CITADEL
StandbyNothing is live right now.
Eval accuracylast measured Sun 22 Nov, 13:49 UTC
%
0.8966399 of 445 eval, 46 wrong
▲ +9.88 pts across 3 evals▲ +3.14 pts on the eval before
0%25%50%75%100%0%25%50%75%100%0%25%50%75%100%

Familiar decider: decider-nano's encoder

finishedRadeon 8060S

One of the small models that decide whether Familiar, a desktop companion, reacts to something on the PC, and how. Trained on the same 6,000 synthetic decision points as every candidate, two ways: the reaction itself, and the plain-case intent that rules adjust for her traits.

started
Sun 22 Nov, 13:32 UTC
ran for
12:04
config
batch 16, seed 1 of 2

run9/10step1032/1032epoch3/3config8/8seed1/2batch16modeltasksource-decider-nano 149Mrev2c7c01fitems5503+445

Loss train 0.0028 ▼ −1.5049 lowest 0.0028 at step 1032 run before, run 8 ▼ −0.1801 step 1030
Timebase
Range
Smoothing
Reference
104 readings, 5 min/div5 min/div2 min/div
0123400:0005:0010:00phase starting, 0.0 s to 20.7 sphase training, 20.7 s to 12:04epoch 1, 04:15epoch 2, 08:09epoch 3, 11:57trainingepoch 1epoch 21.5077Lowest last train loss across this project's 10 runs, run 7lowest last 0.0018, run 70.00280123400:0005:0010:00phase starting, 0.0 s to 20.7 sphase training, 20.7 s to 12:04epoch 1, 04:15epoch 2, 08:09epoch 3, 11:57trainingepoch 1epoch 21.5077Lowest last train loss across this project's 10 runs, run 7lowest last 0.0018, run 70.0028▼ −1.50490123400:0002:0004:0006:0008:0010:0012:00phase starting, 0.0 s to 20.7 sphase training, 20.7 s to 12:04epoch 1, 04:15epoch 2, 08:09epoch 3, 11:57trainingepoch 1epoch 21.5077Lowest last train loss across this project's 10 runs, run 7lowest last 0.0018, run 70.0028▼ −1.5049

Eval measures

Elapsed

231.2 s

epoch interval

1032 of 1032 steps, 3 of 3 epochs, 100%

0.67 s per step, 89.9 steps a minute

Eval accuracy by eval

+0.0314

▲ on the eval before

eval 1 0.7978, eval 3 0.8966

Against the other runs

7 of 10

by last eval accuracy

best so far 0.9360, run 2

▼ −0.0214 against run 8

Set-up

5503 train, 445 eval

synthetic, batch size 16, epochs 3, max length 512, precision bf16-mixed, loss ce, optimizer adamw, scheduler linear, seed 1

Seed spread, observed

0.0112
0.850.90

2 finished seeds

Learning rate

0.00e+0

▼ −1.64e-5 since first

On one axis

phase starting, 0.0 s to 20.7 sphase training, 20.7 s to 12:04E1E2E3traininglowest last 0.0018Train loss0.0028▼ −1.50493.000.001.50770.0028last ▼ −1.5049best 0.9360, run 2Last-value range, 10 runs, 0.8854 to 0.9360Eval accuracy0.8966▲ +0.09880.950.750.79780.8966last ▲ +0.0988Learning rate0.00e+0▼ −1.64e-51.0e-40.0e+01.64e-5last 0.00e+0Power, whole chip, CPU and GPU117 W▼ −3 W130 W110 W120 Wlast 117 WHottest, one board zone92 °C▲ +1 °C92 °C86 °C91 °Clast 92 °C00:0002:0004:0006:0008:0010:0012:00
GPUpower: whole chip, CPU and GPUtemperature: one board zone

GPU 0

Wwhole chip, CPU and GPU

117 W whole chip, CPU and GPU

use84%
memory, whole GPU1.1 GB

peak 13.5 GB, whole GPU

92 °C one board zone

history 0 to 130 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
0.950.850.950.85finished, eval accuracy 0.9100, config 1 of 8 (learning rate 0.0001, batch 16), seed 0, Sun 22 Nov, 10:33 UTCfinished, eval accuracy 0.9360, config 2 of 8 (learning rate 0.00002, batch 16), seed 0, Sun 22 Nov, 10:51 UTCfinished, eval accuracy 0.9220, config 3 of 8 (learning rate 0.00005, batch 16), seed 0, Sun 22 Nov, 11:09 UTCfinished, eval accuracy 0.8966, config 4 of 8 (learning rate 0.0001, batch 16), seed 0, Sun 22 Nov, 11:27 UTCfinished, eval accuracy 0.8944, config 5 of 8 (learning rate 0.00002, batch 16), seed 0, Sun 22 Nov, 11:44 UTCfinished, eval accuracy 0.8944, config 6 of 8 (learning rate 0.00005, batch 16), seed 0, Sun 22 Nov, 12:01 UTCfinished, eval accuracy 0.9140, config 7 of 8 (batch 16), seed 1 of 2, Sun 22 Nov, 12:56 UTCfinished, eval accuracy 0.9180, config 7 of 8 (batch 16), seed 2 of 2, Sun 22 Nov, 13:14 UTCfinished, eval accuracy 0.8966, config 8 of 8 (batch 16), seed 1 of 2, Sun 22 Nov, 13:32 UTCfinished, eval accuracy 0.8854, config 8 of 8 (batch 16), seed 2 of 2, Sun 22 Nov, 13:49 UTC0.950.85finished, eval accuracy 0.9100, config 1 of 8 (learning rate 0.0001, batch 16), seed 0, Sun 22 Nov, 10:33 UTCfinished, eval accuracy 0.9360, config 2 of 8 (learning rate 0.00002, batch 16), seed 0, Sun 22 Nov, 10:51 UTCfinished, eval accuracy 0.9220, config 3 of 8 (learning rate 0.00005, batch 16), seed 0, Sun 22 Nov, 11:09 UTCfinished, eval accuracy 0.8966, config 4 of 8 (learning rate 0.0001, batch 16), seed 0, Sun 22 Nov, 11:27 UTCfinished, eval accuracy 0.8944, config 5 of 8 (learning rate 0.00002, batch 16), seed 0, Sun 22 Nov, 11:44 UTCfinished, eval accuracy 0.8944, config 6 of 8 (learning rate 0.00005, batch 16), seed 0, Sun 22 Nov, 12:01 UTCfinished, eval accuracy 0.9140, config 7 of 8 (batch 16), seed 1 of 2, Sun 22 Nov, 12:56 UTCfinished, eval accuracy 0.9180, config 7 of 8 (batch 16), seed 2 of 2, Sun 22 Nov, 13:14 UTCfinished, eval accuracy 0.8966, config 8 of 8 (batch 16), seed 1 of 2, Sun 22 Nov, 13:32 UTCfinished, eval accuracy 0.8854, config 8 of 8 (batch 16), seed 2 of 2, Sun 22 Nov, 13:49 UTC

9 of 10 finished, 0.8966, config 8 of 8 (batch 16), seed 1 of 2, Sun 22 Nov, 13:32 UTC

Runs as a table
StartedStatusEval accuracyStepsConfig
Sun 22 Nov, 13:49 UTC| finished0.88541032 of 1032batch 16, seed 2 of 2
Sun 22 Nov, 13:32 UTC| finished0.89661032 of 1032batch 16, seed 1 of 2
Sun 22 Nov, 13:14 UTC| finished0.91801125 of 1125batch 16, seed 2 of 2
Sun 22 Nov, 12:56 UTC| finished0.91401125 of 1125batch 16, seed 1 of 2
Sun 22 Nov, 12:01 UTC| finished0.89441032 of 1032learning rate 0.00005, batch 16, seed 0
Sun 22 Nov, 11:44 UTC| finished0.89441032 of 1032learning rate 0.00002, batch 16, seed 0
Sun 22 Nov, 11:27 UTC| finished0.89661032 of 1032learning rate 0.0001, batch 16, seed 0
Sun 22 Nov, 11:09 UTC| finished0.92201125 of 1125learning rate 0.00005, batch 16, seed 0
Sun 22 Nov, 10:51 UTC| finished0.93601125 of 1125learning rate 0.00002, batch 16, seed 0
Sun 22 Nov, 10:33 UTC| finished0.91001125 of 1125learning rate 0.0001, batch 16, seed 0

Plate

Model
tasksource/tasksource-decider-nano, 149M parameters revision 2c7c01f
Data
Familiar decisions, shared draw, plain-case intents, synthetic5503 train445 eval
Config
batch size 16, epochs 3, max length 512, precision bf16-mixed, loss ce, optimizer adamw, scheduler linear, seed 1
Hardware
Radeon 8060S, ROCm

Log

Since startMark
0.0 sphase starting
2.7 swent public
20.7 sphase training
250.1 seval 1
255.1 sepoch 1
487.4 seval 2
489.9 sepoch 2
717.4 sepoch 3
724.7 seval 3

Seeds

SeedStatusEval accuracy
seed 1| finished0.8966
seed 2| finished0.8854

observed spread 0.0112, 2 finished seeds

Readings

106 rows as a table

rows 1 to 100 of 106

StepSecondsTrain lossEval accuracyLearning rate
1020.71.5077-1.64e-5
2027.41.8104-3.28e-5
3033.91.4948-4.92e-5
4040.61.4314-6.56e-5
5047.30.9028-8.20e-5
6053.91.1829-9.84e-5
7060.61.5594-9.91e-5
8067.21.3075-9.80e-5
9073.80.5606-9.70e-5
10080.40.8723-9.60e-5
11086.91.0362-9.50e-5
12093.61.0893-9.39e-5
130100.30.9356-9.29e-5
140107.01.1296-9.19e-5
150113.70.8664-9.08e-5
160120.30.7511-8.98e-5
170127.00.3690-8.88e-5
180133.50.9781-8.77e-5
190140.30.6556-8.67e-5
200147.00.7328-8.57e-5
210153.71.0140-8.47e-5
220160.30.6123-8.36e-5
230166.90.3406-8.26e-5
240173.50.7160-8.16e-5
250180.20.5660-8.05e-5
260186.90.3289-7.95e-5
270193.50.7838-7.85e-5
280200.20.7268-7.74e-5
290206.90.2067-7.64e-5
300213.60.3300-7.54e-5
310220.21.0474-7.44e-5
320226.90.0979-7.33e-5
330233.50.8010-7.23e-5
340240.20.4727-7.13e-5
344250.1-0.7978-
350255.10.4344-7.02e-5
360261.80.2274-6.92e-5
370268.40.6672-6.82e-5
380275.10.2867-6.71e-5
390281.80.1644-6.61e-5
400288.40.1935-6.51e-5
410295.00.1686-6.41e-5
420301.70.3206-6.30e-5
430308.40.4307-6.20e-5
440315.10.0981-6.10e-5
450321.80.7209-5.99e-5
460328.30.3866-5.89e-5
470335.00.7593-5.79e-5
480341.71.1153-5.68e-5
490348.30.7581-5.58e-5
500354.90.4794-5.48e-5
510361.60.6460-5.38e-5
520368.20.0730-5.27e-5
530374.81.2314-5.17e-5
540381.50.5343-5.07e-5
550388.10.3960-4.96e-5
560394.90.4131-4.86e-5
570401.50.1695-4.76e-5
580408.10.0343-4.65e-5
590414.80.7703-4.55e-5
600421.50.0790-4.45e-5
610428.20.5540-4.35e-5
620434.90.1613-4.24e-5
630441.60.0836-4.14e-5
640448.30.1075-4.04e-5
650454.90.2235-3.93e-5
660461.50.2779-3.83e-5
670468.10.2227-3.73e-5
680474.80.3270-3.63e-5
688487.4-0.8652-
690489.91.0125-3.52e-5
700496.30.1272-3.42e-5
710503.00.0380-3.32e-5
720509.60.2120-3.21e-5
730516.20.4470-3.11e-5
740522.90.4632-3.01e-5
750529.50.2304-2.90e-5
760536.20.1985-2.80e-5
770542.90.3154-2.70e-5
780549.60.0908-2.60e-5
790556.30.0223-2.49e-5
800563.00.2498-2.39e-5
810569.70.4814-2.29e-5
820576.20.1088-2.18e-5
830582.90.0993-2.08e-5
840589.50.2436-1.98e-5
850596.10.0437-1.87e-5
860602.90.2475-1.77e-5
870609.40.1504-1.67e-5
880616.10.2018-1.57e-5
890622.80.0037-1.46e-5
900629.50.2168-1.36e-5
910636.00.1125-1.26e-5
920642.70.1318-1.15e-5
930649.50.1019-1.05e-5
940656.20.0535-9.47e-6
950662.90.1556-8.44e-6
960669.60.6268-7.42e-6
970676.30.4561-6.39e-6
980682.90.4594-5.36e-6

GPU readings

106 rows as a table
StepGPUUse %Power W, whole chip, CPU and GPUMemory GB, whole GPUTemp C, one board zone
10063120.09.590.8
600100114.713.586.8
1100100115.013.587.8
160064120.413.588.8
2100100115.413.588.8
260067116.013.589.8
310064120.513.589.8
3440100129.513.591.8
390067118.213.589.8
4400100116.713.588.8
4900100116.513.588.8
5400100115.713.589.8
590065116.913.589.8
6400100116.613.589.8
6800100118.413.589.8
720066120.413.589.8
7700100115.813.589.8
820063117.913.589.8
870067120.113.588.8
9200100118.413.588.8
960063119.413.588.8
1010065118.913.588.8
1032084117.41.191.8

Worked out

ReadoutValueWorked out from
Run complete100%1032 steps over 1032
Step time0.67 s672.8 s over 1008 steps, between readings. Startup and intervals that hold an eval are left out
Steps per minute89.960 s over the step time of 0.67 s
Epoch interval231.2 sMean of 2 gaps between epoch marks
Train loss since first reading−1.50490.0028 at step 1032 minus 1.5077 at step 10
Lowest train loss0.0028Lowest of 104 readings, at step 1032
Eval accuracy, as shown89.7%0.8966 times 100, to one decimal. Changes on that plate are percentage points
Eval accuracy since first eval+0.09880.8966 minus 0.7978
Eval change+0.0314Eval 3 at 0.8966 minus eval 2 at 0.8652
Wrong answers46445 eval items minus 399 correct. 399 over 445 is the only count that gives 0.8966
Rank7 of 10By last eval accuracy, across this project's 10 runs with an eval
Best so far0.9360Highest last eval accuracy, run 2
Against the run before−0.02140.8966 minus 0.9180 on run 8
Lowest last train loss0.0018Lowest last train loss across this project's 10 runs, run 7
Seed spread, observed0.0112High 0.8966 minus low 0.8854 across 2 finished seeds
Items per second24.01 step over 0.67 s, times batch size 16
Train loss against run 8−0.18010.0353 minus 0.2154 on run 8, both at step 1030
Learning rate since first reading−1.64e-50.00e+0 minus 1.64e-5
Peak memory per GPU, whole GPU13.5 GBHighest sensor reading of any GPU, GPU 0, counting every process on it