AbhishekGhosh1998 commited on
Commit
93d82e9
·
verified ·
1 Parent(s): 8d25378

granularity results echo/tierbmem @ 2026-08-26T19:10:38Z from sp-aiinfra-interactive-aghosh-b5xzj-8xb200-moe-cg-0 (part 4)

Browse files
Files changed (13) hide show
  1. echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0006_iter000004_20260826T191025709574Z.pickle +3 -0
  2. echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000000_20260826T191024263867Z.json +79 -0
  3. echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000000_20260826T191024263867Z.pickle +3 -0
  4. echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000001_20260826T191024742492Z.json +79 -0
  5. echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000001_20260826T191024742492Z.pickle +3 -0
  6. echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000002_20260826T191025066187Z.json +79 -0
  7. echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000002_20260826T191025066187Z.pickle +3 -0
  8. echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000003_20260826T191025387991Z.json +79 -0
  9. echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000003_20260826T191025387991Z.pickle +3 -0
  10. echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000004_20260826T191025709649Z.json +79 -0
  11. echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000004_20260826T191025709649Z.pickle +3 -0
  12. echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/run.env +61 -0
  13. echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/train.log +1262 -0
echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0006_iter000004_20260826T191025709574Z.pickle ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:741a43ff5e57b55a136a1d1fbd6da0945a1233e0f152abae1aa45c5e1ccdbc3a
3
+ size 1689648
echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000000_20260826T191024263867Z.json ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "active_bytes_current": 72123770880,
3
+ "active_bytes_current_delta": 31437802496,
4
+ "active_bytes_peak": 73614477312,
5
+ "active_bytes_peak_delta": 32890760192,
6
+ "after": {
7
+ "active_bytes_current": 72123770880,
8
+ "active_bytes_peak": 73614477312,
9
+ "allocated_bytes_current": 72123770880,
10
+ "allocated_bytes_peak": 73614477312,
11
+ "device": 7,
12
+ "inactive_split_bytes_current": 2113312768,
13
+ "inactive_split_bytes_peak": 4687297024,
14
+ "iteration": 0,
15
+ "mode": "iteration",
16
+ "num_alloc_retries": 0,
17
+ "num_ooms": 0,
18
+ "phase": "iteration_end",
19
+ "rank": 7,
20
+ "reserved_bytes_current": 83766542336,
21
+ "reserved_bytes_peak": 83766542336,
22
+ "segment_count_current": 1015,
23
+ "segment_count_peak": 1015,
24
+ "timestamp_utc": "2026-08-26T19:10:24.263381Z",
25
+ "world_size": 8
26
+ },
27
+ "allocated_bytes_current": 72123770880,
28
+ "allocated_bytes_current_delta": 31437802496,
29
+ "allocated_bytes_peak": 73614477312,
30
+ "allocated_bytes_peak_delta": 32890760192,
31
+ "before": {
32
+ "active_bytes_current": 40685968384,
33
+ "active_bytes_peak": 40723717120,
34
+ "allocated_bytes_current": 40685968384,
35
+ "allocated_bytes_peak": 40723717120,
36
+ "device": 7,
37
+ "inactive_split_bytes_current": 508388352,
38
+ "inactive_split_bytes_peak": 728589312,
39
+ "iteration": 0,
40
+ "mode": "iteration",
41
+ "num_alloc_retries": 0,
42
+ "num_ooms": 0,
43
+ "phase": "iteration_start",
44
+ "rank": 7,
45
+ "reserved_bytes_current": 43864031232,
46
+ "reserved_bytes_peak": 43864031232,
47
+ "segment_count_current": 261,
48
+ "segment_count_peak": 261,
49
+ "timestamp_utc": "2026-08-26T19:09:42.057144Z",
50
+ "world_size": 8
51
+ },
52
+ "device": 7,
53
+ "device_delta": 0,
54
+ "inactive_split_bytes_current": 2113312768,
55
+ "inactive_split_bytes_current_delta": 1604924416,
56
+ "inactive_split_bytes_peak": 4687297024,
57
+ "inactive_split_bytes_peak_delta": 3958707712,
58
+ "iteration": 0,
59
+ "iteration_delta": 0,
60
+ "mode": "iteration",
61
+ "num_alloc_retries": 0,
62
+ "num_alloc_retries_delta": 0,
63
+ "num_ooms": 0,
64
+ "num_ooms_delta": 0,
65
+ "rank": 7,
66
+ "rank_delta": 0,
67
+ "reserved_bytes_current": 83766542336,
68
+ "reserved_bytes_current_delta": 39902511104,
69
+ "reserved_bytes_peak": 83766542336,
70
+ "reserved_bytes_peak_delta": 39902511104,
71
+ "segment_count_current": 1015,
72
+ "segment_count_current_delta": 754,
73
+ "segment_count_peak": 1015,
74
+ "segment_count_peak_delta": 754,
75
+ "snapshot_pickle": "snapshot_iteration_rank0007_iter000000_20260826T191024263867Z.pickle",
76
+ "timestamp_utc": "2026-08-26T19:10:24.299582Z",
77
+ "world_size": 8,
78
+ "world_size_delta": 0
79
+ }
echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000000_20260826T191024263867Z.pickle ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d37f7c5b92a93fd6f2411cec191f4784809b327150444e5b5e6726ef64dd20f9
3
+ size 1778379
echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000001_20260826T191024742492Z.json ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "active_bytes_current": 72123770880,
3
+ "active_bytes_current_delta": 0,
4
+ "active_bytes_peak": 94548111360,
5
+ "active_bytes_peak_delta": 20933634048,
6
+ "after": {
7
+ "active_bytes_current": 72123770880,
8
+ "active_bytes_peak": 94548111360,
9
+ "allocated_bytes_current": 72123770880,
10
+ "allocated_bytes_peak": 94548111360,
11
+ "device": 7,
12
+ "inactive_split_bytes_current": 2406914048,
13
+ "inactive_split_bytes_peak": 4687297024,
14
+ "iteration": 1,
15
+ "mode": "iteration",
16
+ "num_alloc_retries": 0,
17
+ "num_ooms": 0,
18
+ "phase": "iteration_end",
19
+ "rank": 7,
20
+ "reserved_bytes_current": 97913929728,
21
+ "reserved_bytes_peak": 97913929728,
22
+ "segment_count_current": 1103,
23
+ "segment_count_peak": 1103,
24
+ "timestamp_utc": "2026-08-26T19:10:24.742069Z",
25
+ "world_size": 8
26
+ },
27
+ "allocated_bytes_current": 72123770880,
28
+ "allocated_bytes_current_delta": 0,
29
+ "allocated_bytes_peak": 94548111360,
30
+ "allocated_bytes_peak_delta": 20933634048,
31
+ "before": {
32
+ "active_bytes_current": 72123770880,
33
+ "active_bytes_peak": 73614477312,
34
+ "allocated_bytes_current": 72123770880,
35
+ "allocated_bytes_peak": 73614477312,
36
+ "device": 7,
37
+ "inactive_split_bytes_current": 2113312768,
38
+ "inactive_split_bytes_peak": 4687297024,
39
+ "iteration": 1,
40
+ "mode": "iteration",
41
+ "num_alloc_retries": 0,
42
+ "num_ooms": 0,
43
+ "phase": "iteration_start",
44
+ "rank": 7,
45
+ "reserved_bytes_current": 83766542336,
46
+ "reserved_bytes_peak": 83766542336,
47
+ "segment_count_current": 1015,
48
+ "segment_count_peak": 1015,
49
+ "timestamp_utc": "2026-08-26T19:10:24.303854Z",
50
+ "world_size": 8
51
+ },
52
+ "device": 7,
53
+ "device_delta": 0,
54
+ "inactive_split_bytes_current": 2406914048,
55
+ "inactive_split_bytes_current_delta": 293601280,
56
+ "inactive_split_bytes_peak": 4687297024,
57
+ "inactive_split_bytes_peak_delta": 0,
58
+ "iteration": 1,
59
+ "iteration_delta": 0,
60
+ "mode": "iteration",
61
+ "num_alloc_retries": 0,
62
+ "num_alloc_retries_delta": 0,
63
+ "num_ooms": 0,
64
+ "num_ooms_delta": 0,
65
+ "rank": 7,
66
+ "rank_delta": 0,
67
+ "reserved_bytes_current": 97913929728,
68
+ "reserved_bytes_current_delta": 14147387392,
69
+ "reserved_bytes_peak": 97913929728,
70
+ "reserved_bytes_peak_delta": 14147387392,
71
+ "segment_count_current": 1103,
72
+ "segment_count_current_delta": 88,
73
+ "segment_count_peak": 1103,
74
+ "segment_count_peak_delta": 88,
75
+ "snapshot_pickle": "snapshot_iteration_rank0007_iter000001_20260826T191024742492Z.pickle",
76
+ "timestamp_utc": "2026-08-26T19:10:24.774378Z",
77
+ "world_size": 8,
78
+ "world_size_delta": 0
79
+ }
echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000001_20260826T191024742492Z.pickle ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6c1875753af1433dcbfa15180da305cde01c4a818cfd0d7491c8b8e0b913549b
3
+ size 1688761
echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000002_20260826T191025066187Z.json ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "active_bytes_current": 72123770880,
3
+ "active_bytes_current_delta": 0,
4
+ "active_bytes_peak": 94548111360,
5
+ "active_bytes_peak_delta": 0,
6
+ "after": {
7
+ "active_bytes_current": 72123770880,
8
+ "active_bytes_peak": 94548111360,
9
+ "allocated_bytes_current": 72123770880,
10
+ "allocated_bytes_peak": 94548111360,
11
+ "device": 7,
12
+ "inactive_split_bytes_current": 2406914048,
13
+ "inactive_split_bytes_peak": 4687297024,
14
+ "iteration": 2,
15
+ "mode": "iteration",
16
+ "num_alloc_retries": 0,
17
+ "num_ooms": 0,
18
+ "phase": "iteration_end",
19
+ "rank": 7,
20
+ "reserved_bytes_current": 98738110464,
21
+ "reserved_bytes_peak": 98738110464,
22
+ "segment_count_current": 1104,
23
+ "segment_count_peak": 1104,
24
+ "timestamp_utc": "2026-08-26T19:10:25.065802Z",
25
+ "world_size": 8
26
+ },
27
+ "allocated_bytes_current": 72123770880,
28
+ "allocated_bytes_current_delta": 0,
29
+ "allocated_bytes_peak": 94548111360,
30
+ "allocated_bytes_peak_delta": 0,
31
+ "before": {
32
+ "active_bytes_current": 72123770880,
33
+ "active_bytes_peak": 94548111360,
34
+ "allocated_bytes_current": 72123770880,
35
+ "allocated_bytes_peak": 94548111360,
36
+ "device": 7,
37
+ "inactive_split_bytes_current": 2406914048,
38
+ "inactive_split_bytes_peak": 4687297024,
39
+ "iteration": 2,
40
+ "mode": "iteration",
41
+ "num_alloc_retries": 0,
42
+ "num_ooms": 0,
43
+ "phase": "iteration_start",
44
+ "rank": 7,
45
+ "reserved_bytes_current": 97913929728,
46
+ "reserved_bytes_peak": 97913929728,
47
+ "segment_count_current": 1103,
48
+ "segment_count_peak": 1103,
49
+ "timestamp_utc": "2026-08-26T19:10:24.777959Z",
50
+ "world_size": 8
51
+ },
52
+ "device": 7,
53
+ "device_delta": 0,
54
+ "inactive_split_bytes_current": 2406914048,
55
+ "inactive_split_bytes_current_delta": 0,
56
+ "inactive_split_bytes_peak": 4687297024,
57
+ "inactive_split_bytes_peak_delta": 0,
58
+ "iteration": 2,
59
+ "iteration_delta": 0,
60
+ "mode": "iteration",
61
+ "num_alloc_retries": 0,
62
+ "num_alloc_retries_delta": 0,
63
+ "num_ooms": 0,
64
+ "num_ooms_delta": 0,
65
+ "rank": 7,
66
+ "rank_delta": 0,
67
+ "reserved_bytes_current": 98738110464,
68
+ "reserved_bytes_current_delta": 824180736,
69
+ "reserved_bytes_peak": 98738110464,
70
+ "reserved_bytes_peak_delta": 824180736,
71
+ "segment_count_current": 1104,
72
+ "segment_count_current_delta": 1,
73
+ "segment_count_peak": 1104,
74
+ "segment_count_peak_delta": 1,
75
+ "snapshot_pickle": "snapshot_iteration_rank0007_iter000002_20260826T191025066187Z.pickle",
76
+ "timestamp_utc": "2026-08-26T19:10:25.094053Z",
77
+ "world_size": 8,
78
+ "world_size_delta": 0
79
+ }
echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000002_20260826T191025066187Z.pickle ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dfabf42ef90a4a1b62095e1172def894cb8994043c09a37ebf741244cd7c6b9b
3
+ size 1683136
echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000003_20260826T191025387991Z.json ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "active_bytes_current": 72123770880,
3
+ "active_bytes_current_delta": 0,
4
+ "active_bytes_peak": 94566120960,
5
+ "active_bytes_peak_delta": 18009600,
6
+ "after": {
7
+ "active_bytes_current": 72123770880,
8
+ "active_bytes_peak": 94566120960,
9
+ "allocated_bytes_current": 72123770880,
10
+ "allocated_bytes_peak": 94566120960,
11
+ "device": 7,
12
+ "inactive_split_bytes_current": 2406914048,
13
+ "inactive_split_bytes_peak": 5174400000,
14
+ "iteration": 3,
15
+ "mode": "iteration",
16
+ "num_alloc_retries": 0,
17
+ "num_ooms": 0,
18
+ "phase": "iteration_end",
19
+ "rank": 7,
20
+ "reserved_bytes_current": 99562291200,
21
+ "reserved_bytes_peak": 99562291200,
22
+ "segment_count_current": 1105,
23
+ "segment_count_peak": 1105,
24
+ "timestamp_utc": "2026-08-26T19:10:25.387592Z",
25
+ "world_size": 8
26
+ },
27
+ "allocated_bytes_current": 72123770880,
28
+ "allocated_bytes_current_delta": 0,
29
+ "allocated_bytes_peak": 94566120960,
30
+ "allocated_bytes_peak_delta": 18009600,
31
+ "before": {
32
+ "active_bytes_current": 72123770880,
33
+ "active_bytes_peak": 94548111360,
34
+ "allocated_bytes_current": 72123770880,
35
+ "allocated_bytes_peak": 94548111360,
36
+ "device": 7,
37
+ "inactive_split_bytes_current": 2406914048,
38
+ "inactive_split_bytes_peak": 4687297024,
39
+ "iteration": 3,
40
+ "mode": "iteration",
41
+ "num_alloc_retries": 0,
42
+ "num_ooms": 0,
43
+ "phase": "iteration_start",
44
+ "rank": 7,
45
+ "reserved_bytes_current": 98738110464,
46
+ "reserved_bytes_peak": 98738110464,
47
+ "segment_count_current": 1104,
48
+ "segment_count_peak": 1104,
49
+ "timestamp_utc": "2026-08-26T19:10:25.097332Z",
50
+ "world_size": 8
51
+ },
52
+ "device": 7,
53
+ "device_delta": 0,
54
+ "inactive_split_bytes_current": 2406914048,
55
+ "inactive_split_bytes_current_delta": 0,
56
+ "inactive_split_bytes_peak": 5174400000,
57
+ "inactive_split_bytes_peak_delta": 487102976,
58
+ "iteration": 3,
59
+ "iteration_delta": 0,
60
+ "mode": "iteration",
61
+ "num_alloc_retries": 0,
62
+ "num_alloc_retries_delta": 0,
63
+ "num_ooms": 0,
64
+ "num_ooms_delta": 0,
65
+ "rank": 7,
66
+ "rank_delta": 0,
67
+ "reserved_bytes_current": 99562291200,
68
+ "reserved_bytes_current_delta": 824180736,
69
+ "reserved_bytes_peak": 99562291200,
70
+ "reserved_bytes_peak_delta": 824180736,
71
+ "segment_count_current": 1105,
72
+ "segment_count_current_delta": 1,
73
+ "segment_count_peak": 1105,
74
+ "segment_count_peak_delta": 1,
75
+ "snapshot_pickle": "snapshot_iteration_rank0007_iter000003_20260826T191025387991Z.pickle",
76
+ "timestamp_utc": "2026-08-26T19:10:25.415386Z",
77
+ "world_size": 8,
78
+ "world_size_delta": 0
79
+ }
echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000003_20260826T191025387991Z.pickle ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e1049f2836c081c59fd668ef9a8004baf4c6fb6a93d6c39db72236459e662abe
3
+ size 1683234
echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000004_20260826T191025709649Z.json ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "active_bytes_current": 72123770880,
3
+ "active_bytes_current_delta": 0,
4
+ "active_bytes_peak": 94566120960,
5
+ "active_bytes_peak_delta": 0,
6
+ "after": {
7
+ "active_bytes_current": 72123770880,
8
+ "active_bytes_peak": 94566120960,
9
+ "allocated_bytes_current": 72123770880,
10
+ "allocated_bytes_peak": 94566120960,
11
+ "device": 7,
12
+ "inactive_split_bytes_current": 2406914048,
13
+ "inactive_split_bytes_peak": 5414726656,
14
+ "iteration": 4,
15
+ "mode": "iteration",
16
+ "num_alloc_retries": 0,
17
+ "num_ooms": 0,
18
+ "phase": "iteration_end",
19
+ "rank": 7,
20
+ "reserved_bytes_current": 99562291200,
21
+ "reserved_bytes_peak": 99562291200,
22
+ "segment_count_current": 1105,
23
+ "segment_count_peak": 1105,
24
+ "timestamp_utc": "2026-08-26T19:10:25.709261Z",
25
+ "world_size": 8
26
+ },
27
+ "allocated_bytes_current": 72123770880,
28
+ "allocated_bytes_current_delta": 0,
29
+ "allocated_bytes_peak": 94566120960,
30
+ "allocated_bytes_peak_delta": 0,
31
+ "before": {
32
+ "active_bytes_current": 72123770880,
33
+ "active_bytes_peak": 94566120960,
34
+ "allocated_bytes_current": 72123770880,
35
+ "allocated_bytes_peak": 94566120960,
36
+ "device": 7,
37
+ "inactive_split_bytes_current": 2406914048,
38
+ "inactive_split_bytes_peak": 5174400000,
39
+ "iteration": 4,
40
+ "mode": "iteration",
41
+ "num_alloc_retries": 0,
42
+ "num_ooms": 0,
43
+ "phase": "iteration_start",
44
+ "rank": 7,
45
+ "reserved_bytes_current": 99562291200,
46
+ "reserved_bytes_peak": 99562291200,
47
+ "segment_count_current": 1105,
48
+ "segment_count_peak": 1105,
49
+ "timestamp_utc": "2026-08-26T19:10:25.418641Z",
50
+ "world_size": 8
51
+ },
52
+ "device": 7,
53
+ "device_delta": 0,
54
+ "inactive_split_bytes_current": 2406914048,
55
+ "inactive_split_bytes_current_delta": 0,
56
+ "inactive_split_bytes_peak": 5414726656,
57
+ "inactive_split_bytes_peak_delta": 240326656,
58
+ "iteration": 4,
59
+ "iteration_delta": 0,
60
+ "mode": "iteration",
61
+ "num_alloc_retries": 0,
62
+ "num_alloc_retries_delta": 0,
63
+ "num_ooms": 0,
64
+ "num_ooms_delta": 0,
65
+ "rank": 7,
66
+ "rank_delta": 0,
67
+ "reserved_bytes_current": 99562291200,
68
+ "reserved_bytes_current_delta": 0,
69
+ "reserved_bytes_peak": 99562291200,
70
+ "reserved_bytes_peak_delta": 0,
71
+ "segment_count_current": 1105,
72
+ "segment_count_current_delta": 0,
73
+ "segment_count_peak": 1105,
74
+ "segment_count_peak_delta": 0,
75
+ "snapshot_pickle": "snapshot_iteration_rank0007_iter000004_20260826T191025709649Z.pickle",
76
+ "timestamp_utc": "2026-08-26T19:10:25.736708Z",
77
+ "world_size": 8,
78
+ "world_size_delta": 0
79
+ }
echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot_iteration_rank0007_iter000004_20260826T191025709649Z.pickle ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f9b5646886da31a34628a607ddb8221313a57f9b526bc6065bc314249176790b
3
+ size 1687743
echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/run.env ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ backend=echo
2
+ config=QWEN2_57B
3
+ family=tierB
4
+ repo=/workspace/Megatron-MoE-Echo
5
+ branch=better-memory-dump-support
6
+ commit=b7c52442
7
+ num_layers=10
8
+ hidden=3584
9
+ ffn_hidden=2560
10
+ moe_ffn_hidden=2560
11
+ num_experts=64
12
+ topk=8
13
+ experts_per_rank=8
14
+ expert_params=27525120
15
+ heads=28
16
+ kv_groups=4
17
+ kv_channels=
18
+ qk_layernorm=0
19
+ shared_ffn=20480
20
+ shared_expert_overlap=0
21
+ mla=0
22
+ dense_first_layers=0
23
+ moe_layers=10
24
+ rotary_base=1000000
25
+ rotary_scaling=
26
+ qkv_bias=1
27
+ router_balancing=aux_loss
28
+ aux_loss_coeff=1e-2
29
+ router_pre_softmax=1
30
+ router_fusion=0
31
+ router_dtype=fp32
32
+ device_initiated_grouped_gemm=0
33
+ norm_epsilon=1e-6
34
+ seq_length=4096
35
+ micro_batch_size=1
36
+ global_batch_size=8
37
+ num_microbatches=1
38
+ assignments_per_rank=32768
39
+ mean_tokens_per_expert=4096
40
+ slot_policy=matched
41
+ slots_per_rank=8
42
+ capacity_mode=dynamic
43
+ capacity=dynamic
44
+ cuda_graph=0
45
+ probe_mode=0
46
+ async_grad_reduce=0
47
+ branch_role=memory
48
+ memory_dump_mode=iteration
49
+ memory_snapshot_path=/workspace/tierb_study_20260826_144644/echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot.pickle
50
+ cuda_graph_warmup_steps=3
51
+ recompute=none
52
+ force_balance=0
53
+ train_iters=5
54
+ ep_size=8
55
+ echo_dispatch_overlap=1
56
+ echo_dispatch_recompute=0
57
+ ultraep_quota_min_tokens=
58
+ ultraep_weight_sync_join=
59
+ profile_enable=0
60
+ kernel_trace=0
61
+ fp8_recipe=mxfp8
echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/train.log ADDED
@@ -0,0 +1,1262 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ==============================================================
2
+ QWEN2_57B / echo (slots/rank=8 policy=matched)
3
+ E=64 K=8 H'=2560 m=8
4
+ L=10 seq=4096 gbs=8 (1 microbatches)
5
+ cap_mode=dynamic cap=dynamic cg=0 recompute=none
6
+ memory_dump=iteration snapshots=/workspace/tierb_study_20260826_144644/echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot.pickle
7
+ repo=/workspace/Megatron-MoE-Echo @ b7c52442
8
+ ==============================================================
9
+ W0826 19:09:17.777000 621555 torch/distributed/run.py:766]
10
+ W0826 19:09:17.777000 621555 torch/distributed/run.py:766] *****************************************
11
+ W0826 19:09:17.777000 621555 torch/distributed/run.py:766] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed.
12
+ W0826 19:09:17.777000 621555 torch/distributed/run.py:766] *****************************************
13
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
14
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
15
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
16
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
17
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
18
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
19
+ 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
20
+ WARNING: overriding default arguments for tokenizer_type:GPT2BPETokenizer with tokenizer_type:GPT2BPETokenizer
21
+ Number of virtual stages per pipeline stage: None
22
+ accumulate and all-reduce gradients in fp32 for bfloat16 data type.
23
+ using torch.bfloat16 for parameters ...
24
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
25
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
26
+ ------------------------ arguments ------------------------
27
+ account_for_embedding_in_pipeline_split ......... False
28
+ account_for_loss_in_pipeline_split .............. False
29
+ accumulate_allreduce_grads_in_fp32 .............. True
30
+ adam_beta1 ...................................... 0.9
31
+ adam_beta2 ...................................... 0.999
32
+ adam_eps ........................................ 1e-08
33
+ add_bias_linear ................................. False
34
+ add_position_embedding .......................... False
35
+ add_qkv_bias .................................... True
36
+ adlr_autoresume ................................. False
37
+ adlr_autoresume_interval ........................ 1000
38
+ align_grad_reduce ............................... True
39
+ align_param_gather .............................. False
40
+ app_tag_run_name ................................ None
41
+ app_tag_run_version ............................. 0.0.0
42
+ apply_layernorm_1p .............................. False
43
+ apply_query_key_layer_scaling ................... False
44
+ apply_residual_connection_post_layernorm ........ False
45
+ apply_rope_fusion ............................... True
46
+ async_save ...................................... None
47
+ async_tensor_model_parallel_allreduce ........... True
48
+ attention_backend ............................... AttnBackend.auto
49
+ attention_dropout ............................... 0.0
50
+ attention_softmax_in_fp32 ....................... False
51
+ auto_detect_ckpt_format ......................... False
52
+ barrier_with_L1_time ............................ True
53
+ bert_binary_head ................................ True
54
+ bert_embedder_type .............................. megatron
55
+ bert_load ....................................... None
56
+ bf16 ............................................ True
57
+ bias_dropout_fusion ............................. True
58
+ bias_gelu_fusion ................................ False
59
+ bias_swiglu_fusion .............................. True
60
+ biencoder_projection_dim ........................ 0
61
+ biencoder_shared_query_context_model ............ False
62
+ block_data_path ................................. None
63
+ calc_ft_timeouts ................................ False
64
+ calculate_per_token_loss ........................ False
65
+ check_for_large_grads ........................... False
66
+ check_for_nan_in_loss_and_grad .................. True
67
+ check_for_spiky_loss ............................ False
68
+ check_weight_hash_across_dp_replicas_interval ... None
69
+ ckpt_assume_constant_structure .................. False
70
+ ckpt_convert_format ............................. None
71
+ ckpt_convert_save ............................... None
72
+ ckpt_convert_update_legacy_dist_opt_format ...... False
73
+ ckpt_format ..................................... torch_dist
74
+ ckpt_fully_parallel_load ........................ False
75
+ ckpt_fully_parallel_save ........................ True
76
+ ckpt_fully_parallel_save_deprecated ............. False
77
+ ckpt_step ....................................... None
78
+ classes_fraction ................................ 1.0
79
+ clip_grad ....................................... 1.0
80
+ clone_scatter_output_in_embedding ............... True
81
+ config_logger_dir ...............................
82
+ consumed_train_samples .......................... 0
83
+ consumed_valid_samples .......................... 0
84
+ context_parallel_size ........................... 1
85
+ cp_comm_type .................................... ['p2p']
86
+ create_attention_mask_in_dataloader ............. True
87
+ cross_entropy_fusion_impl ....................... native
88
+ cross_entropy_loss_fusion ....................... False
89
+ cuda_graph_scope ................................ full
90
+ cuda_graph_warmup_steps ......................... 3
91
+ data_args_path .................................. None
92
+ data_cache_path ................................. None
93
+ data_parallel_random_init ....................... False
94
+ data_parallel_sharding_strategy ................. no_shard
95
+ data_parallel_size .............................. 8
96
+ data_path ....................................... ['/workspace/data/wikitext/wikitext103_gpt2_text_document']
97
+ data_per_class_fraction ......................... 1.0
98
+ data_sharding ................................... True
99
+ dataloader_type ................................. single
100
+ ddp_average_in_collective ....................... False
101
+ ddp_bucket_size ................................. None
102
+ ddp_num_buckets ................................. None
103
+ ddp_pad_buckets_for_high_nccl_busbw ............. False
104
+ decoder_first_pipeline_num_layers ............... None
105
+ decoder_last_pipeline_num_layers ................ None
106
+ decoder_num_layers .............................. None
107
+ decoder_seq_length .............................. None
108
+ decoupled_lr .................................... None
109
+ decoupled_min_lr ................................ None
110
+ decrease_batch_size_if_needed ................... False
111
+ defer_embedding_wgrad_compute ................... False
112
+ delay_wgrad_compute ............................. False
113
+ deprecated_use_mcore_models ..................... True
114
+ deterministic_mode .............................. False
115
+ dino_bottleneck_size ............................ 256
116
+ dino_freeze_last_layer .......................... 1
117
+ dino_head_hidden_size ........................... 2048
118
+ dino_local_crops_number ......................... 10
119
+ dino_local_img_size ............................. 96
120
+ dino_norm_last_layer ............................ False
121
+ dino_teacher_temp ............................... 0.07
122
+ dino_warmup_teacher_temp ........................ 0.04
123
+ dino_warmup_teacher_temp_epochs ................. 30
124
+ disable_bf16_reduced_precision_matmul ........... False
125
+ disable_mamba_mem_eff_path ...................... False
126
+ disable_straggler_on_startup .................... False
127
+ dist_ckpt_format_deprecated ..................... None
128
+ dist_ckpt_strictness ............................ assume_ok_unexpected
129
+ distribute_saved_activations .................... False
130
+ distributed_backend ............................. nccl
131
+ distributed_timeout_minutes ..................... 10
132
+ embedding_init_method_std ....................... None
133
+ embedding_path .................................. None
134
+ empty_unused_memory_level ....................... 0
135
+ enable_cuda_graph ............................... False
136
+ enable_experimental ............................. False
137
+ enable_ft_package ............................... False
138
+ enable_gloo_process_groups ...................... True
139
+ enable_msc ...................................... True
140
+ enable_one_logger ............................... True
141
+ encoder_num_layers .............................. 10
142
+ encoder_seq_length .............................. 4096
143
+ end_weight_decay ................................ 0.1
144
+ eod_mask_loss ................................... False
145
+ error_injection_rate ............................ 0
146
+ error_injection_type ............................ transient_error
147
+ eval_interval ................................... 1000000
148
+ eval_iters ...................................... 0
149
+ evidence_data_path .............................. None
150
+ exit_duration_in_mins ........................... None
151
+ exit_interval ................................... None
152
+ exit_on_missing_checkpoint ...................... False
153
+ exit_signal_handler ............................. False
154
+ exp_avg_dtype ................................... torch.float32
155
+ exp_avg_sq_dtype ................................ torch.float32
156
+ expert_model_parallel_size ...................... 8
157
+ expert_tensor_parallel_size ..................... 1
158
+ export_force_local_attention .................... False
159
+ export_kd_cfg ................................... None
160
+ export_kd_teacher_ckpt_format ................... None
161
+ export_kd_teacher_load .......................... None
162
+ export_kv_cache_quant ........................... False
163
+ export_legacy_megatron .......................... False
164
+ export_model_type ............................... GPTModel
165
+ export_moe_apply_probs_on_input ................. False
166
+ export_qk_l2_norm ............................... False
167
+ export_quant_cfg ................................ None
168
+ export_real_quant_cfg ........................... None
169
+ export_te_mcore_model ........................... False
170
+ external_cuda_graph ............................. False
171
+ ffn_hidden_size ................................. 2560
172
+ finetune ........................................ False
173
+ finetune_data_split ............................. train
174
+ finetune_hf_dataset ............................. None
175
+ first_last_layers_bf16 .......................... False
176
+ flash_decode .................................... False
177
+ fp16 ............................................ False
178
+ fp16_lm_cross_entropy ........................... False
179
+ fp32_residual_connection ........................ False
180
+ fp8 ............................................. e4m3
181
+ fp8_amax_compute_algo ........................... most_recent
182
+ fp8_amax_history_len ............................ 1
183
+ fp8_interval .................................... 1
184
+ fp8_margin ...................................... 0
185
+ fp8_param_gather ................................ False
186
+ fp8_recipe ...................................... mxfp8
187
+ fp8_wgrad ....................................... True
188
+ fsdp_double_buffer .............................. False
189
+ global_batch_size ............................... 8
190
+ grad_reduce_in_bf16 ............................. False
191
+ gradient_accumulation_fusion .................... True
192
+ gradient_reduce_div_fusion ...................... True
193
+ group_query_attention ........................... True
194
+ head_lr_mult .................................... 1.0
195
+ heterogeneous_layers_config_encoded_json ........ None
196
+ heterogeneous_layers_config_path ................ None
197
+ hidden_dropout .................................. 0.0
198
+ hidden_size ..................................... 3584
199
+ hierarchical_context_parallel_sizes ............. None
200
+ high_priority_stream_groups ..................... []
201
+ hybrid_attention_ratio .......................... 0.0
202
+ hybrid_mlp_ratio ................................ 0.0
203
+ hybrid_override_pattern ......................... None
204
+ hysteresis ...................................... 2
205
+ ict_head_size ................................... None
206
+ ict_load ........................................ None
207
+ img_h ........................................... 224
208
+ img_w ........................................... 224
209
+ indexer_batch_size .............................. 128
210
+ indexer_log_interval ............................ 1000
211
+ inference_batch_times_seqlen_threshold .......... -1
212
+ inference_dynamic_batching ...................... False
213
+ inference_dynamic_batching_buffer_guaranteed_fraction 0.2
214
+ inference_dynamic_batching_buffer_overflow_factor None
215
+ inference_dynamic_batching_buffer_size_gb ....... 40.0
216
+ inference_dynamic_batching_chunk_size ........... 256
217
+ inference_dynamic_batching_max_requests_override None
218
+ inference_dynamic_batching_max_tokens_override .. None
219
+ inference_dynamic_batching_num_cuda_graphs ...... 16
220
+ inference_max_batch_size ........................ 8
221
+ inference_max_seq_length ........................ 2560
222
+ inference_rng_tracker ........................... False
223
+ init_method_std ................................. 0.01
224
+ init_method_xavier_uniform ...................... False
225
+ init_model_with_meta_device ..................... False
226
+ initial_loss_scale .............................. 4294967296
227
+ inprocess_active_world_size ..................... 8
228
+ inprocess_barrier_timeout ....................... 120
229
+ inprocess_completion_timeout .................... 120
230
+ inprocess_empty_cuda_cache ...................... False
231
+ inprocess_granularity ........................... node
232
+ inprocess_hard_timeout .......................... 90
233
+ inprocess_heartbeat_interval .................... 30
234
+ inprocess_heartbeat_timeout ..................... 60
235
+ inprocess_last_call_wait ........................ 1
236
+ inprocess_max_iterations ........................ None
237
+ inprocess_monitor_process_interval .............. 1.0
238
+ inprocess_monitor_thread_interval ............... 1.0
239
+ inprocess_progress_watchdog_interval ............ 1.0
240
+ inprocess_restart ............................... False
241
+ inprocess_soft_timeout .......................... 60
242
+ inprocess_termination_grace_time ................ 1
243
+ is_hybrid_model ................................. False
244
+ iter_per_epoch .................................. 1250
245
+ iterations_to_skip .............................. []
246
+ keep_fp8_transpose_cache_when_using_custom_fsdp . False
247
+ kitchen_config_file ............................. None
248
+ kitchen_recipe_number ........................... None
249
+ kv_channels ..................................... 128
250
+ kv_lora_rank .................................... 32
251
+ lazy_mpu_init ................................... None
252
+ load ............................................ None
253
+ load_main_params_from_ckpt ...................... None
254
+ load_model_opt_format ........................... False
255
+ local_rank ...................................... 0
256
+ log_energy ...................................... False
257
+ log_interval .................................... 1
258
+ log_loss_scale_to_tensorboard ................... True
259
+ log_memory_to_tensorboard ....................... False
260
+ log_num_zeros_in_grad ........................... False
261
+ log_params_norm ................................. False
262
+ log_progress .................................... False
263
+ log_straggler ................................... False
264
+ log_throughput .................................. True
265
+ log_timers_to_tensorboard ....................... False
266
+ log_validation_ppl_to_tensorboard ............... False
267
+ log_world_size_to_tensorboard ................... False
268
+ logging_level ................................... None
269
+ loss_scale ...................................... None
270
+ loss_scale_window ............................... 1000
271
+ lr .............................................. 0.0001
272
+ lr_decay_iters .................................. 8
273
+ lr_decay_samples ................................ None
274
+ lr_decay_style .................................. cosine
275
+ lr_warmup_fraction .............................. None
276
+ lr_warmup_init .................................. 0.0
277
+ lr_warmup_iters ................................. 5
278
+ lr_warmup_samples ............................... 0
279
+ lr_wsd_decay_iters .............................. None
280
+ lr_wsd_decay_samples ............................ None
281
+ lr_wsd_decay_style .............................. exponential
282
+ main_grads_dtype ................................ torch.float32
283
+ main_params_dtype ............................... torch.float32
284
+ make_vocab_size_divisible_by .................... 128
285
+ mamba_head_dim .................................. 64
286
+ mamba_num_groups ................................ 8
287
+ mamba_num_heads ................................. None
288
+ mamba_state_dim ................................. 128
289
+ manual_gc ....................................... False
290
+ manual_gc_eval .................................. True
291
+ manual_gc_interval .............................. 0
292
+ mask_factor ..................................... 1.0
293
+ mask_prob ....................................... 0.15
294
+ mask_type ....................................... random
295
+ masked_softmax_fusion ........................... False
296
+ max_position_embeddings ......................... 32768
297
+ max_tokens_to_oom ............................... 12000
298
+ memory_snapshot_path ............................ /workspace/tierb_study_20260826_144644/echo/tierbmem/tierbmem_gbs8/QWEN2_57B__echo__matched__seq4096__gbs8__capdynamic__nocg__memiteration__it5/rep1/memory_snapshots/snapshot.pickle
299
+ merge_file ...................................... /workspace/data/wikitext/gpt2-merges.txt
300
+ micro_batch_size ................................ 1
301
+ microbatch_group_size_per_vp_stage .............. None
302
+ mid_level_dataset_surplus ....................... 0.005
303
+ min_loss_scale .................................. 1.0
304
+ min_lr .......................................... 1e-05
305
+ mlp_chunks_for_prefill .......................... 1
306
+ mmap_bin_files .................................. True
307
+ mock_data ....................................... False
308
+ moe_apply_probs_on_input ........................ False
309
+ moe_aux_loss_coeff .............................. 0.01
310
+ moe_deepep_num_sms .............................. 20
311
+ moe_echo_enable_random_offloading ............... False
312
+ moe_echo_expert_dispatch_overlap ................ True
313
+ moe_echo_expert_dispatcher_type ................. hybridep
314
+ moe_echo_recompute_expert_dispatch .............. False
315
+ moe_enable_deepep ............................... False
316
+ moe_enable_echo ................................. True
317
+ moe_enable_hybridep ............................. True
318
+ moe_expert_capacity_factor ...................... None
319
+ moe_extended_tp ................................. False
320
+ moe_ffn_hidden_size ............................. 2560
321
+ moe_grouped_gemm ................................ True
322
+ moe_hybridep_num_sms ............................ 16
323
+ moe_input_jitter_eps ............................ None
324
+ moe_layer_freq .................................. 1
325
+ moe_layer_recompute ............................. False
326
+ moe_num_echo_experts ............................ 64
327
+ moe_pad_expert_input_to_capacity ................ False
328
+ moe_per_layer_logging ........................... False
329
+ moe_permute_fusion .............................. False
330
+ moe_received_token_capacity ..................... None
331
+ moe_router_bias_update_rate ..................... 0.001
332
+ moe_router_dtype ................................ fp32
333
+ moe_router_enable_expert_bias ................... False
334
+ moe_router_force_load_balancing ................. False
335
+ moe_router_group_topk ........................... None
336
+ moe_router_load_balancing_type .................. aux_loss
337
+ moe_router_num_groups ........................... None
338
+ moe_router_padding_for_fp8 ...................... False
339
+ moe_router_pre_softmax .......................... True
340
+ moe_router_score_function ....................... softmax
341
+ moe_router_topk ................................. 8
342
+ moe_router_topk_scaling_factor .................. None
343
+ moe_shared_expert_intermediate_size ............. 20480
344
+ moe_shared_expert_overlap ....................... False
345
+ moe_token_dispatcher_type ....................... flex
346
+ moe_token_drop_policy ........................... probs
347
+ moe_upcycling_granularity ....................... 1
348
+ moe_use_device_initiated_grouped_gemm ........... False
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-06
373
+ normalization ................................... RMSNorm
374
+ num_attention_heads ............................. 28
375
+ num_channels .................................... 3
376
+ num_classes ..................................... 1000
377
+ num_dataset_builder_threads ..................... 1
378
+ num_distributed_optimizer_instances ............. 1
379
+ num_experts ..................................... 64
380
+ num_layers ...................................... 10
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 ................................ 4
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 ......................................... False
417
+ profile_ranks ................................... [0]
418
+ profile_step_end ................................ 12
419
+ profile_step_start .............................. 10
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_cuda_graph_memory_history ................ False
433
+ record_full_program_memory_history .............. False
434
+ record_iteration_memory_history ................. True
435
+ record_memory_history ........................... False
436
+ relative_attention_max_distance ................. 128
437
+ relative_attention_num_buckets .................. 32
438
+ replication ..................................... False
439
+ replication_factor .............................. 2
440
+ replication_jump ................................ None
441
+ rerun_mode ...................................... validate_results
442
+ reset_attention_mask ............................ False
443
+ reset_position_ids .............................. False
444
+ result_rejected_tracker_filename ................ None
445
+ retriever_report_topk_accuracies ................ []
446
+ retriever_score_scaling ......................... False
447
+ retriever_seq_length ............................ 256
448
+ retro_add_retriever ............................. False
449
+ retro_attention_gate ............................ 1
450
+ retro_cyclic_train_iters ........................ None
451
+ retro_encoder_attention_dropout ................. 0.1
452
+ retro_encoder_hidden_dropout .................... 0.1
453
+ retro_encoder_layers ............................ 2
454
+ retro_num_neighbors ............................. 2
455
+ retro_num_retrieved_chunks ...................... 2
456
+ retro_project_dir ............................... None
457
+ retro_verify_neighbor_count ..................... True
458
+ reuse_grad_buf_for_mxfp8_param_ag ............... False
459
+ rope_scaling_factor ............................. 8.0
460
+ rope_type ....................................... None
461
+ rotary_base ..................................... 1000000
462
+ rotary_interleaved .............................. False
463
+ rotary_percent .................................. 1.0
464
+ rotary_scaling_factor ........................... 1.0
465
+ rotary_seq_len_interpolation_factor ............. None
466
+ run_workload_inspector_server ................... False
467
+ sample_rate ..................................... 1.0
468
+ save ............................................ None
469
+ save_interval ................................... None
470
+ save_retain_interval ............................ None
471
+ scatter_gather_tensors_in_pipeline .............. True
472
+ seed ............................................ 1234
473
+ seq_length ...................................... 4096
474
+ sequence_parallel ............................... False
475
+ sft ............................................. False
476
+ sft_tokenizer_prompt_format ..................... nemotron-h-aligned
477
+ sgd_momentum .................................... 0.9
478
+ short_seq_prob .................................. 0.1
479
+ skip_train ...................................... False
480
+ skipped_train_samples ........................... 0
481
+ spec ............................................ None
482
+ split ........................................... 100,0,0
483
+ squared_relu .................................... False
484
+ start_weight_decay .............................. 0.1
485
+ straggler_ctrlr_port ............................ 65535
486
+ straggler_minmax_count .......................... 1
487
+ suggested_communication_unit_size ............... None
488
+ swiglu .......................................... True
489
+ swin_backbone_type .............................. tiny
490
+ symmetric_ar_type ............................... None
491
+ te_rng_tracker .................................. False
492
+ tensor_model_parallel_size ...................... 1
493
+ tensorboard_dir ................................. None
494
+ tensorboard_log_interval ........................ 1
495
+ tensorboard_queue_size .......................... 1000
496
+ test_data_path .................................. None
497
+ test_mode ....................................... False
498
+ tiktoken_num_special_tokens ..................... 1000
499
+ tiktoken_pattern ................................ None
500
+ tiktoken_special_tokens ......................... None
501
+ timing_log_level ................................ 0
502
+ timing_log_option ............................... minmax
503
+ titles_data_path ................................ None
504
+ tokenizer_model ................................. None
505
+ tokenizer_type .................................. GPT2BPETokenizer
506
+ torch_fsdp2_reshard_after_forward ............... True
507
+ tp_comm_bootstrap_backend ....................... nccl
508
+ tp_comm_bulk_dgrad .............................. True
509
+ tp_comm_bulk_wgrad .............................. True
510
+ tp_comm_overlap ................................. False
511
+ tp_comm_overlap_ag .............................. True
512
+ tp_comm_overlap_cfg ............................. None
513
+ tp_comm_overlap_rs .............................. True
514
+ tp_comm_overlap_rs_dgrad ........................ False
515
+ tp_comm_split_ag ................................ True
516
+ tp_comm_split_rs ................................ True
517
+ train_data_path ................................. None
518
+ train_iters ..................................... 5
519
+ train_samples ................................... None
520
+ train_sync_interval ............................. None
521
+ transformer_impl ................................ transformer_engine
522
+ transformer_pipeline_model_parallel_size ........ 1
523
+ untie_embeddings_and_output_weights ............. True
524
+ use_checkpoint_args ............................. False
525
+ use_checkpoint_opt_param_scheduler .............. False
526
+ use_cpu_initialization .......................... None
527
+ use_custom_fsdp ................................. False
528
+ use_dist_ckpt ................................... True
529
+ use_dist_ckpt_deprecated ........................ False
530
+ use_distributed_optimizer ....................... True
531
+ use_flash_attn .................................. False
532
+ use_legacy_models ............................... False
533
+ use_mp_args_from_checkpoint_args ................ False
534
+ use_one_sent_docs ............................... False
535
+ use_persistent_ckpt_worker ...................... False
536
+ use_precision_aware_optimizer ................... False
537
+ use_pytorch_profiler ............................ False
538
+ use_ring_exchange_p2p ........................... False
539
+ use_rope_scaling ................................ False
540
+ use_rotary_position_embeddings .................. False
541
+ use_sharp ....................................... False
542
+ use_tokenizer_model_from_checkpoint_args ........ True
543
+ use_torch_fsdp2 ................................. False
544
+ use_torch_optimizer_for_cpu_offload ............. False
545
+ use_tp_pp_dp_mapping ............................ False
546
+ v_head_dim ...................................... 128
547
+ valid_data_path ................................. None
548
+ variable_seq_lengths ............................ False
549
+ virtual_pipeline_model_parallel_size ............ None
550
+ vision_backbone_type ............................ vit
551
+ vision_pretraining .............................. False
552
+ vision_pretraining_type ......................... classify
553
+ vocab_extra_ids ................................. 0
554
+ vocab_file ...................................... /workspace/data/wikitext/gpt2-vocab.json
555
+ vocab_size ...................................... None
556
+ wandb_exp_name ..................................
557
+ wandb_project ...................................
558
+ wandb_save_dir ..................................
559
+ weight_decay .................................... 0.1
560
+ weight_decay_incr_style ......................... constant
561
+ wgrad_deferral_limit ............................ 0
562
+ world_size ...................................... 8
563
+ yaml_cfg ........................................ None
564
+ -------------------- end of arguments ---------------------
565
+ INFO:megatron.core.num_microbatches_calculator:setting number of microbatches to constant 1
566
+ > building GPT2BPETokenizer tokenizer ...
567
+ > padded vocab (size: 50257) with 47 dummy tokens (new size: 50304)
568
+ WARNING:megatron.core.rerun_state_machine:RerunStateMachine initialized in mode RerunMode.VALIDATE_RESULTS
569
+ > initializing torch distributed ...
570
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
571
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
572
+ 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
573
+ /workspace/Megatron-MoE-Echo/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-MoE-Echo/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-MoE-Echo/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
+ [Gloo] Rank [Gloo] Rank 7 is connected to 07 is connected to peer ranks. 7Expected number of connected peer ranks is : peer ranks. 7Expected number of connected peer ranks is :
580
+ 7
581
+ [Gloo] Rank 2 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
582
+ [Gloo] Rank 1 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
583
+ [Gloo] Rank 5 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
584
+ [Gloo] Rank 3 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
585
+ [Gloo] Rank [Gloo] Rank 64 is connected to is connected to 77 peer ranks. peer ranks. Expected number of connected peer ranks is : Expected number of connected peer ranks is : 77
586
+
587
+ [Gloo] Rank 0 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
588
+ [Gloo] Rank 1 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
589
+ [Gloo] Rank 2 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
590
+ [Gloo] Rank 4 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
591
+ [Gloo] Rank [Gloo] Rank [Gloo] Rank [Gloo] Rank 3657 is connected to is connected to is connected to is connected to 7777 peer ranks. peer ranks. peer ranks. peer ranks. Expected number of connected peer ranks is : Expected number of connected peer ranks is : Expected number of connected peer ranks is : Expected number of connected peer ranks is : 7777
592
+
593
+
594
+
595
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
596
+ > initialized tensor model parallel with size 1
597
+ > initialized pipeline model parallel with size 1
598
+ > setting random seeds to 1234 ...
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 [Gloo] Rank 00 is connected to is connected to 00 peer ranks. peer ranks. Expected number of connected peer ranks is : Expected number of connected peer ranks is : 00
603
+
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
+ [rank1]:[W826 19:09:35.838691010 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.
607
+ [rank5]:[W826 19:09:35.840925251 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.
608
+ [rank2]:[W826 19:09:35.840936489 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.
609
+ [rank3]:[W826 19:09:35.841124532 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.
610
+ [rank6]:[W826 19:09:35.841240830 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.
611
+ [rank7]:[W826 19:09:35.844868914 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.
612
+ [rank4]:[W826 19:09:35.847662248 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.
613
+ > compiling dataset index builder ...
614
+ make: Entering directory '/workspace/Megatron-MoE-Echo/megatron/core/datasets'
615
+ make: Nothing to be done for 'default'.
616
+ make: Leaving directory '/workspace/Megatron-MoE-Echo/megatron/core/datasets'
617
+ >>> done with dataset index builder. Compilation time: 0.035 seconds
618
+ WARNING: constraints for invoking optimized fused softmax kernel are not met. We default back to unfused kernel invocations.
619
+ > compiling and loading fused kernels ...
620
+ [rank0]:[W826 19:09:35.011935029 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.
621
+ >>> done with compiling and loading fused kernels. Compilation time: 3.652 seconds
622
+ time to initialize megatron (seconds): 6.972
623
+ [after megatron is initialized] datetime: 2026-08-26 19:09:41
624
+ building GPT model ...
625
+ /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.
626
+ warnings.warn(
627
+ /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.
628
+ warnings.warn(
629
+ /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.
630
+ warnings.warn(
631
+ /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.
632
+ warnings.warn(
633
+ /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.
634
+ warnings.warn(
635
+ /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.
636
+ warnings.warn(
637
+ /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.
638
+ warnings.warn(
639
+ /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.
640
+ warnings.warn(
641
+ > number of parameters on (tensor, pipeline) model parallel rank (0, 0): 5060614656
642
+ 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=True, 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)
643
+ INFO:megatron.core.distributed.param_and_grad_buffer:Number of buckets for gradient all-reduce / reduce-scatter: 22
644
+ Params for bucket 1 (180289536 elements, 180289536 padded size):
645
+ module.output_layer.weight
646
+ Params for bucket 2 (73403904 elements, 73403904 padded size):
647
+ module.decoder.final_layernorm.weight
648
+ module.decoder.layers.9.mlp.shared_experts.linear_fc2.weight
649
+ Params for bucket 3 (146800640 elements, 146800640 padded size):
650
+ module.decoder.layers.9.mlp.shared_experts.linear_fc1.weight
651
+ Params for bucket 4 (103001600 elements, 103001600 padded size):
652
+ module.decoder.layers.9.mlp.router.weight
653
+ module.decoder.layers.9.self_attention.linear_qkv.layer_norm_weight
654
+ module.decoder.layers.9.pre_mlp_layernorm.weight
655
+ module.decoder.layers.9.self_attention.linear_qkv.bias
656
+ module.decoder.layers.9.self_attention.linear_proj.weight
657
+ module.decoder.layers.9.self_attention.linear_qkv.weight
658
+ module.decoder.layers.8.mlp.shared_experts.linear_fc2.weight
659
+ Params for bucket 5 (146800640 elements, 146800640 padded size):
660
+ module.decoder.layers.8.mlp.shared_experts.linear_fc1.weight
661
+ Params for bucket 6 (103001600 elements, 103001600 padded size):
662
+ module.decoder.layers.8.self_attention.linear_qkv.bias
663
+ module.decoder.layers.8.self_attention.linear_qkv.layer_norm_weight
664
+ module.decoder.layers.8.pre_mlp_layernorm.weight
665
+ module.decoder.layers.8.self_attention.linear_qkv.weight
666
+ module.decoder.layers.8.self_attention.linear_proj.weight
667
+ module.decoder.layers.8.mlp.router.weight
668
+ module.decoder.layers.7.mlp.shared_experts.linear_fc2.weight
669
+ Params for bucket 7 (146800640 elements, 146800640 padded size):
670
+ module.decoder.layers.7.mlp.shared_experts.linear_fc1.weight
671
+ Params for bucket 8 (103001600 elements, 103001600 padded size):
672
+ module.decoder.layers.7.self_attention.linear_qkv.weight
673
+ module.decoder.layers.7.self_attention.linear_qkv.layer_norm_weight
674
+ module.decoder.layers.7.pre_mlp_layernorm.weight
675
+ module.decoder.layers.7.self_attention.linear_qkv.bias
676
+ module.decoder.layers.7.self_attention.linear_proj.weight
677
+ module.decoder.layers.7.mlp.router.weight
678
+ module.decoder.layers.6.mlp.shared_experts.linear_fc2.weight
679
+ Params for bucket 9 (146800640 elements, 146800640 padded size):
680
+ module.decoder.layers.6.mlp.shared_experts.linear_fc1.weight
681
+ Params for bucket 10 (103001600 elements, 103001600 padded size):
682
+ module.decoder.layers.6.self_attention.linear_qkv.weight
683
+ module.decoder.layers.5.mlp.shared_experts.linear_fc2.weight
684
+ module.decoder.layers.6.mlp.router.weight
685
+ module.decoder.layers.6.self_attention.linear_qkv.layer_norm_weight
686
+ module.decoder.layers.6.pre_mlp_layernorm.weight
687
+ module.decoder.layers.6.self_attention.linear_qkv.bias
688
+ module.decoder.layers.6.self_attention.linear_proj.weight
689
+ Params for bucket 11 (146800640 elements, 146800640 padded size):
690
+ module.decoder.layers.5.mlp.shared_experts.linear_fc1.weight
691
+ Params for bucket 12 (103001600 elements, 103001600 padded size):
692
+ module.decoder.layers.5.mlp.router.weight
693
+ module.decoder.layers.5.self_attention.linear_qkv.layer_norm_weight
694
+ module.decoder.layers.5.self_attention.linear_qkv.weight
695
+ module.decoder.layers.5.pre_mlp_layernorm.weight
696
+ module.decoder.layers.5.self_attention.linear_qkv.bias
697
+ module.decoder.layers.5.self_attention.linear_proj.weight
698
+ module.decoder.layers.4.mlp.shared_experts.linear_fc2.weight
699
+ Params for bucket 13 (146800640 elements, 146800640 padded size):
700
+ module.decoder.layers.4.mlp.shared_experts.linear_fc1.weight
701
+ Params for bucket 14 (103001600 elements, 103001600 padded size):
702
+ module.decoder.layers.4.mlp.router.weight
703
+ module.decoder.layers.4.self_attention.linear_qkv.layer_norm_weight
704
+ module.decoder.layers.4.self_attention.linear_qkv.bias
705
+ module.decoder.layers.4.pre_mlp_layernorm.weight
706
+ module.decoder.layers.4.self_attention.linear_qkv.weight
707
+ module.decoder.layers.4.self_attention.linear_proj.weight
708
+ module.decoder.layers.3.mlp.shared_experts.linear_fc2.weight
709
+ Params for bucket 15 (146800640 elements, 146800640 padded size):
710
+ module.decoder.layers.3.mlp.shared_experts.linear_fc1.weight
711
+ Params for bucket 16 (103001600 elements, 103001600 padded size):
712
+ module.decoder.layers.2.mlp.shared_experts.linear_fc2.weight
713
+ module.decoder.layers.3.mlp.router.weight
714
+ module.decoder.layers.3.self_attention.linear_qkv.layer_norm_weight
715
+ module.decoder.layers.3.pre_mlp_layernorm.weight
716
+ module.decoder.layers.3.self_attention.linear_qkv.bias
717
+ module.decoder.layers.3.self_attention.linear_qkv.weight
718
+ module.decoder.layers.3.self_attention.linear_proj.weight
719
+ Params for bucket 17 (146800640 elements, 146800640 padded size):
720
+ module.decoder.layers.2.mlp.shared_experts.linear_fc1.weight
721
+ Params for bucket 18 (103001600 elements, 103001600 padded size):
722
+ module.decoder.layers.1.mlp.shared_experts.linear_fc2.weight
723
+ module.decoder.layers.2.mlp.router.weight
724
+ module.decoder.layers.2.pre_mlp_layernorm.weight
725
+ module.decoder.layers.2.self_attention.linear_qkv.layer_norm_weight
726
+ module.decoder.layers.2.self_attention.linear_qkv.weight
727
+ module.decoder.layers.2.self_attention.linear_qkv.bias
728
+ module.decoder.layers.2.self_attention.linear_proj.weight
729
+ Params for bucket 19 (146800640 elements, 146800640 padded size):
730
+ module.decoder.layers.1.mlp.shared_experts.linear_fc1.weight
731
+ Params for bucket 20 (103001600 elements, 103001600 padded size):
732
+ module.decoder.layers.1.self_attention.linear_qkv.weight
733
+ module.decoder.layers.1.self_attention.linear_qkv.layer_norm_weight
734
+ module.decoder.layers.1.self_attention.linear_qkv.bias
735
+ module.decoder.layers.1.pre_mlp_layernorm.weight
736
+ module.decoder.layers.0.mlp.shared_experts.linear_fc2.weight
737
+ module.decoder.layers.1.mlp.router.weight
738
+ module.decoder.layers.1.self_attention.linear_proj.weight
739
+ Params for bucket 21 (146800640 elements, 146800640 padded size):
740
+ module.decoder.layers.0.mlp.shared_experts.linear_fc1.weight
741
+ Params for bucket 22 (209890816 elements, 209890816 padded size):
742
+ module.decoder.layers.0.self_attention.linear_qkv.layer_norm_weight
743
+ module.decoder.layers.0.pre_mlp_layernorm.weight
744
+ module.decoder.layers.0.self_attention.linear_qkv.bias
745
+ module.decoder.layers.0.self_attention.linear_proj.weight
746
+ module.embedding.word_embeddings.weight
747
+ module.decoder.layers.0.mlp.router.weight
748
+ module.decoder.layers.0.self_attention.linear_qkv.weight
749
+ INFO:megatron.core.distributed.param_and_grad_buffer:Number of buckets for gradient all-reduce / reduce-scatter: 44
750
+ Params for bucket 1 (45875200 elements, 45875200 padded size):
751
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight7
752
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight5
753
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight3
754
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight6
755
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight4
756
+ Params for bucket 2 (45875200 elements, 45875200 padded size):
757
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight2
758
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight1
759
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight7
760
+ module.decoder.layers.9.mlp.experts.linear_fc2.weight0
761
+ Params for bucket 3 (55050240 elements, 55050240 padded size):
762
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight6
763
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight5
764
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight4
765
+ Params for bucket 4 (55050240 elements, 55050240 padded size):
766
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight3
767
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight1
768
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight2
769
+ Params for bucket 5 (45875200 elements, 45875200 padded size):
770
+ module.decoder.layers.9.mlp.experts.linear_fc1.weight0
771
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight7
772
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight5
773
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight6
774
+ Params for bucket 6 (45875200 elements, 45875200 padded size):
775
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight0
776
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight3
777
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight4
778
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight2
779
+ module.decoder.layers.8.mlp.experts.linear_fc2.weight1
780
+ Params for bucket 7 (55050240 elements, 55050240 padded size):
781
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight7
782
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight6
783
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight5
784
+ Params for bucket 8 (55050240 elements, 55050240 padded size):
785
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight4
786
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight2
787
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight3
788
+ Params for bucket 9 (45875200 elements, 45875200 padded size):
789
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight1
790
+ module.decoder.layers.8.mlp.experts.linear_fc1.weight0
791
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight7
792
+ Params for bucket 10 (45875200 elements, 45875200 padded size):
793
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight6
794
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight4
795
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight5
796
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight2
797
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight3
798
+ Params for bucket 11 (55050240 elements, 55050240 padded size):
799
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight1
800
+ module.decoder.layers.7.mlp.experts.linear_fc2.weight0
801
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight7
802
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight6
803
+ Params for bucket 12 (55050240 elements, 55050240 padded size):
804
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight5
805
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight4
806
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight3
807
+ Params for bucket 13 (55050240 elements, 55050240 padded size):
808
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight2
809
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight0
810
+ module.decoder.layers.7.mlp.experts.linear_fc1.weight1
811
+ Params for bucket 14 (45875200 elements, 45875200 padded size):
812
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight6
813
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight4
814
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight7
815
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight3
816
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight5
817
+ Params for bucket 15 (45875200 elements, 45875200 padded size):
818
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight2
819
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight7
820
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight0
821
+ module.decoder.layers.6.mlp.experts.linear_fc2.weight1
822
+ Params for bucket 16 (55050240 elements, 55050240 padded size):
823
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight6
824
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight4
825
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight5
826
+ Params for bucket 17 (55050240 elements, 55050240 padded size):
827
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight3
828
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight1
829
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight2
830
+ Params for bucket 18 (45875200 elements, 45875200 padded size):
831
+ module.decoder.layers.6.mlp.experts.linear_fc1.weight0
832
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight7
833
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight6
834
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight5
835
+ Params for bucket 19 (45875200 elements, 45875200 padded size):
836
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight3
837
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight0
838
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight2
839
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight4
840
+ module.decoder.layers.5.mlp.experts.linear_fc2.weight1
841
+ Params for bucket 20 (55050240 elements, 55050240 padded size):
842
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight7
843
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight5
844
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight6
845
+ Params for bucket 21 (55050240 elements, 55050240 padded size):
846
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight4
847
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight2
848
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight3
849
+ Params for bucket 22 (45875200 elements, 45875200 padded size):
850
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight1
851
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight7
852
+ module.decoder.layers.5.mlp.experts.linear_fc1.weight0
853
+ Params for bucket 23 (45875200 elements, 45875200 padded size):
854
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight5
855
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight3
856
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight6
857
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight2
858
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight4
859
+ Params for bucket 24 (55050240 elements, 55050240 padded size):
860
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight1
861
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight6
862
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight7
863
+ module.decoder.layers.4.mlp.experts.linear_fc2.weight0
864
+ Params for bucket 25 (55050240 elements, 55050240 padded size):
865
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight5
866
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight4
867
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight3
868
+ Params for bucket 26 (55050240 elements, 55050240 padded size):
869
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight2
870
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight1
871
+ module.decoder.layers.4.mlp.experts.linear_fc1.weight0
872
+ Params for bucket 27 (45875200 elements, 45875200 padded size):
873
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight6
874
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight4
875
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight7
876
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight5
877
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight3
878
+ Params for bucket 28 (45875200 elements, 45875200 padded size):
879
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight2
880
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight1
881
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight7
882
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight0
883
+ Params for bucket 29 (55050240 elements, 55050240 padded size):
884
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight6
885
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight5
886
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight4
887
+ Params for bucket 30 (55050240 elements, 55050240 padded size):
888
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight3
889
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight1
890
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight2
891
+ Params for bucket 31 (45875200 elements, 45875200 padded size):
892
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight0
893
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight6
894
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight5
895
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight7
896
+ Params for bucket 32 (45875200 elements, 45875200 padded size):
897
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight3
898
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight0
899
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight4
900
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight1
901
+ module.decoder.layers.2.mlp.experts.linear_fc2.weight2
902
+ Params for bucket 33 (55050240 elements, 55050240 padded size):
903
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight7
904
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight6
905
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight5
906
+ Params for bucket 34 (55050240 elements, 55050240 padded size):
907
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight4
908
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight2
909
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight3
910
+ Params for bucket 35 (45875200 elements, 45875200 padded size):
911
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight1
912
+ module.decoder.layers.2.mlp.experts.linear_fc1.weight0
913
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight7
914
+ Params for bucket 36 (45875200 elements, 45875200 padded size):
915
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight2
916
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight3
917
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight5
918
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight6
919
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight4
920
+ Params for bucket 37 (55050240 elements, 55050240 padded size):
921
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight1
922
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight6
923
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight7
924
+ module.decoder.layers.1.mlp.experts.linear_fc2.weight0
925
+ Params for bucket 38 (55050240 elements, 55050240 padded size):
926
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight5
927
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight3
928
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight4
929
+ Params for bucket 39 (55050240 elements, 55050240 padded size):
930
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight2
931
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight1
932
+ module.decoder.layers.1.mlp.experts.linear_fc1.weight0
933
+ Params for bucket 40 (45875200 elements, 45875200 padded size):
934
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight7
935
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight5
936
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight4
937
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight3
938
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight6
939
+ Params for bucket 41 (45875200 elements, 45875200 padded size):
940
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight2
941
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight7
942
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight0
943
+ module.decoder.layers.0.mlp.experts.linear_fc2.weight1
944
+ Params for bucket 42 (55050240 elements, 55050240 padded size):
945
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight6
946
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight4
947
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight5
948
+ Params for bucket 43 (55050240 elements, 55050240 padded size):
949
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight3
950
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight2
951
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight1
952
+ Params for bucket 44 (18350080 elements, 18350080 padded size):
953
+ module.decoder.layers.0.mlp.experts.linear_fc1.weight0
954
+ 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 0x7fe14ce206e0>, config_logger_dir='')
955
+ INFO:megatron.core.optimizer_param_scheduler:> learning rate decay style: cosine
956
+ [after model, optimizer, and learning rate scheduler are built] datetime: 2026-08-26 19:09:41
957
+ > building train, validation, and test datasets ...
958
+ > datasets target sizes (minimum size):
959
+ train: 40
960
+ validation: 0
961
+ test: 0
962
+ INFO:megatron.core.datasets.blended_megatron_dataset_config:Let split_matrix = [(0, 1.0), None, None]
963
+ > building train, validation, and test datasets for GPT ...
964
+ INFO:megatron.core.datasets.blended_megatron_dataset_builder:Building GPTDataset splits with sizes=(40, 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 0x7fe14cfb2f60>, 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)
965
+ INFO:megatron.core.datasets.indexed_dataset:Load the _IndexReader from /workspace/data/wikitext/wikitext103_gpt2_text_document.idx
966
+ INFO:megatron.core.datasets.indexed_dataset: Extract the sequence lengths
967
+ INFO:megatron.core.datasets.indexed_dataset: Extract the sequence pointers
968
+ INFO:megatron.core.datasets.indexed_dataset: Extract the document indices
969
+ INFO:megatron.core.datasets.indexed_dataset:> total number of sequences: 1165029
970
+ INFO:megatron.core.datasets.indexed_dataset:> total number of documents: 1801350
971
+ INFO:megatron.core.datasets.gpt_dataset:Load the GPTDataset train indices
972
+ INFO:megatron.core.datasets.gpt_dataset: Load the document index from 0152a0378ab8ae1561633a6718ca974c-GPTDataset-train-document_index.npy
973
+ INFO:megatron.core.datasets.gpt_dataset: Load the sample index from 0152a0378ab8ae1561633a6718ca974c-GPTDataset-train-sample_index.npy
974
+ INFO:megatron.core.datasets.gpt_dataset: Load the shuffle index from 0152a0378ab8ae1561633a6718ca974c-GPTDataset-train-shuffle_index.npy
975
+ INFO:megatron.core.datasets.gpt_dataset:> total number of samples: 29073
976
+ > finished creating GPT datasets ...
977
+ [after dataloaders are built] datetime: 2026-08-26 19:09:41
978
+ done with setup ...
979
+ (min, max) time across ranks (ms):
980
+ model-and-optimizer-setup ......................: (346.13, 350.38)
981
+ train/valid/test-data-iterators-setup ..........: (61.74, 70.33)
982
+ training ...
983
+ Overwriting rerun_state_machine.current_iteration from -1 to 0...
984
+ [before the start of training step] datetime: 2026-08-26 19:09:42
985
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
986
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
987
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
988
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
989
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
990
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
991
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
992
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
993
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
994
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
995
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
996
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
997
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
998
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
999
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1000
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1001
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1002
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1003
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1004
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1005
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1006
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1007
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1008
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1009
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1010
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1011
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1012
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1013
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1014
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1015
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1016
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1017
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1018
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1019
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1020
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1021
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1022
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1023
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1024
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1025
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1026
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1027
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1028
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1029
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1030
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1031
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1032
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1033
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1034
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1035
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1036
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1037
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1038
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1039
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1040
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1041
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1042
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1043
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1044
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1045
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1046
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1047
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1048
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1049
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1050
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1051
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1052
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1053
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1054
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1055
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1056
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1057
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1058
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1059
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1060
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1061
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1062
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1063
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1064
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1065
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1066
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1067
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1068
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1069
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1070
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1071
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1072
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1073
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1074
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1075
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1076
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1077
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1078
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1079
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1080
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1081
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1082
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1083
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1084
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1085
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1086
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1087
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1088
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1089
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1090
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1091
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1092
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1093
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1094
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1095
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1096
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1097
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1098
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1099
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1100
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1101
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1102
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1103
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1104
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1105
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1106
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1107
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1108
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1109
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1110
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1111
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1112
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1113
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1114
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1115
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1116
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1117
+ WARNING:megatron.core.rerun_state_machine:Result validation enabled
1118
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1119
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1120
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1121
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1122
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1123
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1124
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1125
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1126
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1127
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1128
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1129
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1130
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1131
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1132
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1133
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1134
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1135
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1136
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1137
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1138
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1139
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1140
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1141
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1142
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1143
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1144
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1145
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1146
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1147
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1148
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1149
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1150
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1151
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1152
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1153
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1154
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1155
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1156
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1157
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1158
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1159
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1160
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1161
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1162
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1163
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1164
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1165
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1166
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1167
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1168
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1169
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1170
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1171
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1172
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1173
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1174
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1175
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1176
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1177
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1178
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1179
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1180
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1181
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1182
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1183
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1184
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1185
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1186
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1187
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1188
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1189
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1190
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1191
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1192
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1193
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1194
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1195
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1196
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
1197
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
1198
+ [2026-08-26 19:10:24] iteration 1/ 5 | consumed samples: 8 | elapsed time per iteration (ms): 42205.2 | throughput per GPU (TFLOP/s/GPU): 2.9 | learning rate: 2.000000E-05 | global batch size: 8 | lm loss: 1.097931E+01 | loss scale: 1.0 | grad norm: 29.877 | number of skipped iterations: 0 | number of nan iterations: 0 |Number of parameters in transformer block in billions: 20.11
1199
+
1200
+ Number of parameters in embedding layers in billions: 0.36
1201
+ Total number of parameters in billions: 20.47
1202
+ Number of parameters in most loaded shard in billions: 20.4724
1203
+ Theoretical memory footprints: weight and optimizer=146430.14 MB
1204
+ [Rank 0] (after 1 iterations) memory (MB) | allocated: 68783.849609375 | max allocated: 70218.63623046875 | reserved: 79584.0 | max reserved: 79584.0
1205
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1206
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1207
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1208
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1209
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1210
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1211
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1212
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1213
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1214
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1215
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1216
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1217
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1218
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1219
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
1220
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
1221
+ [2026-08-26 19:10:24] iteration 2/ 5 | consumed samples: 16 | elapsed time per iteration (ms): 479.5 | throughput per GPU (TFLOP/s/GPU): 257.5 | learning rate: 4.000000E-05 | global batch size: 8 | lm loss: 1.097906E+01 | loss scale: 1.0 | grad norm: 30.626 | number of skipped iterations: 0 | number of nan iterations: 0 |
1222
+ [2026-08-26 19:10:25] iteration 3/ 5 | consumed samples: 24 | elapsed time per iteration (ms): 323.7 | throughput per GPU (TFLOP/s/GPU): 381.5 | learning rate: 6.000000E-05 | global batch size: 8 | lm loss: 9.538996E+00 | loss scale: 1.0 | grad norm: 32.403 | number of skipped iterations: 0 | number of nan iterations: 0 |
1223
+ [2026-08-26 19:10:25] iteration 4/ 5 | consumed samples: 32 | elapsed time per iteration (ms): 321.7 | throughput per GPU (TFLOP/s/GPU): 383.9 | learning rate: 8.000000E-05 | global batch size: 8 | lm loss: 1.108690E+01 | loss scale: 1.0 | grad norm: 106.969 | number of skipped iterations: 0 | number of nan iterations: 0 |
1224
+ [2026-08-26 19:10:25] iteration 5/ 5 | consumed samples: 40 | elapsed time per iteration (ms): 321.6 | throughput per GPU (TFLOP/s/GPU): 384.0 | learning rate: 1.000000E-04 | global batch size: 8 | lm loss: 1.000203E+01 | loss scale: 1.0 | grad norm: 26.454 | number of skipped iterations: 0 | number of nan iterations: 0 |
1225
+ [after training is done] datetime: 2026-08-26 19:10:25
1226
+ W0826 19:10:26.462000 621555 torch/distributed/elastic/agent/server/api.py:719] Received 15 death signal, shutting down workers
1227
+ W0826 19:10:26.463000 621555 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 621648 closing signal SIGTERM
1228
+ W0826 19:10:26.463000 621555 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 621649 closing signal SIGTERM
1229
+ W0826 19:10:26.464000 621555 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 621650 closing signal SIGTERM
1230
+ W0826 19:10:26.467000 621555 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 621651 closing signal SIGTERM
1231
+ W0826 19:10:26.470000 621555 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 621652 closing signal SIGTERM
1232
+ W0826 19:10:26.472000 621555 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 621653 closing signal SIGTERM
1233
+ W0826 19:10:26.475000 621555 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 621654 closing signal SIGTERM
1234
+ W0826 19:10:26.477000 621555 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 621655 closing signal SIGTERM
1235
+ Traceback (most recent call last):
1236
+ File "/usr/local/bin/torchrun", line 8, in <module>
1237
+ sys.exit(main())
1238
+ ^^^^^^
1239
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 355, in wrapper
1240
+ return f(*args, **kwargs)
1241
+ ^^^^^^^^^^^^^^^^^^
1242
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/run.py", line 892, in main
1243
+ run(args)
1244
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/run.py", line 883, in run
1245
+ elastic_launch(
1246
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/launcher/api.py", line 139, in __call__
1247
+ return launch_agent(self._config, self._entrypoint, list(args))
1248
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1249
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/launcher/api.py", line 261, in launch_agent
1250
+ result = agent.run()
1251
+ ^^^^^^^^^^^
1252
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/metrics/api.py", line 138, in wrapper
1253
+ result = f(*args, **kwargs)
1254
+ ^^^^^^^^^^^^^^^^^^
1255
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/agent/server/api.py", line 711, in run
1256
+ result = self._invoke_run(role)
1257
+ ^^^^^^^^^^^^^^^^^^^^^^
1258
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/agent/server/api.py", line 870, in _invoke_run
1259
+ time.sleep(monitor_interval)
1260
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/multiprocessing/api.py", line 84, in _terminate_process_handler
1261
+ raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval)
1262
+ torch.distributed.elastic.multiprocessing.api.SignalException: Process 621555 got signal: 15