CITADEL
StandbyNothing is live right now.
Eval accuracylast measured Sun 22 Nov, 12:01 UTC
%
0.8944398 of 445 eval, 47 wrong
▲ +6.74 pts across 3 evals▲ +0.68 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, 11:44 UTC
ran for
12:04
config
learning rate 0.00002, batch 16, seed 0

run5/10step1032/1032epoch3/3config5/8seed0lr0.00002batch16modeltasksource-decider-nano 149Mrev2c7c01fitems5503+445

Loss train 0.0428 ▼ −2.3889 lowest 0.0170 at step 980 run before, run 4 ▼ −0.2192 step 1032
Timebase
Range
Smoothing
Reference
104 readings, 5 min/div5 min/div2 min/div
012300:0005:0010:00phase starting, 0.0 s to 21.0 sphase training, 21.0 s to 12:04epoch 1, 04:15epoch 2, 08:10epoch 3, 11:57trainingepoch 1epoch 2−2.3889since first2.4317Lowest last train loss across this project's 10 runs, run 7lowest last 0.0018, run 70.0428012300:0005:0010:00phase starting, 0.0 s to 21.0 sphase training, 21.0 s to 12:04epoch 1, 04:15epoch 2, 08:10epoch 3, 11:57trainingepoch 1epoch 22.4317Lowest last train loss across this project's 10 runs, run 7lowest last 0.0018, run 70.0428▼ −2.3889012300:0002:0004:0006:0008:0010:0012:00phase starting, 0.0 s to 21.0 sphase training, 21.0 s to 12:04epoch 1, 04:15epoch 2, 08:10epoch 3, 11:57trainingepoch 1epoch 22.4317Lowest last train loss across this project's 10 runs, run 7lowest last 0.0018, run 70.0428▼ −2.3889

Eval measures

Elapsed

231.0 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

+0.0068

▲ on the eval before

eval 1 0.8270, eval 3 0.8944

Against the other runs

8 of 10

by last eval accuracy

best so far 0.9360, run 2

▼ −0.0022 against run 4

Set-up

5503 train, 445 eval

synthetic, learning rate 0.00002, batch size 16, epochs 3, max length 512, precision bf16-mixed, loss ce, optimizer adamw, scheduler linear, seed 0

Learning rate

0.00e+0

▼ −3.28e-6 since first

On one axis

phase starting, 0.0 s to 21.0 sphase training, 21.0 s to 12:04E1E2E3traininglowest last 0.0018Train loss0.0428▼ −2.38893.000.002.43170.0428last ▼ −2.3889best 0.9360, run 2Last-value range, 10 runs, 0.8854 to 0.9360Eval accuracy0.8944▲ +0.06740.950.800.82700.8944last ▲ +0.0674Learning rate0.00e+0▼ −3.28e-62.0e-50.0e+03.28e-6last 0.00e+0Power, whole chip, CPU and GPU117 W▲ +1 W140 W110 W117 Wlast 117 WHottest, one board zone91 °C0 °C92 °C84 °C91 °Clast 91 °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

use88%
memory, whole GPU1.1 GB

peak 12.5 GB, whole GPU

91 °C one board zone

history 0 to 131 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

5 of 10 finished, 0.8944, config 5 of 8 (learning rate 0.00002, batch 16), seed 0, Sun 22 Nov, 11:44 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
learning rate 0.00002, batch size 16, epochs 3, 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
9.6 swent public
21.0 sphase training
250.1 seval 1
255.2 sepoch 1
487.7 seval 2
490.2 sepoch 2
717.3 sepoch 3
724.6 seval 3

Readings

106 rows as a table

rows 1 to 100 of 106

StepSecondsTrain lossEval accuracyLearning rate
1021.02.4317-3.28e-6
2027.72.1823-6.56e-6
3034.22.1611-9.84e-6
4040.91.5510-1.31e-5
5047.60.9914-1.64e-5
6054.21.6705-1.97e-5
7060.91.2778-1.98e-5
8067.51.5211-1.96e-5
9074.11.5380-1.94e-5
10080.80.7588-1.92e-5
11087.40.9922-1.90e-5
12093.90.8055-1.88e-5
130100.60.5962-1.86e-5
140107.20.9123-1.84e-5
150113.70.7988-1.82e-5
160120.30.3718-1.80e-5
170127.00.8343-1.78e-5
180133.80.5443-1.75e-5
190140.50.9502-1.73e-5
200147.20.2086-1.71e-5
210153.80.7842-1.69e-5
220160.50.6410-1.67e-5
230167.11.0682-1.65e-5
240173.80.2824-1.63e-5
250180.40.3912-1.61e-5
260187.01.1094-1.59e-5
270193.70.4582-1.57e-5
280200.21.1277-1.55e-5
290206.80.4768-1.53e-5
300213.50.2497-1.51e-5
310220.20.8488-1.49e-5
320226.80.6988-1.47e-5
330233.50.3944-1.45e-5
340240.20.3907-1.43e-5
344250.1-0.8270-
350255.20.6792-1.40e-5
360261.80.1033-1.38e-5
370268.50.9072-1.36e-5
380275.20.4078-1.34e-5
390281.90.4242-1.32e-5
400288.50.9455-1.30e-5
410295.20.6750-1.28e-5
420301.70.0210-1.26e-5
430308.40.2891-1.24e-5
440315.10.1956-1.22e-5
450321.70.1579-1.20e-5
460328.40.1835-1.18e-5
470335.00.0812-1.16e-5
480341.70.0469-1.14e-5
490348.40.5105-1.12e-5
500355.10.1158-1.10e-5
510361.80.7206-1.08e-5
520368.40.3766-1.05e-5
530375.20.4601-1.03e-5
540381.80.9628-1.01e-5
550388.50.6284-9.93e-6
560395.10.7309-9.72e-6
570401.80.2142-9.52e-6
580408.40.2034-9.31e-6
590415.10.3544-9.10e-6
600421.80.6253-8.90e-6
610428.50.1965-8.69e-6
620435.20.1152-8.49e-6
630441.80.2727-8.28e-6
640448.50.1933-8.07e-6
650455.10.2680-7.87e-6
660461.70.3507-7.66e-6
670468.40.2208-7.46e-6
680475.10.2686-7.25e-6
688487.7-0.8876-
690490.20.1473-7.04e-6
700496.70.2640-6.84e-6
710503.40.1320-6.63e-6
720509.90.0388-6.43e-6
730516.60.2400-6.22e-6
740523.30.1001-6.01e-6
750530.00.1256-5.81e-6
760536.70.5434-5.60e-6
770543.40.1139-5.40e-6
780550.00.1221-5.19e-6
790556.70.3560-4.98e-6
800563.30.0256-4.78e-6
810570.10.0817-4.57e-6
820576.70.0224-4.37e-6
830583.30.0317-4.16e-6
840589.90.0558-3.95e-6
850596.60.2283-3.75e-6
860603.10.0395-3.54e-6
870609.80.0535-3.34e-6
880616.40.2394-3.13e-6
890623.10.0251-2.92e-6
900629.80.3460-2.72e-6
910636.40.0754-2.51e-6
920643.00.1034-2.31e-6
930649.70.0701-2.10e-6
940656.30.0705-1.89e-6
950662.90.1393-1.69e-6
960669.50.0415-1.48e-6
970676.10.1012-1.28e-6
980682.80.0170-1.07e-6

GPU readings

106 rows as a table
StepGPUUse %Power W, whole chip, CPU and GPUMemory GB, whole GPUTemp C, one board zone
100100116.58.590.8
600100115.48.585.8
1100100115.610.686.8
1600100116.410.687.8
210063119.812.587.8
260067120.412.588.8
310062121.512.588.8
344067131.412.590.8
3900100118.212.588.8
4400100116.712.588.8
490070115.712.588.8
540067120.912.589.8
5900100116.112.589.8
630066119.412.589.8
680086118.912.589.8
7200100115.912.588.8
7700100117.012.588.8
8200100115.112.588.8
8700100114.412.588.8
910066119.812.587.8
960065120.012.587.8
1010065119.912.587.8
1032088117.31.190.8

Worked out

ReadoutValueWorked out from
Run complete100%1032 steps over 1032
Step time0.67 s672.4 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 interval231.0 sMean of 2 gaps between epoch marks
Train loss since first reading−2.38890.0428 at step 1032 minus 2.4317 at step 10
Lowest train loss0.0170Lowest of 104 readings, at step 980
Eval accuracy, as shown89.4%0.8944 times 100, to one decimal. Changes on that plate are percentage points
Eval accuracy since first eval+0.06740.8944 minus 0.8270
Eval change+0.0068Eval 3 at 0.8944 minus eval 2 at 0.8876
Wrong answers47445 eval items minus 398 correct. 398 over 445 is the only count that gives 0.8944
Rank8 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−0.00220.8944 minus 0.8966 on run 4
Lowest last train loss0.0018Lowest last train loss across this project's 10 runs, run 7
Items per second24.01 step over 0.67 s, times batch size 16
Train loss against run 4−0.21920.0428 minus 0.2620 on run 4, both at step 1032
Learning rate since first reading−3.28e-60.00e+0 minus 3.28e-6
Peak memory per GPU, whole GPU12.5 GBHighest sensor reading of any GPU, GPU 0, counting every process on it