IRSANet_PathMNIST v0.1.0 Pretrained Weights (Seed3)
Model Details
| Field |
Value |
| Model |
IRSANet_PathMNIST |
| Parameters |
351,623 |
| Modules |
124 |
Weight Details
| Field |
Value |
| Experiment |
Validation |
| Run |
Run4 |
| Seed |
3 |
| Fold |
single |
| Epoch |
99 |
| Checkpoint |
bestAP |
Evaluation Details
| Field |
Value |
| Dataset |
PathMNIST |
| Test samples |
7,180 |
| Evaluated |
2026-06-05 10:52 |
| TTA |
False |
| Classes |
9 |
Key Metrics (Macro)
| Metric |
Test |
Val |
| AUC (ROC, OVR) |
0.997 |
1.000 |
| Average Precision |
0.974 |
0.999 |
| F1 Score |
0.951 |
0.993 |
| Balanced Accuracy |
0.952 |
0.993 |
Usage
Reconstruction
from irsanet import load_checkpoint
model, ckpt = load_checkpoint("IRSANet_PathMNIST_seed3_bestAP.pt")Evaluation
from irsanet import evaluate_model
from irsanet_pathmnist import build_datasets
_,_,test_loader,_,_ = build_datasets([],
eval_mode = True,
seed = 3,
)
results = evaluate_model(
weights = "IRSANet_PathMNIST_seed3_bestAP.pt",
test_loader = test_loader,
)
print(results["test_results"])Including a full visual dashboard:
from irsanet.plotting import plot_eval_panel_multi
from irsanet_pathmnist import build_datasets
_,_,test_loader,_,_ = build_datasets([],
eval_mode = True,
seed = 3,
)
results = plot_eval_panel_multi(
weights = "IRSANet_PathMNIST_seed3_bestAP.pt",
test_loader = test_loader,
)
print(results["test_results"])