CITADEL
StandbyNothing is live right now.
Eval accuracylast measured Sat 21 Nov, 05:11 UTC
%
0.9600480 of 500 eval, 20 wrong
▲ +1.40 pts across 2 evals▲ +1.40 pts on the eval before
0%25%50%75%100%0%25%50%75%100%0%25%50%75%100%

Familiar decider: ModernBERT-large

finishedRadeon 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
Sat 21 Nov, 01:57 UTC
ran for
184:20
config
batch 16, seed 0

run7/12step7274/7274epoch2/2config5/9seed0batch16modelModernBERT-large 395Mrev45bb465items58183+500

Loss train 0.0121 ▼ −2.8835 lowest 0.0002 at step 6820 run before, run 6 ▲ +0.0000 step 470
Timebase
Range
Smoothing
Reference
728 readings, 60 min/div60 min/div30 min/div
0123400:0060:00120:00180:00phase starting, 0.0 s to 60.1 sphase training, 60.1 s to 184:20epoch 1, 92:37epoch 2, 184:02trainingepoch 1−2.8835since first2.8956Lowest last train loss across this project's 12 runs, run 11lowest last 0.0007, run 110.01210123400:0060:00120:00180:00phase starting, 0.0 s to 60.1 sphase training, 60.1 s to 184:20epoch 1, 92:37epoch 2, 184:02trainingepoch 12.8956Lowest last train loss across this project's 12 runs, run 11lowest last 0.0007, run 110.0121▼ −2.88350123400:0030:0060:0090:00120:00150:00180:00phase starting, 0.0 s to 60.1 sphase training, 60.1 s to 184:20epoch 1, 92:37epoch 2, 184:02trainingepoch 12.8956Lowest last train loss across this project's 12 runs, run 11lowest last 0.0007, run 110.0121▼ −2.8835

Eval measures

Elapsed

5484.8 s

epoch interval

5551.4 sto eval 15491.4 sto epoch 2

5551.4 s to eval 1, 5491.4 s to epoch 2

7274 of 7274 steps, 2 of 2 epochs, 100%

1.51 s per step, 39.7 steps a minute

Eval accuracy by eval

+0.0140

▲ on the eval before

eval 1 0.9460, eval 2 0.9600

Against the other runs

1 of 11

by last eval accuracy

best so far 0.9600, run 7

▲ +0.0440 against run 5

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

Learning rate

0.00e+0

▼ −4.59e-7 since first

On one axis

phase starting, 0.0 s to 60.1 sphase training, 60.1 s to 184:20E1E2traininglowest last 0.0007Train loss0.0121▼ −2.88353.000.002.89560.0121last ▼ −2.8835best 0.9600, run 7Last-value range, 11 runs, 0.8854 to 0.9600Eval accuracy0.9600▲ +0.01400.980.880.94600.9600last ▲ +0.0140Learning rate0.00e+0▼ −4.59e-72.0e-50.0e+04.59e-7last 0.00e+0Power, whole chip, CPU and GPU118 W▼ −4 W125 W110 W122 Wlast 118 WHottest, one board zone97 °C▲ +6 °C100 °C85 °C91 °Clast 97 °C00:0030:0060:0090:00120:00150:00180:00
GPUpower: whole chip, CPU and GPUtemperature: one board zone

GPU 0

Wwhole chip, CPU and GPU

118 W whole chip, CPU and GPU

use93%
memory, whole GPU2.1 GB

peak 42.9 GB, whole GPU

97 °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

7 of 12 finished, 0.9600, config 5 of 9 (batch 16), seed 0, Sat 21 Nov, 01:57 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
12.7 swent public
60.1 sphase training
5551.4 seval 1
5558.0 sepoch 1
11042.8 sepoch 2
11060.9 seval 2

Seeds

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

Readings

729 rows as a table

rows 1 to 100 of 729

StepSecondsTrain lossEval accuracyLearning rate
1060.12.8956-4.59e-7
2075.12.2427-9.17e-7
3089.81.2195-1.38e-6
40104.71.5202-1.83e-6
50119.61.4519-2.29e-6
60134.71.5058-2.75e-6
70150.02.1543-3.21e-6
80165.11.8057-3.67e-6
90180.21.5225-4.13e-6
100195.31.4257-4.59e-6
110210.41.7275-5.05e-6
120225.21.3977-5.50e-6
130240.20.9098-5.96e-6
140255.41.1020-6.42e-6
150270.50.9815-6.88e-6
160285.51.2700-7.34e-6
170300.41.6029-7.80e-6
180315.72.0976-8.26e-6
190330.81.5964-8.72e-6
200346.11.5246-9.17e-6
210361.01.6213-9.63e-6
220376.11.0311-1.01e-5
230391.11.1013-1.06e-5
240406.10.5720-1.10e-5
250421.10.9923-1.15e-5
260436.01.0750-1.19e-5
270450.91.1787-1.24e-5
280466.10.3852-1.28e-5
290481.11.1324-1.33e-5
300496.21.0966-1.38e-5
310511.21.1032-1.42e-5
320526.20.6567-1.47e-5
330541.31.0348-1.51e-5
340556.20.5876-1.56e-5
350571.40.3380-1.61e-5
360586.40.6670-1.65e-5
370601.50.6134-1.70e-5
380616.60.8531-1.74e-5
390631.50.5519-1.79e-5
400646.51.7125-1.83e-5
410661.60.6274-1.88e-5
420676.50.8189-1.93e-5
430691.61.0996-1.97e-5
440706.60.6528-2.00e-5
450721.90.2444-2.00e-5
460736.90.4687-1.99e-5
470752.00.2214-1.99e-5
480766.60.5642-1.99e-5
490781.80.5921-1.98e-5
500797.01.1310-1.98e-5
510812.30.5477-1.98e-5
520827.50.6262-1.98e-5
530842.50.1518-1.97e-5
540857.70.6119-1.97e-5
550873.00.6506-1.97e-5
560888.20.5905-1.96e-5
570903.30.5444-1.96e-5
580918.50.3649-1.96e-5
590933.70.4588-1.95e-5
600949.00.4510-1.95e-5
610963.90.6316-1.95e-5
620978.90.6979-1.95e-5
630994.00.2160-1.94e-5
6401009.20.2132-1.94e-5
6501024.20.4570-1.94e-5
6601039.20.3722-1.93e-5
6701054.40.3964-1.93e-5
6801069.40.0415-1.93e-5
6901084.60.4324-1.93e-5
7001099.50.5242-1.92e-5
7101114.60.5988-1.92e-5
7201129.80.2344-1.92e-5
7301145.11.1297-1.91e-5
7401160.30.7289-1.91e-5
7501175.30.8532-1.91e-5
7601190.40.1736-1.91e-5
7701205.30.2725-1.90e-5
7801220.41.0530-1.90e-5
7901235.50.4456-1.90e-5
8001251.00.2091-1.89e-5
8101266.30.1486-1.89e-5
8201281.50.5553-1.89e-5
8301296.50.4042-1.88e-5
8401311.70.3172-1.88e-5
8501326.60.3160-1.88e-5
8601341.70.7014-1.88e-5
8701356.70.3582-1.87e-5
8801371.70.3020-1.87e-5
8901386.90.6878-1.87e-5
9001402.40.5309-1.86e-5
9101417.30.0879-1.86e-5
9201432.30.2383-1.86e-5
9301447.60.0632-1.86e-5
9401462.60.3409-1.85e-5
9501477.61.7666-1.85e-5
9601492.80.2015-1.85e-5
9701507.90.2122-1.84e-5
9801523.00.3941-1.84e-5
9901538.40.2914-1.84e-5
10001553.80.9442-1.84e-5

GPU readings

729 rows as a table

rows 1 to 100 of 341

StepGPUUse %Power W, whole chip, CPU and GPUMemory GB, whole GPUTemp C, one board zone
1000122.220.690.8
200100122.520.685.8
500100114.923.988.8
7000122.031.590.8
900100113.231.591.8
1100100123.431.592.8
1300100114.131.593.8
1500100119.031.593.8
1700100113.731.594.8
2000100115.731.594.8
22000120.531.594.8
2400100117.431.594.8
26000114.131.595.8
2800100116.931.594.8
30000119.531.594.8
3200100116.431.594.8
35000119.731.595.8
37000119.931.595.8
3900100117.231.595.8
41000120.131.595.8
4300100116.431.595.8
45000120.831.595.8
4800100117.531.595.8
50000119.431.595.8
5200100113.531.595.8
5400100117.331.595.8
56000121.731.595.8
58000116.731.595.8
6100100116.231.595.8
63000114.531.595.8
6500100116.031.595.8
67000121.131.595.8
6900100115.831.595.8
71000122.131.595.8
73000118.931.595.8
7600100116.531.595.8
78000115.131.595.8
8000100112.831.595.8
8200100115.831.596.8
84000120.931.595.8
8600100116.631.596.8
88000119.631.595.8
9100100114.235.296.8
9300100117.935.296.8
95000114.935.295.8
9700100116.635.295.8
9900100122.635.295.8
101000118.738.996.8
10300100113.438.996.8
10500100115.938.995.8
108000115.438.995.8
11000100116.938.995.8
112000114.438.995.8
11400100116.038.995.8
116000121.838.996.8
11800100114.738.995.8
12000100122.938.995.8
12300100114.638.995.8
125000121.638.995.8
127000119.238.995.8
129000115.138.996.8
131000121.838.996.8
13300100115.838.996.8
135000122.638.996.8
13800100116.738.995.8
140000123.538.995.8
14200100113.738.995.8
14400100116.438.995.8
146000117.738.996.8
14800100116.238.995.8
150000118.938.995.8
15300100114.638.995.8
155000122.238.995.8
15700100114.538.995.8
15900100117.338.995.8
161000119.038.995.8
16300100116.138.995.8
165000120.038.996.8
16800100115.738.995.8
170000122.638.996.8
17200100114.438.995.8
174000122.838.996.8
17600100114.638.995.8
178000120.938.995.8
18100100115.438.995.8
183000120.238.996.8
18500100115.838.995.8
187000121.538.995.8
18900100115.438.995.8
191000121.138.995.8
19300100114.038.995.8
19600100116.638.995.8
198000123.438.995.8
20000100115.638.995.8
202000123.138.995.8
204000118.438.995.8
20600100122.738.995.8
20800100114.238.995.8
21100100119.238.995.8
213000120.238.995.8

Worked out

ReadoutValueWorked out from
Run complete100%7274 steps over 7274
Step time1.51 s10949.5 s over 7253 steps, between readings. Startup and intervals that hold an eval are left out
Steps per minute39.760 s over the step time of 1.51 s
Epoch interval5484.8 sThe one gap between epoch marks
Train loss since first reading−2.88350.0121 at step 7274 minus 2.8956 at step 10
Lowest train loss0.0002Lowest of 728 readings, at step 6820
Eval accuracy, as shown96.0%0.9600 times 100, to one decimal. Changes on that plate are percentage points
Eval accuracy since first eval+0.01400.9600 minus 0.9460
Eval change+0.0140Eval 2 at 0.9600 minus eval 1 at 0.9460
Wrong answers20500 eval items minus 480 correct. 480 over 500 is the only count that gives 0.9600
Rank1 of 11By last eval accuracy, across this project's 11 runs with an eval
Best so far0.9600Highest last eval accuracy, run 7
Against the run before+0.04400.9600 minus 0.9160 on run 5
Lowest last train loss0.0007Lowest last train loss across this project's 12 runs, run 11
Items per second10.61 step over 1.51 s, times batch size 16
Train loss against run 6+0.00000.2214 minus 0.2214 on run 6, both at step 470
Learning rate since first reading−4.59e-70.00e+0 minus 4.59e-7
Peak memory per GPU, whole GPU42.9 GBHighest sensor reading of any GPU, GPU 0, counting every process on it