CITADEL
StandbyNothing is live right now.
Eval accuracy
%
no eval

Familiar decider: ModernBERT-large

failedRadeon 8060S

One of the small models that decide whether Familiar, a desktop companion, reacts to something happening on the PC, and how. Every candidate trains on the same 6,000 decision points from synthetic timelines. The encoders also get one run on the full pool.

started
Fri 20 Nov, 13:18 UTC
ran for
12:31
config
batch 16, seed 0

run6/12step470/7274epoch0/2config5/9seed0batch16modelModernBERT-large 395Mrev45bb465items58183+500

Loss train 0.2214 ▼ −2.6742 lowest 0.2214 at step 470 run before, run 5 ▼ −0.0397 step 470
Timebase
Range
Smoothing
Reference
47 readings, 60 min/div60 min/div30 min/div
0123400:0060:00120:00180:00phase starting, 0.0 s to 59.6 sphase training, 59.6 s to 12:31−2.6742since first2.8956Lowest last train loss across this project's 12 runs, run 11lowest last 0.0007, run 11failed here7274 steps0.22140123400:0060:00120:00180:00phase starting, 0.0 s to 59.6 sphase training, 59.6 s to 12:312.8956Lowest last train loss across this project's 12 runs, run 11lowest last 0.0007, run 11failed here7274 steps0.2214▼ −2.67420123400:0030:0060:0090:00120:00150:00180:00phase starting, 0.0 s to 59.6 sphase training, 59.6 s to 12:31training2.8956Lowest last train loss across this project's 12 runs, run 11lowest last 0.0007, run 11failed here7274 stepsrun 50.2214▼ −2.6742

Elapsed

751.6 sto last reading

751.6 s to last reading

Set-up

58183 train, 500 eval

synthetic, batch size 16, epochs 2, max length 512, precision bf16-mixed, loss ce, optimizer adamw, scheduler linear, seed 0

Progress

6.5%

470 of 7274 steps, 0 of 2 epochs

Pace

1.50 s

per step

39.9 steps a minute

10.6 items a second

Learning rate

1.99e-5

▲ +1.94e-5 since first

On one axis

phase starting, 0.0 s to 59.6 sphase training, 59.6 s to 12:31lowest last 0.0007Train loss0.2214▼ −2.67423.000.002.89560.2214last ▼ −2.6742Learning rate1.99e-5▲ +1.94e-52.0e-50.0e+04.59e-71.99e-5last ▲ +1.94e-5Power, whole chip, CPU and GPU41 W▼ −81 W150 W0 W123 W41 Wlast ▼ −81 WHottest, one board zone98 °C▲ +9 °C100 °C80 °C89 °C98 °Clast ▲ +9 °C00:0030:0060:0090:00120:00150:00180:00
GPUpower: whole chip, CPU and GPUtemperature: one board zone

GPU 0

Wwhole chip, CPU and GPU

41 W whole chip, CPU and GPU

use0%
memory, whole GPU0.1 GB

peak 31.5 GB, whole GPU

98 °C one board zone

history 0 to 124 W

Input

  • Runs128 runs
  • Agentsnone showing
  • Treesnone showing
  • Projectsnone showing

Check

Every entry is checked on its way out; some are held for review.

Not shown on any run: data samples, predictions, raw logs, file paths, machine names

Delay

2 min, set

Watching the agents work, about 2 min behind.

One way

Nothing here reaches the agents.

This page

This page updates by itself.

visits counted with Google Analytics · choose

Runsheight is eval accuracyall recordings ▸ | finished× failed■ stopped○ lost● live
1.000.851.000.85finished, eval accuracy 0.9000, config 1 of 9 (learning rate 0.0001, batch 16), seed 0, Wed 18 Nov, 12:57 UTCfinished, eval accuracy 0.9200, config 2 of 9 (learning rate 0.00002, batch 16), seed 0, Wed 18 Nov, 14:41 UTCfinished, eval accuracy 0.9020, config 3 of 9 (learning rate 0.00005, batch 16), seed 0, Wed 18 Nov, 16:41 UTCfinished, eval accuracy 0.9080, config 4 of 9 (batch 16), seed 1 of 2, Thu 19 Nov, 00:21 UTCfinished, eval accuracy 0.9160, config 4 of 9 (batch 16), seed 2 of 2, Thu 19 Nov, 02:14 UTCfailed at step 470 of 7274, eval accuracy -, config 5 of 9 (batch 16), seed 0, Fri 20 Nov, 13:18 UTCfinished, eval accuracy 0.9600, config 5 of 9 (batch 16), seed 0, Sat 21 Nov, 01:57 UTCfinished, eval accuracy 0.8876, config 6 of 9 (learning rate 0.0001, batch 16), seed 0, Sat 21 Nov, 12:36 UTCfinished, eval accuracy 0.9191, config 7 of 9 (learning rate 0.00002, batch 16), seed 0, Sat 21 Nov, 13:15 UTCfinished, eval accuracy 0.8944, config 8 of 9 (learning rate 0.00005, batch 16), seed 0, Sat 21 Nov, 13:54 UTCfinished, eval accuracy 0.8854, config 9 of 9 (batch 16), seed 1 of 2, Sat 21 Nov, 19:01 UTCfinished, eval accuracy 0.8921, config 9 of 9 (batch 16), seed 2 of 2, Sat 21 Nov, 19:40 UTC1.000.85finished, eval accuracy 0.9000, config 1 of 9 (learning rate 0.0001, batch 16), seed 0, Wed 18 Nov, 12:57 UTCfinished, eval accuracy 0.9200, config 2 of 9 (learning rate 0.00002, batch 16), seed 0, Wed 18 Nov, 14:41 UTCfinished, eval accuracy 0.9020, config 3 of 9 (learning rate 0.00005, batch 16), seed 0, Wed 18 Nov, 16:41 UTCfinished, eval accuracy 0.9080, config 4 of 9 (batch 16), seed 1 of 2, Thu 19 Nov, 00:21 UTCfinished, eval accuracy 0.9160, config 4 of 9 (batch 16), seed 2 of 2, Thu 19 Nov, 02:14 UTCfailed at step 470 of 7274, eval accuracy -, config 5 of 9 (batch 16), seed 0, Fri 20 Nov, 13:18 UTCfinished, eval accuracy 0.9600, config 5 of 9 (batch 16), seed 0, Sat 21 Nov, 01:57 UTCfinished, eval accuracy 0.8876, config 6 of 9 (learning rate 0.0001, batch 16), seed 0, Sat 21 Nov, 12:36 UTCfinished, eval accuracy 0.9191, config 7 of 9 (learning rate 0.00002, batch 16), seed 0, Sat 21 Nov, 13:15 UTCfinished, eval accuracy 0.8944, config 8 of 9 (learning rate 0.00005, batch 16), seed 0, Sat 21 Nov, 13:54 UTCfinished, eval accuracy 0.8854, config 9 of 9 (batch 16), seed 1 of 2, Sat 21 Nov, 19:01 UTCfinished, eval accuracy 0.8921, config 9 of 9 (batch 16), seed 2 of 2, Sat 21 Nov, 19:40 UTC

6 of 12 failed at step 470 of 7274, -, config 5 of 9 (batch 16), seed 0, Fri 20 Nov, 13:18 UTC

Runs as a table
StartedStatusEval accuracyStepsConfig
Sat 21 Nov, 19:40 UTC| finished0.89211032 of 1032batch 16, seed 2 of 2
Sat 21 Nov, 19:01 UTC| finished0.88541032 of 1032batch 16, seed 1 of 2
Sat 21 Nov, 13:54 UTC| finished0.89441032 of 1032learning rate 0.00005, batch 16, seed 0
Sat 21 Nov, 13:15 UTC| finished0.91911032 of 1032learning rate 0.00002, batch 16, seed 0
Sat 21 Nov, 12:36 UTC| finished0.88761032 of 1032learning rate 0.0001, batch 16, seed 0
Sat 21 Nov, 01:57 UTC| finished0.96007274 of 7274batch 16, seed 0
Fri 20 Nov, 13:18 UTC× failed-470 of 7274batch 16, seed 0
Thu 19 Nov, 02:14 UTC| finished0.91601125 of 1125batch 16, seed 2 of 2
Thu 19 Nov, 00:21 UTC| finished0.90801125 of 1125batch 16, seed 1 of 2
Wed 18 Nov, 16:41 UTC| finished0.90201125 of 1125learning rate 0.00005, batch 16, seed 0
Wed 18 Nov, 14:41 UTC| finished0.92001125 of 1125learning rate 0.00002, batch 16, seed 0
Wed 18 Nov, 12:57 UTC| finished0.90001125 of 1125learning rate 0.0001, batch 16, seed 0

Plate

Model
answerdotai/ModernBERT-large, 395M parameters revision 45bb465
Data
Familiar decisions, full pool, synthetic58183 train500 eval
Config
batch size 16, epochs 2, max length 512, precision bf16-mixed, loss ce, optimizer adamw, scheduler linear, seed 0
Hardware
Radeon 8060S, ROCm

Log

Since startMark
0.0 sphase starting
3.7 swent public
59.6 sphase training

Seeds

SeedStatusEval accuracy
seed 0× failed-
seed 0| finished0.9600

Readings

47 rows as a table
StepSecondsTrain lossLearning rate
1059.62.89564.59e-7
2074.52.24279.17e-7
3089.21.21951.38e-6
40104.11.52021.83e-6
50118.91.45192.29e-6
60134.01.50582.75e-6
70149.22.15433.21e-6
80164.31.80573.67e-6
90179.31.52254.13e-6
100194.41.42574.59e-6
110209.51.72755.05e-6
120224.31.39775.50e-6
130239.30.90985.96e-6
140254.41.10206.42e-6
150269.50.98156.88e-6
160284.51.27007.34e-6
170299.41.60297.80e-6
180314.62.09768.26e-6
190329.61.59648.72e-6
200344.91.52469.17e-6
210359.81.62139.63e-6
220374.81.03111.01e-5
230389.91.10131.06e-5
240404.80.57201.10e-5
250419.80.99231.15e-5
260434.71.07501.19e-5
270449.51.17871.24e-5
280464.70.38521.28e-5
290479.81.13241.33e-5
300494.81.09661.38e-5
310509.91.10321.42e-5
320524.80.65671.47e-5
330540.01.03481.51e-5
340554.90.58761.56e-5
350570.10.33801.61e-5
360585.20.66701.65e-5
370600.30.61341.70e-5
380615.70.85311.74e-5
390630.90.55191.79e-5
400645.91.71251.83e-5
410660.90.62741.88e-5
420675.90.81891.93e-5
430691.11.09961.97e-5
440706.10.65282.00e-5
450721.50.24442.00e-5
460736.50.46871.99e-5
470751.60.22141.99e-5

GPU readings

47 rows as a table
StepGPUUse %Power W, whole chip, CPU and GPUMemory GB, whole GPUTemp C, one board zone
100100122.620.688.8
200100118.520.684.8
500100114.223.987.8
7000123.731.588.8
900100113.631.590.8
11000122.831.591.8
1300100115.131.592.8
15000122.731.593.8
1800100115.431.593.8
2000100123.531.594.8
2200100115.431.594.8
24000119.331.594.8
2600100118.531.594.8
28000119.231.593.8
3000100117.231.594.8
33000119.131.594.8
3500100119.031.594.8
3700100119.331.594.8
3900100116.031.594.8
41000116.431.594.8
43000119.331.595.8
4600100113.831.595.8
4700041.30.197.8

Worked out

ReadoutValueWorked out from
Run complete6.5%470 steps over 7274
Step time1.50 s692.0 s over 460 steps, between readings. Startup and intervals that hold an eval are left out
Steps per minute39.960 s over the step time of 1.50 s
Train loss since first reading−2.67420.2214 at step 470 minus 2.8956 at step 10
Lowest train loss0.2214Lowest of 47 readings, at step 470
Lowest last train loss0.0007Lowest last train loss across this project's 12 runs, run 11
Items per second10.61 step over 1.50 s, times batch size 16
Train loss against run 5−0.03970.2214 minus 0.2610 on run 5, both at step 470
Learning rate since first reading+1.94e-51.99e-5 minus 4.59e-7
Peak memory per GPU, whole GPU31.5 GBHighest sensor reading of any GPU, GPU 0, counting every process on it