r2s-cur-doodle-ft-p09-s4-24d6aad-9dc35f11-47646604
completed —
Hypothesis
Fine-tuned for 30k steps on renders degraded to look hand-drawn (probability 0.9), the final model turns degraded validation renders into programs closer to the clean icon than the parent model does, and closer than the degraded input itself (what a faithful tracer returns), while its error on clean validation renders rises by at most 0.010. Scored by doodle_eval on the sets declared in state/CURRENT.md (2026-10-31 16:00): clean, the training kinds and the held-out kinds pencil, hatched, affine, thumb.
Visual output
Written result
Hypothesis. Fine-tuned for 30k steps on renders degraded to look hand-drawn (probability 0.9), the final model turns degraded validation renders into programs closer to the clean icon than the parent model does, and closer than the degraded input itself (what a faithful tracer returns), while its error on clean validation renders rises by at most 0.010. Scored by doodle_eval on the sets declared in state/CURRENT.md (2026-10-31 16:00): clean, the training kinds and the held-out kinds pencil, hatched, affine, thumb.
Result
| measure | value |
|---|---|
| icons evaluated | 339 (primary/validation) |
| model pixel error, mean [95% CI] | 0.0941 [0.0889, 0.0996] |
| model pixel error, median | 0.0901 |
| exact program rate | 0.003 |
| 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 | -0.0044 [-0.0090, -0.0000] |
| icons where model beats nearest icon | 175 |
| CLIP top-1, model greedy (32 Gate N icons) | 0.625 |
| 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 | 1715 |
| peak VRAM GiB | 4.388660430908203 |
| latency ms/icon, batch 1, median | 3353.5 |
| latency ms/icon, batch 1, p95 | 3359.3 |
| throughput ms/icon, batch 64 | 118.9 |
| model calls per icon (decoding) | 688.0 |
| 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
- running2026-10-31T15:03:58Z
- completed2026-10-31T15:34:46Z
Run record
Verbatim from runs/r2s-cur-doodle-ft-p09-s4-24d6aad-9dc35f11-47646604/run.yaml, the record committed before launch.
schema_version
2
run_id
r2s-cur-doodle-ft-p09-s4-24d6aad-9dc35f11-47646604
state
completed
planned_at
2026-10-31T15:03:58Z
hypothesis
Fine-tuned for 30k steps on renders degraded to look hand-drawn (probability 0.9), the final model turns degraded validation renders into programs closer to the clean icon than the parent model does, and closer than the degraded input itself (what a faithful tracer returns), while its error on clean validation renders rises by at most 0.010. Scored by doodle_eval on the sets declared in state/CURRENT.md (2026-10-31 16:00): clean, the training kinds and the held-out kinds pencil, hatched, affine, thumb.
parent_run
r2s-cur-d02-s120k-seed2-53b3108-46294452-47646604
git_commit
24d6aad826891dbb213a392fd687d549c70f7dd0
dirty_patch_sha256
config
configs/render2svg/cur-doodle-ft-p09-s4.yaml
config_sha256
9dc35f11ad3aca52…config_resolved
model
image_size
144
d_model
256
heads
8
encoder_layers
2
decoder_layers
6
feedforward
1024
dropout
0.2
metric
true
fourier
10
order
path
sources
0
training
steps
30000
batch_size
32
learning_rate
0.0003
weight_decay
0.01
warmup_steps
500
seed
7004
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
degrade_probability
0.9
degrade_kinds
lineart, sloppy, flat
degrade_eval_seed
0
init_run
r2s-cur-d02-s120k-seed2-53b3108-46294452-47646604
init_sha256
38daeca4dd27fb4c…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
evaluation_inputs
renders degraded by doodle.degrade_split, kinds ['lineart', 'sloppy', 'flat'], seed 0; scored against the clean targets
init
run
r2s-cur-d02-s120k-seed2-53b3108-46294452-47646604
checkpoint
/home/eai-jdkxz/eai-team11/mojidiff/cache/runs/r2s-cur-d02-s120k-seed2-53b3108-46294452-47646604/best.pt
sha256
38daeca4dd27fb4c…step
64000
seed
7004
determinism
seeded; cuDNN and SDPA kernels not forced deterministic
model_parameters
9025826
command
.venv/bin/python -m mojidiff.learning.render2svg --config configs/render2svg/cur-doodle-ft-p09-s4.yaml
outputs
run_dir
runs/r2s-cur-doodle-ft-p09-s4-24d6aad-9dc35f11-47646604
checkpoints
/home/eai-jdkxz/eai-team11/mojidiff/cache/runs/r2s-cur-doodle-ft-p09-s4-24d6aad-9dc35f11-47646604
notes
completed_at
2026-10-31T15:34:46Z
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
1714.8550694240257
inference_ms_per_icon
3353.5152785480022
inference_p95_ms_per_icon
3359.2972779879346
icons_per_second
0.29819455614139256
batched_ms_per_icon
118.86458965636848
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
0.09409951904165671, 0.08890426586583242, 0.09956553022949913
model_pixel_error_median
0.09006672352552414
exact_program_rate
0.0029498525073746312
rendered_rate
1.0
blank_pixel_error
0.1720480552069557
selected_step
8000
nearest_training_icon_pixel_error
0.08973124639576185, 0.0848205749792233, 0.09466393787711656
model_minus_baseline_error_reduction
-0.004368272645894855, -0.008964583077156335, -3.912900538018181e-05
icons_model_beats_baseline
175
clip_retrieval
model_greedy
top1_rate
0.625
top5_rate
0.84375
mean_rank
3.40625
clip_to_reference_mean
0.9168222770094872
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
688