Frank-PVG commited on
Commit
5b3edd4
·
verified ·
1 Parent(s): 54492ad

Upload Schatten MoE checkpoints archive (part 2)

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +70 -0
  2. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/base/checkpoints/step-00015625/manifest.json +1 -0
  3. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/base/checkpoints/step-00015625/state.pt +3 -0
  4. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/launch.json +1 -0
  5. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/run_manifest.json +1 -0
  6. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/run_status.json +1 -0
  7. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/base-only-moe-final/COMMIT +1 -0
  8. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata +3 -0
  9. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp +3 -0
  10. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/base-only-moe-final/manifest.json +1 -0
  11. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/base-only-moe-final/state.pt +3 -0
  12. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000100/COMMIT +1 -0
  13. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000100/dcp/.metadata +3 -0
  14. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000100/dcp/__0_0.distcp +3 -0
  15. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000100/manifest.json +1 -0
  16. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000100/state.pt +3 -0
  17. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000200/COMMIT +1 -0
  18. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000200/dcp/.metadata +3 -0
  19. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000200/dcp/__0_0.distcp +3 -0
  20. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000200/manifest.json +1 -0
  21. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000200/state.pt +3 -0
  22. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/launch.json +1 -0
  23. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/run_manifest.json +1 -0
  24. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/run_status.json +1 -0
  25. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/base-only-moe-final/COMMIT +1 -0
  26. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata +3 -0
  27. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp +3 -0
  28. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/base-only-moe-final/manifest.json +1 -0
  29. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/base-only-moe-final/state.pt +3 -0
  30. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000100/COMMIT +1 -0
  31. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000100/dcp/.metadata +3 -0
  32. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000100/dcp/__0_0.distcp +3 -0
  33. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000100/manifest.json +1 -0
  34. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000100/state.pt +3 -0
  35. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000200/COMMIT +1 -0
  36. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000200/dcp/.metadata +3 -0
  37. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000200/dcp/__0_0.distcp +3 -0
  38. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000200/manifest.json +1 -0
  39. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000200/state.pt +3 -0
  40. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/launch.json +1 -0
  41. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/run_manifest.json +1 -0
  42. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/run_status.json +1 -0
  43. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/base-only-moe-final/COMMIT +1 -0
  44. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata +3 -0
  45. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp +3 -0
  46. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/base-only-moe-final/manifest.json +1 -0
  47. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/base-only-moe-final/state.pt +3 -0
  48. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/step-00000020/COMMIT +1 -0
  49. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/step-00000020/dcp/.metadata +3 -0
  50. artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/step-00000020/dcp/__0_0.distcp +3 -0
.gitattributes CHANGED
@@ -97,3 +97,73 @@ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta0
97
  artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/base/checkpoints/step-00007813/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
98
  artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/base/checkpoints/step-00015625/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
99
  artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/base/checkpoints/step-00015625/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
97
  artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/base/checkpoints/step-00007813/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
98
  artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/base/checkpoints/step-00015625/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
99
  artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/base/checkpoints/step-00015625/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
100
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
101
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
102
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000100/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
103
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000100/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
104
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000200/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
105
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000200/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
106
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
107
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
108
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000100/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
109
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000100/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
110
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000200/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
111
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000200/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
112
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
113
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
114
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/step-00000020/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
115
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/step-00000020/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
116
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized100-batch1600-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
117
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized100-batch1600-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
118
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized100-batch1600-20260807a/base/checkpoints/step-00000100/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
119
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized100-batch1600-20260807a/base/checkpoints/step-00000100/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
120
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
121
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
122
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-batch1000-20260807a/base/checkpoints/step-00000100/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
123
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-batch1000-20260807a/base/checkpoints/step-00000100/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
124
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-batch1000-20260807a/base/checkpoints/step-00000200/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
125
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-batch1000-20260807a/base/checkpoints/step-00000200/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
126
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-batch640-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
127
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-batch640-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
128
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-batch640-20260807a/base/checkpoints/step-00000100/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
129
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-batch640-20260807a/base/checkpoints/step-00000100/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
130
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-batch640-20260807a/base/checkpoints/step-00000200/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
131
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-batch640-20260807a/base/checkpoints/step-00000200/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
132
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-routerstats-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
133
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-routerstats-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
134
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-routerstats-20260807a/base/checkpoints/step-00000100/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
135
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-routerstats-20260807a/base/checkpoints/step-00000100/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
136
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-routerstats-20260807a/base/checkpoints/step-00000200/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
137
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-routerstats-20260807a/base/checkpoints/step-00000200/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
138
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-routerstats-20260807b/base/checkpoints/base-only-moe-final/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
139
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-routerstats-20260807b/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
140
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-routerstats-20260807b/base/checkpoints/step-00000100/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
141
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-routerstats-20260807b/base/checkpoints/step-00000100/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
142
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-routerstats-20260807b/base/checkpoints/step-00000200/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
143
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-optimized200-routerstats-20260807b/base/checkpoints/step-00000200/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
144
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-prefetch-nosync100-batch2000-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
145
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-prefetch-nosync100-batch2000-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
146
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-prefetch-nosync100-batch2000-20260807a/base/checkpoints/step-00000100/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
147
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-prefetch-nosync100-batch2000-20260807a/base/checkpoints/step-00000100/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
148
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-prefetch-nosync200-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
149
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-prefetch-nosync200-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
150
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-prefetch-nosync200-batch1000-20260807a/base/checkpoints/step-00000100/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
151
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-prefetch-nosync200-batch1000-20260807a/base/checkpoints/step-00000100/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
152
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-prefetch-nosync200-batch1000-20260807a/base/checkpoints/step-00000200/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
153
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-prefetch-nosync200-batch1000-20260807a/base/checkpoints/step-00000200/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
154
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-resumed200-exact/base/checkpoints/step-00000200/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
155
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-resumed200-exact/base/checkpoints/step-00000200/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
156
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/runs/dev-e15-base-continuous200/base/checkpoints/step-00000100/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
157
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/runs/dev-e15-base-continuous200/base/checkpoints/step-00000100/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
158
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/runs/dev-e15-base-continuous200/base/checkpoints/step-00000200/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
159
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/runs/dev-e15-base-continuous200/base/checkpoints/step-00000200/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
160
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/runs/dev-e15-base-resume100/base/checkpoints/step-00000100/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
161
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/runs/dev-e15-base-resume100/base/checkpoints/step-00000100/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
162
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/runs/dev-e15-base-resumed200-exact/base/checkpoints/step-00000200/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
163
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/runs/dev-e15-base-resumed200-exact/base/checkpoints/step-00000200/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
164
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/runs/dev-e15-base-throughput10/base/checkpoints/step-00000010/dcp/.metadata filter=lfs diff=lfs merge=lfs -text
165
+ artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/runs/dev-e15-base-throughput10/base/checkpoints/step-00000010/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
166
+ artifacts/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/continuation/e2-full-v2-attempt2/runs/E2-fullv2-base1701-split1001-twin_proxy-p046-3abed5e/proxy/checkpoints/shadow-bank-proxy/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
167
+ artifacts/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/continuation/e2-full-v2-attempt2/runs/E2-fullv2-base1701-split1001-twin_proxy-p061-3abed5e/proxy/checkpoints/shadow-bank-proxy/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
168
+ artifacts/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/continuation/e2-full-v2-attempt2/runs/E2-fullv2-base1701-split1001-twin_proxy-p063-3abed5e/proxy/checkpoints/shadow-bank-proxy/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
169
+ artifacts/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/continuation/e2-full-v2-attempt2/runs/E2-fullv2-base1701-split1001-twin_proxy-p071-3abed5e/proxy/checkpoints/shadow-bank-proxy/dcp/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/base/checkpoints/step-00015625/manifest.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"checkpoint_id":"step-00015625","component_names":["ema","engine","model","optimizer","sampler","scheduler"],"content_hash":"59f16f5aec7130bc28a410bb74669f878134e1297875b3edafab4caf7ff0e5a2","files":{"dcp/.metadata":"0bf44a6538d04e9ec2e002c93ef032f9da618c488b3eadc13718e8df245487a8","dcp/__0_0.distcp":"22d18487dc182949ab8270f354beda8ad7b299ddb76805e391b673c9c4b1d08f","state.pt":"45f49958fdb8800b45daff2653633db66deea3a468a66e19470124a9f96827c1"},"identity":{"checkpoint_layout":"dcp_with_local_payload_v1","config_hash":"6c1f1cbf03f8cae3d068caf1fad0f4246b5ed4774aefe80abfe6ddee9b2244b6","data_manifest_hash":"4ec206f8c63b542a5210ffa77d3cfa122379bf97b7e87437a2ff34912f59a7cc","edge_manifest_hash":"dc0cb6347196f976c34e2e0f5aacca353b9e55582fa5240b662f2f57a87ca167","mesh":{"data_parallel":1,"expert_mapping_version":"contiguous_global_expert_v1","expert_parallel":1,"world_size":1},"model_compatibility":{"payload":{"interface_structure":{"formula_version":"directed_cross_gate_v1","max_interfaces_per_token":1,"norm_epsilon":1e-05,"null_logit_mode":"per_layer_fp32_scalar_v1","pair_router_beta0":null,"pair_router_feature_norm_epsilon":1e-05,"pair_router_has_null":true,"pair_router_null_logit_init":2.0,"pair_router_probability_epsilon":1e-08,"pair_router_temperature":1.0,"pair_router_temperature_end":0.1,"pair_router_temperature_schedule":"linear","rank":16,"separable_rank":12,"slot_capacity":2},"model":{"depth":8,"expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_width":320,"hidden_size":320,"in_channels":4,"learn_sigma":false,"moe_layer_pattern":[0,1,2,3,4,5,6,7],"num_classes":100,"num_heads":5,"out_channels":4,"patch_size":4,"variant":"mini"},"router":{"aux_loss_weight":0.0,"correction_bias_clip":0.1,"correction_bias_eta":0.002,"correction_bias_update_interval":10,"correction_bias_warmup_steps":0,"gate_init_std":0.01,"group_score_top_n":2,"load_balancer":"correction_bias","num_experts":4,"num_groups":1,"route_scale":1.0,"shared_expert":true,"top_k":2,"topk_group":1},"runtime_layout":{"canonical_weight_alignment_version":"expert_swiglu_gate_up_down_v1","compute_dtype":"bfloat16","expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_mapping_version":"contiguous_global_expert_v1","precision":"bf16","profile_id":"h100_sm90_te_bf16_nv24_10_v1"}},"signature":"25a79bd702c41ddc230e9da49950848570ded039885929998395fc7a2838924d","version":"model_compatibility_signature_v1"},"parent_checkpoint_hash":"f1e793ec41c52c7dba5f05d83a20fddabe98738f713c4301942f589c11395889","role_spec_hash":"23e60b4260edafc09b8808cf330d989a08de08579276dda5dc7db777b7c5c315","schema_version":"1.0.0","shadow_commit_hash":null,"split_manifest_hash":"f2d99809dd21966190a823e684aaadda994281f5ccc2ba348db027fa6b43d716"},"train_state":{"image_exposures":4000000,"latent_tokens":4096000000,"microstep":0,"optimizer_step":15625,"phase":"base","phase_step":15625,"sampler_cursor":32000},"world_size":1}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/base/checkpoints/step-00015625/state.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:45f49958fdb8800b45daff2653633db66deea3a468a66e19470124a9f96827c1
3
+ size 168490858
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/launch.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"adapter":"__main__:base_stage_service","argv":["--run-id","dev-e15-base-formal4m-fullbatch","--runs-root","/workspace/gsh/tmp/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs","--config-dir","/workspace/gsh/tmp/Schatten_MoE-continuation/configs","--config-name","e15","--input","base_train_plan=/workspace/gsh/tmp/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/plans/formal_4m","--input","data_manifest=/workspace/gsh/tmp/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/data/preprocess/data/manifest.json","--input","split_manifest=/workspace/gsh/tmp/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/data/splits/manifest.json"],"input_hashes":{"base_train_plan":"ce3557a158b8073ad07e8e1a00c61779ce7e237379eb2b3ad35132c8c2b90983","data_manifest":"4ec206f8c63b542a5210ffa77d3cfa122379bf97b7e87437a2ff34912f59a7cc","split_manifest":"f2d99809dd21966190a823e684aaadda994281f5ccc2ba348db027fa6b43d716"},"inputs":{"base_train_plan":"/workspace/gsh/tmp/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/plans/formal_4m","data_manifest":"/workspace/gsh/tmp/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/data/preprocess/data/manifest.json","split_manifest":"/workspace/gsh/tmp/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/data/splits/manifest.json"},"overrides":[],"stage":"train_base"}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/run_manifest.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"artifacts":[{"hash":"90da573af3c8db0369df582cd32bf72f265c7ef2c0b7ab170f8f6ba96b14a05a","name":"base_metrics","parent_hash":null,"path":"base/base_metrics.json","schema_version":"1.0.0"},{"hash":"8419e64309ae8865369daebb310189cca9b4332683af3391cf730d4748e8bfd3","name":"checkpoint","parent_hash":null,"path":"base/checkpoints/step-00015625/manifest.json","schema_version":"1.0.0"},{"hash":"6ed75f274938edd114e198f05996706c4bc2ae2c6037b6dc4567a3df96ab207a","name":"launch","parent_hash":null,"path":"launch.json","schema_version":"1.0.0"}],"code":{"contract_version":"contracts_v1","dirty":true,"dirty_patch_hash":"6cbca79fd0c465da4aaaff4c6a22cf82b092b9ed44b4094ce2861bc586a387e6","git_commit":"3abed5e6d10894ea1551b5d2b97afa194db72ff3"},"config_hash":"6c1f1cbf03f8cae3d068caf1fad0f4246b5ed4774aefe80abfe6ddee9b2244b6","content_hash":"28556d43689be4e5ab9c667252eb69df1da4ece085b569726ace84f38a9725b9","created_at":"2026-08-07T05:11:02.355586+00:00","data_manifest_hash":"4ec206f8c63b542a5210ffa77d3cfa122379bf97b7e87437a2ff34912f59a7cc","edge_manifest_hash":null,"environment":{"cuda":"12.6","gpus":[{"compute_capability":"9.0","driver":"590.48.01","memory_bytes":85019328512,"name":"NVIDIA H100 80GB HBM3"}],"nccl":"2.22.3","python":"3.10.12","pytorch":"2.5.0a0+e000cf0ad9.nv24.10"},"flops":{"estimated_active":0.0,"measured":null,"method_version":"not_measured_preflight_v1"},"parameters":{"active":0,"logical":0},"parent_checkpoint_hash":null,"run_id":"dev-e15-base-formal4m-fullbatch","runtime":{"canonical_weight_alignment_version":"expert_swiglu_gate_up_down_v1","compute_dtype":"bfloat16","cuda_version":"12.6","driver_version":"590.48.01","expert_backend":"transformer_engine_grouped_linear_bf16_v1","profile_hash":"7930764e0384323c317c49db511e455f3392b9e7b73cac7be06927b515007802","profile_id":"h100_sm90_te_bf16_nv24_10_v1","python_version":"3.10.12","pytorch_build":"2.5.0a0+e000cf0ad9.nv24.10","sm":"9.0","transformer_engine_version":"1.11.0+4df8488"},"schema_version":"2.0.0","seeds":{"arm_scheduling":1731,"augmentation":1706,"candidate_audit":1741,"capture_sampling":1711,"cfg":1705,"data_order":1701,"diffusion":1703,"generation":1751,"global_seed":1701,"posterior":1702,"shadow_init":1721,"timestep":1704},"split_manifest_hash":"f2d99809dd21966190a823e684aaadda994281f5ccc2ba348db027fa6b43d716","stage":"train_base","vae_hash":null,"world_mesh":{"data_parallel":1,"expert_mapping_version":"rank_major_expert_contiguous_v1","expert_parallel":1,"layout_hash":"bbb880479a3d8f7e12e020ecdcb9c55f1ae88d077cbc01b940f1e9c5277207db","world_size":1}}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-formal4m-fullbatch/run_status.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"manifest_hash":"28556d43689be4e5ab9c667252eb69df1da4ece085b569726ace84f38a9725b9","reason":"BASE_ROUTER_GATE_FAILED","status":"FAILED_SCIENCE","updated_at":"2026-08-07T05:26:36.968461+00:00"}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/base-only-moe-final/COMMIT ADDED
@@ -0,0 +1 @@
 
 
1
+ b0c969f89539538ca8e26843963d582e9da1522203a8647fcf46a77d29497a67
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e5fcdd4cfae52b672c2b60e805680be4877a3c3594efc5128c72c0b0076c4013
3
+ size 296504
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9a50107787b2a4d4da6452e8baf293fce5fc015cb3135b36363dc5e7310a3086
3
+ size 169113792
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/base-only-moe-final/manifest.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"checkpoint_id":"base-only-moe-final","component_names":["ema","engine","model","optimizer","sampler","scheduler"],"content_hash":"b0c969f89539538ca8e26843963d582e9da1522203a8647fcf46a77d29497a67","files":{"dcp/.metadata":"e5fcdd4cfae52b672c2b60e805680be4877a3c3594efc5128c72c0b0076c4013","dcp/__0_0.distcp":"9a50107787b2a4d4da6452e8baf293fce5fc015cb3135b36363dc5e7310a3086","state.pt":"406384e5c65b176dff4d8ca0c218795d268dc749b66ab93132f1d5269385c368"},"identity":{"checkpoint_layout":"dcp_with_local_payload_v1","config_hash":"6c1f1cbf03f8cae3d068caf1fad0f4246b5ed4774aefe80abfe6ddee9b2244b6","data_manifest_hash":"4ec206f8c63b542a5210ffa77d3cfa122379bf97b7e87437a2ff34912f59a7cc","edge_manifest_hash":"dc0cb6347196f976c34e2e0f5aacca353b9e55582fa5240b662f2f57a87ca167","mesh":{"data_parallel":1,"expert_mapping_version":"contiguous_global_expert_v1","expert_parallel":1,"world_size":1},"model_compatibility":{"payload":{"interface_structure":{"formula_version":"directed_cross_gate_v1","max_interfaces_per_token":1,"norm_epsilon":1e-05,"null_logit_mode":"per_layer_fp32_scalar_v1","pair_router_beta0":null,"pair_router_feature_norm_epsilon":1e-05,"pair_router_has_null":true,"pair_router_null_logit_init":2.0,"pair_router_probability_epsilon":1e-08,"pair_router_temperature":1.0,"pair_router_temperature_end":0.1,"pair_router_temperature_schedule":"linear","rank":16,"separable_rank":12,"slot_capacity":2},"model":{"depth":8,"expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_width":320,"hidden_size":320,"in_channels":4,"learn_sigma":false,"moe_layer_pattern":[0,1,2,3,4,5,6,7],"num_classes":100,"num_heads":5,"out_channels":4,"patch_size":4,"variant":"mini"},"router":{"aux_loss_weight":0.0,"correction_bias_clip":0.1,"correction_bias_eta":0.002,"correction_bias_update_interval":10,"correction_bias_warmup_steps":0,"gate_init_std":0.01,"group_score_top_n":2,"load_balancer":"correction_bias","num_experts":4,"num_groups":1,"route_scale":1.0,"shared_expert":true,"top_k":2,"topk_group":1},"runtime_layout":{"canonical_weight_alignment_version":"expert_swiglu_gate_up_down_v1","compute_dtype":"bfloat16","expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_mapping_version":"contiguous_global_expert_v1","precision":"bf16","profile_id":"h100_sm90_te_bf16_nv24_10_v1"}},"signature":"25a79bd702c41ddc230e9da49950848570ded039885929998395fc7a2838924d","version":"model_compatibility_signature_v1"},"parent_checkpoint_hash":"e34b4caddd1e30ecd149c7a9d72d4c2fe57e6599ec43685535be06b5e8b258cc","role_spec_hash":"23e60b4260edafc09b8808cf330d989a08de08579276dda5dc7db777b7c5c315","schema_version":"1.0.0","shadow_commit_hash":null,"split_manifest_hash":"f2d99809dd21966190a823e684aaadda994281f5ccc2ba348db027fa6b43d716"},"train_state":{"image_exposures":200000,"latent_tokens":204800000,"microstep":0,"optimizer_step":200,"phase":"base","phase_step":200,"sampler_cursor":72000},"world_size":1}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/base-only-moe-final/state.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:406384e5c65b176dff4d8ca0c218795d268dc749b66ab93132f1d5269385c368
3
+ size 168498858
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000100/COMMIT ADDED
@@ -0,0 +1 @@
 
 
1
+ 656a7d0e2de8d89c98321dc46847b5430c83dc03e930c4d762da3ce91d6f7a66
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000100/dcp/.metadata ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:263c4d49dd483a7bac59899d9307c3c11d2be31cac9326754c0c2c0f57f99078
3
+ size 296498
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000100/dcp/__0_0.distcp ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:263770b4a881c9dc4f4e94af16ebebe75304de722a36718ffd29744db9cb276f
3
+ size 169113792
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000100/manifest.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"checkpoint_id":"step-00000100","component_names":["ema","engine","model","optimizer","sampler","scheduler"],"content_hash":"656a7d0e2de8d89c98321dc46847b5430c83dc03e930c4d762da3ce91d6f7a66","files":{"dcp/.metadata":"263c4d49dd483a7bac59899d9307c3c11d2be31cac9326754c0c2c0f57f99078","dcp/__0_0.distcp":"263770b4a881c9dc4f4e94af16ebebe75304de722a36718ffd29744db9cb276f","state.pt":"22e5baafe2fdd4a146f4852faa1a37cc8698eab2cc721b87c9fd6c722b97720b"},"identity":{"checkpoint_layout":"dcp_with_local_payload_v1","config_hash":"6c1f1cbf03f8cae3d068caf1fad0f4246b5ed4774aefe80abfe6ddee9b2244b6","data_manifest_hash":"4ec206f8c63b542a5210ffa77d3cfa122379bf97b7e87437a2ff34912f59a7cc","edge_manifest_hash":"dc0cb6347196f976c34e2e0f5aacca353b9e55582fa5240b662f2f57a87ca167","mesh":{"data_parallel":1,"expert_mapping_version":"contiguous_global_expert_v1","expert_parallel":1,"world_size":1},"model_compatibility":{"payload":{"interface_structure":{"formula_version":"directed_cross_gate_v1","max_interfaces_per_token":1,"norm_epsilon":1e-05,"null_logit_mode":"per_layer_fp32_scalar_v1","pair_router_beta0":null,"pair_router_feature_norm_epsilon":1e-05,"pair_router_has_null":true,"pair_router_null_logit_init":2.0,"pair_router_probability_epsilon":1e-08,"pair_router_temperature":1.0,"pair_router_temperature_end":0.1,"pair_router_temperature_schedule":"linear","rank":16,"separable_rank":12,"slot_capacity":2},"model":{"depth":8,"expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_width":320,"hidden_size":320,"in_channels":4,"learn_sigma":false,"moe_layer_pattern":[0,1,2,3,4,5,6,7],"num_classes":100,"num_heads":5,"out_channels":4,"patch_size":4,"variant":"mini"},"router":{"aux_loss_weight":0.0,"correction_bias_clip":0.1,"correction_bias_eta":0.002,"correction_bias_update_interval":10,"correction_bias_warmup_steps":0,"gate_init_std":0.01,"group_score_top_n":2,"load_balancer":"correction_bias","num_experts":4,"num_groups":1,"route_scale":1.0,"shared_expert":true,"top_k":2,"topk_group":1},"runtime_layout":{"canonical_weight_alignment_version":"expert_swiglu_gate_up_down_v1","compute_dtype":"bfloat16","expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_mapping_version":"contiguous_global_expert_v1","precision":"bf16","profile_id":"h100_sm90_te_bf16_nv24_10_v1"}},"signature":"25a79bd702c41ddc230e9da49950848570ded039885929998395fc7a2838924d","version":"model_compatibility_signature_v1"},"parent_checkpoint_hash":null,"role_spec_hash":"23e60b4260edafc09b8808cf330d989a08de08579276dda5dc7db777b7c5c315","schema_version":"1.0.0","shadow_commit_hash":null,"split_manifest_hash":"f2d99809dd21966190a823e684aaadda994281f5ccc2ba348db027fa6b43d716"},"train_state":{"image_exposures":100000,"latent_tokens":102400000,"microstep":0,"optimizer_step":100,"phase":"base","phase_step":100,"sampler_cursor":100000},"world_size":1}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000100/state.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:22e5baafe2fdd4a146f4852faa1a37cc8698eab2cc721b87c9fd6c722b97720b
3
+ size 168491178
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000200/COMMIT ADDED
@@ -0,0 +1 @@
 
 
1
+ e34b4caddd1e30ecd149c7a9d72d4c2fe57e6599ec43685535be06b5e8b258cc
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000200/dcp/.metadata ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e6a4691a378d48cb917bb1ff213434660d49a88d6d37309178f8dd450d8835da
3
+ size 296498
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000200/dcp/__0_0.distcp ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9a50107787b2a4d4da6452e8baf293fce5fc015cb3135b36363dc5e7310a3086
3
+ size 169113792
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000200/manifest.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"checkpoint_id":"step-00000200","component_names":["ema","engine","model","optimizer","sampler","scheduler"],"content_hash":"e34b4caddd1e30ecd149c7a9d72d4c2fe57e6599ec43685535be06b5e8b258cc","files":{"dcp/.metadata":"e6a4691a378d48cb917bb1ff213434660d49a88d6d37309178f8dd450d8835da","dcp/__0_0.distcp":"9a50107787b2a4d4da6452e8baf293fce5fc015cb3135b36363dc5e7310a3086","state.pt":"5bc66fa76c2e48dff8bf67ccc5d714d3a866d3c97bf79609825d7bcbfe10f960"},"identity":{"checkpoint_layout":"dcp_with_local_payload_v1","config_hash":"6c1f1cbf03f8cae3d068caf1fad0f4246b5ed4774aefe80abfe6ddee9b2244b6","data_manifest_hash":"4ec206f8c63b542a5210ffa77d3cfa122379bf97b7e87437a2ff34912f59a7cc","edge_manifest_hash":"dc0cb6347196f976c34e2e0f5aacca353b9e55582fa5240b662f2f57a87ca167","mesh":{"data_parallel":1,"expert_mapping_version":"contiguous_global_expert_v1","expert_parallel":1,"world_size":1},"model_compatibility":{"payload":{"interface_structure":{"formula_version":"directed_cross_gate_v1","max_interfaces_per_token":1,"norm_epsilon":1e-05,"null_logit_mode":"per_layer_fp32_scalar_v1","pair_router_beta0":null,"pair_router_feature_norm_epsilon":1e-05,"pair_router_has_null":true,"pair_router_null_logit_init":2.0,"pair_router_probability_epsilon":1e-08,"pair_router_temperature":1.0,"pair_router_temperature_end":0.1,"pair_router_temperature_schedule":"linear","rank":16,"separable_rank":12,"slot_capacity":2},"model":{"depth":8,"expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_width":320,"hidden_size":320,"in_channels":4,"learn_sigma":false,"moe_layer_pattern":[0,1,2,3,4,5,6,7],"num_classes":100,"num_heads":5,"out_channels":4,"patch_size":4,"variant":"mini"},"router":{"aux_loss_weight":0.0,"correction_bias_clip":0.1,"correction_bias_eta":0.002,"correction_bias_update_interval":10,"correction_bias_warmup_steps":0,"gate_init_std":0.01,"group_score_top_n":2,"load_balancer":"correction_bias","num_experts":4,"num_groups":1,"route_scale":1.0,"shared_expert":true,"top_k":2,"topk_group":1},"runtime_layout":{"canonical_weight_alignment_version":"expert_swiglu_gate_up_down_v1","compute_dtype":"bfloat16","expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_mapping_version":"contiguous_global_expert_v1","precision":"bf16","profile_id":"h100_sm90_te_bf16_nv24_10_v1"}},"signature":"25a79bd702c41ddc230e9da49950848570ded039885929998395fc7a2838924d","version":"model_compatibility_signature_v1"},"parent_checkpoint_hash":"656a7d0e2de8d89c98321dc46847b5430c83dc03e930c4d762da3ce91d6f7a66","role_spec_hash":"23e60b4260edafc09b8808cf330d989a08de08579276dda5dc7db777b7c5c315","schema_version":"1.0.0","shadow_commit_hash":null,"split_manifest_hash":"f2d99809dd21966190a823e684aaadda994281f5ccc2ba348db027fa6b43d716"},"train_state":{"image_exposures":200000,"latent_tokens":204800000,"microstep":0,"optimizer_step":200,"phase":"base","phase_step":200,"sampler_cursor":72000},"world_size":1}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/base/checkpoints/step-00000200/state.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5bc66fa76c2e48dff8bf67ccc5d714d3a866d3c97bf79609825d7bcbfe10f960
3
+ size 168491178
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/launch.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"adapter":"__main__:base_stage_service","argv":["--run-id","dev-e15-base-nosync200-batch1000-20260807a","--runs-root","/workspace/gsh/tmp/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs","--config-dir","/workspace/gsh/tmp/Schatten_MoE-continuation/configs","--config-name","e15","--input","base_train_plan=/workspace/gsh/tmp/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/perf/plans/continuous_200_batch1000","--input","data_manifest=/workspace/gsh/tmp/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/data/preprocess/data/manifest.json","--input","split_manifest=/workspace/gsh/tmp/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/data/splits/manifest.json"],"input_hashes":{"base_train_plan":"cc2cfec0ffa7713c69f9932a3ce1785eaaabf50d26405e91469829b8c15b757b","data_manifest":"4ec206f8c63b542a5210ffa77d3cfa122379bf97b7e87437a2ff34912f59a7cc","split_manifest":"f2d99809dd21966190a823e684aaadda994281f5ccc2ba348db027fa6b43d716"},"inputs":{"base_train_plan":"/workspace/gsh/tmp/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/perf/plans/continuous_200_batch1000","data_manifest":"/workspace/gsh/tmp/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/data/preprocess/data/manifest.json","split_manifest":"/workspace/gsh/tmp/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/data/splits/manifest.json"},"overrides":[],"stage":"train_base"}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/run_manifest.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"artifacts":[{"hash":"521d9b82197e8b46a2d980ce232e73f02c4d36f8c36573cb3f2d05cd9a1718bb","name":"base_artifact_manifest","parent_hash":null,"path":"base/base_artifact_manifest.json","schema_version":"1.0.0"},{"hash":"dfda0b3d74049fc178b1fff77938997ce24803b516f835e8aeaef529be29caa3","name":"base_loss_history","parent_hash":null,"path":"base/base_loss_history.json","schema_version":"1.0.0"},{"hash":"9f4a3bd78c512d43270d7d9130f0a22793864f0ac2784f598bae0db679d8cab2","name":"base_metrics","parent_hash":null,"path":"base/base_metrics.json","schema_version":"1.0.0"},{"hash":"5183bb961b9ce47ba50b5d307b5cd783995a12c9e1ffd05234009a54d411f272","name":"base_training_config","parent_hash":null,"path":"base/base_training_config.json","schema_version":"1.0.0"},{"hash":"880a38edc0d1a6ced983df88033c039b1eb49abcf9f9507b789f421579dc185f","name":"base_training_log","parent_hash":null,"path":"base/base_training_log.json","schema_version":"1.0.0"},{"hash":"753ba930fb09348142d5331c1a5fdf59fc6b64b76da17f2ea5386f7d65093e24","name":"checkpoint","parent_hash":null,"path":"base/checkpoints/base-only-moe-final/manifest.json","schema_version":"1.0.0"},{"hash":"53fd62885e3c576f857c47dca91981e5ad9e19d833b30215afd98f1fec75d570","name":"launch","parent_hash":null,"path":"launch.json","schema_version":"1.0.0"}],"code":{"contract_version":"contracts_v1","dirty":true,"dirty_patch_hash":"09ab2a90c42b9f81cee34424bff3737415e94518922d2256ca3dd4692854c072","git_commit":"3abed5e6d10894ea1551b5d2b97afa194db72ff3"},"config_hash":"6c1f1cbf03f8cae3d068caf1fad0f4246b5ed4774aefe80abfe6ddee9b2244b6","content_hash":"1d9dfbf6a18945a361f04bcce1ef5d825def337d10133b94a9ee64480066508f","created_at":"2026-08-07T08:11:16.520742+00:00","data_manifest_hash":"4ec206f8c63b542a5210ffa77d3cfa122379bf97b7e87437a2ff34912f59a7cc","edge_manifest_hash":null,"environment":{"cuda":"12.6","gpus":[{"compute_capability":"9.0","driver":"590.48.01","memory_bytes":85019328512,"name":"NVIDIA H100 80GB HBM3"}],"nccl":"2.22.3","python":"3.10.12","pytorch":"2.5.0a0+e000cf0ad9.nv24.10"},"flops":{"estimated_active":0.0,"measured":null,"method_version":"not_measured_preflight_v1"},"parameters":{"active":0,"logical":0},"parent_checkpoint_hash":null,"run_id":"dev-e15-base-nosync200-batch1000-20260807a","runtime":{"canonical_weight_alignment_version":"expert_swiglu_gate_up_down_v1","compute_dtype":"bfloat16","cuda_version":"12.6","driver_version":"590.48.01","expert_backend":"transformer_engine_grouped_linear_bf16_v1","profile_hash":"7930764e0384323c317c49db511e455f3392b9e7b73cac7be06927b515007802","profile_id":"h100_sm90_te_bf16_nv24_10_v1","python_version":"3.10.12","pytorch_build":"2.5.0a0+e000cf0ad9.nv24.10","sm":"9.0","transformer_engine_version":"1.11.0+4df8488"},"schema_version":"2.0.0","seeds":{"arm_scheduling":1731,"augmentation":1706,"candidate_audit":1741,"capture_sampling":1711,"cfg":1705,"data_order":1701,"diffusion":1703,"generation":1751,"global_seed":1701,"posterior":1702,"shadow_init":1721,"timestep":1704},"split_manifest_hash":"f2d99809dd21966190a823e684aaadda994281f5ccc2ba348db027fa6b43d716","stage":"train_base","vae_hash":null,"world_mesh":{"data_parallel":1,"expert_mapping_version":"rank_major_expert_contiguous_v1","expert_parallel":1,"layout_hash":"bbb880479a3d8f7e12e020ecdcb9c55f1ae88d077cbc01b940f1e9c5277207db","world_size":1}}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nosync200-batch1000-20260807a/run_status.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"manifest_hash":"1d9dfbf6a18945a361f04bcce1ef5d825def337d10133b94a9ee64480066508f","reason":null,"status":"COMMITTED","updated_at":"2026-08-07T08:11:16.804569+00:00"}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/base-only-moe-final/COMMIT ADDED
@@ -0,0 +1 @@
 
 
1
+ e26b841ab4af97ce641e60d47c7a2b0b0545176b9e285a60151ad905f87ff6c4
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d132d2b57758e886bf63d9e9807b8672279299a06834de28b0069ef7a71ff50
3
+ size 296502
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9a50107787b2a4d4da6452e8baf293fce5fc015cb3135b36363dc5e7310a3086
3
+ size 169113792
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/base-only-moe-final/manifest.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"checkpoint_id":"base-only-moe-final","component_names":["ema","engine","model","optimizer","sampler","scheduler"],"content_hash":"e26b841ab4af97ce641e60d47c7a2b0b0545176b9e285a60151ad905f87ff6c4","files":{"dcp/.metadata":"5d132d2b57758e886bf63d9e9807b8672279299a06834de28b0069ef7a71ff50","dcp/__0_0.distcp":"9a50107787b2a4d4da6452e8baf293fce5fc015cb3135b36363dc5e7310a3086","state.pt":"0f06f8c80dc7f99ceecc88454b0bb2e2fac081602b6903f484a4a0d57cebfbb5"},"identity":{"checkpoint_layout":"dcp_with_local_payload_v1","config_hash":"6c1f1cbf03f8cae3d068caf1fad0f4246b5ed4774aefe80abfe6ddee9b2244b6","data_manifest_hash":"4ec206f8c63b542a5210ffa77d3cfa122379bf97b7e87437a2ff34912f59a7cc","edge_manifest_hash":"dc0cb6347196f976c34e2e0f5aacca353b9e55582fa5240b662f2f57a87ca167","mesh":{"data_parallel":1,"expert_mapping_version":"contiguous_global_expert_v1","expert_parallel":1,"world_size":1},"model_compatibility":{"payload":{"interface_structure":{"formula_version":"directed_cross_gate_v1","max_interfaces_per_token":1,"norm_epsilon":1e-05,"null_logit_mode":"per_layer_fp32_scalar_v1","pair_router_beta0":null,"pair_router_feature_norm_epsilon":1e-05,"pair_router_has_null":true,"pair_router_null_logit_init":2.0,"pair_router_probability_epsilon":1e-08,"pair_router_temperature":1.0,"pair_router_temperature_end":0.1,"pair_router_temperature_schedule":"linear","rank":16,"separable_rank":12,"slot_capacity":2},"model":{"depth":8,"expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_width":320,"hidden_size":320,"in_channels":4,"learn_sigma":false,"moe_layer_pattern":[0,1,2,3,4,5,6,7],"num_classes":100,"num_heads":5,"out_channels":4,"patch_size":4,"variant":"mini"},"router":{"aux_loss_weight":0.0,"correction_bias_clip":0.1,"correction_bias_eta":0.002,"correction_bias_update_interval":10,"correction_bias_warmup_steps":0,"gate_init_std":0.01,"group_score_top_n":2,"load_balancer":"correction_bias","num_experts":4,"num_groups":1,"route_scale":1.0,"shared_expert":true,"top_k":2,"topk_group":1},"runtime_layout":{"canonical_weight_alignment_version":"expert_swiglu_gate_up_down_v1","compute_dtype":"bfloat16","expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_mapping_version":"contiguous_global_expert_v1","precision":"bf16","profile_id":"h100_sm90_te_bf16_nv24_10_v1"}},"signature":"25a79bd702c41ddc230e9da49950848570ded039885929998395fc7a2838924d","version":"model_compatibility_signature_v1"},"parent_checkpoint_hash":"f70c283cba852e6cc69717f0f6de13b7de359bbd2fc3d34293e98d60e4d58295","role_spec_hash":"23e60b4260edafc09b8808cf330d989a08de08579276dda5dc7db777b7c5c315","schema_version":"1.0.0","shadow_commit_hash":null,"split_manifest_hash":"f2d99809dd21966190a823e684aaadda994281f5ccc2ba348db027fa6b43d716"},"train_state":{"image_exposures":200000,"latent_tokens":204800000,"microstep":0,"optimizer_step":200,"phase":"base","phase_step":200,"sampler_cursor":72000},"world_size":1}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/base-only-moe-final/state.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0f06f8c80dc7f99ceecc88454b0bb2e2fac081602b6903f484a4a0d57cebfbb5
3
+ size 168499114
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000100/COMMIT ADDED
@@ -0,0 +1 @@
 
 
1
+ 491cfdc2a1293b902f780998db6a4caa278db47e681c54124a2fec0975fa7f5f
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000100/dcp/.metadata ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:17edd1b9bbeda95db0446d89ffc877cd26c068f7a08d6df756d836438b51554b
3
+ size 296496
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000100/dcp/__0_0.distcp ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:263770b4a881c9dc4f4e94af16ebebe75304de722a36718ffd29744db9cb276f
3
+ size 169113792
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000100/manifest.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"checkpoint_id":"step-00000100","component_names":["ema","engine","model","optimizer","sampler","scheduler"],"content_hash":"491cfdc2a1293b902f780998db6a4caa278db47e681c54124a2fec0975fa7f5f","files":{"dcp/.metadata":"17edd1b9bbeda95db0446d89ffc877cd26c068f7a08d6df756d836438b51554b","dcp/__0_0.distcp":"263770b4a881c9dc4f4e94af16ebebe75304de722a36718ffd29744db9cb276f","state.pt":"31f53017152d3e32da70eb059848df3dde7ff2ae58fa9d904229ebcf6e66485f"},"identity":{"checkpoint_layout":"dcp_with_local_payload_v1","config_hash":"6c1f1cbf03f8cae3d068caf1fad0f4246b5ed4774aefe80abfe6ddee9b2244b6","data_manifest_hash":"4ec206f8c63b542a5210ffa77d3cfa122379bf97b7e87437a2ff34912f59a7cc","edge_manifest_hash":"dc0cb6347196f976c34e2e0f5aacca353b9e55582fa5240b662f2f57a87ca167","mesh":{"data_parallel":1,"expert_mapping_version":"contiguous_global_expert_v1","expert_parallel":1,"world_size":1},"model_compatibility":{"payload":{"interface_structure":{"formula_version":"directed_cross_gate_v1","max_interfaces_per_token":1,"norm_epsilon":1e-05,"null_logit_mode":"per_layer_fp32_scalar_v1","pair_router_beta0":null,"pair_router_feature_norm_epsilon":1e-05,"pair_router_has_null":true,"pair_router_null_logit_init":2.0,"pair_router_probability_epsilon":1e-08,"pair_router_temperature":1.0,"pair_router_temperature_end":0.1,"pair_router_temperature_schedule":"linear","rank":16,"separable_rank":12,"slot_capacity":2},"model":{"depth":8,"expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_width":320,"hidden_size":320,"in_channels":4,"learn_sigma":false,"moe_layer_pattern":[0,1,2,3,4,5,6,7],"num_classes":100,"num_heads":5,"out_channels":4,"patch_size":4,"variant":"mini"},"router":{"aux_loss_weight":0.0,"correction_bias_clip":0.1,"correction_bias_eta":0.002,"correction_bias_update_interval":10,"correction_bias_warmup_steps":0,"gate_init_std":0.01,"group_score_top_n":2,"load_balancer":"correction_bias","num_experts":4,"num_groups":1,"route_scale":1.0,"shared_expert":true,"top_k":2,"topk_group":1},"runtime_layout":{"canonical_weight_alignment_version":"expert_swiglu_gate_up_down_v1","compute_dtype":"bfloat16","expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_mapping_version":"contiguous_global_expert_v1","precision":"bf16","profile_id":"h100_sm90_te_bf16_nv24_10_v1"}},"signature":"25a79bd702c41ddc230e9da49950848570ded039885929998395fc7a2838924d","version":"model_compatibility_signature_v1"},"parent_checkpoint_hash":null,"role_spec_hash":"23e60b4260edafc09b8808cf330d989a08de08579276dda5dc7db777b7c5c315","schema_version":"1.0.0","shadow_commit_hash":null,"split_manifest_hash":"f2d99809dd21966190a823e684aaadda994281f5ccc2ba348db027fa6b43d716"},"train_state":{"image_exposures":100000,"latent_tokens":102400000,"microstep":0,"optimizer_step":100,"phase":"base","phase_step":100,"sampler_cursor":100000},"world_size":1}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000100/state.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:31f53017152d3e32da70eb059848df3dde7ff2ae58fa9d904229ebcf6e66485f
3
+ size 168491178
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000200/COMMIT ADDED
@@ -0,0 +1 @@
 
 
1
+ f70c283cba852e6cc69717f0f6de13b7de359bbd2fc3d34293e98d60e4d58295
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000200/dcp/.metadata ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:726c22a54fdb64e3855550759236c781ca0d5ab0b8979033f8859d480efc77b4
3
+ size 296496
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000200/dcp/__0_0.distcp ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9a50107787b2a4d4da6452e8baf293fce5fc015cb3135b36363dc5e7310a3086
3
+ size 169113792
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000200/manifest.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"checkpoint_id":"step-00000200","component_names":["ema","engine","model","optimizer","sampler","scheduler"],"content_hash":"f70c283cba852e6cc69717f0f6de13b7de359bbd2fc3d34293e98d60e4d58295","files":{"dcp/.metadata":"726c22a54fdb64e3855550759236c781ca0d5ab0b8979033f8859d480efc77b4","dcp/__0_0.distcp":"9a50107787b2a4d4da6452e8baf293fce5fc015cb3135b36363dc5e7310a3086","state.pt":"f7eef57c77ac4685861066db5bf0c2efad3671cd2362536e63920ded21d32b3d"},"identity":{"checkpoint_layout":"dcp_with_local_payload_v1","config_hash":"6c1f1cbf03f8cae3d068caf1fad0f4246b5ed4774aefe80abfe6ddee9b2244b6","data_manifest_hash":"4ec206f8c63b542a5210ffa77d3cfa122379bf97b7e87437a2ff34912f59a7cc","edge_manifest_hash":"dc0cb6347196f976c34e2e0f5aacca353b9e55582fa5240b662f2f57a87ca167","mesh":{"data_parallel":1,"expert_mapping_version":"contiguous_global_expert_v1","expert_parallel":1,"world_size":1},"model_compatibility":{"payload":{"interface_structure":{"formula_version":"directed_cross_gate_v1","max_interfaces_per_token":1,"norm_epsilon":1e-05,"null_logit_mode":"per_layer_fp32_scalar_v1","pair_router_beta0":null,"pair_router_feature_norm_epsilon":1e-05,"pair_router_has_null":true,"pair_router_null_logit_init":2.0,"pair_router_probability_epsilon":1e-08,"pair_router_temperature":1.0,"pair_router_temperature_end":0.1,"pair_router_temperature_schedule":"linear","rank":16,"separable_rank":12,"slot_capacity":2},"model":{"depth":8,"expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_width":320,"hidden_size":320,"in_channels":4,"learn_sigma":false,"moe_layer_pattern":[0,1,2,3,4,5,6,7],"num_classes":100,"num_heads":5,"out_channels":4,"patch_size":4,"variant":"mini"},"router":{"aux_loss_weight":0.0,"correction_bias_clip":0.1,"correction_bias_eta":0.002,"correction_bias_update_interval":10,"correction_bias_warmup_steps":0,"gate_init_std":0.01,"group_score_top_n":2,"load_balancer":"correction_bias","num_experts":4,"num_groups":1,"route_scale":1.0,"shared_expert":true,"top_k":2,"topk_group":1},"runtime_layout":{"canonical_weight_alignment_version":"expert_swiglu_gate_up_down_v1","compute_dtype":"bfloat16","expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_mapping_version":"contiguous_global_expert_v1","precision":"bf16","profile_id":"h100_sm90_te_bf16_nv24_10_v1"}},"signature":"25a79bd702c41ddc230e9da49950848570ded039885929998395fc7a2838924d","version":"model_compatibility_signature_v1"},"parent_checkpoint_hash":"491cfdc2a1293b902f780998db6a4caa278db47e681c54124a2fec0975fa7f5f","role_spec_hash":"23e60b4260edafc09b8808cf330d989a08de08579276dda5dc7db777b7c5c315","schema_version":"1.0.0","shadow_commit_hash":null,"split_manifest_hash":"f2d99809dd21966190a823e684aaadda994281f5ccc2ba348db027fa6b43d716"},"train_state":{"image_exposures":200000,"latent_tokens":204800000,"microstep":0,"optimizer_step":200,"phase":"base","phase_step":200,"sampler_cursor":72000},"world_size":1}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/base/checkpoints/step-00000200/state.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f7eef57c77ac4685861066db5bf0c2efad3671cd2362536e63920ded21d32b3d
3
+ size 168491178
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/launch.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"adapter":"__main__:base_stage_service","argv":["--run-id","dev-e15-base-nsys100-batch1000-20260807a","--runs-root","/workspace/gsh/tmp/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs","--config-dir","/workspace/gsh/tmp/Schatten_MoE-continuation/configs","--config-name","e15","--input","base_train_plan=/workspace/gsh/tmp/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/perf/plans/continuous_200_batch1000","--input","data_manifest=/workspace/gsh/tmp/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/data/preprocess/data/manifest.json","--input","split_manifest=/workspace/gsh/tmp/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/data/splits/manifest.json"],"input_hashes":{"base_train_plan":"cc2cfec0ffa7713c69f9932a3ce1785eaaabf50d26405e91469829b8c15b757b","data_manifest":"4ec206f8c63b542a5210ffa77d3cfa122379bf97b7e87437a2ff34912f59a7cc","split_manifest":"f2d99809dd21966190a823e684aaadda994281f5ccc2ba348db027fa6b43d716"},"inputs":{"base_train_plan":"/workspace/gsh/tmp/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/perf/plans/continuous_200_batch1000","data_manifest":"/workspace/gsh/tmp/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/data/preprocess/data/manifest.json","split_manifest":"/workspace/gsh/tmp/Schatten_MoE-singlecard-experiments/E2-SC40-imagenet100-tiny-main-f49af44/data/splits/manifest.json"},"overrides":[],"stage":"train_base"}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/run_manifest.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"artifacts":[{"hash":"a609ade5352d66fab474125b6b1f5eb9445bcf9aa84c3df52089f56d051f25c6","name":"base_artifact_manifest","parent_hash":null,"path":"base/base_artifact_manifest.json","schema_version":"1.0.0"},{"hash":"40d9c36b99f52bd5c52d54776806d8a065d3902ddc1409b9f2af52fa99215652","name":"base_loss_history","parent_hash":null,"path":"base/base_loss_history.json","schema_version":"1.0.0"},{"hash":"1d81737b2c255ec5d4e96b387e64ea274e473f05b8c06d1df4c133afd56c4e78","name":"base_metrics","parent_hash":null,"path":"base/base_metrics.json","schema_version":"1.0.0"},{"hash":"5183bb961b9ce47ba50b5d307b5cd783995a12c9e1ffd05234009a54d411f272","name":"base_training_config","parent_hash":null,"path":"base/base_training_config.json","schema_version":"1.0.0"},{"hash":"d85c008a8de3eb0dfc9207edc615a23614bdcc69675923aad1f10bbd04af48eb","name":"base_training_log","parent_hash":null,"path":"base/base_training_log.json","schema_version":"1.0.0"},{"hash":"ccb85c9ffbf48013242132e6d1a27013e2ae3aec37552bbb008b676a324c55d7","name":"checkpoint","parent_hash":null,"path":"base/checkpoints/base-only-moe-final/manifest.json","schema_version":"1.0.0"},{"hash":"420a01f3482ec5b7c51ff3f83c298ee27234b31b9d9f4cd57c6ca32f4f6ed02e","name":"launch","parent_hash":null,"path":"launch.json","schema_version":"1.0.0"}],"code":{"contract_version":"contracts_v1","dirty":true,"dirty_patch_hash":"eea5705f122676b591892482ba41fa956952cda2260f61e590d1ed070de4ed29","git_commit":"3abed5e6d10894ea1551b5d2b97afa194db72ff3"},"config_hash":"6c1f1cbf03f8cae3d068caf1fad0f4246b5ed4774aefe80abfe6ddee9b2244b6","content_hash":"049073201163d6c4cb82143db55f1abda74459115af775c0983efb4d4fd1e1ef","created_at":"2026-08-07T08:23:18.272068+00:00","data_manifest_hash":"4ec206f8c63b542a5210ffa77d3cfa122379bf97b7e87437a2ff34912f59a7cc","edge_manifest_hash":null,"environment":{"cuda":"12.6","gpus":[{"compute_capability":"9.0","driver":"590.48.01","memory_bytes":85019328512,"name":"NVIDIA H100 80GB HBM3"}],"nccl":"2.22.3","python":"3.10.12","pytorch":"2.5.0a0+e000cf0ad9.nv24.10"},"flops":{"estimated_active":0.0,"measured":null,"method_version":"not_measured_preflight_v1"},"parameters":{"active":0,"logical":0},"parent_checkpoint_hash":null,"run_id":"dev-e15-base-nsys100-batch1000-20260807a","runtime":{"canonical_weight_alignment_version":"expert_swiglu_gate_up_down_v1","compute_dtype":"bfloat16","cuda_version":"12.6","driver_version":"590.48.01","expert_backend":"transformer_engine_grouped_linear_bf16_v1","profile_hash":"7930764e0384323c317c49db511e455f3392b9e7b73cac7be06927b515007802","profile_id":"h100_sm90_te_bf16_nv24_10_v1","python_version":"3.10.12","pytorch_build":"2.5.0a0+e000cf0ad9.nv24.10","sm":"9.0","transformer_engine_version":"1.11.0+4df8488"},"schema_version":"2.0.0","seeds":{"arm_scheduling":1731,"augmentation":1706,"candidate_audit":1741,"capture_sampling":1711,"cfg":1705,"data_order":1701,"diffusion":1703,"generation":1751,"global_seed":1701,"posterior":1702,"shadow_init":1721,"timestep":1704},"split_manifest_hash":"f2d99809dd21966190a823e684aaadda994281f5ccc2ba348db027fa6b43d716","stage":"train_base","vae_hash":null,"world_mesh":{"data_parallel":1,"expert_mapping_version":"rank_major_expert_contiguous_v1","expert_parallel":1,"layout_hash":"bbb880479a3d8f7e12e020ecdcb9c55f1ae88d077cbc01b940f1e9c5277207db","world_size":1}}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys100-batch1000-20260807a/run_status.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"manifest_hash":"049073201163d6c4cb82143db55f1abda74459115af775c0983efb4d4fd1e1ef","reason":null,"status":"COMMITTED","updated_at":"2026-08-07T08:23:18.645735+00:00"}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/base-only-moe-final/COMMIT ADDED
@@ -0,0 +1 @@
 
 
1
+ ac4c577e75586773fc50cd0578e898b0845d7644ec11784d427ec004403ce014
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/.metadata ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:73d7ecf06dfb57f04714375dabc5448dcaf7b06823511b53fab4b3db2c06fede
3
+ size 296501
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/base-only-moe-final/dcp/__0_0.distcp ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8e5134b1b7d2f4d1597d5fad576639d83e354ca86c6ac7cee1453d44959468f
3
+ size 169113792
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/base-only-moe-final/manifest.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"checkpoint_id":"base-only-moe-final","component_names":["ema","engine","model","optimizer","sampler","scheduler"],"content_hash":"ac4c577e75586773fc50cd0578e898b0845d7644ec11784d427ec004403ce014","files":{"dcp/.metadata":"73d7ecf06dfb57f04714375dabc5448dcaf7b06823511b53fab4b3db2c06fede","dcp/__0_0.distcp":"a8e5134b1b7d2f4d1597d5fad576639d83e354ca86c6ac7cee1453d44959468f","state.pt":"dfb149fc1feb6a87ca09a2844f10c28302085048e46cb4d0b2b53a9719721626"},"identity":{"checkpoint_layout":"dcp_with_local_payload_v1","config_hash":"6c1f1cbf03f8cae3d068caf1fad0f4246b5ed4774aefe80abfe6ddee9b2244b6","data_manifest_hash":"4ec206f8c63b542a5210ffa77d3cfa122379bf97b7e87437a2ff34912f59a7cc","edge_manifest_hash":"dc0cb6347196f976c34e2e0f5aacca353b9e55582fa5240b662f2f57a87ca167","mesh":{"data_parallel":1,"expert_mapping_version":"contiguous_global_expert_v1","expert_parallel":1,"world_size":1},"model_compatibility":{"payload":{"interface_structure":{"formula_version":"directed_cross_gate_v1","max_interfaces_per_token":1,"norm_epsilon":1e-05,"null_logit_mode":"per_layer_fp32_scalar_v1","pair_router_beta0":null,"pair_router_feature_norm_epsilon":1e-05,"pair_router_has_null":true,"pair_router_null_logit_init":2.0,"pair_router_probability_epsilon":1e-08,"pair_router_temperature":1.0,"pair_router_temperature_end":0.1,"pair_router_temperature_schedule":"linear","rank":16,"separable_rank":12,"slot_capacity":2},"model":{"depth":8,"expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_width":320,"hidden_size":320,"in_channels":4,"learn_sigma":false,"moe_layer_pattern":[0,1,2,3,4,5,6,7],"num_classes":100,"num_heads":5,"out_channels":4,"patch_size":4,"variant":"mini"},"router":{"aux_loss_weight":0.0,"correction_bias_clip":0.1,"correction_bias_eta":0.002,"correction_bias_update_interval":10,"correction_bias_warmup_steps":0,"gate_init_std":0.01,"group_score_top_n":2,"load_balancer":"correction_bias","num_experts":4,"num_groups":1,"route_scale":1.0,"shared_expert":true,"top_k":2,"topk_group":1},"runtime_layout":{"canonical_weight_alignment_version":"expert_swiglu_gate_up_down_v1","compute_dtype":"bfloat16","expert_backend":"transformer_engine_grouped_linear_bf16_v1","expert_mapping_version":"contiguous_global_expert_v1","precision":"bf16","profile_id":"h100_sm90_te_bf16_nv24_10_v1"}},"signature":"25a79bd702c41ddc230e9da49950848570ded039885929998395fc7a2838924d","version":"model_compatibility_signature_v1"},"parent_checkpoint_hash":"7a035664561e3c5a60dcc8a5d19413e430764ffa1b25294f31dd22878680f662","role_spec_hash":"23e60b4260edafc09b8808cf330d989a08de08579276dda5dc7db777b7c5c315","schema_version":"1.0.0","shadow_commit_hash":null,"split_manifest_hash":"f2d99809dd21966190a823e684aaadda994281f5ccc2ba348db027fa6b43d716"},"train_state":{"image_exposures":20000,"latent_tokens":20480000,"microstep":0,"optimizer_step":20,"phase":"base","phase_step":20,"sampler_cursor":20000},"world_size":1}
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/base-only-moe-final/state.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dfb149fc1feb6a87ca09a2844f10c28302085048e46cb4d0b2b53a9719721626
3
+ size 168499114
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/step-00000020/COMMIT ADDED
@@ -0,0 +1 @@
 
 
1
+ 7a035664561e3c5a60dcc8a5d19413e430764ffa1b25294f31dd22878680f662
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/step-00000020/dcp/.metadata ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:99c4aa8137c259349d13a1149ff7d1421f070c08f54b2b90f0bd8e0a2ff57f87
3
+ size 296495
artifacts/Schatten_MoE-e15-experiments/E15-mini-imagenet100-base1701/formal-eta002/runs/dev-e15-base-nsys20-batch1000-20260807a/base/checkpoints/step-00000020/dcp/__0_0.distcp ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8e5134b1b7d2f4d1597d5fad576639d83e354ca86c6ac7cee1453d44959468f
3
+ size 169113792