Training Metrics

as donto believed on

From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-10.)

Training Metrics has 47 facts recorded in Dontopedia across 11 references, with 8 live disagreements.

47 facts·27 predicates·11 sources·8 in dispute

Mostly:includes(7), rdf:type(3), recorded at step(3)

Inbound mentions (11)

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.

`partOf`Part of(2)

Grad Normex:grad-norm
Val Lossex:val-loss

`announcesProgress`Announces Progress(1)

Xenonfunex:xenonfun

`containsLogData`Contains Log Data(1)

Log Entry 2026 03 08 04 45ex:log-entry-2026-03-08-04-45

`logs`Logs(1)

Logging Infoex:logging-info

`outputs`Outputs(1)

Print Statementex:print-statement

`performsReporting`Performs Reporting(1)

Seq Len 2048 Resonance Training Logex:seq-len-2048-resonance-training-log

`presentsData`Presents Data(1)

Xenonfun Postex:xenonfun-post

`records`Records(1)

Logging Sectionex:logging-section

`reportsObservation`Reports Observation(1)

Xenonfunex:xenonfun

`sourceOfMetrics`Source of Metrics(1)

Phase Metrics Lohe Spherical 20260313 023945 Jsonlex:phase-metrics-lohe-spherical-20260313-023945-jsonl

Other facts (47)

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.

47 facts

Predicate	Value	Ref
Includes	Epoch Number	[9]
Includes	Batch Number	[9]
Includes	Loss Value	[9]
Includes	timestamp	[10]
Includes	log-level	[10]
Includes	batch-size	[10]
Includes	loss-value	[10]
Rdf:type	Measurement	[8]
Rdf:type	Metrics	[9]
Rdf:type	Metrics	[11]
Recorded at Step	1	[8]
Recorded at Step	50	[8]
Recorded at Step	100	[8]
Has Bpb	5.766	[8]
Has Bpb	4.737	[8]
Has Bpb	3.979	[8]
Has Loss	3.997	[8]
Has Loss	3.2834	[8]
Has Loss	2.7579	[8]
Has Tok Per Sec	409	[8]
Has Tok Per Sec	6860	[8]
Has Tok Per Sec	8339	[8]
Has Iter Per Sec	0.02	[8]
Has Iter Per Sec	0.42	[8]
Has Iter Per Sec	0.51	[8]
Has Elapsed Time	0m40s	[8]
Has Elapsed Time	1m59s	[8]
Has Elapsed Time	3m16s	[8]
Exist for Mapping	true	[1]
References Best Global	Best 2 0151	[2]
Indicate No Instability	true	[3]
Indicate Progress	Decreasing Bpb	[4]
Indicate Ongoing Process	Experiment	[5]
Show Steady Tok S	True	[6]
Step Number	500	[7]
Total Steps	2000	[7]
Progress Percentage	25	[7]
Bpb	5.739	[7]
Correlation	0.263	[7]
Snr	-11.3	[7]
Capacity	2.7	[7]
DC	0.08	[7]
Temperature	0.34	[7]
Learning Rate	0.0027	[7]
Tokens Per Second	49436	[7]
Eta	8min	[7]
Validation Bpb	5.732	[7]

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.

47 undated

—

existForMappingblah/training-and-evals/part-25

true

—

referencesBestGlobalblah/watt-activation/part-39

ex:best-2-0151

—

indicateNoInstabilityblah/watt-activation/part-267

true

—

indicateProgressblah/watt-activation/part-372

ex:decreasing-bpb

—

indicateOngoingProcessblah/watt-activation/part-402

ex:experiment

—

showSteadyTokSblah/watt-activation/part-701

ex:true

—

stepNumberblah/watt-activation/400

500

—

totalStepsblah/watt-activation/400

2000

—

progressPercentageblah/watt-activation/400

—

bpbblah/watt-activation/400

5.739

—

correlationblah/watt-activation/400

0.263

—

snrblah/watt-activation/400

-11.3

—

capacityblah/watt-activation/400

2.7

—

dcblah/watt-activation/400

0.08

—

temperatureblah/watt-activation/400

0.34

—

learningRateblah/watt-activation/400

0.0027

—

tokensPerSecondblah/watt-activation/400

49436

—

etablah/watt-activation/400

8min

—

validationBpbblah/watt-activation/400

5.732

—

typeblah/watt-activation/696

ex:Measurement

—

recordedAtStepblah/watt-activation/696

—

hasBpbblah/watt-activation/696

5.766

—

hasLossblah/watt-activation/696

3.997

—

hasTokPerSecblah/watt-activation/696

409

—

hasIterPerSecblah/watt-activation/696

0.02

—

hasElapsedTimeblah/watt-activation/696

0m40s

—

recordedAtStepblah/watt-activation/696

—

hasBpbblah/watt-activation/696

4.737

—

hasLossblah/watt-activation/696

3.2834

—

hasTokPerSecblah/watt-activation/696

6860

—

hasIterPerSecblah/watt-activation/696

0.42

—

hasElapsedTimeblah/watt-activation/696

1m59s

—

recordedAtStepblah/watt-activation/696

100

—

hasBpbblah/watt-activation/696

3.979

—

hasLossblah/watt-activation/696

2.7579

—

hasTokPerSecblah/watt-activation/696

8339

—

hasIterPerSecblah/watt-activation/696

0.51

—

hasElapsedTimeblah/watt-activation/696

3m16s

—

typebeam/f5a5540b-3c9d-4103-85d7-7db7b8ea25d3

ex:Metrics

—

includesbeam/f5a5540b-3c9d-4103-85d7-7db7b8ea25d3

ex:epoch-number

—

includesbeam/f5a5540b-3c9d-4103-85d7-7db7b8ea25d3

ex:batch-number

—

includesbeam/f5a5540b-3c9d-4103-85d7-7db7b8ea25d3

ex:loss-value

—

includesbeam/d9a80d69-c4c9-47c5-8393-2eaf674f6563

timestamp

—

includesbeam/d9a80d69-c4c9-47c5-8393-2eaf674f6563

log-level

—

includesbeam/d9a80d69-c4c9-47c5-8393-2eaf674f6563

batch-size

—

includesbeam/d9a80d69-c4c9-47c5-8393-2eaf674f6563

loss-value

—

typebeam/c8102774-0736-45ab-8d51-87fae35d0377

ex:Metrics

References (11)

11 references

[1]Part 251 fact
ctx:discord/blah/training-and-evals/part-25
[2]Part 391 fact
ctx:discord/blah/watt-activation/part-39
[3]Part 2671 fact
ctx:discord/blah/watt-activation/part-267
[4]Part 3721 fact
ctx:discord/blah/watt-activation/part-372
[5]Part 4021 fact
ctx:discord/blah/watt-activation/part-402
[6]Part 7011 fact
ctx:discord/blah/watt-activation/part-701
[7]40013 facts
ctx:discord/blah/watt-activation/400
- full textwatt-activation-400
  text/plain2 KBdoc:agent/watt-activation-400/bfd3ae1a-a87b-4ef9-bd0b-548fd78cc0cb
  Show excerpt
  [2026-03-19 05:11] xenonfun: ⏺ The ConstellationDecoder is 94% of the model's parameters (32K of 27K dynamics). That's a design smell. ``` The most elegant option: use the encoding table itself as the decoder. The BPSK table maps each byt…
[8]69619 facts
ctx:discord/blah/watt-activation/696
- full textwatt-activation-696
  text/plain2 KBdoc:agent/watt-activation-696/6bc363f6-e780-4242-9e13-32e8d01a4dd8
  Show excerpt
  [2026-05-01 02:47] xenonfun: It wants 150M or so so think the chatgpt-2 in 24hr is looking achievable. we do need to get more dataset here. [2026-05-01 02:50] lisamegawatts: mines downloading datasets and parsing now, added a 4096 option [2…
[9]F5a5540b 3c9d 4103 85d7 7db7b8ea25d34 facts
ctx:claims/beam/f5a5540b-3c9d-4103-85d7-7db7b8ea25d3
[10]D9a80d69 C4c9 47c5 8393 2eaf674f65634 facts
ctx:claims/beam/d9a80d69-c4c9-47c5-8393-2eaf674f6563
- full textbeam-chunk
  text/plain1 KBdoc:beam/d9a80d69-c4c9-47c5-8393-2eaf674f6563
  Show excerpt
  inputs = torch.tensor(decrypted_batch['query'], dtype=torch.float32).to(device) labels = torch.tensor(decrypted_batch['label'], dtype=torch.long).to(device) # Forward pass outputs = model(inputs) los…
[11]C8102774 0736 45ab 8d51 87fae35d03771 fact
ctx:claims/beam/c8102774-0736-45ab-8d51-87fae35d0377
- full textbeam-chunk
  text/plain1 KBdoc:beam/c8102774-0736-45ab-8d51-87fae35d0377
  Show excerpt
  for epoch in range(100): for batch in data_loader: inputs = batch['query'].float().to(device) labels = batch['label'].long().to(device) optimizer.zero_grad() outputs = model(input…

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.

Training Metrics

Inbound mentions (11)

partOfPart of(2)

announcesProgressAnnounces Progress(1)

containsLogDataContains Log Data(1)

logsLogs(1)

outputsOutputs(1)

performsReportingPerforms Reporting(1)

presentsDataPresents Data(1)

recordsRecords(1)

reportsObservationReports Observation(1)

sourceOfMetricsSource of Metrics(1)

Other facts (47)

Timeline

References (11)

See also

Keep researching

`partOf`Part of(2)

`announcesProgress`Announces Progress(1)

`containsLogData`Contains Log Data(1)

`logs`Logs(1)

`outputs`Outputs(1)

`performsReporting`Performs Reporting(1)

`presentsData`Presents Data(1)

`records`Records(1)

`reportsObservation`Reports Observation(1)

`sourceOfMetrics`Source of Metrics(1)