r2s-cur-diag-aug9004-3d24194-afac26b4-47646604
completed —
Hypothesis
Diagnostic, not a candidate: if seed 4's augmentation stream is what makes its runs bad, this v9 run scores near seed 4's 0.0861 rather than v9's 0.072; it is reported, not passed.
Visual output
Written result
Hypothesis. Diagnostic, not a candidate: if seed 4's augmentation stream is what makes its runs bad, this v9 run scores near seed 4's 0.0861 rather than v9's 0.072; it is reported, not passed.
Result
| measure | value |
|---|---|
| icons evaluated | 339 (primary/validation) |
| model pixel error, mean [95% CI] | 0.0727 [0.0679, 0.0776] |
| model pixel error, median | 0.0681 |
| exact program rate | 0.024 |
| rendered rate | 1.000 |
| blank canvas pixel error | 0.1720 |
| nearest training icon pixel error | 0.0897 [0.0848, 0.0947] |
| error reduction vs nearest icon | 0.0170 [0.0130, 0.0209] |
| icons where model beats nearest icon | 240 |
| CLIP top-1, model greedy (32 Gate N icons) | 0.500 |
| CLIP top-1, nearest training icon | 0.344 |
| CLIP top-1, OmniSVG 4B zero-shot (Gate N) | 0.609 |
Resources
| measure | value |
|---|---|
| device | NVIDIA B300 SXM6 AC |
| parameters | 9,025,826 |
| train seconds | 2878 |
| peak VRAM GiB | 4.388660430908203 |
| latency ms/icon, batch 1, median | 3799.0 |
| latency ms/icon, batch 1, p95 | 3800.5 |
| throughput ms/icon, batch 64 | 112.3 |
| model calls per icon (decoding) | 775.0 |
| torch / CUDA | 2.14.1+cu130 / 13.0 |
Latency covers encoding and decoding to a validated token program; it excludes rasterising the SVG. Samples: samples.png, rows are reference, model, nearest training icon.
State transitions
- running2026-10-30T22:36:58Z
- completed2026-10-30T23:28:01Z
Run record
Verbatim from runs/r2s-cur-diag-aug9004-3d24194-afac26b4-47646604/run.yaml, the record committed before launch.
schema_version
2
run_id
r2s-cur-diag-aug9004-3d24194-afac26b4-47646604
state
completed
planned_at
2026-10-30T22:36:58Z
hypothesis
Diagnostic, not a candidate: if seed 4's augmentation stream is what makes its runs bad, this v9 run scores near seed 4's 0.0861 rather than v9's 0.072; it is reported, not passed.
parent_run
r2s-full-v9-colour
git_commit
3d2419434735250afeee5b6014c99d281b1acdfc
dirty_patch_sha256
config
configs/render2svg/cur-diag-aug9004.yaml
config_sha256
afac26b4681d058a…config_resolved
model
image_size
144
d_model
256
heads
8
encoder_layers
2
decoder_layers
6
feedforward
1024
dropout
0.1
metric
true
fourier
10
order
path
sources
0
training
steps
60000
batch_size
32
learning_rate
0.0005
weight_decay
0.01
warmup_steps
1000
seed
7001
eval_every
2000
eval_icons
64
train_icons
None
evaluate_on_train
false
augment_variants
0
augment_seed
9004
augment_mirror
0.5
augment_max_shift
48
augment_colour
0.1
augment_online
true
compose_probability
0.5
compose_parts
4
augment_original
0.1
loader_workers
16
extra_training
bf16
true
final_eval_icons
339
clip_icons
32
dataset
pilot_config
configs/learning/openmoji-g1-geometric-gate-v16.yaml
cache_sha256
476466042da98d72…train_icons
2681
augmented_variants
0
extra_training
None
augmented_sha256
None
evaluated_on
primary/validation
seed
7001
determinism
seeded; cuDNN and SDPA kernels not forced deterministic
model_parameters
9025826
command
.venv/bin/python -m mojidiff.learning.render2svg --config configs/render2svg/cur-diag-aug9004.yaml
outputs
run_dir
runs/r2s-cur-diag-aug9004-3d24194-afac26b4-47646604
checkpoints
/home/eai-jdkxz/eai-team11/mojidiff/cache/runs/r2s-cur-diag-aug9004-3d24194-afac26b4-47646604
notes
completed_at
2026-10-30T23:28:01Z
resource
device
NVIDIA B300 SXM6 AC
torch_version
2.14.1+cu130
cuda_version
13.0
python_version
3.12.3
peak_vram_gib
4.388660430908203
train_seconds
2878.0806376470136
inference_ms_per_icon
3799.02235098416
inference_p95_ms_per_icon
3800.4854380269535
icons_per_second
0.26322561638547454
batched_ms_per_icon
112.27360873454018
batched_icons_per_call
64
excludes
rendering the output SVG to pixels
baselines
blank_canvas_pixel_error
0.1720480552069557
nearest_training_icon_pixel_error
0.08973124639576185
gate_n_omnisvg_zero_shot_top1
0.609375
nearest_training_icon_clip_top1
0.34375
result
icons
339
model_pixel_error
0.07269772243125235, 0.0678699091282445, 0.07757447120761203
model_pixel_error_median
0.06805908679962158
exact_program_rate
0.02359882005899705
rendered_rate
1.0
blank_pixel_error
0.1720480552069557
selected_step
52000
nearest_training_icon_pixel_error
0.08973124639576185, 0.0848205749792233, 0.09466393787711656
model_minus_baseline_error_reduction
0.017033523964509505, 0.013011373827702406, 0.02091945528058027
icons_model_beats_baseline
240
clip_retrieval
model_greedy
top1_rate
0.5
top5_rate
0.9375
mean_rank
3.0
clip_to_reference_mean
0.9089534990489483
chance_top1
0.03125
icons
32
nearest_training_icon
top1_rate
0.34375
top5_rate
0.59375
mean_rank
7.40625
clip_to_reference_mean
0.8630617223680019
chance_top1
0.03125
icons
32
gate_n_omnisvg_zero_shot_top1
0.609375
gate_n_omnisvg_best_of_12_top1
0.828125
model_calls_per_icon
775