CLIP · ViT-L/14 · Foundation Model
Metrics & Analytics
Production-grade evaluation suite: ROC, PR, confusion, loss curves and live training telemetry.
⌘K
GPU78%
Loss0.184
Throughput312/s
Accuracy
94.1%
0.80% vs last epoch
Precision
0.927
0.50% vs last epoch
Recall
0.915
0.60% vs last epoch
F1 Score
0.921
0.50% vs last epoch
ROC-AUC
0.974
0.30% vs last epoch
Training vs validation
Accuracy & loss curves
Contrastive loss
InfoNCE over epochs
ROC curve
AUC = 0.974
Precision-Recall
AP = 0.918
Confusion matrix
5-class evaluation
| actual ↓ / pred → | cat | dog | car | ship | bird |
|---|---|---|---|---|---|
| cat | 92 | 3 | 1 | 1 | 3 |
| dog | 4 | 90 | 2 | 1 | 3 |
| car | 1 | 2 | 94 | 2 | 1 |
| ship | 1 | 1 | 3 | 93 | 2 |
| bird | 3 | 2 | 1 | 2 | 92 |