Euler View
1 pending

Run Dossier

Training run for UniFuse

71 metrics · 37,168 train · 484 val

Chart Overlays

Compare this run against its peers

0/59
Add peer runs here to layer their train and val traces onto the current metric charts.
interrupted Apr 11, 22:50 Duration: 11h 58m Python 3.11.7 src/train.py --config configs/model/inverted_joint_metric.yaml --config configs/model/backbone_dinov2_vitl14.yaml --config configs/data/real_drive_sim_euler_loading.yaml --config configs/training/phase_b_joint_structure.yaml --init-from /cluster/scratch/drothenpiele/euler_train/unifuse/checkpoints/geometry-prompt/checkpoint-epoch-2.pt
lr 2e-4batch 4epochs 12precision automodel weather_metric_system wd 1e-4
ID 63055282 Job metric_diffusion_pty Part gpupr.24h Node eu-a65-05 CPU 8 GPU 1
unifuse/runs/2026-04-11_22-50-53_32e9
Error
Signal: SIGINT
Metrics 71/71
depth.train 49
diag 26
consolidation 1
delta_vs_prior
final 5
depth_abs_rel
depth_mae
depth_rmse
log_radius_l1
xyz_l1
pred 5
depth_abs_rel
depth_mae
depth_rmse
log_radius_l1
xyz_l1
prior 5
depth_abs_rel
depth_mae
depth_rmse
log_radius_l1
xyz_l1
structure 10
depth_l1
edge_l1
eval_ensemble_size
first_nonfinite_step_index
first_nonfinite_timestep
had_nonfinite
mode_is_sampled
mode_is_teacher_forced
sampled_debug_enabled
used_fp32_sampler
loss 18
consolidation 3
low_frequency
structure_align
structure_align_edge
final 4
confidence
depth
log_radius
xyz
pred 1
log_radius
prior 4
aux_metric_probe
aux_metric_probe_confidence
confidence
log_radius
structure 5
confidence
depth
depth_denoise_mse
edge
rgb_denoise_mse
no stage 1
total
stat 5
final 1
confidence
prior 1
confidence
structure 3
confidence
pred_finite_fraction
pred_has_nonfinite
rgb.train 6
diag 2
image_mae
image_psnr
loss 3
clear
confidence
rgb_denoise_mse
stat 1
visibility
sys.train 16
data_mean_depth
data_median_depth
data_valid_fraction
flag_refiner_enabled
flag_structure_enabled
gpu_mem_total_gb
gpu_mem_used_gb
gpu_mem_util_pct
gpu_util_pct
lr
lr.default
lr.geometry_encoder_adapter
lr.joint_structure_expert_heads
lr.joint_structure_expert_model
lr.prompt_encoder
ray_source_is_predicted
71 selected
X-Axis
Y-Scale
Series
Group
Namespace

Delta Vs Prior

↓ kind=diagstage=consolidation
T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Depth Abs Rel

↓ kind=diagstage=final
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Depth Mae

↓ meters kind=diagstage=final
T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Depth Rmse

↓ meters kind=diagstage=final
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Log Radius L1

↓ kind=diagstage=final
T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Xyz L1

↓ meters kind=diagstage=final
T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Depth Abs Rel

↓ kind=diagstage=pred
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Depth Mae

↓ meters kind=diagstage=pred
T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Depth Rmse

↓ meters kind=diagstage=pred
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Log Radius L1

↓ kind=diagstage=pred
T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Xyz L1

↓ meters kind=diagstage=pred
T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Depth Abs Rel

↓ kind=diagstage=prior
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Depth Mae

↓ meters kind=diagstage=prior
T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Depth Rmse

↓ meters kind=diagstage=prior
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Log Radius L1

↓ kind=diagstage=prior
T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Xyz L1

↓ meters kind=diagstage=prior
T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Depth L1

↓ kind=diagstage=structure
T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Edge L1

↓ kind=diagstage=structure
T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Eval Ensemble Size

kind=diagstage=structure
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

First Nonfinite Step Index

kind=diagstage=structure
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

First Nonfinite Timestep

kind=diagstage=structure
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Had Nonfinite

kind=diagstage=structure
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Mode Is Sampled

kind=diagstage=structure
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Mode Is Teacher Forced

kind=diagstage=structure
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Sampled Debug Enabled

kind=diagstage=structure
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Used Fp32 Sampler

kind=diagstage=structure
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Low Frequency

↓ kind=lossstage=consolidation
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Structure Align

↓ kind=lossstage=consolidation
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Structure Align Edge

↓ kind=lossstage=consolidation
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Confidence

↑ kind=lossstage=final
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Depth

↓ meters kind=lossstage=final
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Log Radius

↓ kind=lossstage=final
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Xyz

↓ meters kind=lossstage=final
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Log Radius

↓ kind=lossstage=pred
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Aux Metric Probe

kind=lossstage=prior
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Aux Metric Probe Confidence

kind=lossstage=prior
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Confidence

↑ kind=lossstage=prior
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Log Radius

↓ kind=lossstage=prior
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Confidence

↑ kind=lossstage=structure
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Depth

↓ meters kind=lossstage=structure
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Depth Denoise Mse

kind=lossstage=structure
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Edge

↓ kind=lossstage=structure
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Rgb Denoise Mse

kind=lossstage=structure
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Total

↓ kind=loss
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Confidence

↑ kind=statstage=final
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Confidence

↑ kind=statstage=prior
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Confidence

↑ kind=statstage=structure
T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Pred Finite Fraction

kind=statstage=structure
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Pred Has Nonfinite

kind=statstage=structure
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Image Mae

↓ kind=diagstage=structure
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Image Psnr

↑ dB kind=diagstage=structure
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Clear

kind=lossstage=structure
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Confidence

↑ kind=lossstage=structure
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Rgb Denoise Mse

kind=lossstage=structure
T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Visibility

↑ kind=statstage=weather
V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Data Mean Depth

V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Data Median Depth

V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Data Valid Fraction

V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Flag Refiner Enabled

V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Flag Structure Enabled

V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

Gpu Mem Total Gb

T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Gpu Mem Used Gb

T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Gpu Mem Util Pct

T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Gpu Util Pct

T V
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —
2026-04-11_22-50-53_32e9 val (raw) —

Lr

T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Lr.Default

T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Lr.Geometry Encoder Adapter

T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Lr.Joint Structure Expert Heads

T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Lr.Joint Structure Expert Model

T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Lr.Prompt Encoder

T
2026-04-11_22-50-53_32e9 train (smoothed) —
2026-04-11_22-50-53_32e9 train (raw) —

Ray Source Is Predicted

V
2026-04-11_22-50-53_32e9 val (smoothed) —
2026-04-11_22-50-53_32e9 val (raw) —

No output snapshots found for this run.

Outputs are generated during training and saved to outputs/epoch_N_step_M/ directories.

epoch 1 / step 800 Checkpoint #2689
4/11/2026, 10:34:39 PM
/cluster/scratch/drothenpiele/euler_train/unifuse/checkpoints/joint-structure-expert-2026-04-11-22-50-53-32e9/checkpoint-epoch-1.pt
epoch 2 / step 1600 Checkpoint #2693
4/11/2026, 11:58:16 PM
/cluster/scratch/drothenpiele/euler_train/unifuse/checkpoints/joint-structure-expert-2026-04-11-22-50-53-32e9/checkpoint-epoch-2.pt
epoch 3 / step 2400 Checkpoint #2703
4/12/2026, 8:42:37 AM
/cluster/scratch/drothenpiele/euler_train/unifuse/checkpoints/joint-structure-expert-2026-04-11-22-50-53-32e9/checkpoint-epoch-3.pt
epoch 4 / step 3200 Checkpoint #2705
4/12/2026, 8:42:37 AM
/cluster/scratch/drothenpiele/euler_train/unifuse/checkpoints/joint-structure-expert-2026-04-11-22-50-53-32e9/checkpoint-epoch-4.pt
epoch 5 / step 4000 Checkpoint #2707
4/12/2026, 8:42:37 AM
/cluster/scratch/drothenpiele/euler_train/unifuse/checkpoints/joint-structure-expert-2026-04-11-22-50-53-32e9/checkpoint-epoch-5.pt
epoch 6 / step 4800 Checkpoint #2708
4/12/2026, 8:42:37 AM
/cluster/scratch/drothenpiele/euler_train/unifuse/checkpoints/joint-structure-expert-2026-04-11-22-50-53-32e9/checkpoint-epoch-6.pt
epoch 7 / step 5600 Checkpoint #2709
4/12/2026, 8:42:37 AM
/cluster/scratch/drothenpiele/euler_train/unifuse/checkpoints/joint-structure-expert-2026-04-11-22-50-53-32e9/checkpoint-epoch-7.pt
epoch 8 / step 6400 Checkpoint #2710
4/12/2026, 8:42:37 AM
/cluster/scratch/drothenpiele/euler_train/unifuse/checkpoints/joint-structure-expert-2026-04-11-22-50-53-32e9/checkpoint-epoch-8.pt
epoch 9 / step 7200 Checkpoint #2711
4/12/2026, 8:42:37 AM
/cluster/scratch/drothenpiele/euler_train/unifuse/checkpoints/joint-structure-expert-2026-04-11-22-50-53-32e9/checkpoint-epoch-9.pt
epoch 10 / step 8000 Checkpoint #2712
4/12/2026, 8:42:37 AM
/cluster/scratch/drothenpiele/euler_train/unifuse/checkpoints/joint-structure-expert-2026-04-11-22-50-53-32e9/checkpoint-epoch-10.pt
epoch 11 / step 8800 Checkpoint #2713
4/12/2026, 8:42:37 AM
/cluster/scratch/drothenpiele/euler_train/unifuse/checkpoints/joint-structure-expert-2026-04-11-22-50-53-32e9/checkpoint-epoch-11.pt

Run-Owned Exports

Publish direct filesystem paths here.

Published

No run-owned exports are published yet.

Raw Artifacts

Run-owned exports are typically direct paths, so there are no captured artifacts to publish from here.

Euler View - ML Experiment Monitor