AbstractPhil commited on
Commit
1ab5886
·
verified ·
1 Parent(s): 46dc53c

Upload mini-beatrix-2s-control-attn-fp16-r10k/manifest.json with huggingface_hub

Browse files
mini-beatrix-2s-control-attn-fp16-r10k/manifest.json CHANGED
@@ -27,9 +27,9 @@
27
  "attn_kernel": "fp16"
28
  },
29
  "created_utc": "2026-10-10T03:52:07Z",
30
- "updated_utc": "2026-10-11T00:01:59Z",
31
- "steps": 32000,
32
- "tokens_seen": 8388608000,
33
  "wall_hours": 18.408296005129813,
34
  "phases": [
35
  {
@@ -71,14 +71,14 @@
71
  "name": "curriculum_s3",
72
  "dataset": "curriculum-s3",
73
  "planned_tokens": 1200000000,
74
- "status": "active",
75
- "tokens_done": 687865856
76
  },
77
  {
78
  "name": "curriculum_s4",
79
  "dataset": "curriculum-s4",
80
  "planned_tokens": 1200000000,
81
- "status": "planned",
82
  "tokens_done": 0
83
  },
84
  {
@@ -532,12 +532,28 @@
532
  "path": "resume/latest.pt",
533
  "val_bpb": null,
534
  "utc": "2026-10-11T00:01:30Z"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
535
  }
536
  ],
537
  "data_state": {
538
  "dataset": "curriculum-s3",
539
- "rows_consumed": 2314834,
540
- "epoch": 1
541
  },
542
  "notes": [
543
  {
@@ -583,6 +599,10 @@
583
  {
584
  "utc": "2026-10-10T22:23:29Z",
585
  "msg": "fp16 kernel this process: 2,730,260 calls, 0 backward retries (the gradient scale lowered after an overflow), 0 fp32 fallbacks (a forward overflow)"
 
 
 
 
586
  }
587
  ],
588
  "halt": null,
 
27
  "attn_kernel": "fp16"
28
  },
29
  "created_utc": "2026-10-10T03:52:07Z",
30
+ "updated_utc": "2026-10-11T02:30:41Z",
31
+ "steps": 33954,
32
+ "tokens_seen": 8900837376,
33
  "wall_hours": 18.408296005129813,
34
  "phases": [
35
  {
 
71
  "name": "curriculum_s3",
72
  "dataset": "curriculum-s3",
73
  "planned_tokens": 1200000000,
74
+ "status": "done",
75
+ "tokens_done": 1200095232
76
  },
77
  {
78
  "name": "curriculum_s4",
79
  "dataset": "curriculum-s4",
80
  "planned_tokens": 1200000000,
81
+ "status": "active",
82
  "tokens_done": 0
83
  },
84
  {
 
532
  "path": "resume/latest.pt",
533
  "val_bpb": null,
534
  "utc": "2026-10-11T00:01:30Z"
535
+ },
536
+ {
537
+ "step": 33954,
538
+ "tokens": 8900837376,
539
+ "kind": "safetensors",
540
+ "path": "checkpoints/step_00033954.safetensors",
541
+ "val_bpb": 1.1610089243602937,
542
+ "utc": "2026-10-11T02:30:06Z"
543
+ },
544
+ {
545
+ "step": 33954,
546
+ "tokens": 8900837376,
547
+ "kind": "resume",
548
+ "path": "resume/latest.pt",
549
+ "val_bpb": null,
550
+ "utc": "2026-10-11T02:30:06Z"
551
  }
552
  ],
553
  "data_state": {
554
  "dataset": "curriculum-s3",
555
+ "rows_consumed": 4032872,
556
+ "epoch": 2
557
  },
558
  "notes": [
559
  {
 
599
  {
600
  "utc": "2026-10-10T22:23:29Z",
601
  "msg": "fp16 kernel this process: 2,730,260 calls, 0 backward retries (the gradient scale lowered after an overflow), 0 fp32 fallbacks (a forward overflow)"
602
+ },
603
+ {
604
+ "utc": "2026-10-11T02:29:57Z",
605
+ "msg": "phase 'curriculum_s3' complete at 1,200,095,232 tokens"
606
  }
607
  ],
608
  "halt": null,