AbhishekGhosh1998 commited on
Commit
6979dd0
·
verified ·
1 Parent(s): 612b6f6

granularity results granularity_next_b200/tiera_topk_trace @ 2026-09-21T05:22:23Z from sp-aiinfra-interactive-aghosh-8-mb99k-8xb200-moe-cg-0

Browse files
Files changed (13) hide show
  1. granularity_next_b200/tiera_topk_trace/STATUS.txt +1 -1
  2. granularity_next_b200/tiera_topk_trace/hf_watcher.log +5 -0
  3. granularity_next_b200/tiera_topk_trace/hf_watcher.pid +1 -1
  4. granularity_next_b200/tiera_topk_trace/study.log +23 -0
  5. granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/provenance.txt +10 -71
  6. granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G3_K8__ultraep__matched__seq4096__gbs16__cap1p09375/nsys/train.log +44 -0
  7. granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G3_K8__ultraep__matched__seq4096__gbs16__cap1p09375/nsys/verify_trace_ranks.log +6 -0
  8. granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G4_K2__ultraep__matched__seq4096__gbs16__cap1p265625/nsys/run.env +102 -0
  9. granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G4_K2__ultraep__matched__seq4096__gbs16__cap1p265625/nsys/train.log +0 -0
  10. granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G4_K2__ultraep__matched__seq4096__gbs16__cap1p265625/nsys/verify_trace_ranks.log +6 -0
  11. granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G4_K4__ultraep__matched__seq4096__gbs16__cap1p171875/nsys/run.env +102 -0
  12. granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G4_K4__ultraep__matched__seq4096__gbs16__cap1p171875/nsys/train.log +1458 -0
  13. granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/trace_results.tsv +2 -0
granularity_next_b200/tiera_topk_trace/STATUS.txt CHANGED
@@ -1,6 +1,6 @@
1
  study_root: /workspace/granularity_next_b200/tiera_topk_trace
2
  gbs: 16
3
- updated: 2026-09-17T03:22:27+00:00
4
  current: ultraep__tierbtrace__gbs16
5
 
6
  completed steps:
 
1
  study_root: /workspace/granularity_next_b200/tiera_topk_trace
2
  gbs: 16
3
+ updated: 2026-09-21T05:22:22+00:00
4
  current: ultraep__tierbtrace__gbs16
5
 
6
  completed steps:
granularity_next_b200/tiera_topk_trace/hf_watcher.log CHANGED
@@ -31,3 +31,8 @@ https://huggingface.co/MoE-CUDA-Graph-Analysis/tierb-b200-gbs8/tree/main
31
  uploaded: /workspace/granularity_next_b200/tiera_topk_trace -> MoE-CUDA-Graph-Analysis/tierb-b200-gbs8/granularity_next_b200/tiera_topk_trace
32
  https://huggingface.co/MoE-CUDA-Graph-Analysis/tierb-b200-gbs8/tree/main
33
  [2026-09-17T03:35:04+00:00] backup pass
 
 
 
 
 
 
31
  uploaded: /workspace/granularity_next_b200/tiera_topk_trace -> MoE-CUDA-Graph-Analysis/tierb-b200-gbs8/granularity_next_b200/tiera_topk_trace
32
  https://huggingface.co/MoE-CUDA-Graph-Analysis/tierb-b200-gbs8/tree/main
33
  [2026-09-17T03:35:04+00:00] backup pass
34
+ uploaded: /workspace/granularity_next_b200/tiera_topk_trace -> MoE-CUDA-Graph-Analysis/tierb-b200-gbs8/granularity_next_b200/tiera_topk_trace
35
+ https://huggingface.co/MoE-CUDA-Graph-Analysis/tierb-b200-gbs8/tree/main
36
+ Already logged in to Hugging Face as: AbhishekGhosh1998
37
+ Watching /workspace/granularity_next_b200/tiera_topk_trace -> HF every 600s (kill to stop)
38
+ [2026-09-21T05:22:22+00:00] backup pass
granularity_next_b200/tiera_topk_trace/hf_watcher.pid CHANGED
@@ -1 +1 @@
1
- 750811
 
1
+ 18040
granularity_next_b200/tiera_topk_trace/study.log CHANGED
@@ -120,3 +120,26 @@ https://huggingface.co/MoE-CUDA-Graph-Analysis/tierb-b200-gbs8/tree/main
120
  [03:32:56] trace verified -> G3_K2__ultraep__matched__seq4096__gbs16__cap1p1875.nsys-rep (12M, all 8 GPUs + ultraep markers)
121
  [03:32:56] TRACE-ONLY G3_K8__ultraep__matched__seq4096__gbs16__cap1p09375 (ranks: 0 1 2 3 4 5 6 7)
122
  [guard] iter 8/10 (70s)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
120
  [03:32:56] trace verified -> G3_K2__ultraep__matched__seq4096__gbs16__cap1p1875.nsys-rep (12M, all 8 GPUs + ultraep markers)
121
  [03:32:56] TRACE-ONLY G3_K8__ultraep__matched__seq4096__gbs16__cap1p09375 (ranks: 0 1 2 3 4 5 6 7)
122
  [guard] iter 8/10 (70s)
123
+ [guard] no progress for 60s at iter 8/10 (past profile window); treating as done
124
+ [guard] nsys stop (flushing report)
125
+ [guard] tearing down process group (outcome=STALLED)
126
+ [03:35:54] trace verified -> G3_K8__ultraep__matched__seq4096__gbs16__cap1p09375.nsys-rep (12M, all 8 GPUs + ultraep markers)
127
+ [03:35:54] TRACE-ONLY G4_K2__ultraep__matched__seq4096__gbs16__cap1p265625 (ranks: 0 1 2 3 4 5 6 7)
128
+ [guard] iter 8/10 (70s)
129
+ [guard] no progress for 60s at iter 8/10 (past profile window); treating as done
130
+ [guard] nsys stop (flushing report)
131
+ [guard] tearing down process group (outcome=STALLED)
132
+ [03:38:52] trace verified -> G4_K2__ultraep__matched__seq4096__gbs16__cap1p265625.nsys-rep (14M, all 8 GPUs + ultraep markers)
133
+ [03:38:52] TRACE-ONLY G4_K4__ultraep__matched__seq4096__gbs16__cap1p171875 (ranks: 0 1 2 3 4 5 6 7)
134
+ study_root: /workspace/granularity_next_b200/tiera_topk_trace
135
+ gbs: 16
136
+ stages: 9
137
+ nsys: 1
138
+ hf_repo: MoE-CUDA-Graph-Analysis/tierb-b200-gbs8
139
+ hf_remote_root: granularity_next_b200/tiera_topk_trace
140
+ started: 2026-09-21T05:22:22+00:00
141
+ driver_repo: /workspace/blackwell_run_scripts @ b8bd100
142
+
143
+ === STAGE 9 -- Tier B nsys traces (final, trace-only)
144
+
145
+ === STEP ultraep__tierbtrace__gbs16
granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/provenance.txt CHANGED
@@ -1,8 +1,8 @@
1
  ==============================================================
2
  PROVENANCE: ultraep__tierbtrace__gbs16
3
  ==============================================================
4
- captured_at : 2026-09-17T03:22:27+00:00
5
- hostname : hp-aiinfra-interactive-aghosh-8-xzthl-8xb200-moe-cg-0
6
  kernel : 6.8.0-1043-nvidia
7
  cwd : /workspace/blackwell_run_scripts
8
  command : run_ladder.sh --phase tierbtrace --gbs 16 --out /workspace/granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace --nsys
@@ -17,74 +17,13 @@ command : run_ladder.sh --phase tierbtrace --gbs 16 --out /workspace/granu
17
  6, NVIDIA B200, 10.0, 183359 MiB, 580.95.05
18
  7, NVIDIA B200, 10.0, 183359 MiB, 580.95.05
19
  persistence/clocks:
20
- 0, 120 MHz, 1965 MHz, 1000.00 W, 29
21
- 1, 120 MHz, 1965 MHz, 1000.00 W, 32
22
- 2, 1965 MHz, 1965 MHz, 1000.00 W, 34
23
- 3, 1965 MHz, 1965 MHz, 1000.00 W, 30
24
- 4, 120 MHz, 1965 MHz, 1000.00 W, 29
25
- 5, 120 MHz, 1965 MHz, 1000.00 W, 32
26
- 6, 1965 MHz, 1965 MHz, 1000.00 W, 33
27
- 7, 120 MHz, 1965 MHz, 1000.00 W, 29
28
 
29
  NVLink topology:
30
- GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 NIC0 NIC1 NIC2 NIC3 NIC4 NIC5 NIC6 NIC7 NIC8 CPU Affinity NUMA Affinity GPU NUMA ID
31
- GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 PIX NODE SYS SYS SYS SYS SYS SYS NODE 0-27 0 N/A
32
- GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 NODE PIX SYS SYS SYS SYS SYS SYS NODE 0-27 0 N/A
33
- GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 SYS SYS PIX NODE SYS SYS SYS SYS SYS 28-55 1 N/A
34
- GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 SYS SYS NODE PIX SYS SYS SYS SYS SYS 28-55 1 N/A
35
- GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 SYS SYS SYS SYS PIX NODE SYS SYS SYS 56-83 2 N/A
36
- GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 SYS SYS SYS SYS NODE PIX SYS SYS SYS 56-83 2 N/A
37
- GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 SYS SYS SYS SYS SYS SYS PIX NODE SYS 84-111 3 N/A
38
- GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X SYS SYS SYS SYS SYS SYS NODE PIX SYS 84-111 3 N/A
39
- NIC0 PIX NODE SYS SYS SYS SYS SYS SYS X NODE SYS SYS SYS SYS SYS SYS NODE
40
- NIC1 NODE PIX SYS SYS SYS SYS SYS SYS NODE X SYS SYS SYS SYS SYS SYS NODE
41
- NIC2 SYS SYS PIX NODE SYS SYS SYS SYS SYS SYS X NODE SYS SYS SYS SYS SYS
42
- NIC3 SYS SYS NODE PIX SYS SYS SYS SYS SYS SYS NODE X SYS SYS SYS SYS SYS
43
- NIC4 SYS SYS SYS SYS PIX NODE SYS SYS SYS SYS SYS SYS X NODE SYS SYS SYS
44
- unavailable
45
-
46
- ---- toolchain -----------------------------------------------
47
- nvcc : release 12.9, V12.9.86
48
- gcc : gcc (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0
49
- python : 3.12.3
50
- torch : 2.8.0a0+5228986c39.nv25.06 cuda=12.9 nccl=2.27.3
51
- torch arch : ['sm_75', 'sm_80', 'sm_86', 'sm_90', 'sm_100', 'sm_120', 'compute_120']
52
- transformer_engine: 2.10.0.dev0+10f6128d [/usr/local/lib/python3.12/dist-packages/transformer_engine/__init__.py]
53
- deep_ep : 1.2.1+34152ae [/workspace/DeepEP/deep_ep/__init__.py]
54
- ultra_ep : 1.0.0 [/workspace/UltraEP/ultra_ep/__init__.py]
55
- moonep : 0.0.1 [/workspace/MoonEP/moonep/__init__.py]
56
- megatron : 0.14.0rc5 [None]
57
- flash_attn : 2.7.4.post1 [/usr/local/lib/python3.12/dist-packages/flash_attn/__init__.py]
58
- numpy : 1.26.4 [/usr/local/lib/python3.12/dist-packages/numpy/__init__.py]
59
- huggingface_hub: 1.31.0 [/usr/local/lib/python3.12/dist-packages/huggingface_hub/__init__.py]
60
-
61
- ---- repositories --------------------------------------------
62
- Megatron-MoE-Echo bracing-regions-through-kernels 9a23679390d7
63
- Megatron-UltraEP ultraep-bracing-regions-through-kernels 03d57ef7d476
64
- UltraEP main 94cab099b44f
65
- DeepEP HEAD 34152ae28f80 (+3 untracked: build output)
66
- MoonEP megatron-ddp-symmetric-params d8bc6667fad7
67
- blackwell_run_scripts granularity-ultraep 78aa53abad55
68
-
69
- ---- environment ---------------------------------------------
70
- CUDA_ARCH_LIST=7.5 8.0 8.6 9.0 10.0 12.0
71
- CUDA_DRIVER_VERSION=575.57.08
72
- CUDA_HOME=/usr/local/cuda
73
- CUDA_MODULE_LOADING=LAZY
74
- CUDA_VERSION=12.9.1.010
75
- HF_REMOTE_ROOT=granularity_next_b200/tiera_topk_trace
76
- NCCL_VERSION=2.27.3
77
- NVSHMEM_VERSION=3.2.5
78
- PYTORCH_BUILD_NUMBER=0
79
- PYTORCH_BUILD_VERSION=2.8.0a0+5228986
80
- PYTORCH_HOME=/opt/pytorch/pytorch
81
- PYTORCH_VERSION=2.8.0a0+5228986
82
- TORCH_ALLOW_TF32_CUBLAS_OVERRIDE=1
83
- TORCH_CUDA_ARCH_LIST=7.5 8.0 8.6 9.0 10.0 12.0+PTX
84
- TORCH_NCCL_USE_COMM_NONBLOCKING=0
85
-
86
- ---- shared library resolution -------------------------------
87
- libnvshmem_host.so: not in ldconfig
88
- libnccl.so: /lib/x86_64-linux-gnu/libnccl.so.2
89
- lrwxrwxrwx 1 root root 20 Sep 16 20:54 /usr/local/lib/python3.12/dist-packages/nvidia/nvshmem/lib/libnvshmem_host.so -> libnvshmem_host.so.3
90
- -rw-r--r-- 1 root root 179045744 Sep 16 20:52 /usr/local/lib/python3.12/dist-packages/nvidia/nvshmem/lib/libnvshmem_host.so.3
 
1
  ==============================================================
2
  PROVENANCE: ultraep__tierbtrace__gbs16
3
  ==============================================================
4
+ captured_at : 2026-09-21T05:22:22+00:00
5
+ hostname : sp-aiinfra-interactive-aghosh-8-mb99k-8xb200-moe-cg-0
6
  kernel : 6.8.0-1043-nvidia
7
  cwd : /workspace/blackwell_run_scripts
8
  command : run_ladder.sh --phase tierbtrace --gbs 16 --out /workspace/granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace --nsys
 
17
  6, NVIDIA B200, 10.0, 183359 MiB, 580.95.05
18
  7, NVIDIA B200, 10.0, 183359 MiB, 580.95.05
19
  persistence/clocks:
20
+ 0, 120 MHz, 1965 MHz, 1000.00 W, 28
21
+ 1, 1965 MHz, 1965 MHz, 1000.00 W, 34
22
+ 2, 120 MHz, 1965 MHz, 1000.00 W, 33
23
+ 3, 120 MHz, 1965 MHz, 1000.00 W, 29
24
+ 4, 1965 MHz, 1965 MHz, 1000.00 W, 30
25
+ 5, 120 MHz, 1965 MHz, 1000.00 W, 33
26
+ 6, 120 MHz, 1965 MHz, 1000.00 W, 32
27
+ 7, 120 MHz, 1965 MHz, 1000.00 W, 28
28
 
29
  NVLink topology:
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G3_K8__ultraep__matched__seq4096__gbs16__cap1p09375/nsys/train.log CHANGED
@@ -1266,3 +1266,47 @@ Capture range ended in the application.
1266
  Generating '/tmp/nsys-report-2835.qdstrm'
1267
 
1268
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1266
  Generating '/tmp/nsys-report-2835.qdstrm'
1267
 
1268
 
1269
+ Collection stop is not allowed in this state.
1270
+ W0917 03:35:37.900000 969559 torch/distributed/elastic/agent/server/api.py:719] Received 15 death signal, shutting down workers
1271
+ W0917 03:35:37.901000 969559 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 969792 closing signal SIGTERM
1272
+ Generated:
1273
+ /workspace/granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G3_K8__ultraep__matched__seq4096__gbs16__cap1p09375/nsys/G3_K8__ultraep__matched__seq4096__gbs16__cap1p09375.nsys-rep
1274
+ W0917 03:35:37.906000 969559 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 969793 closing signal SIGTERM
1275
+ W0917 03:35:37.908000 969559 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 969795 closing signal SIGTERM
1276
+ W0917 03:35:37.919000 969559 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 969796 closing signal SIGTERM
1277
+ W0917 03:35:37.921000 969559 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 969798 closing signal SIGTERM
1278
+ W0917 03:35:37.933000 969559 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 969799 closing signal SIGTERM
1279
+ W0917 03:35:37.935000 969559 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 969800 closing signal SIGTERM
1280
+ W0917 03:35:37.939000 969559 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 969801 closing signal SIGTERM
1281
+ Traceback (most recent call last):
1282
+ File "/usr/local/bin/torchrun", line 8, in <module>
1283
+ sys.exit(main())
1284
+ ^^^^^^
1285
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 355, in wrapper
1286
+ return f(*args, **kwargs)
1287
+ ^^^^^^^^^^^^^^^^^^
1288
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/run.py", line 892, in main
1289
+ run(args)
1290
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/run.py", line 883, in run
1291
+ elastic_launch(
1292
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/launcher/api.py", line 139, in __call__
1293
+ return launch_agent(self._config, self._entrypoint, list(args))
1294
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1295
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/launcher/api.py", line 261, in launch_agent
1296
+ result = agent.run()
1297
+ ^^^^^^^^^^^
1298
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/metrics/api.py", line 138, in wrapper
1299
+ result = f(*args, **kwargs)
1300
+ ^^^^^^^^^^^^^^^^^^
1301
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/agent/server/api.py", line 711, in run
1302
+ result = self._invoke_run(role)
1303
+ ^^^^^^^^^^^^^^^^^^^^^^
1304
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/agent/server/api.py", line 870, in _invoke_run
1305
+ time.sleep(monitor_interval)
1306
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/multiprocessing/api.py", line 84, in _terminate_process_handler
1307
+ raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval)
1308
+ torch.distributed.elastic.multiprocessing.api.SignalException: Process 969559 got signal: 15
1309
+ The target application terminated. One or more process it created re-parented.
1310
+ Waiting for termination of re-parented processes.
1311
+ Use the `--wait` option to modify this behavior.
1312
+ There are no active sessions.
granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G3_K8__ultraep__matched__seq4096__gbs16__cap1p09375/nsys/verify_trace_ranks.log ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ report : /workspace/granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G3_K8__ultraep__matched__seq4096__gbs16__cap1p09375/nsys/G3_K8__ultraep__matched__seq4096__gbs16__cap1p09375.nsys-rep
2
+ kernels : 157846
3
+ devices : 8 of 8 -> [0, 1, 2, 3, 4, 5, 6, 7]
4
+ processes: 1
5
+ markers : 26624 ultraep_trace__ kernels on devices [0, 1, 2, 3, 4, 5, 6, 7]
6
+ PASS: all ranks present and backend markers emitted
granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G4_K2__ultraep__matched__seq4096__gbs16__cap1p265625/nsys/run.env ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ backend=ultraep
2
+ config=G4_K2
3
+ family=tierA
4
+ source_model=
5
+ standard_num_layers=
6
+ base_rung=G4
7
+ repo=/workspace/Megatron-UltraEP
8
+ branch=ultraep-bracing-regions-through-kernels
9
+ commit=03d57ef7
10
+ num_layers=16
11
+ hidden=4096
12
+ ffn_hidden=14336
13
+ moe_ffn_hidden=896
14
+ num_experts=128
15
+ topk=2
16
+ experts_per_rank=16
17
+ expert_params=11010048
18
+ heads=32
19
+ kv_groups=8
20
+ kv_channels=
21
+ qk_layernorm=0
22
+ shared_ffn=0
23
+ shared_expert_overlap=0
24
+ moe_token_dispatcher=flex
25
+ mla=0
26
+ q_lora_rank=
27
+ kv_lora_rank=512
28
+ qk_head_dim=128
29
+ qk_pos_emb_head_dim=64
30
+ v_head_dim=128
31
+ dense_first_layers=0
32
+ moe_layers=16
33
+ rotary_base=1000000
34
+ rotary_percent=1.0
35
+ rotary_scaling=
36
+ mscale=
37
+ mscale_all_dim=
38
+ qkv_bias=0
39
+ router_balancing=aux_loss
40
+ aux_loss_coeff=1e-2
41
+ router_pre_softmax=0
42
+ router_fusion=0
43
+ router_dtype=fp32
44
+ router_groups=
45
+ router_group_topk=
46
+ router_score_function=softmax
47
+ router_expert_bias=0
48
+ router_bias_update_rate=
49
+ router_topk_scaling_factor=
50
+ shared_expert_gate=0
51
+ apply_layernorm_1p=0
52
+ attention_output_gate=0
53
+ linear_attention_type=
54
+ linear_attention_freq=
55
+ linear_conv_kernel_dim=
56
+ linear_key_head_dim=
57
+ linear_value_head_dim=
58
+ linear_num_key_heads=
59
+ linear_num_value_heads=
60
+ mtp_num_layers=0
61
+ mtp_loss_scaling_factor=0.1
62
+ no_weight_decay_cond_type=
63
+ max_position_embeddings=32768
64
+ init_method_std=0.01
65
+ device_initiated_grouped_gemm=1
66
+ norm_epsilon=1e-5
67
+ seq_length=4096
68
+ micro_batch_size=1
69
+ global_batch_size=16
70
+ num_microbatches=2
71
+ assignments_per_rank=8192
72
+ mean_tokens_per_expert=512
73
+ slot_policy=matched
74
+ slots_per_rank=16
75
+ capacity_mode=static
76
+ capacity=1.265625
77
+ cuda_graph=1
78
+ probe_mode=0
79
+ async_grad_reduce=0
80
+ branch_role=timing
81
+ memory_dump_mode=none
82
+ memory_snapshot_path=
83
+ cuda_graph_warmup_steps=3
84
+ recompute=none
85
+ force_balance=0
86
+ train_iters=10
87
+ ep_size=8
88
+ echo_dispatch_overlap=1
89
+ echo_dispatch_recompute=0
90
+ ultraep_quota_min_tokens=64
91
+ ultraep_weight_sync_join=
92
+ profile_enable=1
93
+ kernel_trace=1
94
+ fp8_mode=mxfp8
95
+ fp8_recipe=mxfp8
96
+ training_seed=1234
97
+ learning_rate=1e-4
98
+ min_learning_rate=1.0e-5
99
+ lr_warmup_iters=5
100
+ state_dump_enable=0
101
+ state_dump_dir=
102
+ moonep_assert_no_pending_reduce=0
granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G4_K2__ultraep__matched__seq4096__gbs16__cap1p265625/nsys/train.log ADDED
The diff for this file is too large to render. See raw diff
 
granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G4_K2__ultraep__matched__seq4096__gbs16__cap1p265625/nsys/verify_trace_ranks.log ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ report : /workspace/granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G4_K2__ultraep__matched__seq4096__gbs16__cap1p265625/nsys/G4_K2__ultraep__matched__seq4096__gbs16__cap1p265625.nsys-rep
2
+ kernels : 186566
3
+ devices : 8 of 8 -> [0, 1, 2, 3, 4, 5, 6, 7]
4
+ processes: 1
5
+ markers : 26624 ultraep_trace__ kernels on devices [0, 1, 2, 3, 4, 5, 6, 7]
6
+ PASS: all ranks present and backend markers emitted
granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G4_K4__ultraep__matched__seq4096__gbs16__cap1p171875/nsys/run.env ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ backend=ultraep
2
+ config=G4_K4
3
+ family=tierA
4
+ source_model=
5
+ standard_num_layers=
6
+ base_rung=G4
7
+ repo=/workspace/Megatron-UltraEP
8
+ branch=ultraep-bracing-regions-through-kernels
9
+ commit=03d57ef7
10
+ num_layers=16
11
+ hidden=4096
12
+ ffn_hidden=14336
13
+ moe_ffn_hidden=896
14
+ num_experts=128
15
+ topk=4
16
+ experts_per_rank=16
17
+ expert_params=11010048
18
+ heads=32
19
+ kv_groups=8
20
+ kv_channels=
21
+ qk_layernorm=0
22
+ shared_ffn=0
23
+ shared_expert_overlap=0
24
+ moe_token_dispatcher=flex
25
+ mla=0
26
+ q_lora_rank=
27
+ kv_lora_rank=512
28
+ qk_head_dim=128
29
+ qk_pos_emb_head_dim=64
30
+ v_head_dim=128
31
+ dense_first_layers=0
32
+ moe_layers=16
33
+ rotary_base=1000000
34
+ rotary_percent=1.0
35
+ rotary_scaling=
36
+ mscale=
37
+ mscale_all_dim=
38
+ qkv_bias=0
39
+ router_balancing=aux_loss
40
+ aux_loss_coeff=1e-2
41
+ router_pre_softmax=0
42
+ router_fusion=0
43
+ router_dtype=fp32
44
+ router_groups=
45
+ router_group_topk=
46
+ router_score_function=softmax
47
+ router_expert_bias=0
48
+ router_bias_update_rate=
49
+ router_topk_scaling_factor=
50
+ shared_expert_gate=0
51
+ apply_layernorm_1p=0
52
+ attention_output_gate=0
53
+ linear_attention_type=
54
+ linear_attention_freq=
55
+ linear_conv_kernel_dim=
56
+ linear_key_head_dim=
57
+ linear_value_head_dim=
58
+ linear_num_key_heads=
59
+ linear_num_value_heads=
60
+ mtp_num_layers=0
61
+ mtp_loss_scaling_factor=0.1
62
+ no_weight_decay_cond_type=
63
+ max_position_embeddings=32768
64
+ init_method_std=0.01
65
+ device_initiated_grouped_gemm=1
66
+ norm_epsilon=1e-5
67
+ seq_length=4096
68
+ micro_batch_size=1
69
+ global_batch_size=16
70
+ num_microbatches=2
71
+ assignments_per_rank=16384
72
+ mean_tokens_per_expert=1024
73
+ slot_policy=matched
74
+ slots_per_rank=16
75
+ capacity_mode=static
76
+ capacity=1.171875
77
+ cuda_graph=1
78
+ probe_mode=0
79
+ async_grad_reduce=0
80
+ branch_role=timing
81
+ memory_dump_mode=none
82
+ memory_snapshot_path=
83
+ cuda_graph_warmup_steps=3
84
+ recompute=none
85
+ force_balance=0
86
+ train_iters=10
87
+ ep_size=8
88
+ echo_dispatch_overlap=1
89
+ echo_dispatch_recompute=0
90
+ ultraep_quota_min_tokens=128
91
+ ultraep_weight_sync_join=
92
+ profile_enable=1
93
+ kernel_trace=1
94
+ fp8_mode=mxfp8
95
+ fp8_recipe=mxfp8
96
+ training_seed=1234
97
+ learning_rate=1e-4
98
+ min_learning_rate=1.0e-5
99
+ lr_warmup_iters=5
100
+ state_dump_enable=0
101
+ state_dump_dir=
102
+ moonep_assert_no_pending_reduce=0
granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G4_K4__ultraep__matched__seq4096__gbs16__cap1p171875/nsys/train.log ADDED
@@ -0,0 +1,1458 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ==============================================================
2
+ G4_K4 / ultraep (slots/rank=16 policy=matched)
3
+ E=128 K=4 H'=896 m=16
4
+ L=16 seq=4096 gbs=16 (2 microbatches)
5
+ cap_mode=static cap=1.171875 cg=1 recompute=none
6
+ memory_dump=none snapshots=/workspace/granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G4_K4__ultraep__matched__seq4096__gbs16__cap1p171875/nsys/memory_snapshots/snapshot.pickle
7
+ repo=/workspace/Megatron-UltraEP @ 03d57ef7
8
+ ==============================================================
9
+ W0917 03:39:00.688000 989911 torch/distributed/run.py:766]
10
+ W0917 03:39:00.688000 989911 torch/distributed/run.py:766] *****************************************
11
+ W0917 03:39:00.688000 989911 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
+ W0917 03:39:00.688000 989911 torch/distributed/run.py:766] *****************************************
13
+ 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
14
+ WARNING: overriding default arguments for tokenizer_type:GPT2BPETokenizer with tokenizer_type:GPT2BPETokenizer
15
+ Number of virtual stages per pipeline stage: None
16
+ accumulate and all-reduce gradients in fp32 for bfloat16 data type.
17
+ using torch.bfloat16 for parameters ...
18
+ /workspace/Megatron-UltraEP/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
19
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
20
+ ------------------------ arguments ------------------------
21
+ account_for_embedding_in_pipeline_split ......... False
22
+ account_for_loss_in_pipeline_split .............. False
23
+ accumulate_allreduce_grads_in_fp32 .............. True
24
+ adam_beta1 ...................................... 0.9
25
+ adam_beta2 ...................................... 0.999
26
+ adam_eps ........................................ 1e-08
27
+ add_bias_linear ................................. False
28
+ add_position_embedding .......................... False
29
+ add_qkv_bias .................................... False
30
+ adlr_autoresume ................................. False
31
+ adlr_autoresume_interval ........................ 1000
32
+ align_grad_reduce ............................... True
33
+ align_param_gather .............................. False
34
+ app_tag_run_name ................................ None
35
+ app_tag_run_version ............................. 0.0.0
36
+ apply_layernorm_1p .............................. False
37
+ apply_query_key_layer_scaling ................... False
38
+ apply_residual_connection_post_layernorm ........ False
39
+ apply_rope_fusion ............................... True
40
+ async_save ...................................... None
41
+ async_tensor_model_parallel_allreduce ........... True
42
+ attention_backend ............................... AttnBackend.auto
43
+ attention_dropout ............................... 0.0
44
+ attention_softmax_in_fp32 ....................... False
45
+ auto_detect_ckpt_format ......................... False
46
+ barrier_with_L1_time ............................ True
47
+ bert_binary_head ................................ True
48
+ bert_embedder_type .............................. megatron
49
+ bert_load ....................................... None
50
+ bf16 ............................................ True
51
+ bias_dropout_fusion ............................. True
52
+ bias_gelu_fusion ................................ False
53
+ bias_swiglu_fusion .............................. True
54
+ biencoder_projection_dim ........................ 0
55
+ biencoder_shared_query_context_model ............ False
56
+ block_data_path ................................. None
57
+ calc_ft_timeouts ................................ False
58
+ calculate_per_token_loss ........................ False
59
+ check_for_large_grads ........................... False
60
+ check_for_nan_in_loss_and_grad .................. False
61
+ check_for_spiky_loss ............................ False
62
+ check_weight_hash_across_dp_replicas_interval ... None
63
+ ckpt_assume_constant_structure .................. False
64
+ ckpt_convert_format ............................. None
65
+ ckpt_convert_save ............................... None
66
+ ckpt_convert_update_legacy_dist_opt_format ...... False
67
+ ckpt_format ..................................... torch_dist
68
+ ckpt_fully_parallel_load ........................ False
69
+ ckpt_fully_parallel_save ........................ True
70
+ ckpt_fully_parallel_save_deprecated ............. False
71
+ ckpt_step ....................................... None
72
+ classes_fraction ................................ 1.0
73
+ clip_grad ....................................... 1.0
74
+ clone_scatter_output_in_embedding ............... True
75
+ config_logger_dir ...............................
76
+ consumed_train_samples .......................... 0
77
+ consumed_valid_samples .......................... 0
78
+ context_parallel_size ........................... 1
79
+ cp_comm_type .................................... ['p2p']
80
+ create_attention_mask_in_dataloader ............. True
81
+ cross_entropy_fusion_impl ....................... native
82
+ cross_entropy_loss_fusion ....................... False
83
+ cuda_graph_scope ................................ full_iteration
84
+ cuda_graph_warmup_steps ......................... 3
85
+ data_args_path .................................. None
86
+ data_cache_path ................................. None
87
+ data_parallel_random_init ....................... False
88
+ data_parallel_sharding_strategy ................. no_shard
89
+ data_parallel_size .............................. 8
90
+ data_path ....................................... ['/workspace/data/wikitext/wikitext103_gpt2_text_document']
91
+ data_per_class_fraction ......................... 1.0
92
+ data_sharding ................................... True
93
+ dataloader_type ................................. single
94
+ ddp_average_in_collective ....................... False
95
+ ddp_bucket_size ................................. None
96
+ ddp_num_buckets ................................. None
97
+ ddp_pad_buckets_for_high_nccl_busbw ............. False
98
+ decoder_first_pipeline_num_layers ............... None
99
+ decoder_last_pipeline_num_layers ................ None
100
+ decoder_num_layers .............................. None
101
+ decoder_seq_length .............................. None
102
+ decoupled_lr .................................... None
103
+ decoupled_min_lr ................................ None
104
+ decrease_batch_size_if_needed ................... False
105
+ defer_embedding_wgrad_compute ................... False
106
+ delay_wgrad_compute ............................. False
107
+ deprecated_use_mcore_models ..................... True
108
+ deterministic_mode .............................. False
109
+ dino_bottleneck_size ............................ 256
110
+ dino_freeze_last_layer .......................... 1
111
+ dino_head_hidden_size ........................... 2048
112
+ dino_local_crops_number ......................... 10
113
+ dino_local_img_size ............................. 96
114
+ dino_norm_last_layer ............................ False
115
+ dino_teacher_temp ............................... 0.07
116
+ dino_warmup_teacher_temp ........................ 0.04
117
+ dino_warmup_teacher_temp_epochs ................. 30
118
+ disable_bf16_reduced_precision_matmul ........... False
119
+ disable_mamba_mem_eff_path ...................... False
120
+ disable_straggler_on_startup .................... False
121
+ dist_ckpt_format_deprecated ..................... None
122
+ dist_ckpt_strictness ............................ assume_ok_unexpected
123
+ distribute_saved_activations .................... False
124
+ distributed_backend ............................. nccl
125
+ distributed_timeout_minutes ..................... 10
126
+ embedding_init_method_std ....................... None
127
+ embedding_path .................................. None
128
+ empty_unused_memory_level ....................... 0
129
+ enable_cuda_graph ............................... True
130
+ enable_experimental ............................. False
131
+ enable_ft_package ............................... False
132
+ enable_gloo_process_groups ...................... True
133
+ enable_msc ...................................... True
134
+ enable_one_logger ............................... True
135
+ encoder_num_layers .............................. 16
136
+ encoder_seq_length .............................. 4096
137
+ end_weight_decay ................................ 0.1
138
+ eod_mask_loss ................................... False
139
+ error_injection_rate ............................ 0
140
+ error_injection_type ............................ transient_error
141
+ eval_interval ................................... 1000000
142
+ eval_iters ...................................... 0
143
+ evidence_data_path .............................. None
144
+ exit_duration_in_mins ........................... None
145
+ exit_interval ................................... None
146
+ exit_on_missing_checkpoint ...................... False
147
+ exit_signal_handler ............................. False
148
+ exp_avg_dtype ................................... torch.float32
149
+ exp_avg_sq_dtype ................................ torch.float32
150
+ expert_model_parallel_size ...................... 8
151
+ expert_tensor_parallel_size ..................... 1
152
+ export_force_local_attention .................... False
153
+ export_kd_cfg ................................... None
154
+ export_kd_teacher_ckpt_format ................... None
155
+ export_kd_teacher_load .......................... None
156
+ export_kv_cache_quant ........................... False
157
+ export_legacy_megatron .......................... False
158
+ export_model_type ............................... GPTModel
159
+ export_moe_apply_probs_on_input ................. False
160
+ export_qk_l2_norm ............................... False
161
+ export_quant_cfg ................................ None
162
+ export_real_quant_cfg ........................... None
163
+ export_te_mcore_model ........................... False
164
+ external_cuda_graph ............................. False
165
+ ffn_hidden_size ................................. 14336
166
+ finetune ........................................ False
167
+ finetune_data_split ............................. train
168
+ finetune_hf_dataset ............................. None
169
+ first_last_layers_bf16 .......................... False
170
+ flash_decode .................................... False
171
+ fp16 ............................................ False
172
+ fp16_lm_cross_entropy ........................... False
173
+ fp32_residual_connection ........................ False
174
+ fp8 ............................................. e4m3
175
+ fp8_amax_compute_algo ........................... most_recent
176
+ fp8_amax_history_len ............................ 1
177
+ fp8_interval .................................... 1
178
+ fp8_margin ...................................... 0
179
+ fp8_param_gather ................................ False
180
+ fp8_recipe ...................................... mxfp8
181
+ fp8_wgrad ....................................... True
182
+ fsdp_double_buffer .............................. False
183
+ global_batch_size ............................... 16
184
+ grad_reduce_in_bf16 ............................. False
185
+ gradient_accumulation_fusion .................... True
186
+ gradient_reduce_div_fusion ...................... True
187
+ group_query_attention ........................... True
188
+ head_lr_mult .................................... 1.0
189
+ heterogeneous_layers_config_encoded_json ........ None
190
+ heterogeneous_layers_config_path ................ None
191
+ hidden_dropout .................................. 0.0
192
+ hidden_size ..................................... 4096
193
+ hierarchical_context_parallel_sizes ............. None
194
+ high_priority_stream_groups ..................... []
195
+ hybrid_attention_ratio .......................... 0.0
196
+ hybrid_mlp_ratio ................................ 0.0
197
+ hybrid_override_pattern ......................... None
198
+ hysteresis ...................................... 2
199
+ ict_head_size ................................... None
200
+ ict_load ........................................ None
201
+ img_h ........................................... 224
202
+ img_w ........................................... 224
203
+ indexer_batch_size .............................. 128
204
+ indexer_log_interval ............................ 1000
205
+ inference_batch_times_seqlen_threshold .......... -1
206
+ inference_dynamic_batching ...................... False
207
+ inference_dynamic_batching_buffer_guaranteed_fraction 0.2
208
+ inference_dynamic_batching_buffer_overflow_factor None
209
+ inference_dynamic_batching_buffer_size_gb ....... 40.0
210
+ inference_dynamic_batching_chunk_size ........... 256
211
+ inference_dynamic_batching_max_requests_override None
212
+ inference_dynamic_batching_max_tokens_override .. None
213
+ inference_dynamic_batching_num_cuda_graphs ...... 16
214
+ inference_max_batch_size ........................ 8
215
+ inference_max_seq_length ........................ 2560
216
+ inference_rng_tracker ........................... False
217
+ init_method_std ................................. 0.01
218
+ init_method_xavier_uniform ...................... False
219
+ init_model_with_meta_device ..................... False
220
+ initial_loss_scale .............................. 4294967296
221
+ inprocess_active_world_size ..................... 8
222
+ inprocess_barrier_timeout ....................... 120
223
+ inprocess_completion_timeout .................... 120
224
+ inprocess_empty_cuda_cache ...................... False
225
+ inprocess_granularity ........................... node
226
+ inprocess_hard_timeout .......................... 90
227
+ inprocess_heartbeat_interval .................... 30
228
+ inprocess_heartbeat_timeout ..................... 60
229
+ inprocess_last_call_wait ........................ 1
230
+ inprocess_max_iterations ........................ None
231
+ inprocess_monitor_process_interval .............. 1.0
232
+ inprocess_monitor_thread_interval ............... 1.0
233
+ inprocess_progress_watchdog_interval ............ 1.0
234
+ inprocess_restart ............................... False
235
+ inprocess_soft_timeout .......................... 60
236
+ inprocess_termination_grace_time ................ 1
237
+ is_hybrid_model ................................. False
238
+ iter_per_epoch .................................. 1250
239
+ iterations_to_skip .............................. []
240
+ keep_fp8_transpose_cache_when_using_custom_fsdp . False
241
+ kitchen_config_file ............................. None
242
+ kitchen_recipe_number ........................... None
243
+ kv_channels ..................................... 128
244
+ kv_lora_rank .................................... 32
245
+ lazy_mpu_init ................................... None
246
+ load ............................................ None
247
+ load_main_params_from_ckpt ...................... None
248
+ load_model_opt_format ........................... False
249
+ local_rank ...................................... 0
250
+ log_energy ...................................... False
251
+ log_interval .................................... 1
252
+ log_loss_scale_to_tensorboard ................... True
253
+ log_memory_to_tensorboard ....................... False
254
+ log_num_zeros_in_grad ........................... False
255
+ log_params_norm ................................. False
256
+ log_progress .................................... False
257
+ log_straggler ................................... False
258
+ log_throughput .................................. True
259
+ log_timers_to_tensorboard ....................... False
260
+ log_validation_ppl_to_tensorboard ............... False
261
+ log_world_size_to_tensorboard ................... False
262
+ logging_level ................................... None
263
+ loss_scale ...................................... None
264
+ loss_scale_window ............................... 1000
265
+ lr .............................................. 0.0001
266
+ lr_decay_iters .................................. 8
267
+ lr_decay_samples ................................ None
268
+ lr_decay_style .................................. cosine
269
+ lr_warmup_fraction .............................. None
270
+ lr_warmup_init .................................. 0.0
271
+ lr_warmup_iters ................................. 5
272
+ lr_warmup_samples ............................... 0
273
+ lr_wsd_decay_iters .............................. None
274
+ lr_wsd_decay_samples ............................ None
275
+ lr_wsd_decay_style .............................. exponential
276
+ main_grads_dtype ................................ torch.float32
277
+ main_params_dtype ............................... torch.float32
278
+ make_vocab_size_divisible_by .................... 128
279
+ mamba_head_dim .................................. 64
280
+ mamba_num_groups ................................ 8
281
+ mamba_num_heads ................................. None
282
+ mamba_state_dim ................................. 128
283
+ manual_gc ....................................... False
284
+ manual_gc_eval .................................. True
285
+ manual_gc_interval .............................. 0
286
+ mask_factor ..................................... 1.0
287
+ mask_prob ....................................... 0.15
288
+ mask_type ....................................... random
289
+ masked_softmax_fusion ........................... False
290
+ max_position_embeddings ......................... 32768
291
+ max_tokens_to_oom ............................... 12000
292
+ memory_snapshot_path ............................ snapshot.pickle
293
+ merge_file ...................................... /workspace/data/wikitext/gpt2-merges.txt
294
+ micro_batch_size ................................ 1
295
+ microbatch_group_size_per_vp_stage .............. None
296
+ mid_level_dataset_surplus ....................... 0.005
297
+ min_loss_scale .................................. 1.0
298
+ min_lr .......................................... 1e-05
299
+ mlp_chunks_for_prefill .......................... 1
300
+ mmap_bin_files .................................. True
301
+ mock_data ....................................... False
302
+ moe_apply_probs_on_input ........................ False
303
+ moe_aux_loss_coeff .............................. 0.01
304
+ moe_capacity_monitor ............................ True
305
+ moe_deepep_num_sms .............................. 20
306
+ moe_echo_enable_random_offloading ............... False
307
+ moe_echo_expert_dispatch_overlap ................ False
308
+ moe_echo_expert_dispatcher_type ................. hybridep
309
+ moe_echo_recompute_expert_dispatch .............. False
310
+ moe_enable_deepep ............................... False
311
+ moe_enable_echo ................................. False
312
+ moe_enable_hybridep ............................. True
313
+ moe_enable_ultraep .............................. True
314
+ moe_expert_capacity_factor ...................... None
315
+ moe_extended_tp ................................. False
316
+ moe_ffn_hidden_size ............................. 896
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 ............................ None
323
+ moe_num_redundant_experts_per_rank .............. 16
324
+ moe_pad_expert_input_to_capacity ................ False
325
+ moe_per_layer_logging ........................... False
326
+ moe_permute_fusion .............................. False
327
+ moe_received_token_capacity ..................... 1.171875
328
+ moe_router_bias_update_rate ..................... 0.001
329
+ moe_router_dtype ................................ fp32
330
+ moe_router_enable_expert_bias ................... False
331
+ moe_router_force_load_balancing ................. False
332
+ moe_router_group_topk ........................... None
333
+ moe_router_load_balancing_type .................. aux_loss
334
+ moe_router_num_groups ........................... None
335
+ moe_router_padding_for_fp8 ...................... False
336
+ moe_router_pre_softmax .......................... False
337
+ moe_router_score_function ....................... softmax
338
+ moe_router_topk ................................. 4
339
+ moe_router_topk_scaling_factor .................. None
340
+ moe_shared_expert_intermediate_size ............. None
341
+ moe_shared_expert_overlap ....................... False
342
+ moe_token_dispatcher_type ....................... flex
343
+ moe_token_drop_policy ........................... probs
344
+ moe_ultraep_max_microbatches .................... None
345
+ moe_ultraep_overlap_weight_sync ................. True
346
+ moe_ultraep_weight_sync_join .................... before_dispatch
347
+ moe_upcycling_granularity ....................... 1
348
+ moe_use_device_initiated_grouped_gemm ........... True
349
+ moe_use_legacy_grouped_gemm ..................... False
350
+ moe_use_upcycling ............................... False
351
+ moe_z_loss_coeff ................................ None
352
+ mrope_section ................................... None
353
+ mscale .......................................... 1.0
354
+ mscale_all_dim .................................. 0.0
355
+ mtp_loss_scaling_factor ......................... 0.1
356
+ mtp_num_layers .................................. None
357
+ multi_latent_attention .......................... False
358
+ nccl_all_reduce_for_prefill ..................... False
359
+ nccl_communicator_config_path ................... None
360
+ nccl_ub ......................................... False
361
+ no_load_optim ................................... None
362
+ no_load_rng ..................................... None
363
+ no_persist_layer_norm ........................... False
364
+ no_rope_freq .................................... None
365
+ no_save_optim ................................... None
366
+ no_save_rng ..................................... None
367
+ non_persistent_ckpt_type ........................ None
368
+ non_persistent_global_ckpt_dir .................. None
369
+ non_persistent_local_ckpt_algo .................. fully_parallel
370
+ non_persistent_local_ckpt_dir ................... None
371
+ non_persistent_save_interval .................... None
372
+ norm_epsilon .................................... 1e-05
373
+ normalization ................................... RMSNorm
374
+ num_attention_heads ............................. 32
375
+ num_channels .................................... 3
376
+ num_classes ..................................... 1000
377
+ num_dataset_builder_threads ..................... 1
378
+ num_distributed_optimizer_instances ............. 1
379
+ num_experts ..................................... 128
380
+ num_layers ...................................... 16
381
+ num_layers_at_end_in_bf16 ....................... 1
382
+ num_layers_at_start_in_bf16 ..................... 1
383
+ num_layers_per_virtual_pipeline_stage ........... None
384
+ num_query_groups ................................ 8
385
+ num_virtual_stages_per_pipeline_rank ............ None
386
+ num_workers ..................................... 2
387
+ object_storage_cache_path ....................... None
388
+ one_logger_async ................................ False
389
+ one_logger_project .............................. megatron-lm
390
+ one_logger_run_name ............................. None
391
+ onnx_safe ....................................... None
392
+ openai_gelu ..................................... False
393
+ optimizer ....................................... adam
394
+ optimizer_cpu_offload ........................... False
395
+ optimizer_offload_fraction ...................... 1.0
396
+ output_bert_embeddings .......................... False
397
+ overlap_cpu_optimizer_d2h_h2d ................... False
398
+ overlap_grad_reduce ............................. True
399
+ overlap_moe_expert_parallel_comm ................ False
400
+ overlap_p2p_comm ................................ False
401
+ overlap_p2p_comm_warmup_flush ................... False
402
+ overlap_param_gather ............................ True
403
+ overlap_param_gather_with_optimizer_step ........ False
404
+ override_opt_param_scheduler .................... False
405
+ params_dtype .................................... torch.bfloat16
406
+ patch_dim ....................................... 16
407
+ per_split_data_args_path ........................ None
408
+ perform_initialization .......................... True
409
+ pin_cpu_grads ................................... True
410
+ pin_cpu_params .................................. True
411
+ pipeline_model_parallel_comm_backend ............ None
412
+ pipeline_model_parallel_layout .................. None
413
+ pipeline_model_parallel_size .................... 1
414
+ position_embedding_type ......................... rope
415
+ pretrained_checkpoint ........................... None
416
+ profile ......................................... True
417
+ profile_ranks ................................... [0, 1, 2, 3, 4, 5, 6, 7]
418
+ profile_step_end ................................ 8
419
+ profile_step_start .............................. 6
420
+ q_lora_rank ..................................... None
421
+ qk_head_dim ..................................... 128
422
+ qk_l2_norm ...................................... False
423
+ qk_layernorm .................................... False
424
+ qk_pos_emb_head_dim ............................. 64
425
+ query_in_block_prob ............................. 0.1
426
+ rampup_batch_size ............................... None
427
+ rank ............................................ 0
428
+ recompute_granularity ........................... None
429
+ recompute_method ................................ None
430
+ recompute_modules ............................... None
431
+ recompute_num_layers ............................ None
432
+ record_memory_history ........................... False
433
+ relative_attention_max_distance ................. 128
434
+ relative_attention_num_buckets .................. 32
435
+ replication ..................................... False
436
+ replication_factor .............................. 2
437
+ replication_jump ................................ None
438
+ rerun_mode ...................................... validate_results
439
+ reset_attention_mask ............................ False
440
+ reset_position_ids .............................. False
441
+ result_rejected_tracker_filename ................ None
442
+ retriever_report_topk_accuracies ................ []
443
+ retriever_score_scaling ......................... False
444
+ retriever_seq_length ............................ 256
445
+ retro_add_retriever ............................. False
446
+ retro_attention_gate ............................ 1
447
+ retro_cyclic_train_iters ........................ None
448
+ retro_encoder_attention_dropout ................. 0.1
449
+ retro_encoder_hidden_dropout .................... 0.1
450
+ retro_encoder_layers ............................ 2
451
+ retro_num_neighbors ............................. 2
452
+ retro_num_retrieved_chunks ...................... 2
453
+ retro_project_dir ............................... None
454
+ retro_verify_neighbor_count ..................... True
455
+ reuse_grad_buf_for_mxfp8_param_ag ............... False
456
+ rope_scaling_factor ............................. 8.0
457
+ rope_type ....................................... None
458
+ rotary_base ..................................... 1000000
459
+ rotary_interleaved .............................. False
460
+ rotary_percent .................................. 1.0
461
+ rotary_scaling_factor ........................... 1.0
462
+ rotary_seq_len_interpolation_factor ............. None
463
+ run_workload_inspector_server ................... False
464
+ sample_rate ..................................... 1.0
465
+ save ............................................ None
466
+ save_interval ................................... None
467
+ save_retain_interval ............................ None
468
+ scatter_gather_tensors_in_pipeline .............. True
469
+ seed ............................................ 1234
470
+ seq_length ...................................... 4096
471
+ sequence_parallel ............................... False
472
+ sft ............................................. False
473
+ sft_tokenizer_prompt_format ..................... nemotron-h-aligned
474
+ sgd_momentum .................................... 0.9
475
+ short_seq_prob .................................. 0.1
476
+ skip_train ...................................... False
477
+ skipped_train_samples ........................... 0
478
+ spec ............................................ None
479
+ split ........................................... 100,0,0
480
+ squared_relu .................................... False
481
+ start_weight_decay .............................. 0.1
482
+ straggler_ctrlr_port ............................ 65535
483
+ straggler_minmax_count .......................... 1
484
+ suggested_communication_unit_size ............... None
485
+ swiglu .......................................... True
486
+ swin_backbone_type .............................. tiny
487
+ symmetric_ar_type ............................... None
488
+ te_rng_tracker .................................. True
489
+ tensor_model_parallel_size ...................... 1
490
+ tensorboard_dir ................................. None
491
+ tensorboard_log_interval ........................ 1
492
+ tensorboard_queue_size .......................... 1000
493
+ test_data_path .................................. None
494
+ test_mode ....................................... False
495
+ tiktoken_num_special_tokens ..................... 1000
496
+ tiktoken_pattern ................................ None
497
+ tiktoken_special_tokens ......................... None
498
+ timing_log_level ................................ 0
499
+ timing_log_option ............................... minmax
500
+ titles_data_path ................................ None
501
+ tokenizer_model ................................. None
502
+ tokenizer_type .................................. GPT2BPETokenizer
503
+ torch_fsdp2_reshard_after_forward ............... True
504
+ tp_comm_bootstrap_backend ....................... nccl
505
+ tp_comm_bulk_dgrad .............................. True
506
+ tp_comm_bulk_wgrad .............................. True
507
+ tp_comm_overlap ................................. False
508
+ tp_comm_overlap_ag .............................. True
509
+ tp_comm_overlap_cfg ............................. None
510
+ tp_comm_overlap_rs .............................. True
511
+ tp_comm_overlap_rs_dgrad ........................ False
512
+ tp_comm_split_ag ................................ True
513
+ tp_comm_split_rs ................................ True
514
+ train_data_path ................................. None
515
+ train_iters ..................................... 10
516
+ train_samples ................................... None
517
+ train_sync_interval ............................. None
518
+ transformer_impl ................................ transformer_engine
519
+ transformer_pipeline_model_parallel_size ........ 1
520
+ untie_embeddings_and_output_weights ............. True
521
+ use_checkpoint_args ............................. False
522
+ use_checkpoint_opt_param_scheduler .............. False
523
+ use_cpu_initialization .......................... None
524
+ use_custom_fsdp ................................. False
525
+ use_dist_ckpt ................................... True
526
+ use_dist_ckpt_deprecated ........................ False
527
+ use_distributed_optimizer ....................... True
528
+ use_flash_attn .................................. False
529
+ use_legacy_models ............................... False
530
+ use_mp_args_from_checkpoint_args ................ False
531
+ use_one_sent_docs ............................... False
532
+ use_persistent_ckpt_worker ...................... False
533
+ use_precision_aware_optimizer ................... False
534
+ use_pytorch_profiler ............................ False
535
+ use_ring_exchange_p2p ........................... False
536
+ use_rope_scaling ................................ False
537
+ use_rotary_position_embeddings .................. False
538
+ use_sharp ....................................... False
539
+ use_tokenizer_model_from_checkpoint_args ........ True
540
+ use_torch_fsdp2 ................................. False
541
+ use_torch_optimizer_for_cpu_offload ............. False
542
+ use_tp_pp_dp_mapping ............................ False
543
+ v_head_dim ...................................... 128
544
+ valid_data_path ................................. None
545
+ variable_seq_lengths ............................ False
546
+ virtual_pipeline_model_parallel_size ............ None
547
+ vision_backbone_type ............................ vit
548
+ vision_pretraining .............................. False
549
+ vision_pretraining_type ......................... classify
550
+ vocab_extra_ids ................................. 0
551
+ vocab_file ...................................... /workspace/data/wikitext/gpt2-vocab.json
552
+ vocab_size ...................................... None
553
+ wandb_exp_name ..................................
554
+ wandb_project ...................................
555
+ wandb_save_dir ..................................
556
+ weight_decay .................................... 0.1
557
+ weight_decay_incr_style ......................... constant
558
+ wgrad_deferral_limit ............................ 0
559
+ world_size ...................................... 8
560
+ yaml_cfg ........................................ None
561
+ -------------------- end of arguments ---------------------
562
+ INFO:megatron.core.num_microbatches_calculator:setting number of microbatches to constant 2
563
+ > building GPT2BPETokenizer tokenizer ...
564
+ > padded vocab (size: 50257) with 47 dummy tokens (new size: 50304)
565
+ WARNING:megatron.core.rerun_state_machine:RerunStateMachine initialized in mode RerunMode.VALIDATE_RESULTS
566
+ > initializing torch distributed ...
567
+ /workspace/Megatron-UltraEP/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
+ /workspace/Megatron-UltraEP/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
570
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
571
+ /workspace/Megatron-UltraEP/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
572
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
573
+ /workspace/Megatron-UltraEP/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
574
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
575
+ /workspace/Megatron-UltraEP/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
576
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
577
+ /workspace/Megatron-UltraEP/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
578
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
579
+ /workspace/Megatron-UltraEP/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
580
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
581
+ 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
582
+ [Gloo] Rank 5 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
583
+ [Gloo] Rank 7 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
584
+ [Gloo] Rank 2 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
585
+ [Gloo] Rank 1 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
586
+ [Gloo] Rank 0 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
587
+ [Gloo] Rank 4 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
588
+ [Gloo] Rank 3 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
589
+ [Gloo] Rank 6 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
590
+ [Gloo] Rank 5 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
591
+ [Gloo] Rank 4 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
592
+ [Gloo] Rank 1 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
593
+ [Gloo] Rank 0 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
594
+ [Gloo] Rank 2 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
595
+ [Gloo] Rank 3 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
596
+ [Gloo] Rank 6 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
597
+ [Gloo] Rank 7 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
598
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
599
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
600
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
601
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
602
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
603
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
604
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
605
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
606
+ > initialized tensor model parallel with size 1
607
+ > initialized pipeline model parallel with size 1
608
+ > setting random seeds to 1234 ...
609
+ [rank5]:[W917 03:39:13.781925838 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.
610
+ [rank7]:[W917 03:39:13.782013200 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.
611
+ [rank6]:[W917 03:39:13.782017160 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.
612
+ [rank1]:[W917 03:39:13.782380098 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.
613
+ [rank4]:[W917 03:39:13.782414970 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.
614
+ [rank2]:[W917 03:39:13.782547558 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.
615
+ [rank3]:[W917 03:39:13.785685160 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.
616
+ > compiling dataset index builder ...
617
+ make: Entering directory '/workspace/Megatron-UltraEP/megatron/core/datasets'
618
+ make: Nothing to be done for 'default'.
619
+ make: Leaving directory '/workspace/Megatron-UltraEP/megatron/core/datasets'
620
+ >>> done with dataset index builder. Compilation time: 0.431 seconds
621
+ WARNING: constraints for invoking optimized fused softmax kernel are not met. We default back to unfused kernel invocations.
622
+ > compiling and loading fused kernels ...
623
+ [rank0]:[W917 03:39:14.411528119 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.
624
+ >>> done with compiling and loading fused kernels. Compilation time: 3.909 seconds
625
+ time to initialize megatron (seconds): 7.647
626
+ [after megatron is initialized] datetime: 2026-09-17 03:39:20
627
+ building GPT model ...
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
+ /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.
639
+ warnings.warn(
640
+ /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.
641
+ warnings.warn(
642
+ /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.
643
+ warnings.warn(
644
+ [INFO] Use MNNVL fabric: False
645
+ [WARN] MAX_NUM_NVL_PEERS is not set. Using detected value 8.
646
+ NVSHMEM v3.7.2
647
+ /dvs/p4/build/sw/rel/gpgpu/toolkit/r12.9/main_nvshmem/src/modules/transport/ibrc/ibrc.cpp:1754: error status: 7 (NVSHMEMX_ERROR_INTERNAL) Device enumeration failed.
648
+
649
+ /dvs/p4/build/sw/rel/gpgpu/toolkit/r12.9/main_nvshmem/src/modules/transport/ibrc/ibrc.cpp:1754: error status: 7 (NVSHMEMX_ERROR_INTERNAL) Device enumeration failed.
650
+
651
+ /dvs/p4/build/sw/rel/gpgpu/toolkit/r12.9/main_nvshmem/src/modules/transport/ibrc/ibrc.cpp:1754: error status: 7 (NVSHMEMX_ERROR_INTERNAL) Device enumeration failed.
652
+
653
+ /dvs/p4/build/sw/rel/gpgpu/toolkit/r12.9/main_nvshmem/src/modules/transport/ibrc/ibrc.cpp:1754: error status: 7 (NVSHMEMX_ERROR_INTERNAL) Device enumeration failed.
654
+
655
+ /dvs/p4/build/sw/rel/gpgpu/toolkit/r12.9/main_nvshmem/src/modules/transport/ibrc/ibrc.cpp:1754: error status: 7 (NVSHMEMX_ERROR_INTERNAL) Device enumeration failed.
656
+
657
+ /dvs/p4/build/sw/rel/gpgpu/toolkit/r12.9/main_nvshmem/src/modules/transport/ibrc/ibrc.cpp:1754: error status: 7 (NVSHMEMX_ERROR_INTERNAL) Device enumeration failed.
658
+
659
+ /dvs/p4/build/sw/rel/gpgpu/toolkit/r12.9/main_nvshmem/src/modules/transport/ibrc/ibrc.cpp:1754: error status: 7 (NVSHMEMX_ERROR_INTERNAL) Device enumeration failed.
660
+
661
+ /dvs/p4/build/sw/rel/gpgpu/toolkit/r12.9/main_nvshmem/src/modules/transport/ibrc/ibrc.cpp:1754: error status: 7 (NVSHMEMX_ERROR_INTERNAL) Device enumeration failed.
662
+
663
+ UltraEP: NVL domain size <= 8; forcing weight_sync plan mode to direct (was adaptive).
664
+ > number of parameters on (tensor, pipeline) model parallel rank (0, 0): 6728847360
665
+ 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)
666
+ INFO:megatron.core.distributed.param_and_grad_buffer:Number of buckets for gradient all-reduce / reduce-scatter: 18
667
+ Params for bucket 1 (206045184 elements, 206045184 padded size):
668
+ module.output_layer.weight
669
+ Params for bucket 2 (42479616 elements, 42479616 padded size):
670
+ module.decoder.layers.15.self_attention.linear_qkv.weight
671
+ module.decoder.layers.15.pre_mlp_layernorm.weight
672
+ module.decoder.final_layernorm.weight
673
+ module.decoder.layers.15.mlp.router.weight
674
+ module.decoder.layers.15.self_attention.linear_qkv.layer_norm_weight
675
+ module.decoder.layers.15.self_attention.linear_proj.weight
676
+ Params for bucket 3 (42475520 elements, 42475520 padded size):
677
+ module.decoder.layers.14.self_attention.linear_qkv.weight
678
+ module.decoder.layers.14.pre_mlp_layernorm.weight
679
+ module.decoder.layers.14.mlp.router.weight
680
+ module.decoder.layers.14.self_attention.linear_qkv.layer_norm_weight
681
+ module.decoder.layers.14.self_attention.linear_proj.weight
682
+ Params for bucket 4 (42475520 elements, 42475520 padded size):
683
+ module.decoder.layers.13.mlp.router.weight
684
+ module.decoder.layers.13.self_attention.linear_qkv.layer_norm_weight
685
+ module.decoder.layers.13.self_attention.linear_proj.weight
686
+ module.decoder.layers.13.self_attention.linear_qkv.weight
687
+ module.decoder.layers.13.pre_mlp_layernorm.weight
688
+ Params for bucket 5 (42475520 elements, 42475520 padded size):
689
+ module.decoder.layers.12.self_attention.linear_qkv.weight
690
+ module.decoder.layers.12.pre_mlp_layernorm.weight
691
+ module.decoder.layers.12.mlp.router.weight
692
+ module.decoder.layers.12.self_attention.linear_proj.weight
693
+ module.decoder.layers.12.self_attention.linear_qkv.layer_norm_weight
694
+ Params for bucket 6 (42475520 elements, 42475520 padded size):
695
+ module.decoder.layers.11.self_attention.linear_qkv.weight
696
+ module.decoder.layers.11.pre_mlp_layernorm.weight
697
+ module.decoder.layers.11.mlp.router.weight
698
+ module.decoder.layers.11.self_attention.linear_proj.weight
699
+ module.decoder.layers.11.self_attention.linear_qkv.layer_norm_weight
700
+ Params for bucket 7 (42475520 elements, 42475520 padded size):
701
+ module.decoder.layers.10.mlp.router.weight
702
+ module.decoder.layers.10.self_attention.linear_qkv.layer_norm_weight
703
+ module.decoder.layers.10.self_attention.linear_proj.weight
704
+ module.decoder.layers.10.self_attention.linear_qkv.weight
705
+ module.decoder.layers.10.pre_mlp_layernorm.weight
706
+ Params for bucket 8 (42475520 elements, 42475520 padded size):
707
+ module.decoder.layers.9.mlp.router.weight
708
+ module.decoder.layers.9.self_attention.linear_proj.weight
709
+ module.decoder.layers.9.self_attention.linear_qkv.layer_norm_weight
710
+ module.decoder.layers.9.pre_mlp_layernorm.weight
711
+ module.decoder.layers.9.self_attention.linear_qkv.weight
712
+ Params for bucket 9 (42475520 elements, 42475520 padded size):
713
+ module.decoder.layers.8.mlp.router.weight
714
+ module.decoder.layers.8.self_attention.linear_qkv.layer_norm_weight
715
+ module.decoder.layers.8.self_attention.linear_proj.weight
716
+ module.decoder.layers.8.self_attention.linear_qkv.weight
717
+ module.decoder.layers.8.pre_mlp_layernorm.weight
718
+ Params for bucket 10 (42475520 elements, 42475520 padded size):
719
+ module.decoder.layers.7.mlp.router.weight
720
+ module.decoder.layers.7.self_attention.linear_proj.weight
721
+ module.decoder.layers.7.self_attention.linear_qkv.layer_norm_weight
722
+ module.decoder.layers.7.pre_mlp_layernorm.weight
723
+ module.decoder.layers.7.self_attention.linear_qkv.weight
724
+ Params for bucket 11 (42475520 elements, 42475520 padded size):
725
+ module.decoder.layers.6.pre_mlp_layernorm.weight
726
+ module.decoder.layers.6.self_attention.linear_qkv.weight
727
+ module.decoder.layers.6.mlp.router.weight
728
+ module.decoder.layers.6.self_attention.linear_proj.weight
729
+ module.decoder.layers.6.self_attention.linear_qkv.layer_norm_weight
730
+ Params for bucket 12 (42475520 elements, 42475520 padded size):
731
+ module.decoder.layers.5.self_attention.linear_qkv.weight
732
+ module.decoder.layers.5.self_attention.linear_proj.weight
733
+ module.decoder.layers.5.pre_mlp_layernorm.weight
734
+ module.decoder.layers.5.mlp.router.weight
735
+ module.decoder.layers.5.self_attention.linear_qkv.layer_norm_weight
736
+ Params for bucket 13 (42475520 elements, 42475520 padded size):
737
+ module.decoder.layers.4.mlp.router.weight
738
+ module.decoder.layers.4.self_attention.linear_proj.weight
739
+ module.decoder.layers.4.self_attention.linear_qkv.layer_norm_weight
740
+ module.decoder.layers.4.self_attention.linear_qkv.weight
741
+ module.decoder.layers.4.pre_mlp_layernorm.weight
742
+ Params for bucket 14 (42475520 elements, 42475520 padded size):
743
+ module.decoder.layers.3.mlp.router.weight
744
+ module.decoder.layers.3.self_attention.linear_qkv.layer_norm_weight
745
+ module.decoder.layers.3.self_attention.linear_proj.weight
746
+ module.decoder.layers.3.self_attention.linear_qkv.weight
747
+ module.decoder.layers.3.pre_mlp_layernorm.weight
748
+ Params for bucket 15 (42475520 elements, 42475520 padded size):
749
+ module.decoder.layers.2.mlp.router.weight
750
+ module.decoder.layers.2.self_attention.linear_proj.weight
751
+ module.decoder.layers.2.self_attention.linear_qkv.layer_norm_weight
752
+ module.decoder.layers.2.pre_mlp_layernorm.weight
753
+ module.decoder.layers.2.self_attention.linear_qkv.weight
754
+ Params for bucket 16 (42475520 elements, 42475520 padded size):
755
+ module.decoder.layers.1.self_attention.linear_qkv.weight
756
+ module.decoder.layers.1.pre_mlp_layernorm.weight
757
+ module.decoder.layers.1.mlp.router.weight
758
+ module.decoder.layers.1.self_attention.linear_qkv.layer_norm_weight
759
+ module.decoder.layers.1.self_attention.linear_proj.weight
760
+ Params for bucket 17 (42475520 elements, 42475520 padded size):
761
+ module.decoder.layers.0.mlp.router.weight
762
+ module.decoder.layers.0.pre_mlp_layernorm.weight
763
+ module.decoder.layers.0.self_attention.linear_proj.weight
764
+ module.decoder.layers.0.self_attention.linear_qkv.weight
765
+ module.decoder.layers.0.self_attention.linear_qkv.layer_norm_weight
766
+ Params for bucket 18 (206045184 elements, 206045184 padded size):
767
+ module.embedding.word_embeddings.weight
768
+ INFO:megatron.core.distributed.param_and_grad_buffer:Number of buckets for gradient all-reduce / reduce-scatter: 67
769
+ Params for bucket 1 (40370176 elements, 40370176 padded size):
770
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight12
771
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight14
772
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight10
773
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight8
774
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight6
775
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight15
776
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight13
777
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight11
778
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight9
779
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight7
780
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight5
781
+ Params for bucket 2 (40370176 elements, 40370176 padded size):
782
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight4
783
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight2
784
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight0
785
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight15
786
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight13
787
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight3
788
+ module.decoder.layers.15.mlp.experts.linear_fc2.weight1
789
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight14
790
+ Params for bucket 3 (44040192 elements, 44040192 padded size):
791
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight11
792
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight9
793
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight7
794
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight12
795
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight8
796
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight10
797
+ Params for bucket 4 (44040192 elements, 44040192 padded size):
798
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight5
799
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight3
800
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight1
801
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight6
802
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight2
803
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight4
804
+ Params for bucket 5 (40370176 elements, 40370176 padded size):
805
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight14
806
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight12
807
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight10
808
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight8
809
+ module.decoder.layers.15.mlp.experts.linear_fc1.weight0
810
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight15
811
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight13
812
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight11
813
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight9
814
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight7
815
+ Params for bucket 6 (40370176 elements, 40370176 padded size):
816
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight6
817
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight4
818
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight2
819
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight0
820
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight15
821
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight5
822
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight3
823
+ module.decoder.layers.14.mlp.experts.linear_fc2.weight1
824
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight14
825
+ Params for bucket 7 (44040192 elements, 44040192 padded size):
826
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight13
827
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight11
828
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight9
829
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight10
830
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight12
831
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight8
832
+ Params for bucket 8 (44040192 elements, 44040192 padded size):
833
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight7
834
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight5
835
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight3
836
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight4
837
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight6
838
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight2
839
+ Params for bucket 9 (40370176 elements, 40370176 padded size):
840
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight1
841
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight14
842
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight12
843
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight10
844
+ module.decoder.layers.14.mlp.experts.linear_fc1.weight0
845
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight15
846
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight13
847
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight11
848
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight9
849
+ Params for bucket 10 (40370176 elements, 40370176 padded size):
850
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight8
851
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight4
852
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight6
853
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight2
854
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight0
855
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight7
856
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight5
857
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight3
858
+ module.decoder.layers.13.mlp.experts.linear_fc2.weight1
859
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight15
860
+ Params for bucket 11 (44040192 elements, 44040192 padded size):
861
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight14
862
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight10
863
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight12
864
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight11
865
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight13
866
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight9
867
+ Params for bucket 12 (44040192 elements, 44040192 padded size):
868
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight8
869
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight4
870
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight6
871
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight5
872
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight7
873
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight3
874
+ Params for bucket 13 (40370176 elements, 40370176 padded size):
875
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight2
876
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight15
877
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight0
878
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight13
879
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight11
880
+ module.decoder.layers.13.mlp.experts.linear_fc1.weight1
881
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight14
882
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight12
883
+ Params for bucket 14 (40370176 elements, 40370176 padded size):
884
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight9
885
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight7
886
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight5
887
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight3
888
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight1
889
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight10
890
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight8
891
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight6
892
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight4
893
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight2
894
+ module.decoder.layers.12.mlp.experts.linear_fc2.weight0
895
+ Params for bucket 15 (44040192 elements, 44040192 padded size):
896
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight12
897
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight14
898
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight10
899
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight15
900
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight11
901
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight13
902
+ Params for bucket 16 (44040192 elements, 44040192 padded size):
903
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight6
904
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight8
905
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight4
906
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight9
907
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight7
908
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight5
909
+ Params for bucket 17 (40370176 elements, 40370176 padded size):
910
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight0
911
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight2
912
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight14
913
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight3
914
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight15
915
+ module.decoder.layers.12.mlp.experts.linear_fc1.weight1
916
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight13
917
+ Params for bucket 18 (40370176 elements, 40370176 padded size):
918
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight12
919
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight10
920
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight8
921
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight6
922
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight4
923
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight2
924
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight9
925
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight11
926
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight7
927
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight5
928
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight3
929
+ Params for bucket 19 (44040192 elements, 44040192 padded size):
930
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight13
931
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight0
932
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight15
933
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight11
934
+ module.decoder.layers.11.mlp.experts.linear_fc2.weight1
935
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight14
936
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight12
937
+ Params for bucket 20 (44040192 elements, 44040192 padded size):
938
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight7
939
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight9
940
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight5
941
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight10
942
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight8
943
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight6
944
+ Params for bucket 21 (40370176 elements, 40370176 padded size):
945
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight3
946
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight1
947
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight15
948
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight4
949
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight0
950
+ module.decoder.layers.11.mlp.experts.linear_fc1.weight2
951
+ Params for bucket 22 (40370176 elements, 40370176 padded size):
952
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight13
953
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight11
954
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight9
955
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight7
956
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight5
957
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight14
958
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight10
959
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight12
960
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight8
961
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight6
962
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight4
963
+ Params for bucket 23 (44040192 elements, 44040192 padded size):
964
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight3
965
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight1
966
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight14
967
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight12
968
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight0
969
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight15
970
+ module.decoder.layers.10.mlp.experts.linear_fc2.weight2
971
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight13
972
+ Params for bucket 24 (44040192 elements, 44040192 padded size):
973
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight10
974
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight8
975
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight6
976
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight11
977
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight7
978
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight9
979
+ Params for bucket 25 (44040192 elements, 44040192 padded size):
980
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight4
981
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight2
982
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight0
983
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight5
984
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight1
985
+ module.decoder.layers.10.mlp.experts.linear_fc1.weight3
986
+ Params for bucket 26 (40370176 elements, 40370176 padded size):
987
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight15
988
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight11
989
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight13
990
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight9
991
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight7
992
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight5
993
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight12
994
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight14
995
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight10
996
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight8
997
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight6
998
+ Params for bucket 27 (40370176 elements, 40370176 padded size):
999
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight1
1000
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight3
1001
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight14
1002
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight4
1003
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight2
1004
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight0
1005
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight15
1006
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight13
1007
+ Params for bucket 28 (44040192 elements, 44040192 padded size):
1008
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight12
1009
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight8
1010
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight10
1011
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight9
1012
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight11
1013
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight7
1014
+ Params for bucket 29 (44040192 elements, 44040192 padded size):
1015
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight6
1016
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight2
1017
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight4
1018
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight3
1019
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight5
1020
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight1
1021
+ Params for bucket 30 (40370176 elements, 40370176 padded size):
1022
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight0
1023
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight13
1024
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight15
1025
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight11
1026
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight9
1027
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight7
1028
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight14
1029
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight12
1030
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight10
1031
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight8
1032
+ Params for bucket 31 (40370176 elements, 40370176 padded size):
1033
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight3
1034
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight5
1035
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight1
1036
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight14
1037
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight6
1038
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight4
1039
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight2
1040
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight0
1041
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight15
1042
+ Params for bucket 32 (44040192 elements, 44040192 padded size):
1043
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight10
1044
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight12
1045
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight8
1046
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight13
1047
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight11
1048
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight9
1049
+ Params for bucket 33 (44040192 elements, 44040192 padded size):
1050
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight4
1051
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight6
1052
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight2
1053
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight7
1054
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight3
1055
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight5
1056
+ Params for bucket 34 (40370176 elements, 40370176 padded size):
1057
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight15
1058
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight0
1059
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight13
1060
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight11
1061
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight9
1062
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight1
1063
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight14
1064
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight12
1065
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight10
1066
+ Params for bucket 35 (40370176 elements, 40370176 padded size):
1067
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight5
1068
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight7
1069
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight3
1070
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight1
1071
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight8
1072
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight4
1073
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight6
1074
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight2
1075
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight0
1076
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight15
1077
+ Params for bucket 36 (44040192 elements, 44040192 padded size):
1078
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight14
1079
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight12
1080
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight10
1081
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight13
1082
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight11
1083
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight9
1084
+ Params for bucket 37 (44040192 elements, 44040192 padded size):
1085
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight8
1086
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight6
1087
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight4
1088
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight7
1089
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight5
1090
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight3
1091
+ Params for bucket 38 (40370176 elements, 40370176 padded size):
1092
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight2
1093
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight0
1094
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight15
1095
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight13
1096
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight11
1097
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight14
1098
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight1
1099
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight12
1100
+ Params for bucket 39 (40370176 elements, 40370176 padded size):
1101
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight7
1102
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight9
1103
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight5
1104
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight3
1105
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight1
1106
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight10
1107
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight6
1108
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight8
1109
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight4
1110
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight2
1111
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight0
1112
+ Params for bucket 40 (44040192 elements, 44040192 padded size):
1113
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight15
1114
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight13
1115
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight11
1116
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight12
1117
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight14
1118
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight10
1119
+ Params for bucket 41 (44040192 elements, 44040192 padded size):
1120
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight9
1121
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight7
1122
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight5
1123
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight6
1124
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight8
1125
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight4
1126
+ Params for bucket 42 (40370176 elements, 40370176 padded size):
1127
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight3
1128
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight1
1129
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight14
1130
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight0
1131
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight15
1132
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight2
1133
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight13
1134
+ Params for bucket 43 (40370176 elements, 40370176 padded size):
1135
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight12
1136
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight10
1137
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight8
1138
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight6
1139
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight4
1140
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight2
1141
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight11
1142
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight9
1143
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight7
1144
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight5
1145
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight3
1146
+ Params for bucket 44 (44040192 elements, 44040192 padded size):
1147
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight0
1148
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight15
1149
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight13
1150
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight11
1151
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight1
1152
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight14
1153
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight12
1154
+ Params for bucket 45 (44040192 elements, 44040192 padded size):
1155
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight9
1156
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight7
1157
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight5
1158
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight10
1159
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight6
1160
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight8
1161
+ Params for bucket 46 (40370176 elements, 40370176 padded size):
1162
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight1
1163
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight3
1164
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight4
1165
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight2
1166
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight0
1167
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight15
1168
+ Params for bucket 47 (40370176 elements, 40370176 padded size):
1169
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight14
1170
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight12
1171
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight10
1172
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight8
1173
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight6
1174
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight4
1175
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight13
1176
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight11
1177
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight9
1178
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight7
1179
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight5
1180
+ Params for bucket 48 (44040192 elements, 44040192 padded size):
1181
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight2
1182
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight0
1183
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight14
1184
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight12
1185
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight3
1186
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight15
1187
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight1
1188
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight13
1189
+ Params for bucket 49 (44040192 elements, 44040192 padded size):
1190
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight10
1191
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight8
1192
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight6
1193
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight11
1194
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight9
1195
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight7
1196
+ Params for bucket 50 (44040192 elements, 44040192 padded size):
1197
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight2
1198
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight4
1199
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight0
1200
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight5
1201
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight1
1202
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight3
1203
+ Params for bucket 51 (40370176 elements, 40370176 padded size):
1204
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight15
1205
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight11
1206
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight13
1207
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight9
1208
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight7
1209
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight5
1210
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight14
1211
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight12
1212
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight10
1213
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight8
1214
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight6
1215
+ Params for bucket 52 (40370176 elements, 40370176 padded size):
1216
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight1
1217
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight3
1218
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight14
1219
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight4
1220
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight0
1221
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight2
1222
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight15
1223
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight13
1224
+ Params for bucket 53 (44040192 elements, 44040192 padded size):
1225
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight12
1226
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight10
1227
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight8
1228
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight11
1229
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight9
1230
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight7
1231
+ Params for bucket 54 (44040192 elements, 44040192 padded size):
1232
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight6
1233
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight4
1234
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight2
1235
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight5
1236
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight3
1237
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight1
1238
+ Params for bucket 55 (40370176 elements, 40370176 padded size):
1239
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight0
1240
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight15
1241
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight13
1242
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight11
1243
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight9
1244
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight7
1245
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight12
1246
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight14
1247
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight10
1248
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight8
1249
+ Params for bucket 56 (40370176 elements, 40370176 padded size):
1250
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight3
1251
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight5
1252
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight1
1253
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight14
1254
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight6
1255
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight4
1256
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight2
1257
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight0
1258
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight15
1259
+ Params for bucket 57 (44040192 elements, 44040192 padded size):
1260
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight12
1261
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight10
1262
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight8
1263
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight13
1264
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight9
1265
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight11
1266
+ Params for bucket 58 (44040192 elements, 44040192 padded size):
1267
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight6
1268
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight4
1269
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight2
1270
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight7
1271
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight3
1272
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight5
1273
+ Params for bucket 59 (40370176 elements, 40370176 padded size):
1274
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight0
1275
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight14
1276
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight12
1277
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight10
1278
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight1
1279
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight13
1280
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight15
1281
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight11
1282
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight9
1283
+ Params for bucket 60 (40370176 elements, 40370176 padded size):
1284
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight8
1285
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight6
1286
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight4
1287
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight2
1288
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight0
1289
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight15
1290
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight7
1291
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight5
1292
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight3
1293
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight1
1294
+ Params for bucket 61 (44040192 elements, 44040192 padded size):
1295
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight13
1296
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight11
1297
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight9
1298
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight14
1299
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight10
1300
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight12
1301
+ Params for bucket 62 (44040192 elements, 44040192 padded size):
1302
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight7
1303
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight5
1304
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight3
1305
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight8
1306
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight6
1307
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight4
1308
+ Params for bucket 63 (40370176 elements, 40370176 padded size):
1309
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight14
1310
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight1
1311
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight12
1312
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight2
1313
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight0
1314
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight15
1315
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight13
1316
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight11
1317
+ Params for bucket 64 (40370176 elements, 40370176 padded size):
1318
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight10
1319
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight8
1320
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight6
1321
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight4
1322
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight2
1323
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight0
1324
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight9
1325
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight7
1326
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight5
1327
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight3
1328
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight1
1329
+ Params for bucket 65 (44040192 elements, 44040192 padded size):
1330
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight15
1331
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight13
1332
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight11
1333
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight12
1334
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight14
1335
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight10
1336
+ Params for bucket 66 (44040192 elements, 44040192 padded size):
1337
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight9
1338
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight5
1339
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight7
1340
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight6
1341
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight8
1342
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight4
1343
+ Params for bucket 67 (29360128 elements, 29360128 padded size):
1344
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight3
1345
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight2
1346
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight1
1347
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight0
1348
+ 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 0x78f6c1f6a6f0>, config_logger_dir='')
1349
+ INFO:megatron.core.optimizer_param_scheduler:> learning rate decay style: cosine
1350
+ [after model, optimizer, and learning rate scheduler are built] datetime: 2026-09-17 03:39:24
1351
+ > building train, validation, and test datasets ...
1352
+ > datasets target sizes (minimum size):
1353
+ train: 160
1354
+ validation: 0
1355
+ test: 0
1356
+ INFO:megatron.core.datasets.blended_megatron_dataset_config:Let split_matrix = [(0, 1.0), None, None]
1357
+ > building train, validation, and test datasets for GPT ...
1358
+ INFO:megatron.core.datasets.blended_megatron_dataset_builder:Building GPTDataset splits with sizes=(160, 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 0x78f6c2bda780>, 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)
1359
+ INFO:megatron.core.datasets.indexed_dataset:Load the _IndexReader from /workspace/data/wikitext/wikitext103_gpt2_text_document.idx
1360
+ INFO:megatron.core.datasets.indexed_dataset: Extract the sequence lengths
1361
+ INFO:megatron.core.datasets.indexed_dataset: Extract the sequence pointers
1362
+ INFO:megatron.core.datasets.indexed_dataset: Extract the document indices
1363
+ INFO:megatron.core.datasets.indexed_dataset:> total number of sequences: 1165029
1364
+ INFO:megatron.core.datasets.indexed_dataset:> total number of documents: 1801350
1365
+ INFO:megatron.core.datasets.gpt_dataset:Load the GPTDataset train indices
1366
+ INFO:megatron.core.datasets.gpt_dataset: Load the document index from 68684a515e118253a8572c4517cf2ffe-GPTDataset-train-document_index.npy
1367
+ INFO:megatron.core.datasets.gpt_dataset: Load the sample index from 68684a515e118253a8572c4517cf2ffe-GPTDataset-train-sample_index.npy
1368
+ INFO:megatron.core.datasets.gpt_dataset: Load the shuffle index from 68684a515e118253a8572c4517cf2ffe-GPTDataset-train-shuffle_index.npy
1369
+ INFO:megatron.core.datasets.gpt_dataset:> total number of samples: 29073
1370
+ > finished creating GPT datasets ...
1371
+ [after dataloaders are built] datetime: 2026-09-17 03:39:24
1372
+ done with setup ...
1373
+ (min, max) time across ranks (ms):
1374
+ model-and-optimizer-setup ......................: (4148.18, 4158.05)
1375
+ train/valid/test-data-iterators-setup ..........: (108.26, 139.90)
1376
+ training ...
1377
+ Overwriting rerun_state_machine.current_iteration from -1 to 0...
1378
+ [before the start of training step] datetime: 2026-09-17 03:39:24
1379
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1380
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1381
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1382
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1383
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1384
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1385
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1386
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1387
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1388
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1389
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1390
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1391
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1392
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1393
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1394
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1395
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1396
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1397
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1398
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1399
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1400
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1401
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1402
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1403
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1404
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1405
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1406
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1407
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1408
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1409
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1410
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1411
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1412
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1413
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1414
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1415
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1416
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1417
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1418
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1419
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1420
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1421
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1422
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1423
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1424
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1425
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1426
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1427
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1428
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1429
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1430
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1431
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1432
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1433
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1434
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1435
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1436
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1437
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1438
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1439
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1440
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1441
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1442
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1443
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1444
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1445
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1446
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1447
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1448
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1449
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1450
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1451
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1452
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1453
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1454
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1455
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1456
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1457
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1458
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/trace_results.tsv CHANGED
@@ -1,3 +1,5 @@
1
  model backend capacity status report async_overlap_json
2
  G2_K2 ultraep 1.15625 OK /workspace/granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G2_K2__ultraep__matched__seq4096__gbs16__cap1p15625/nsys/G2_K2__ultraep__matched__seq4096__gbs16__cap1p15625.nsys-rep
3
  G3_K2 ultraep 1.1875 OK /workspace/granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G3_K2__ultraep__matched__seq4096__gbs16__cap1p1875/nsys/G3_K2__ultraep__matched__seq4096__gbs16__cap1p1875.nsys-rep
 
 
 
1
  model backend capacity status report async_overlap_json
2
  G2_K2 ultraep 1.15625 OK /workspace/granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G2_K2__ultraep__matched__seq4096__gbs16__cap1p15625/nsys/G2_K2__ultraep__matched__seq4096__gbs16__cap1p15625.nsys-rep
3
  G3_K2 ultraep 1.1875 OK /workspace/granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G3_K2__ultraep__matched__seq4096__gbs16__cap1p1875/nsys/G3_K2__ultraep__matched__seq4096__gbs16__cap1p1875.nsys-rep
4
+ G3_K8 ultraep 1.09375 OK /workspace/granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G3_K8__ultraep__matched__seq4096__gbs16__cap1p09375/nsys/G3_K8__ultraep__matched__seq4096__gbs16__cap1p09375.nsys-rep
5
+ G4_K2 ultraep 1.265625 OK /workspace/granularity_next_b200/tiera_topk_trace/ultraep/tierbtrace/tierbtrace_gbs16/G4_K2__ultraep__matched__seq4096__gbs16__cap1p265625/nsys/G4_K2__ultraep__matched__seq4096__gbs16__cap1p265625.nsys-rep