CITADEL
StandbyNothing is live right now.
Eval accuracylast measured Mon 2 Nov, 14:29 UTC
0%25%50%75%100%0%25%50%75%100%0%25%50%75%100%

Claviger

finishedRadeon 8060S

A small model that reads one Discord message and decides whether an agent should reply, and which one. Fine-tuned from Laya on invented chat.

started
Mon 2 Nov, 14:22 UTC
ran for
06:29
config
batch 8, seed 1 of 3

run5/22step150/150epoch5/5config2/11seed1/3batch8modellaya 421Mrev7b928d8items479+130

Loss train 0.0044 ▼ −0.6407 eval 0.6370 lowest -0.0320 at step 100 run before, run 4 ▲ +0.0044 step 150
Timebase
Range
Smoothing
Reference
30 readings, 2 min/div2 min/div1 min/div
-1012300:0002:0004:0006:00phase starting, 0.0 s to 36.6 sphase training, 36.6 s to 06:27phase eval, 06:27 to 06:29epoch 2, 02:39epoch 3, 03:46epoch 4, 04:53epoch 5, 06:00trainingE2E3E4E50.6451Lowest last train loss across this project's 22 runs, run 19lowest last -0.0937, run 190.0044-1012300:0002:0004:0006:00phase starting, 0.0 s to 36.6 sphase training, 36.6 s to 06:27phase eval, 06:27 to 06:29epoch 1, 91.5 sepoch 2, 02:39epoch 3, 03:46epoch 4, 04:53epoch 5, 06:00trainingepoch 1epoch 2epoch 3epoch 4E50.6451Lowest last train loss across this project's 22 runs, run 19lowest last -0.0937, run 19eval 0.63700.0044▼ −0.6407-1012300:0001:0002:0003:0004:0005:0006:00phase starting, 0.0 s to 36.6 sphase training, 36.6 s to 06:27phase eval, 06:27 to 06:29epoch 1, 91.5 sepoch 2, 02:39epoch 3, 03:46epoch 4, 04:53epoch 5, 06:00startingtrainingepoch 1epoch 2epoch 3epoch 4epoch 50.6451Lowest last train loss across this project's 22 runs, run 19lowest last -0.0937, run 19eval 0.63700.0044▼ −0.6407

Eval measures

Eval losslower is better
0.6370
02 →
ECEcalibration error, lower is better
0.1450
01
Brierlower is better
0.2969
02 →

Elapsed

67.3 s

epoch interval

150 of 150 steps, 5 of 5 epochs, 100%

2.21 s per step, 27.2 steps a minute

Against the other runs

9 of 22

by last eval accuracy

best so far 0.9231, run 19

▲ +0.0539 against run 4

Train minus eval

−0.6326

step 150

train 0.0044, eval 0.6370

Set-up

479 train, 130 eval

synthetic, batch size 8, grad accum 2, epochs 5, seed 1

Seed spread, observed

0.0539
0.750.85

3 finished seeds

Samples per second

7.33

▲ +2.29 since first

Learning rate

1.32e-6

▼ −2.36e-5 since first

GPU memory, this run

6.5 GB

On one axis

phase starting, 0.0 s to 36.6 sphase training, 36.6 s to 06:27phase eval, 06:27 to 06:29E2E3E4E5traininglowest last -0.0937Train loss0.0044▼ −0.64073.00-1.000.64510.0044last ▼ −0.6407Samples per second7.33▲ +2.29855.03Learning rate1.32e-6▼ −2.36e-53.0e-50.0e+02.50e-5Power, whole chip, CPU and GPU27 W▼ −56 W100 W20 W83 W27 Wlast ▼ −56 WHottest, one board zone85 °C▼ −3 °C88 °C80 °C88 °Clast 85 °C00:0001:0002:0003:0004:0005:0006:00
GPUpower: whole chip, CPU and GPUtemperature: one board zone

GPU 0

Wwhole chip, CPU and GPU

27 W whole chip, CPU and GPU

use0%
memory, whole GPU13.5 GB

this run's peak 8.0 GB

85 °C one board zone

history 0 to 85 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
0.950.750.950.75finished, eval accuracy 0.8077, config 1 of 11 (batch 8), seed 2, Mon 2 Nov, 13:19 UTCfinished, eval accuracy 0.8231, config 2 of 11 (batch 8), seed 2 of 3, Mon 2 Nov, 14:03 UTCfinished, eval accuracy 0.8231, config 3 of 11 (batch 8), seed 2, Mon 2 Nov, 14:09 UTCfinished, eval accuracy 0.7846, config 2 of 11 (batch 8), seed 0 of 3, Mon 2 Nov, 14:16 UTCfinished, eval accuracy 0.8385, config 2 of 11 (batch 8), seed 1 of 3, Mon 2 Nov, 14:22 UTCfinished, eval accuracy 0.7923, config 4 of 11 (batch 8), seed 0 of 3, Mon 2 Nov, 18:15 UTCfinished, eval accuracy 0.8000, config 4 of 11 (batch 8), seed 1 of 3, Mon 2 Nov, 18:22 UTCfinished, eval accuracy 0.8154, config 4 of 11 (batch 8), seed 2 of 3, Mon 2 Nov, 18:28 UTCfinished, eval accuracy 0.7923, config 5 of 11 (batch 8), seed 0 of 3, Mon 2 Nov, 18:46 UTCfinished, eval accuracy 0.8385, config 5 of 11 (batch 8), seed 1 of 3, Mon 2 Nov, 18:52 UTCfinished, eval accuracy 0.8000, config 5 of 11 (batch 8), seed 2 of 3, Mon 2 Nov, 18:59 UTCfinished, eval accuracy 0.8231, config 6 of 11 (batch 8), seed 0 of 3, Tue 3 Nov, 00:59 UTCfinished, eval accuracy 0.8231, config 6 of 11 (batch 8), seed 1 of 3, Tue 3 Nov, 01:13 UTCfinished, eval accuracy 0.8000, config 6 of 11 (batch 8), seed 2 of 3, Tue 3 Nov, 01:26 UTCfinished at step 336 of 339, eval accuracy 0.8769, config 7 of 11 (batch 8), seed 0 of 3, Thu 19 Nov, 21:49 UTCfinished at step 336 of 339, eval accuracy 0.8846, config 7 of 11 (batch 8), seed 1 of 3, Thu 19 Nov, 22:06 UTCfinished at step 336 of 339, eval accuracy 0.8846, config 7 of 11 (batch 8), seed 2 of 3, Thu 19 Nov, 22:21 UTCfinished at step 336 of 339, eval accuracy 0.8615, config 8 of 11 (learning rate 0.0001, batch 8), seed 0, Sat 21 Nov, 23:04 UTCfinished at step 336 of 339, eval accuracy 0.9231, config 9 of 11 (learning rate 0.00002, batch 8), seed 0, Sat 21 Nov, 23:17 UTCfinished at step 336 of 339, eval accuracy 0.8538, config 10 of 11 (learning rate 0.00005, batch 8), seed 0, Sat 21 Nov, 23:29 UTCfinished at step 336 of 339, eval accuracy 0.9231, config 11 of 11 (learning rate 0.00002, batch 8), seed 1 of 2, Sun 22 Nov, 03:49 UTCfinished at step 336 of 339, eval accuracy 0.8846, config 11 of 11 (learning rate 0.00002, batch 8), seed 2 of 2, Sun 22 Nov, 04:01 UTC0.950.75finished, eval accuracy 0.8077, config 1 of 11 (batch 8), seed 2, Mon 2 Nov, 13:19 UTCfinished, eval accuracy 0.8231, config 2 of 11 (batch 8), seed 2 of 3, Mon 2 Nov, 14:03 UTCfinished, eval accuracy 0.8231, config 3 of 11 (batch 8), seed 2, Mon 2 Nov, 14:09 UTCfinished, eval accuracy 0.7846, config 2 of 11 (batch 8), seed 0 of 3, Mon 2 Nov, 14:16 UTCfinished, eval accuracy 0.8385, config 2 of 11 (batch 8), seed 1 of 3, Mon 2 Nov, 14:22 UTCfinished, eval accuracy 0.7923, config 4 of 11 (batch 8), seed 0 of 3, Mon 2 Nov, 18:15 UTCfinished, eval accuracy 0.8000, config 4 of 11 (batch 8), seed 1 of 3, Mon 2 Nov, 18:22 UTCfinished, eval accuracy 0.8154, config 4 of 11 (batch 8), seed 2 of 3, Mon 2 Nov, 18:28 UTCfinished, eval accuracy 0.7923, config 5 of 11 (batch 8), seed 0 of 3, Mon 2 Nov, 18:46 UTCfinished, eval accuracy 0.8385, config 5 of 11 (batch 8), seed 1 of 3, Mon 2 Nov, 18:52 UTCfinished, eval accuracy 0.8000, config 5 of 11 (batch 8), seed 2 of 3, Mon 2 Nov, 18:59 UTCfinished, eval accuracy 0.8231, config 6 of 11 (batch 8), seed 0 of 3, Tue 3 Nov, 00:59 UTCfinished, eval accuracy 0.8231, config 6 of 11 (batch 8), seed 1 of 3, Tue 3 Nov, 01:13 UTCfinished, eval accuracy 0.8000, config 6 of 11 (batch 8), seed 2 of 3, Tue 3 Nov, 01:26 UTCfinished at step 336 of 339, eval accuracy 0.8769, config 7 of 11 (batch 8), seed 0 of 3, Thu 19 Nov, 21:49 UTCfinished at step 336 of 339, eval accuracy 0.8846, config 7 of 11 (batch 8), seed 1 of 3, Thu 19 Nov, 22:06 UTCfinished at step 336 of 339, eval accuracy 0.8846, config 7 of 11 (batch 8), seed 2 of 3, Thu 19 Nov, 22:21 UTCfinished at step 336 of 339, eval accuracy 0.8615, config 8 of 11 (learning rate 0.0001, batch 8), seed 0, Sat 21 Nov, 23:04 UTCfinished at step 336 of 339, eval accuracy 0.9231, config 9 of 11 (learning rate 0.00002, batch 8), seed 0, Sat 21 Nov, 23:17 UTCfinished at step 336 of 339, eval accuracy 0.8538, config 10 of 11 (learning rate 0.00005, batch 8), seed 0, Sat 21 Nov, 23:29 UTCfinished at step 336 of 339, eval accuracy 0.9231, config 11 of 11 (learning rate 0.00002, batch 8), seed 1 of 2, Sun 22 Nov, 03:49 UTCfinished at step 336 of 339, eval accuracy 0.8846, config 11 of 11 (learning rate 0.00002, batch 8), seed 2 of 2, Sun 22 Nov, 04:01 UTC

5 of 22 finished, 0.8385, config 2 of 11 (batch 8), seed 1 of 3, Mon 2 Nov, 14:22 UTC

Runs as a table
StartedStatusEval accuracyStepsConfig
Sun 22 Nov, 04:01 UTC| finished0.8846336 of 339learning rate 0.00002, batch 8, seed 2 of 2
Sun 22 Nov, 03:49 UTC| finished0.9231336 of 339learning rate 0.00002, batch 8, seed 1 of 2
Sat 21 Nov, 23:29 UTC| finished0.8538336 of 339learning rate 0.00005, batch 8, seed 0
Sat 21 Nov, 23:17 UTC| finished0.9231336 of 339learning rate 0.00002, batch 8, seed 0
Sat 21 Nov, 23:04 UTC| finished0.8615336 of 339learning rate 0.0001, batch 8, seed 0
Thu 19 Nov, 22:21 UTC| finished0.8846336 of 339batch 8, seed 2 of 3
Thu 19 Nov, 22:06 UTC| finished0.8846336 of 339batch 8, seed 1 of 3
Thu 19 Nov, 21:49 UTC| finished0.8769336 of 339batch 8, seed 0 of 3
Tue 3 Nov, 01:26 UTC| finished0.8000150 of 150batch 8, seed 2 of 3
Tue 3 Nov, 01:13 UTC| finished0.8231150 of 150batch 8, seed 1 of 3
Tue 3 Nov, 00:59 UTC| finished0.8231150 of 150batch 8, seed 0 of 3
Mon 2 Nov, 18:59 UTC| finished0.8000150 of 150batch 8, seed 2 of 3
Mon 2 Nov, 18:52 UTC| finished0.8385150 of 150batch 8, seed 1 of 3
Mon 2 Nov, 18:46 UTC| finished0.7923150 of 150batch 8, seed 0 of 3
Mon 2 Nov, 18:28 UTC| finished0.8154150 of 150batch 8, seed 2 of 3
Mon 2 Nov, 18:22 UTC| finished0.8000150 of 150batch 8, seed 1 of 3
Mon 2 Nov, 18:15 UTC| finished0.7923150 of 150batch 8, seed 0 of 3
Mon 2 Nov, 14:22 UTC| finished0.8385150 of 150batch 8, seed 1 of 3
Mon 2 Nov, 14:16 UTC| finished0.7846150 of 150batch 8, seed 0 of 3
Mon 2 Nov, 14:09 UTC| finished0.8231150 of 150batch 8, seed 2
Mon 2 Nov, 14:03 UTC| finished0.8231150 of 150batch 8, seed 2 of 3
Mon 2 Nov, 13:19 UTC| finished0.8077150 of 150batch 8, seed 2

Plate

Model
convaiinnovations/laya, 421M parameters revision 7b928d8
Data
Synthetic Discord triage set, synthetic479 train130 eval
Config
batch size 8, grad accum 2, epochs 5, seed 1
Hardware
Radeon 8060S, ROCm

Log

Since startMark
0.0 sphase starting
14.2 swent public
36.6 sphase training
91.5 sepoch 1
159.4 sepoch 2
226.8 sepoch 3
293.8 sepoch 4
360.5 sepoch 5
387.6 sphase eval
387.6 seval 1

Seeds

SeedStatusEval accuracy
seed 0| finished0.7846
seed 1| finished0.8385
seed 2| finished0.8231

observed spread 0.0539, 3 finished seeds

Readings

30 rows as a table
StepSecondsTrain lossEval lossEval accuracyECEBrierSamples per secondLearning rateGPU GB, this run
545.00.6451----5.0312.50e-56.5
1055.80.6451----7.4062.48e-56.5
1566.61.1987----7.4032.45e-56.5
2077.90.0266----7.0702.41e-56.5
2589.30.7189----7.0202.35e-56.5
30101.52.0934----6.5412.29e-56.5
35112.40.0405----7.3242.21e-56.5
40123.8-0.0134----7.0062.12e-56.5
45137.10.5871----6.0242.03e-56.5
50148.50.2916----7.0271.92e-56.5
55159.20.1305----7.4911.81e-56.5
60171.81.3634----6.3371.69e-56.5
65182.70.1934----7.3711.57e-56.5
70182.80.0604-------
75195.50.4954----6.2511.45e-56.5
80206.60.1766----7.1921.33e-56.5
85217.90.3401----7.0791.20e-56.5
90230.30.1733----6.4411.08e-56.5
95241.50.1981----7.1339.53e-66.5
100252.3-0.0320----7.4058.35e-66.5
105263.30.0017----7.2887.22e-66.5
110274.10.0000----7.4176.15e-66.5
115285.00.0000----7.3645.16e-66.5
120297.22.4782----6.5534.25e-66.5
125310.30.0482----6.0723.44e-66.5
130321.40.0000----7.2412.74e-66.5
135332.20.0000----7.4232.14e-66.5
140343.10.1175----7.3201.67e-66.5
145354.00.0000----7.3261.32e-66.5
150389.10.00440.63700.83850.14500.2969---

GPU readings

30 rows as a table
StepGPUUse %Power W, whole chip, CPU and GPUMemory GB, whole GPUTemp C, one board zone
503282.913.187.8
150182.513.482.8
3004483.813.486.8
4506683.313.482.8
600185.013.485.8
7501984.813.482.8
900141.313.583.8
10003083.413.581.8
11503383.313.582.8
13006783.413.582.8
14504984.713.582.8
1500027.013.584.8

Worked out

ReadoutValueWorked out from
Run complete100%150 steps over 150
Step time2.21 s309.1 s over 140 steps, between readings. Startup and intervals that hold an eval are left out
Steps per minute27.260 s over the step time of 2.21 s
Epoch interval67.3 sMean of 4 gaps between epoch marks
Train loss since first reading−0.64070.0044 at step 150 minus 0.6451 at step 5
Lowest train loss-0.0320Lowest of 30 readings, at step 100
Eval accuracy, as shown83.9%0.8385 times 100, to one decimal. Changes on that plate are percentage points
Train minus eval loss−0.63260.0044 minus 0.6370, both at step 150
Wrong answers21130 eval items minus 109 correct. 109 over 130 is the only count that gives 0.8385
Rank9 of 22By last eval accuracy, across this project's 22 runs with an eval
Best so far0.9231Highest last eval accuracy, run 19
Against the run before+0.05390.8385 minus 0.7846 on run 4
Lowest last train loss-0.0937Lowest last train loss across this project's 22 runs, run 19
Seed spread, observed0.0539High 0.8385 minus low 0.7846 across 3 finished seeds
Train loss against run 4+0.00440.0044 minus 0.0000 on run 4, both at step 150
Samples per second since first reading+2.297.33 minus 5.03
Learning rate since first reading−2.36e-51.32e-6 minus 2.50e-5
This run's peak memory per GPU8.0 GBAs the run reports it: its own memory, not the whole GPU's