Dontopedia

step 2000

From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-06.)

step 2000 has 50 facts recorded in Dontopedia across 11 references, with 6 live disagreements.

50 facts·34 predicates·11 sources·6 in dispute

Mostly:rdf:type(5), has loss(4), has learning rate(3)

Maturity scale raw canonical shape-checked rule-derived certified

Inbound mentions (10)

Other subjects in dontopedia point AT this entity as a value. These are inverse relationships — e.g. "X motherOf this subject" — and answer questions the forward facts can't. Grouped by predicate.

completesSuccessfullyCompletes Successfully(1)

correspondsToStepCorresponds to Step(1)

hasLowerLossThanHas Lower Loss Than(1)

hasSlightlyHigherLossThanHas Slightly Higher Loss Than(1)

indicateConvergenceIndicate Convergence(1)

occursAtTrainingStepOccurs at Training Step(1)

precedesPrecedes(1)

recordsEventRecords Event(1)

temporallyAfterTemporally After(1)

validatesStepValidates Step(1)

Other facts (49)

The long tail: predicates that appear too rarely to warrant their own section. Filter or scroll to find a specific one. Each row links to its source.

49 facts
PredicateValueRef
Rdf:typeTraining Step[5]
Rdf:typeTraining Step[6]
Rdf:typeTraining Event[7]
Rdf:typeTraining Step[10]
Rdf:typeTraining Step[11]
Has Loss1.4115[1]
Has Loss1.3944[2]
Has Loss0.9776[4]
Has Loss2.3812[9]
Has Learning Rate0.0000488[3]
Has Learning Rate0.0000488[6]
Has Learning Rate0.00199[10]
Has Loss Value6.0245[3]
Has Loss Value6.0245[6]
Has Loss Value0.9776[10]
Step Number2000[5]
Step Number2000[6]
Step Number2000[10]
Has Tokens Per Second12290[3]
Has Tokens Per Second10607[10]
Has Bpb1.41[4]
Has Bpb3.44[9]
Is Out of2000[1]
PrecedesInference Section[2]
Has Step Number2000[2]
SucceedsTraining Step 2100[3]
Has Total Steps16684[3]
Has Progress Percentage12[3]
Has Ppl Value413.4[3]
Has Current Step Number2000[3]
Has Time Per Step Ms667[3]
Has Eta Minutes163[3]
Has Gnorm0.3589[4]
Has Tok Per S10607[4]
Has Lr0.00199[4]
Reported Loss1.3944[5]
Progress Percentage12[6]
Has Perplexity Value413.4[6]
Has Step Duration667[6]
Tokens Per Second12290[6]
Has Estimated Time Remaining163min[6]
Perplexity31[8]
Utilization0.594[8]
Throughput196457[8]
Has Throughput110,335-B/s[9]
Part ofCross Patch Ar Decoder[9]
Has Bpb Value1.41[10]
Has Gradient Norm0.3589[10]
Performed First Real Weight Updatetrue[11]

Timeline

Timeline axis is valid_time — when each source says the fact was true in the world, not when Dontopedia learned about it. Retracted rows are kept for provenance; coloured stripes indicate the context kind.

50 undated
—
isOutOfblah/vidya/part-4
2000
—
hasLossblah/vidya/part-4
1.4115
—
hasLossblah/vidya/part-3
1.3944
—
precedesblah/vidya/part-3
ex:inference-section
—
hasStepNumberblah/vidya/part-3
2000
—
succeedsblah/watt-activation/part-150
ex:training-step-2100
—
hasTotalStepsblah/watt-activation/part-150
16684
—
hasLearningRateblah/watt-activation/part-150
0.0000488
—
hasTokensPerSecondblah/watt-activation/part-150
12290
—
hasLossValueblah/watt-activation/part-150
6.0245
—
hasProgressPercentageblah/watt-activation/part-150
12
—
hasPplValueblah/watt-activation/part-150
413.4
—
hasCurrentStepNumberblah/watt-activation/part-150
2000
—
hasTimePerStepMsblah/watt-activation/part-150
667
—
hasEtaMinutesblah/watt-activation/part-150
163
—
hasLossblah/watt-activation/part-658
0.9776
—
hasGnormblah/watt-activation/part-658
0.3589
—
hasTokPerSblah/watt-activation/part-658
10607
—
hasLrblah/watt-activation/part-658
0.00199
—
hasBpbblah/watt-activation/part-658
1.41
—
typeblah/vidya/3
ex:TrainingStep
—
stepNumberblah/vidya/3
2000
—
reportedLossblah/vidya/3
1.3944
—
typeblah/watt-activation/150
ex:TrainingStep
—
stepNumberblah/watt-activation/150
2000
—
progressPercentageblah/watt-activation/150
12
—
hasLossValueblah/watt-activation/150
6.0245
—
hasPerplexityValueblah/watt-activation/150
413.4
—
hasLearningRateblah/watt-activation/150
0.0000488
—
hasStepDurationblah/watt-activation/150
667
—
tokensPerSecondblah/watt-activation/150
12290
—
hasEstimatedTimeRemainingblah/watt-activation/150
163min
—
typeblah/watt-activation/297
ex:TrainingEvent
—
perplexityblah/watt-activation/299
31
—
utilizationblah/watt-activation/299
0.594
—
throughputblah/watt-activation/299
196457
—
hasLossblah/watt-activation/300
2.3812
—
hasBpbblah/watt-activation/300
3.44
—
hasThroughputblah/watt-activation/300
110,335-B/s
—
partOfblah/watt-activation/300
ex:cross-patch-ar-decoder
—
typeblah/watt-activation/655
ex:TrainingStep
—
stepNumberblah/watt-activation/655
2000
—
hasLossValueblah/watt-activation/655
0.9776
—
hasBpbValueblah/watt-activation/655
1.41
—
hasGradientNormblah/watt-activation/655
0.3589
—
hasLearningRateblah/watt-activation/655
0.00199
—
hasTokensPerSecondblah/watt-activation/655
10607
—
typeblah/watt-activation/664
ex:TrainingStep
—
labelblah/watt-activation/664
step 2000
—
performedFirstRealWeightUpdateblah/watt-activation/664
true

References (11)

11 references
  1. [1]Part 42 facts
    ctx:discord/blah/vidya/part-4
  2. [2]Part 33 facts
    ctx:discord/blah/vidya/part-3
  3. [3]Part 15010 facts
    ctx:discord/blah/watt-activation/part-150
  4. [4]Part 6585 facts
    ctx:discord/blah/watt-activation/part-658
  5. [5]33 facts
    ctx:discord/blah/vidya/3
    • full textvidya-3
      text/plain2 KBdoc:agent/vidya-3/36dc77aa-7dff-487d-af09-d932dc181732
      Show excerpt
      [2026-02-20 07:37] rolandnsharp7643: roland@cube:~/code/flow/ocaml$ ./microgpt_scaled num docs: 16478 vocab size: 80 num params: 206848 step 500 / 2000 | loss 2.5678 step 1000 / 2000 | loss 2.1392 step 1500 / 2000 | loss 1.7365 step 2000 /…
  6. [6]1509 facts
    ctx:discord/blah/watt-activation/150
    • full textwatt-activation-150
      text/plain2 KBdoc:agent/watt-activation-150/337a9b46-3368-40e6-b1b8-d27104be216d
      Show excerpt
      [2026-03-09 15:48] xenonfun: step 2000/16684 12.0% loss=6.0245 ppl= 413.4 lr=4.88e-05 667ms 12,290tok/s eta=163min VAL step 2000 loss=6.0279 ppl=414.9 ★ best ckpt step 2000 → step_002000 step 2100/16684 12.6% loss=6.01…
  7. [7]2971 fact
    ctx:discord/blah/watt-activation/297
    • full textwatt-activation-297
      text/plain2 KBdoc:agent/watt-activation-297/ad91f718-f038-464f-a6d2-ba91d77fe4e3
      Show excerpt
      [2026-03-14 05:23] xenonfun: 600K context, the UI crazy scroll thing and memory leaks are annoyin claude is sucking up 6GB now. [2026-03-14 05:24] xenonfun: ``` ⏺ Launched in rjs:longgen. This trains with longer context (512) and more steps…
  8. [8]2993 facts
    ctx:discord/blah/watt-activation/299
    • full textwatt-activation-299
      text/plain2 KBdoc:agent/watt-activation-299/3c2f94b1-841d-457c-8bfa-1161d5e80363
      Show excerpt
      [2026-03-14 06:27] xenonfun: ``` Loaded 41.4 MB VQ encoder (5000 steps)... Frozen downstream: 170,329 trainable / 520,945 total step 1000 | ppl=9.0 util=0.234 | 197,267 B/s step 2000 | ppl=31.0 util=0.594 | 196,457 B/s st…
  9. [9]3004 facts
    ctx:discord/blah/watt-activation/300
    • full textwatt-activation-300
      text/plain3 KBdoc:agent/watt-activation-300/3b6edccf-3524-4608-838f-25890efaea15
      Show excerpt
      [2026-03-14 06:34] xenonfun: ``` 3. Manual attention (lines 110-128) — Hand-rolled softmax attention instead of using mx.fast.scaled_dot_product_attention. MLX's fused attention kernel is significantly faster for small sequence lengths. …
  10. [10]6557 facts
    ctx:discord/blah/watt-activation/655
    • full textwatt-activation-655
      text/plain2 KBdoc:agent/watt-activation-655/184fbb55-c5df-4c57-ba7d-a525dd418a7a
      Show excerpt
      [2026-04-19 09:46] xenonfun: loaded 24999946 bytes from /Users/ms/MS/HarmonicRust/data/curriculum/domain_tinystories.bin step 0 loss=6.1092 bpb=8.814 gnorm=10.6885 lr=4.00e-6 9259 tok/s step 50 loss=2.8955 bpb=4.177 gnorm=2.2132…
  11. [11]6643 facts
    ctx:discord/blah/watt-activation/664

See also

Keep researching

Missing something or suspicious of what's here? Kick off a research session — a Claude agent will investigate, cite its sources, and file new facts into a dedicated context you can review before accepting into the shared view.

↑