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

Claviger router: OneJev 0.8B (backfilled)

finishedRadeon 8060S

A model that routes a request to the one destination, of many, that should take it. Backfilled: recorded after the fact from the run's own logs and records. Times inside a run are estimated, and there are no GPU sensor readings.

started
Tue 10 Nov, 16:24 UTC
ran for
45:44
config
learning rate 0.00001, batch 1, seed 1 of 3

run4/5step42/42epoch1/1config1/3seed1/3lr0.00001batch1modelOneJev1.6-0.8B 0.8Brevc89933c

Loss train 0.0014 ▼ −0.1812 lowest 0.0001 at step 39 run before, run 3 ▼ −3.3914 step 42
Timebase
Range
Smoothing
Reference
42 readings, 30 min/div10 min/div5 min/div
024600:0030:00phase starting, 0.0 s to 65.3 sphase training, 65.3 s to 45:44epoch 1, 65.3 straining0.1826Lowest last train loss across this project's 5 runs, run 4lowest last 0.0014, run 40.0014024600:0010:0020:0030:0040:00phase starting, 0.0 s to 65.3 sphase training, 65.3 s to 45:44epoch 1, 65.3 straining0.1826Lowest last train loss across this project's 5 runs, run 4lowest last 0.0014, run 40.0014▼ −0.1812024600:0005:0010:0015:0020:0025:0030:0035:0040:0045:00phase starting, 0.0 s to 65.3 sphase training, 65.3 s to 45:44epoch 1, 65.3 straining0.1826Lowest last train loss across this project's 5 runs, run 4lowest last 0.0014, run 40.0014▼ −0.1812

Elapsed

65.3 s to epoch 1, 2679.3 s to last reading

42 of 42 steps, 1 of 1 epochs, 100%

65.3 s per step, 0.9 steps a minute

Set-up

666 train

synthetic, learning rate 0.00001, batch size 1, grad accum 16, epochs 1, optimizer adamw, seed 1

Learning rate

0.00e+0

▼ −1.00e-5 since first

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

4 of 5 finished, -, config 1 of 3 (learning rate 0.00001, batch 1), seed 1 of 3, Tue 10 Nov, 16:24 UTC

Runs as a table
StartedStatusEval accuracyStepsConfig
Tue 10 Nov, 17:10 UTC| finished-42 of 42learning rate 0.00001, batch 1, seed 2 of 3
Tue 10 Nov, 16:24 UTC| finished-42 of 42learning rate 0.00001, batch 1, seed 1 of 3
Tue 10 Nov, 15:27 UTC| finished0.084742 of 42learning rate 0.00005, batch 1, seed 0
Tue 10 Nov, 14:30 UTC| finished0.912742 of 42learning rate 0.00002, batch 1, seed 0
Tue 10 Nov, 13:33 UTC| finished0.923342 of 42learning rate 0.00001, batch 1, seed 0 of 3

Plate

Model
OmniJev/OneJev1.6-0.8B, 0.8B parameters revision c89933c
Data
Synthetic routing worlds, synthetic
Config
learning rate 0.00001, batch size 1, grad accum 16, epochs 1, optimizer adamw, seed 1
Hardware
Radeon 8060S, ROCm

Log

Since startMark
0.0 sphase starting
65.3 sphase training
65.3 sepoch 1
188 h 27 min laterwent public

Seeds

SeedStatusEval accuracy
seed 0| finished0.9233
seed 1| finished-
seed 2| finished-

Readings

42 rows as a table
StepSecondsTrain lossLearning rate
165.30.18261.00e-5
2130.70.02369.99e-6
3196.00.04629.94e-6
4261.40.01539.87e-6
5326.70.72089.77e-6
6392.10.11819.64e-6
7457.40.21209.48e-6
8522.80.00099.30e-6
9588.10.78199.09e-6
10653.50.25098.86e-6
11718.80.44608.60e-6
12784.20.71638.33e-6
13849.50.01928.03e-6
14914.90.11547.72e-6
15980.20.81617.39e-6
161045.60.02247.05e-6
171110.90.97246.69e-6
181176.30.35056.32e-6
191241.60.00055.95e-6
201307.00.05365.57e-6
211372.30.48375.19e-6
221437.60.00094.81e-6
231503.01.66154.43e-6
241568.30.00104.05e-6
251633.70.13663.68e-6
261699.01.19113.31e-6
271764.40.46372.95e-6
281829.70.00052.61e-6
291895.10.03272.28e-6
301960.40.00031.97e-6
312025.80.00011.67e-6
322091.10.00011.40e-6
332156.50.27721.14e-6
342221.80.00039.10e-7
352287.20.03547.02e-7
362352.50.23025.19e-7
372417.90.00043.62e-7
382483.20.00032.33e-7
392548.60.00011.32e-7
402613.90.00015.86e-8
412679.30.00271.47e-8
422744.60.00140.00e+0

Worked out

ReadoutValueWorked out from
Run complete100%42 steps over 42
Step time65.3 s2679.3 s over 41 steps, between readings. Startup and intervals that hold an eval are left out
Steps per minute0.960 s over the step time of 65.35 s
Train loss since first reading−0.18120.0014 at step 42 minus 0.1826 at step 1
Lowest train loss0.0001Lowest of 42 readings, at step 39
Lowest last train loss0.0014Lowest last train loss across this project's 5 runs, run 4
Items per second0.01 step over 65.35 s, times batch size 1
Train loss against run 3−3.39140.0014 minus 3.3927 on run 3, both at step 42
Learning rate since first reading−1.00e-50.00e+0 minus 1.00e-5