mlp_ratio ................................ 0.0
199
+ hybrid_override_pattern ......................... None
200
+ hysteresis ...................................... 2
201
+ ict_head_size ................................... None
202
+ ict_load ........................................ None
203
+ img_h ........................................... 224
204
+ img_w ........................................... 224
205
+ indexer_batch_size .............................. 128
206
+ indexer_log_interval ............................ 1000
207
+ inference_batch_times_seqlen_threshold .......... -1
208
+ inference_dynamic_batching ...................... False
209
+ inference_dynamic_batching_buffer_guaranteed_fraction 0.2
210
+ inference_dynamic_batching_buffer_overflow_factor None
211
+ inference_dynamic_batching_buffer_size_gb ....... 40.0
212
+ inference_dynamic_batching_chunk_size ........... 256
213
+ inference_dynamic_batching_max_requests_override None
214
+ inference_dynamic_batching_max_tokens_override .. None
215
+ inference_dynamic_batching_num_cuda_graphs ...... 16
216
+ inference_max_batch_size ........................ 8
217
+ inference_max_seq_length ........................ 2560
218
+ inference_rng_tracker ........................... False
219
+ init_method_std ................................. 0.01
220
+ init_method_xavier_uniform ...................... False
221
+ init_model_with_meta_device ..................... False
222
+ initial_loss_scale .............................. 4294967296
223
+ inprocess_active_world_size ..................... 8
224
+ inprocess_barrier_timeout ....................... 120
225
+ inprocess_completion_timeout .................... 120
226
+ inprocess_empty_cuda_cache ...................... False
227
+ inprocess_granularity ........................... node
228
+ inprocess_hard_timeout .......................... 90
229
+ inprocess_heartbeat_interval .................... 30
230
+ inprocess_heartbeat_timeout ..................... 60
231
+ inprocess_last_call_wait ........................ 1
232
+ inprocess_max_iterations ........................ None
233
+ inprocess_monitor_process_interval .............. 1.0
234
+ inprocess_monitor_thread_interval ............... 1.0
235
+ inprocess_progress_watchdog_interval ............ 1.0
236
+ inprocess_restart ............................... False
237
+ inprocess_soft_timeout .......................... 60
238
+ inprocess_termination_grace_time ................ 1
239
+ is_hybrid_model ................................. False
240
+ iter_per_epoch .................................. 1250
241
+ iterations_to_skip .............................. []
242
+ keep_fp8_transpose_cache_when_using_custom_fsdp . False
243
+ kitchen_config_file ............................. None
244
+ kitchen_recipe_number ........................... None
245
+ kv_channels ..................................... 128
246
+ kv_lora_rank .................................... 32
247
+ lazy_mpu_init ................................... None
248
+ load ............................................ None
249
+ load_main_params_from_ckpt ...................... None
250
+ load_model_opt_format ........................... False
251
+ local_rank ...................................... 0
252
+ log_energy ...................................... False
253
+ log_interval .................................... 1
254
+ log_loss_scale_to_tensorboard ................... True
255
+ log_memory_to_tensorboard ....................... False
256
+ log_num_zeros_in_grad ........................... False
257
+ log_params_norm ................................. False
258
+ log_progress .................................... False
259
+ log_straggler ................................... False
260
+ log_throughput .................................. True
261
+ log_timers_to_tensorboard ....................... False
262
+ log_validation_ppl_to_tensorboard ............... False
263
+ log_world_size_to_tensorboard ................... False
264
+ logging_level ................................... None
265
+ loss_scale ...................................... None
266
+ loss_scale_window ............................... 1000
267
+ lr .............................................. 0.0001
268
+ lr_decay_iters .................................. 8
269
+ lr_decay_samples ................................ None
270
+ lr_decay_style .................................. cosine
271
+ lr_warmup_fraction .............................. None
272
+ lr_warmup_init .................................. 0.0
273
+ lr_warmup_iters ................................. 5
274
+ lr_warmup_samples ............................... 0
275
+ lr_wsd_decay_iters .............................. None
276
+ lr_wsd_decay_samples ............................ None
277
+ lr_wsd_decay_style .............................. exponential
278
+ main_grads_dtype ................................ torch.float32
279
+ main_params_dtype ............................... torch.float32
280
+ make_vocab_size_divisible_by .................... 128
281
+ mamba_head_dim .................................. 64
282
+ mamba_num_groups ................................ 8
283
+ mamba_num_heads ................................. None
284
+ mamba_state_dim ................................. 128
285
+ manual_gc ....................................... False
286
+ manual_gc_eval .................................. True
287
+ manual_gc_interval .............................. 0
288
+ mask_factor ..................................... 1.0
289
+ mask_prob ....................................... 0.15
290
+ mask_type ....................................... random
291
+ masked_softmax_fusion ........................... False
292
+ max_position_embeddings ......................... 32768
293
+ max_tokens_to_oom ............................... 12000
294
+ memory_snapshot_path ............................ snapshot.pickle
295
+ merge_file ...................................... /workspace/data/wikitext/gpt2-merges.txt
296
+ micro_batch_size ................................ 1
297
+ microbatch_group_size_per_vp_stage .............. None
298
+ mid_level_dataset_surplus ....................... 0.005
299
+ min_loss_scale .................................. 1.0
300
+ min_lr .......................................... 1e-05
301
+ mlp_chunks_for_prefill .......................... 1
302
+ mmap_bin_files .................................. True
303
+ mock_data ....................................... False
304
+ moe_apply_probs_on_input ........................ False
305
+ moe_aux_loss_coeff .............................. 0.01
306
+ moe_deepep_num_sms .............................. 20
307
+ moe_echo_enable_random_offloading ............... False
308
+ moe_echo_expert_dispatch_overlap ................ True
309
+ moe_echo_expert_dispatcher_type ................. hybridep
310
+ moe_echo_recompute_expert_dispatch .............. False
311
+ moe_enable_deepep ............................... False
312
+ moe_enable_echo ................................. True
313
+ moe_enable_hybridep ............................. True
314
+ moe_expert_capacity_factor ...................... None
315
+ moe_extended_tp ................................. False
316
+ moe_ffn_hidden_size ............................. 2560
317
+ moe_grouped_gemm ................................ True
318
+ moe_hybridep_num_sms ............................ 16
319
+ moe_input_jitter_eps ............................ None
320
+ moe_layer_freq .................................. 1
321
+ moe_layer_recompute ............................. False
322
+ moe_num_echo_experts ............................ 64
323
+ moe_pad_expert_input_to_capacity ................ False
324
+ moe_per_layer_logging ........................... False
325
+ moe_permute_fusion .............................. False
326
+ moe_received_token_capacity ..................... 1.08984375
327
+ moe_router_bias_update_rate ..................... 0.001
328
+ moe_router_dtype ................................ fp32
329
+ moe_router_enable_expert_bias ................... False
330
+ moe_router_force_load_balancing ................. False
331
+ moe_router_group_topk ........................... None
332
+ moe_router_load_balancing_type .................. aux_loss
333
+ moe_router_num_groups ........................... None
334
+ moe_router_padding_for_fp8 ...................... False
335
+ moe_router_pre_softmax .......................... True
336
+ moe_router_score_function ....................... softmax
337
+ moe_router_topk ................................. 8
338
+ moe_router_topk_scaling_factor .................. None
339
+ moe_shared_expert_intermediate_size ............. 20480
340
+ moe_shared_expert_overlap ....................... False
341
+ moe_token_dispatcher_type ....................... flex
342
+ moe_token_drop_policy ........................... probs
343
+ moe_upcycling_granularity ....................... 1
344
+ moe_use_device_initiated_grouped_gemm ........... True
345
+ moe_use_legacy_grouped_gemm ..................... False
346
+ moe_use_upcycling ............................... False
347
+ moe_z_loss_coeff ................................ None
348
+ mrope_section ................................... None
349
+ mscale .......................................... 1.0
350
+ mscale_all_dim .................................. 0.0
351
+ mtp_loss_scaling_factor ......................... 0.1
352
+ mtp_num_layers .................................. None
353
+ multi_latent_attention .......................... False
354
+ nccl_all_reduce_for_prefill ..................... False
355
+ nccl_communicator_config_path ................... None
356
+ nccl_ub ......................................... False
357
+ no_load_optim ................................... None
358
+ no_load_rng ..................................... None
359
+ no_persist_layer_norm ........................... False
360
+ no_rope_freq .................................... None
361
+ no_save_optim ................................... None
362
+ no_save_rng ..................................... None
363
+ non_persistent_ckpt_type ........................ None
364
+ non_persistent_global_ckpt_dir .................. None
365
+ non_persistent_local_ckpt_algo .................. fully_parallel
366
+ non_persistent_local_ckpt_dir ................... None
367
+ non_persistent_save_interval .................... None
368
+ norm_epsilon .................................... 1e-06
369
+ normalization ................................... RMSNorm
370
+ num_attention_heads ............................. 28
371
+ num_channels .................................... 3
372
+ num_classes ..................................... 1000
373
+ num_dataset_builder_threads ..................... 1
374
+ num_distributed_optimizer_instances ............. 1
375
+ num_experts ..................................... 64
376
+ num_layers ...................................... 10
377
+ num_layers_at_end_in_bf16 ....................... 1
378
+ num_layers_at_start_in_bf16 ..................... 1
379
+ num_layers_per_virtual_pipeline_stage ........... None
380
+ num_query_groups ................................ 4
381
+ num_virtual_stages_per_pipeline_rank ............ None
382
+ num_workers ..................................... 2
383
+ object_storage_cache_path ....................... None
384
+ one_logger_async ................................ False
385
+ one_logger_project .............................. megatron-lm
386
+ one_logger_run_name ............................. None
387
+ onnx_safe ....................................... None
388
+ openai_gelu ..................................... False
389
+ optimizer ....................................... adam
390
+ optimizer_cpu_offload ........................... False
391
+ optimizer_offload_fraction ...................... 1.0
392
+ output_bert_embeddings .......................... False
393
+ overlap_cpu_optimizer_d2h_h2d ................... False
394
+ overlap_grad_reduce ............................. True
395
+ overlap_moe_expert_parallel_comm ................ False
396
+ overlap_p2p_comm ................................ False
397
+ overlap_p2p_comm_warmup_flush ................... False
398
+ overlap_param_gather ............................ True
399
+ overlap_param_gather_with_optimizer_step ........ False
400
+ override_opt_param_scheduler .................... False
401
+ params_dtype .................................... torch.bfloat16
402
+ patch_dim ....................................... 16
403
+ per_split_data_args_path ........................ None
404
+ perform_initialization .......................... True
405
+ pin_cpu_grads ................................... True
406
+ pin_cpu_params .................................. True
407
+ pipeline_model_parallel_comm_backend ............ None
408
+ pipeline_model_parallel_layout .................. None
409
+ pipeline_model_parallel_size .................... 1
410
+ position_embedding_type ......................... rope
411
+ pretrained_checkpoint ........................... None
412
+ profile ......................................... True
413
+ profile_ranks ................................... [0, 1, 2, 3, 4, 5, 6, 7]
414
+ profile_step_end ................................ 8
415
+ profile_step_start .............................. 6
416
+ q_lora_rank ..................................... None
417
+ qk_head_dim ..................................... 128
418
+ qk_l2_norm ...................................... False
419
+ qk_layernorm .................................... False
420
+ qk_pos_emb_head_dim ............................. 64
421
+ query_in_block_prob ............................. 0.1
422
+ rampup_batch_size ............................... None
423
+ rank ............................................ 0
424
+ recompute_granularity ........................... None
425
+ recompute_method ................................ None
426
+ recompute_modules ............................... None
427
+ recompute_num_layers ............................ None
428
+ record_memory_history ........................... False
429
+ relative_attention_max_distance ................. 128
430
+ relative_attention_num_buckets .................. 32
431
+ replication ..................................... False
432
+ replication_factor .............................. 2
433
+ replication_jump ................................ None
434
+ rerun_mode ...................................... validate_results
435
+ reset_attention_mask ............................ False
436
+ reset_position_ids .............................. False
437
+ result_rejected_tracker_filename ................ None
438
+ retriever_report_topk_accuracies ................ []
439
+ retriever_score_scaling ......................... False
440
+ retriever_seq_length ............................ 256
441
+ retro_add_retriever ............................. False
442
+ retro_attention_gate ............................ 1
443
+ retro_cyclic_train_iters ........................ None
444
+ retro_encoder_attention_dropout ................. 0.1
445
+ retro_encoder_hidden_dropout .................... 0.1
446
+ retro_encoder_layers ............................ 2
447
+ retro_num_neighbors ............................. 2
448
+ retro_num_retrieved_chunks ...................... 2
449
+ retro_project_dir ............................... None
450
+ retro_verify_neighbor_count ..................... True
451
+ reuse_grad_buf_for_mxfp8_param_ag ............... False
452
+ rope_scaling_factor ............................. 8.0
453
+ rope_type ....................................... None
454
+ rotary_base ..................................... 1000000
455
+ rotary_interleaved .............................. False
456
+ rotary_percent .................................. 1.0
457
+ rotary_scaling_factor ........................... 1.0
458
+ rotary_seq_len_interpolation_factor ............. None
459
+ run_workload_inspector_server ................... False
460
+ sample_rate ..................................... 1.0
461
+ save ............................................ None
462
+ save_interval ................................... None
463
+ save_retain_interval ............................ None
464
+ scatter_gather_tensors_in_pipeline .............. True
465
+ seed ............................................ 1234
466
+ seq_length ...................................... 4096
467
+ sequence_parallel ............................... False
468
+ sft ............................................. False
469
+ sft_tokenizer_prompt_format ..................... nemotron-h-aligned
470
+ sgd_momentum .................................... 0.9
471
+ short_seq_prob .................................. 0.1
472
+ skip_train ...................................... False
473
+ skipped_train_samples ........................... 0
474
+ spec ............................................ None
475
+ split ........................................... 100,0,0
476
+ squared_relu .................................... False
477
+ start_weight_decay .............................. 0.1
478
+ straggler_ctrlr_port ............................ 65535
479
+ straggler_minmax_count .......................... 1
480
+ suggested_communication_unit_size ............... None
481
+ swiglu .......................................... True
482
+ swin_backbone_type .............................. tiny
483
+ symmetric_ar_type ............................... None
484
+ te_rng_tracker .................................. True
485
+ tensor_model_parallel_size ...................... 1
486
+ tensorboard_dir ................................. None
487
+ tensorboard_log_interval ........................ 1
488
+ tensorboard_queue_size .......................... 1000
489
+ test_data_path .................................. None
490
+ test_mode ....................................... False
491
+ tiktoken_num_special_tokens ..................... 1000
492
+ tiktoken_pattern ................................ None
493
+ tiktoken_special_tokens ......................... None
494
+ timing_log_level ................................ 0
495
+ timing_log_option ............................... minmax
496
+ titles_data_path ................................ None
497
+ tokenizer_model ................................. None
498
+ tokenizer_type .................................. GPT2BPETokenizer
499
+ torch_fsdp2_reshard_after_forward ............... True
500
+ tp_comm_bootstrap_backend ....................... nccl
501
+ tp_comm_bulk_dgrad .............................. True
502
+ tp_comm_bulk_wgrad .............................. True
503
+ tp_comm_overlap ................................. False
504
+ tp_comm_overlap_ag .............................. True
505
+ tp_comm_overlap_cfg ............................. None
506
+ tp_comm_overlap_rs .............................. True
507
+ tp_comm_overlap_rs_dgrad ........................ False
508
+ tp_comm_split_ag ................................ True
509
+ tp_comm_split_rs ................................ True
510
+ train_data_path ................................. None
511
+ train_iters ..................................... 10
512
+ train_samples ................................... None
513
+ train_sync_interval ............................. None
514
+ transformer_impl ................................ transformer_engine
515
+ transformer_pipeline_model_parallel_size ........ 1
516
+ untie_embeddings_and_output_weights ............. True
517
+ use_checkpoint_args ............................. False
518
+ use_checkpoint_opt_param_scheduler .............. False
519
+ use_cpu_initialization .......................... None
520
+ use_custom_fsdp ................................. False
521
+ use_dist_ckpt ................................... True
522
+ use_dist_ckpt_deprecated ........................ False
523
+ use_distributed_optimizer ....................... True
524
+ use_flash_attn .................................. False
525
+ use_legacy_models ............................... False
526
+ use_mp_args_from_checkpoint_args ................ False
527
+ use_one_sent_docs ............................... False
528
+ use_persistent_ckpt_worker ...................... False
529
+ use_precision_aware_optimizer ................... False
530
+ use_pytorch_profiler ............................ False
531
+ use_ring_exchange_p2p ........................... False
532
+ use_rope_scaling ................................ False
533
+ use_rotary_position_embeddings .................. False
534
+ use_sharp ....................................... False
535
+ use_tokenizer_model_from_checkpoint_args ........ True
536
+ use_torch_fsdp2 ................................. False
537
+ use_torch_optimizer_for_cpu_offload ............. False
538
+ use_tp_pp_dp_mapping ............................ False
539
+ v_head_dim ...................................... 128
540
+ valid_data_path ................................. None
541
+ variable_seq_lengths ............................ False
542
+ virtual_pipeline_model_parallel_size ............ None
543
+ vision_backbone_type ............................ vit
544
+ vision_pretraining .............................. False
545
+ vision_pretraining_type ......................... classify
546
+ vocab_extra_ids ................................. 0
547
+ vocab_file ...................................... /workspace/data/wikitext/gpt2-vocab.json
548
+ vocab_size ...................................... None
549
+ wandb_exp_name ..................................
550
+ wandb_project ...................................
551
+ wandb_save_dir ..................................
552
+ weight_decay .................................... 0.1
553
+ weight_decay_incr_style ......................... constant
554
+ wgrad_deferral_limit ............................ 0
555
+ world_size ...................................... 8
556
+ yaml_cfg ........................................ None
557
+ -------------------- end of arguments ---------------------
558
+ INFO:megatron.core.num_microbatches_calculator:setting number of microbatches to constant 1
559
+ > building GPT2BPETokenizer tokenizer ...
560
+ > padded vocab (size: 50257) with 47 dummy tokens (new size: 50304)
561
+ WARNING:megatron.core.rerun_state_machine:RerunStateMachine initialized in mode RerunMode.VALIDATE_RESULTS
562
+ > initializing torch distributed ...
563
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
564
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
565
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
566
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
567
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
568
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
569
+ WARNING: one_logger package is required to enable e2e metrics tracking. please go to https://confluence.nvidia.com/display/MLWFO/Package+Repositories for details to install it
570
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
571
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
572
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
573
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
574
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
575
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
576
+ [Gloo] Rank 1 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
577
+ [Gloo] Rank 0 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
578
+ [Gloo] Rank 2 is connected to [Gloo] Rank 7 peer ranks. Expected number of connected peer ranks is : 73 is connected to
579
+ 7 peer ranks. Expected number of connected peer ranks is : 7
580
+ [Gloo] Rank 4 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
581
+ [Gloo] Rank 5 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
582
+ [Gloo] Rank [Gloo] Rank 7 is connected to 67 is connected to peer ranks. 7Expected number of connected peer ranks is : peer ranks. Expected number of connected peer ranks is : 77
583
+
584
+ [Gloo] Rank [Gloo] Rank 0 is connected to 73 peer ranks. is connected to Expected number of connected peer ranks is : 77 peer ranks. Expected number of connected peer ranks is :
585
+ 7
586
+ [Gloo] Rank 1 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7[Gloo] Rank
587
+ 2 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
588
+ [Gloo] Rank 4 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
589
+ [Gloo] Rank [Gloo] Rank 7 is connected to 7 peer ranks. 6Expected number of connected peer ranks is : is connected to 77 peer ranks.
590
+ Expected number of connected peer ranks is : 7
591
+ [Gloo] Rank 5 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
592
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
593
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
594
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
595
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
596
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
597
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
598
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
599
+ > initialized tensor model parallel with size 1
600
+ > initialized pipeline model parallel with size 1
601
+ > setting random seeds to 1234 ...
602
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
603
+ [rank1]:[W826 19:56:14.537263346 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 1] using GPU 1 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device.
604
+ [rank4]:[W826 19:56:14.543374402 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 4] using GPU 4 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device.
605
+ [rank7]:[W826 19:56:14.543463556 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 7] using GPU 7 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device.
606
+ [rank5]:[W826 19:56:14.543625565 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 5] using GPU 5 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device.
607
+ [rank2]:[W826 19:56:14.543657432 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 2] using GPU 2 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device.
608
+ [rank3]:[W826 19:56:14.543820607 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 3] using GPU 3 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device.
609
+ [rank6]:[W826 19:56:14.543905518 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 6] using GPU 6 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device.
610
+ > compiling dataset index builder ...
611
+ make: Entering directory '/workspace/Megatron-MoE-Echo/megatron/core/datasets'
612
+ make: Nothing to be done for 'default'.
613
+ make: Leaving directory '/workspace/Megatron-MoE-Echo/megatron/core/datasets'
614
+ >>> done with dataset index builder. Compilation time: 0.424 seconds
615
+ WARNING: constraints for invoking optimized fused softmax kernel are not met. We default back to unfused kernel invocations.
616
+ > compiling and loading fused kernels ...
617
+ [rank0]:[W826 19:56:15.149085342 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 0] using GPU 0 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device.
618
+ >>> done with compiling and loading fused kernels. Compilation time: 3.782 seconds
619
+ time to initialize megatron (seconds): 9.386
620
+ [after megatron is initialized] datetime: 2026-08-26 19:56:21
621
+ building GPT model ...
622
+ /usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
623
+ warnings.warn(
624
+ /usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
625
+ warnings.warn(
626
+ /usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
627
+ warnings.warn(
628
+ /usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
629
+ warnings.warn(
630
+ /usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
631
+ warnings.warn(
632
+ /usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
633
+ warnings.warn(
634
+ /usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
635
+ warnings.warn(
636
+ /usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect.
637
+ warnings.warn(
638
+ > number of parameters on (tensor, pipeline) model parallel rank (0, 0): 5060614656
639
+ INFO:megatron.core.distributed.distributed_data_parallel:Setting up DistributedDataParallel with config DistributedDataParallelConfig(grad_reduce_in_fp32=True, overlap_grad_reduce=True, overlap_param_gather=True, align_param_gather=False, use_distributed_optimizer=True, num_distributed_optimizer_instances=1, check_for_nan_in_grad=False, check_for_large_grads=False, bucket_size=40000000, pad_buckets_for_high_nccl_busbw=False, average_in_collective=False, fp8_param_gather=False, reuse_grad_buf_for_mxfp8_param_ag=False, use_custom_fsdp=False, data_parallel_sharding_strategy='no_shard', gradient_reduce_div_fusion=True, suggested_communication_unit_size=None, preserve_fp32_weights=True, keep_fp8_transpose_cache_when_using_custom_fsdp=False, nccl_ub=False, fsdp_double_buffer=False)
640
+ INFO:megatron.core.distributed.param_and_grad_buffer:Number of buckets for gradient all-reduce / reduce-scatter: 22
641
+ Params for bucket 1 (180289536 elements, 180289536 padded size):
642
+ module.output_layer.weight
643
+ Params for bucket 2 (73403904 elements, 73403904 padded size):
644
+ module.decoder.final_layernorm.weight
645
+ module.decoder.layers.9.mlp.shared_experts.linear_fc2.weight
646
+ Params for bucket 3 (146800640 elements, 146800640 padded size):
647
+ module.decoder.layers.9.mlp.shared_experts.linear_fc1.weight
648
+ Params for bucket 4 (103001600 elements, 103001600 padded size):
649
+ module.decoder.layers.9.mlp.router.weight
650
+ module.decoder.layers.9.self_attention.linear_qkv.weight
651
+ module.decoder.layers.8.mlp.shared_experts.linear_fc2.weight
652
+ module.decoder.layers.9.self_attention.linear_qkv.layer_norm_weight
653
+ module.decoder.layers.9.pre_mlp_layernorm.weight
654
+ module.decoder.layers.9.self_attention.linear_qkv.bias
655
+ module.decoder.layers.9.self_attention.linear_proj.weight
656
+ Params for bucket 5 (146800640 elements, 146800640 padded size):
657
+ module.decoder.layers.8.mlp.shared_experts.linear_fc1.weight
658
+ Params for bucket 6 (103001600 elements, 103001600 padded size):
659
+ module.decoder.layers.8.mlp.router.weight
660
+ module.decoder.layers.8.self_attention.linear_qkv.layer_norm_weight
661
+ module.decoder.layers.8.pre_mlp_layernorm.weight
662
+ module.decoder.layers.8.self_attention.linear_qkv.bias
663
+ module.decoder.layers.8.self_attention.linear_proj.weight
664
+ module.decoder.layers.7.mlp.shared_experts.linear_fc2.weight
665
+ module.decoder.layers.8.self_attention.linear_qkv.weight
666
+ Params for bucket 7 (146800640 elements, 146800640 padded size):
667
+ module.decoder.layers.7.mlp.shared_experts.linear_fc1.weight
668
+ Params for bucket 8 (103001600 elements, 103001600 padded size):
669
+ module.decoder.layers.7.self_attention.linear_qkv.bias
670
+ module.decoder.layers.7.pre_mlp_layernorm.weight
671
+ module.decoder.layers.6.mlp.shared_experts.linear_fc2.weight
672
+ module.decoder.layers.7.mlp.router.weight
673
+ module.decoder.layers.7.self_attention.linear_qkv.weight
674
+ module.decoder.layers.7.self_attention.linear_qkv.layer_norm_weight
675
+ module.decoder.layers.7.self_attention.linear_proj.weight
676
+ Params for bucket 9 (146800640 elements, 146800640 padded size):
677
+ module.decoder.layers.6.mlp.shared_experts.linear_fc1.weight
678
+ Params for bucket 10 (103001600 elements, 103001600 padded size):
679
+ module.decoder.layers.6.mlp.router.weight
680
+ module.decoder.layers.6.self_attention.linear_qkv.weight
681
+ module.decoder.layers.5.mlp.shared_experts.linear_fc2.weight
682
+ module.decoder.layers.6.self_attention.linear_qkv.layer_norm_weight
683
+ module.decoder.layers.6.pre_mlp_layernorm.weight
684
+ module.decoder.layers.6.self_attention.linear_qkv.bias
685
+ module.decoder.layers.6.self_attention.linear_proj.weight
686
+ Params for bucket 11 (146800640 elements, 146800640 padded size):
687
+ module.decoder.layers.5.mlp.shared_experts.linear_fc1.weight
688
+ Params for bucket 12 (103001600 elements, 103001600 padded size):
689
+ module.decoder.layers.5.mlp.router.weight
690
+ module.decoder.layers.5.self_attention.linear_qkv.weight
691
+ module.decoder.layers.5.self_attention.linear_qkv.layer_norm_weight
692
+ module.decoder.layers.5.self_attention.linear_proj.weight
693
+ module.decoder.layers.5.pre_mlp_layernorm.weight
694
+ module.decoder.layers.5.self_attention.linear_qkv.bias
695
+ module.decoder.layers.4.mlp.shared_experts.linear_fc2.weight
696
+ Params for bucket 13 (146800640 elements, 146800640 padded size):
697
+ module.decoder.layers.4.mlp.shared_experts.linear_fc1.weight
698
+ Params for bucket 14 (103001600 elements, 103001600 padded size):
699
+ module.decoder.layers.4.pre_mlp_layernorm.weight
700
+ module.decoder.layers.4.self_attention.linear_qkv.bias
701
+ module.decoder.layers.4.mlp.router.weight
702
+ module.decoder.layers.4.self_attention.linear_qkv.layer_norm_weight
703
+ module.decoder.layers.4.self_attention.linear_qkv.weight
704
+ module.decoder.layers.4.self_attention.linear_proj.weight
705
+ module.decoder.layers.3.mlp.shared_experts.linear_fc2.weight
706
+ Params for bucket 15 (146800640 elements, 146800640 padded size):
707
+ module.decoder.layers.3.mlp.shared_experts.linear_fc1.weight
708
+ Params for bucket 16 (103001600 elements, 103001600 padded size):
709
+ module.decoder.layers.3.mlp.router.weight
710
+ module.decoder.layers.3.self_attention.linear_qkv.layer_norm_weight
711
+ module.decoder.layers.3.self_attention.linear_qkv.weight
712
+ module.decoder.layers.3.self_attention.linear_proj.weight
713
+ module.decoder.layers.2.mlp.shared_experts.linear_fc2.weight
714
+ module.decoder.layers.3.pre_mlp_layernorm.weight
715
+ module.decoder.layers.3.self_attention.linear_qkv.bias
716
+ Params for bucket 17 (146800640 elements, 146800640 padded size):
717
+ module.decoder.layers.2.mlp.shared_experts.linear_fc1.weight
718
+ Params for bucket 18 (103001600 elements, 103001600 padded size):
719
+ module.decoder.layers.2.mlp.router.weight
720
+ module.decoder.layers.2.self_attention.linear_qkv.bias
721
+ module.decoder.layers.2.self_attention.linear_qkv.layer_norm_weight
722
+ module.decoder.layers.2.pre_mlp_layernorm.weight
723
+ module.decoder.layers.2.self_attention.linear_proj.weight
724
+ module.decoder.layers.1.mlp.shared_experts.linear_fc2.weight
725
+ module.decoder.layers.2.self_attention.linear_qkv.weight
726
+ Params for bucket 19 (146800640 elements, 146800640 padded size):
727
+ module.decoder.layers.1.mlp.shared_experts.linear_fc1.weight
728
+ Params for bucket 20 (103001600 elements, 103001600 padded size):
729
+ module.decoder.layers.1.self_attention.linear_qkv.layer_norm_weight
730
+ module.decoder.layers.1.pre_mlp_layernorm.weight
731
+ module.decoder.layers.1.self_attention.linear_qkv.weight
732
+ module.decoder.layers.1.self_attention.linear_proj.weight
733
+ module.decoder.layers.0.mlp.shared_experts.linear_fc2.weight
734
+ module.decoder.layers.1.mlp.router.weight
735
+ module.decoder.layers.1.self_attention.linear_qkv.bias
736
+ Params for bucket 21 (146800640 elements, 146800640 padded size):
737
+ module.decoder.layers.0.mlp.shared_experts.linear_fc1.weight
738
+ Params for bucket 22 (209890816 elements, 209890816 padded size):
739
+ module.decoder.layers.0.pre_mlp_layernorm.weight
740
+ module.embedding.word_embeddings.weight
741
+ module.decoder.layers.0.mlp.router.weight
742
+ module.decoder.layers.0.self_attention.linear_qkv.layer_norm_weight
743
+ module.decoder.layers.0.self_attention.linear_qkv.weight
744
+ module.decoder.layers.0.self_attention.linear_qkv.bias
745
+ module.decoder.layers.0.self_attention.linear_proj.weight
746
+ INFO:megatron.core.distributed.param_and_grad_buffer:Number of buckets for gradient all-reduce / reduce-scatter: 44
747
+ Params for bucket 1 (45875200 elements, 45875200 padded size):
748
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight6
749
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight4
750
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight7
751
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight3
752
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight5
753
+ Params for bucket 2 (45875200 elements, 45875200 padded size):
754
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight2
755
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight0
756
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight1
757
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight7
758
+ Params for bucket 3 (55050240 elements, 55050240 padded size):
759
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight6
760
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight4
761
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight5
762
+ Params for bucket 4 (55050240 elements, 55050240 padded size):
763
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight3
764
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight1
765
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight2
766
+ Params for bucket 5 (45875200 elements, 45875200 padded size):
767
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight0
768
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight7
769
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight6
770
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight5
771
+ Params for bucket 6 (45875200 elements, 45875200 padded size):
772
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight3
773
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight1
774
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight4
775
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight0
776
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight2
777
+ Params for bucket 7 (55050240 elements, 55050240 padded size):
778
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight7
779
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight5
780
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight6
781
+ Params for bucket 8 (55050240 elements, 55050240 padded size):
782
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight4
783
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight2
784
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight3
785
+ Params for bucket 9 (45875200 elements, 45875200 padded size):
786
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight1
787
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight7
788
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight0
789
+ Params for bucket 10 (45875200 elements, 45875200 padded size):
790
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight6
791
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight4
792
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight2
793
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight3
794
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight5
795
+ Params for bucket 11 (55050240 elements, 55050240 padded size):
796
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight1
797
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight0
798
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight6
799
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight7
800
+ Params for bucket 12 (55050240 elements, 55050240 padded size):
801
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight5
802
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight3
803
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight4
804
+ Params for bucket 13 (55050240 elements, 55050240 padded size):
805
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight2
806
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight0
807
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight1
808
+ Params for bucket 14 (45875200 elements, 45875200 padded size):
809
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight7
810
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight5
811
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight3
812
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight4
813
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight6
814
+ Params for bucket 15 (45875200 elements, 45875200 padded size):
815
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight2
816
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight1
817
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight7
818
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight0
819
+ Params for bucket 16 (55050240 elements, 55050240 padded size):
820
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight6
821
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight4
822
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight5
823
+ Params for bucket 17 (55050240 elements, 55050240 padded size):
824
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight3
825
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight1
826
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight2
827
+ Params for bucket 18 (45875200 elements, 45875200 padded size):
828
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight0
829
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight6
830
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight7
831
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight5
832
+ Params for bucket 19 (45875200 elements, 45875200 padded size):
833
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight3
834
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight1
835
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight4
836
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight0
837
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight2
838
+ Params for bucket 20 (55050240 elements, 55050240 padded size):
839
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight7
840
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight5
841
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight6
842
+ Params for bucket 21 (55050240 elements, 55050240 padded size):
843
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight4
844
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight2
845
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight3
846
+ Params for bucket 22 (45875200 elements, 45875200 padded size):
847
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight1
848
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight0
849
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight7
850
+ Params for bucket 23 (45875200 elements, 45875200 padded size):
851
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight5
852
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight3
853
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight6
854
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight2
855
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight4
856
+ Params for bucket 24 (55050240 elements, 55050240 padded size):
857
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight1
858
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight6
859
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight7
860
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight0
861
+ Params for bucket 25 (55050240 elements, 55050240 padded size):
862
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight5
863
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight4
864
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight3
865
+ Params for bucket 26 (55050240 elements, 55050240 padded size):
866
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight2
867
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight1
868
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight0
869
+ Params for bucket 27 (45875200 elements, 45875200 padded size):
870
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight7
871
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight5
872
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight3
873
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight6
874
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight4
875
+ Params for bucket 28 (45875200 elements, 45875200 padded size):
876
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight2
877
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight1
878
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight7
879
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight0
880
+ Params for bucket 29 (55050240 elements, 55050240 padded size):
881
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight6
882
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight5
883
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight4
884
+ Params for bucket 30 (55050240 elements, 55050240 padded size):
885
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight3
886
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight2
887
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight1
888
+ Params for bucket 31 (45875200 elements, 45875200 padded size):
889
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight0
890
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight5
891
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight7
892
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight6
893
+ Params for bucket 32 (45875200 elements, 45875200 padded size):
894
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight4
895
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight2
896
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight0
897
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight1
898
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight3
899
+ Params for bucket 33 (55050240 elements, 55050240 padded size):
900
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight7
901
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight5
902
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight6
903
+ Params for bucket 34 (55050240 elements, 55050240 padded size):
904
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight4
905
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight2
906
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight3
907
+ Params for bucket 35 (45875200 elements, 45875200 padded size):
908
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight1
909
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight7
910
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight0
911
+ Params for bucket 36 (45875200 elements, 45875200 padded size):
912
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight5
913
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight3
914
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight6
915
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight4
916
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight2
917
+ Params for bucket 37 (55050240 elements, 55050240 padded size):
918
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight1
919
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight7
920
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight6
921
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight0
922
+ Params for bucket 38 (55050240 elements, 55050240 padded size):
923
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight5
924
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight4
925
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight3
926
+ Params for bucket 39 (55050240 elements, 55050240 padded size):
927
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight2
928
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight1
929
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight0
930
+ Params for bucket 40 (45875200 elements, 45875200 padded size):
931
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight4
932
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight6
933
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight7
934
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight3
935
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight5
936
+ Params for bucket 41 (45875200 elements, 45875200 padded size):
937
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight2
938
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight0
939
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight1
940
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight7
941
+ Params for bucket 42 (55050240 elements, 55050240 padded size):
942
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight6
943
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight4
944
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight5
945
+ Params for bucket 43 (55050240 elements, 55050240 padded size):
946
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight3
947
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight1
948
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight2
949
+ Params for bucket 44 (18350080 elements, 18350080 padded size):
950
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight0
951
+ INFO:megatron.core.optimizer:Setting up optimizer with config OptimizerConfig(optimizer='adam', lr=0.0001, min_lr=1e-05, decoupled_lr=None, decoupled_min_lr=None, weight_decay=0.1, fp8_recipe='mxfp8', fp16=False, bf16=True, reuse_grad_buf_for_mxfp8_param_ag=False, params_dtype=torch.bfloat16, use_precision_aware_optimizer=False, store_param_remainders=True, main_grads_dtype=torch.float32, main_params_dtype=torch.float32, exp_avg_dtype=torch.float32, exp_avg_sq_dtype=torch.float32, loss_scale=None, initial_loss_scale=4294967296, min_loss_scale=1.0, loss_scale_window=1000, hysteresis=2, adam_beta1=0.9, adam_beta2=0.999, adam_eps=1e-08, sgd_momentum=0.9, use_distributed_optimizer=True, overlap_param_gather_with_optimizer_step=False, optimizer_cpu_offload=False, optimizer_offload_fraction=1.0, use_torch_optimizer_for_cpu_offload=False, overlap_cpu_optimizer_d2h_h2d=False, pin_cpu_grads=True, pin_cpu_params=True, clip_grad=1.0, log_num_zeros_in_grad=False, barrier_with_L1_time=True, timers=<megatron.core.timers.Timers object at 0x714361ca4620>, config_logger_dir='')
952
+ INFO:megatron.core.optimizer_param_scheduler:> learning rate decay style: cosine
953
+ [after model, optimizer, and learning rate scheduler are built] datetime: 2026-08-26 19:56:21
954
+ > building train, validation, and test datasets ...
955
+ > datasets target sizes (minimum size):
956
+ train: 80
957
+ validation: 0
958
+ test: 0
959
+ INFO:megatron.core.datasets.blended_megatron_dataset_config:Let split_matrix = [(0, 1.0), None, None]
960
+ > building train, validation, and test datasets for GPT ...
961
+ INFO:megatron.core.datasets.blended_megatron_dataset_builder:Building GPTDataset splits with sizes=(80, 0, 0) and config=GPTDatasetConfig(random_seed=1234, sequence_length=4096, blend=(['/workspace/data/wikitext/wikitext103_gpt2_text_document'], None), blend_per_split=None, split='100,0,0', split_matrix=[(0, 1.0), None, None], num_dataset_builder_threads=1, path_to_cache=None, mmap_bin_files=True, mock=False, tokenizer=<megatron.training.tokenizer.tokenizer._GPT2BPETokenizer object at 0x714362c47200>, mid_level_dataset_surplus=0.005, reset_position_ids=False, reset_attention_mask=False, eod_mask_loss=False, create_attention_mask=True, drop_last_partial_validation_sequence=True, add_extra_token_to_sequence=True, object_storage_cache_path=None)
962
+ INFO:megatron.core.datasets.indexed_dataset:Load the _IndexReader from /workspace/data/wikitext/wikitext103_gpt2_text_document.idx
963
+ INFO:megatron.core.datasets.indexed_dataset: Extract the sequence lengths
964
+ INFO:megatron.core.datasets.indexed_dataset: Extract the sequence pointers
965
+ INFO:megatron.core.datasets.indexed_dataset: Extract the document indices
966
+ INFO:megatron.core.datasets.indexed_dataset:> total number of sequences: 1165029
967
+ INFO:megatron.core.datasets.indexed_dataset:> total number of documents: 1801350
968
+ INFO:megatron.core.datasets.gpt_dataset:Load the GPTDataset train indices
969
+ INFO:megatron.core.datasets.gpt_dataset: Load the document index from 71e99bc7c46bc3412b339a272e0f9cf3-GPTDataset-train-document_index.npy
970
+ INFO:megatron.core.datasets.gpt_dataset: Load the sample index from 71e99bc7c46bc3412b339a272e0f9cf3-GPTDataset-train-sample_index.npy
971
+ INFO:megatron.core.datasets.gpt_dataset: Load the shuffle index from 71e99bc7c46bc3412b339a272e0f9cf3-GPTDataset-train-shuffle_index.npy
972
+ INFO:megatron.core.datasets.gpt_dataset:> total number of samples: 29073
973
+ > finished creating GPT datasets ...
974
+ [after dataloaders are built] datetime: 2026-08-26 19:56:21
975
+ done with setup ...
976
+ training ...
977
+ Overwriting rerun_state_machine.current_iteration from -1 to 0...
978
+ (min, max) time across ranks (ms):
979
+ model-and-optimizer-setup ......................: (488.34, 498.57)
980
+ train/valid/test-data-iterators-setup ..........: (87.96, 110.72)
981
+ [before the start of training step] datetime: 2026-08-26 19:56:21
982
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
983
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
984
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
985
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
986
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
987
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
988
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
989
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
990
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
991
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
992
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
993
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
994
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
995
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
996
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
997
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
998
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
999
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1000
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1001
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1002
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1003
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1004
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1005
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1006
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1007
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1008
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1009
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1010
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1011
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1012
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1013
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1014
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1015
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1016
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1017
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1018
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1019
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1020
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1021
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1022
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1023
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1024
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1025
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1026
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1027
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1028
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1029
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1030
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1031
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1032
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1033
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1034
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1035
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1036
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1037
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1038
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1039
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1040
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1041
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1042
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1043
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1044
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1045
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1046
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1047
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1048
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1049
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1050
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1051
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1052
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1053
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1054
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1055
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1056
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1057
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1058
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1059
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1060
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1061
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1062
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1063
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1064
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1065
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1066
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1067
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1068
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1069
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1070
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1071
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1072
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1073
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1074
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1075
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1076
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1077
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1078
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1079
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1080
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1081
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1082
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1083
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1084
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1085
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1086
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1087
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1088
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1089
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1090
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1091
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1092
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1093
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1094
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1095
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1096
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1097
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1098
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1099
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1100
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1101
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1102
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1103
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1104
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1105
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1106
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1107
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1108
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1109
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1110
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1111
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1112
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1113
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1114
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1115
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1116
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1117
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1118
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1119
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1120
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1121
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1122
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1123
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1124
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1125
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1126
+ Number of parameters in transformer block in billions: 20.11
1127
+ Number of parameters in embedding layers in billions: 0.36
1128
+ [2026-08-26 19:57:04] iteration 1/ 10 | consumed samples: 8 | elapsed time per iteration (ms): 42916.0 | throughput per GPU (TFLOP/s/GPU): 2.9 | learning rate: 2.000000E-05 | global batch size: 8 | lm loss: 1.097931E+01 | loss scale: 1.0 | grad norm: 29.871 | number of skipped iterations: 0 | number of nan iterations: 0 |Total number of parameters in billions: 20.47
1129
+
1130
+ Number of parameters in most loaded shard in billions: 20.4724
1131
+ Theoretical memory footprints: weight and optimizer=146430.14 MB
1132
+ [Rank 0] (after 1 iterations) memory (MB) | allocated: 68733.94384765625 | max allocated: 70397.458984375 | reserved: 80296.0 | max reserved: 80296.0
1133
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1134
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1135
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1136
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1137
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1138
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1139
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1140
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1141
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1142
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1143
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1144
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1145
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1146
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1147
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1148
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1149
+ [2026-08-26 19:57:05] iteration 2/ 10 | consumed samples: 16 | elapsed time per iteration (ms): 552.6 | throughput per GPU (TFLOP/s/GPU): 223.5 | learning rate: 4.000000E-05 | global batch size: 8 | lm loss: 1.097906E+01 | loss scale: 1.0 | grad norm: 30.622 | number of skipped iterations: 0 | number of nan iterations: 0 |
1150
+ [2026-08-26 19:57:05] iteration 3/ 10 | consumed samples: 24 | elapsed time per iteration (ms): 234.3 | throughput per GPU (TFLOP/s/GPU): 527.0 | learning rate: 6.000000E-05 | global batch size: 8 | lm loss: 9.538235E+00 | loss scale: 1.0 | grad norm: 32.371 | number of skipped iterations: 0 | number of nan iterations: 0 |
1151
+ Capture CUDA graph for training!!!
1152
+ CUDA graph capture done!!!
1153
+ [2026-08-26 19:57:07] iteration 4/ 10 | consumed samples: 32 | elapsed time per iteration (ms): 1852.0 | throughput per GPU (TFLOP/s/GPU): 66.7 | learning rate: 8.000000E-05 | global batch size: 8 | lm loss: 1.109051E+01 | loss scale: 1.0 | grad norm: 107.094 | number of skipped iterations: 0 | number of nan iterations: 0 |
1154
+ [2026-08-26 19:57:07] iteration 5/ 10 | consumed samples: 40 | elapsed time per iteration (ms): 181.5 | throughput per GPU (TFLOP/s/GPU): 680.3 | learning rate: 1.000000E-04 | global batch size: 8 | lm loss: 9.997239E+00 | loss scale: 1.0 | grad norm: 26.207 | number of skipped iterations: 0 | number of nan iterations: 0 |
1155
+ [2026-08-26 19:57:07] iteration 6/ 10 | consumed samples: 48 | elapsed time per iteration (ms): 180.2 | throughput per GPU (TFLOP/s/GPU): 685.1 | learning rate: 7.750000E-05 | global batch size: 8 | lm loss: 1.049711E+01 | loss scale: 1.0 | grad norm: 8.737 | number of skipped iterations: 0 | number of nan iterations: 0 |
1156
+ [2026-08-26 19:57:10] iteration 7/ 10 | consumed samples: 56 | elapsed time per iteration (ms): 3015.6 | throughput per GPU (TFLOP/s/GPU): 40.9 | learning rate: 3.250000E-05 | global batch size: 8 | lm loss: 9.440758E+00 | loss scale: 1.0 | grad norm: 5.370 | number of skipped iterations: 0 | number of nan iterations: 0 |
1157
+ [2026-08-26 19:57:10] iteration 8/ 10 | consumed samples: 64 | elapsed time per iteration (ms): 187.9 | throughput per GPU (TFLOP/s/GPU): 657.2 | learning rate: 1.000000E-05 | global batch size: 8 | lm loss: 8.961362E+00 | loss scale: 1.0 | grad norm: 4.547 | number of skipped iterations: 0 | number of nan iterations: 0 |
1158
+ Capture range started in the application.
1159
+ Capture range ended in the application.
1160
+ Generating '/tmp/nsys-report-c5e9.qdstrm'
1161
+
1162
+
1163
+ [2026-08-26 19:57:51] iteration 9/ 10 | consumed samples: 72 | elapsed time per iteration (ms): 40414.4 | throughput per GPU (TFLOP/s/GPU): 3.1 | learning rate: 1.000000E-05 | global batch size: 8 | lm loss: 8.745535E+00 | loss scale: 1.0 | grad norm: 4.292 | number of skipped iterations: 0 | number of nan iterations: 0 |
1164
+ [2026-08-26 19:57:51] iteration 10/ 10 | consumed samples: 80 | elapsed time per iteration (ms): 185.0 | throughput per GPU (TFLOP/s/GPU): 667.3 | learning rate: 1.000000E-05 | global batch size: 8 | lm loss: 8.684624E+00 | loss scale: 1.0 | grad norm: 4.417 | number of skipped iterations: 0 | number of nan iterations: 0 |
1165
+ [after training is done] datetime: 2026-08-26 19:57:51
1166
+ Collection stop is not allowed in this state.
1167
+ [rank0]:[W826 19:57:52.986251685 ProcessGroupNCCL.cpp:1505] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
1168
+ [rank6]:[W826 19:57:53.271259684 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
1169
+ pure virtual method called
1170
+ terminate called without an active exception
1171
+ [rank2]:[W826 19:57:53.373840644 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
1172
+ [rank4]:[W826 19:57:53.379758564 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
1173
+ [rank7]:[W826 19:57:53.381224533 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
1174
+ pure virtual method called
1175
+ terminate called without an active exception
1176
+ pure virtual method called
1177
+ terminate called without an active exception
1178
+ pure virtual method called
1179
+ terminate called without an active exception
1180
+ [rank0]:[W826 19:57:53.843816444 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
1181
+ pure virtual method called
1182
+ terminate called without an active exception
1183
+ [rank5]:[W826 19:57:53.899605230 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
1184
+ pure virtual method called
1185
+ terminate called without an active exception
1186
+ [rank1]:[W826 19:57:53.926550389 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
1187
+ pure virtual method called
1188
+ terminate called without an active exception
1189
+ [rank3]:[W826 19:57:53.956212118 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
1190
+ W0826 19:57:53.921000 723080 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 723190 closing signal SIGTERM
1191
+ W0826 19:57:53.922000 723080 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 723191 closing signal SIGTERM
1192
+ W0826 19:57:53.922000 723080 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 723192 closing signal SIGTERM
1193
+ W0826 19:57:53.922000 723080 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 723193 closing signal SIGTERM
1194
+ W0826 19:57:53.923000 723080 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 723194 closing signal SIGTERM
1195
+ W0826 19:57:53.924000 723080 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 723195 closing signal SIGTERM
1196
+ W0826 19:57:53.925000 723080 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 723197 closing signal SIGTERM
1197
+ E0826 19:57:56.293000 723080 torch/distributed/elastic/multiprocessing/api.py:874] failed (exitcode: -6) local_rank: 6 (pid: 723196) of binary: /usr/bin/python
1198
+ Traceback (most recent call last):
1199
+ File "/usr/local/bin/torchrun", line 8, in <module>
1200
+ sys.exit(main())
1201
+ ^^^^^^
1202
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 355, in wrapper
1203
+ return f(*args, **kwargs)
1204
+ ^^^^^^^^^^^^^^^^^^
1205
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/run.py", line 892, in main
1206
+ run(args)
1207
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/run.py", line 883, in run
1208
+ elastic_launch(
1209
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/launcher/api.py", line 139, in __call__
1210
+ return launch_agent(self._config, self._entrypoint, list(args))
1211
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1212
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/launcher/api.py", line 270, in launch_agent
1213
+ raise ChildFailedError(
1214
+ torch.distributed.elastic.multiprocessing.errors.ChildFailedError:
1215
+ ============================================================
1216
+ pretrain_gpt.py FAILED
1217
+ ------------------------------------------------------------
1218
+ Failures:
1219
+ <NO_OTHER_FAILURES>
1220
+ ------------------------------------------------------------
1221
+ Root Cause (first observed failure):
1222
+ [0]:
1223
+ time : 2026-08-26_19:57:53
1224
+ host : sp-aiinfra-interactive-aghosh-b5xzj-8xb200-moe-cg-0
1225
+ rank : 6 (local_rank: 6)
1226
+ exitcode : -6 (pid: 723196)
1227
+ error_file: <N/A>
1228
+ traceback : Signal 6 (SIGABRT) received by PID 723196
1229
+ ============================================================
1230
+ Generated:
1231
+ /workspace/tierb_study_20260826_144644/echo/tierbtrace/tierbtrace_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__cap1p08984375/nsys/QWEN2_57B__echo__matched__seq4096__gbs8__cap1p08984375.nsys-rep
echo/tierbtrace/tierbtrace_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__cap1p08984375/nsys/verify_trace_ranks.log ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ report : /workspace/tierb_study_20260826_144644/echo/tierbtrace/tierbtrace_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__cap1p08984375/nsys/QWEN2_57B__echo__matched__seq4096__gbs8__cap1p08984375.nsys-rep
2
+ kernels : 81790
3
+ devices : 8 of 8 -> [0, 1, 2, 3, 4, 5, 6, 7]
4
+ processes: 1
5
+ PASS: all ranks present
echo/tierbtrace/tierbtrace_gbs8/trace_results.tsv CHANGED
@@ -1,3 +1,5 @@
1
  model backend capacity status report
2
  QWEN3_235B echo 1.083984375 OK /workspace/tierb_study_20260826_144644/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep
3
  QWEN3_30B echo 1.11328125 OK /workspace/tierb_study_20260826_144644/echo/tierbtrace/tierbtrace_gbs8/QWEN3_30B__echo__matched__seq4096__gbs8__cap1p11328125/nsys/QWEN3_30B__echo__matched__seq4096__gbs8__cap1p11328125.nsys-rep
 
 
 
1
  model backend capacity status report
2
  QWEN3_235B echo 1.083984375 OK /workspace/tierb_study_20260826_144644/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep
3
  QWEN3_30B echo 1.11328125 OK /workspace/tierb_study_20260826_144644/echo/tierbtrace/tierbtrace_gbs8/QWEN3_30B__echo__matched__seq4096__gbs8__cap1p11328125/nsys/QWEN3_30B__echo__matched__seq4096__gbs8__cap1p11328125.nsys-rep
4
+ DSV2_LITE_L24 echo 1.0755208333333333 OK /workspace/tierb_study_20260826_144644/echo/tierbtrace/tierbtrace_gbs8/DSV2_LITE_L24__echo__matched__seq8192__gbs8__cap1p0755208333333333/nsys/DSV2_LITE_L24__echo__matched__seq8192__gbs8__cap1p0755208333333333.nsys-rep
5
+ QWEN2_57B echo 1.08984375 OK /workspace/tierb_study_20260826_144644/echo/tierbtrace/tierbtrace_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__cap1p08984375/nsys/QWEN2_57B__echo__matched__seq4096__gbs8__cap1p08984375.nsys-rep