gtn_surrogate_lamagic_5comp_v1

Stabilized LaMAGIC GTN surrogate (loop A/B B run)

This is the selected inference checkpoint from the stabilized B arm of the independent GTN loop-attention A/B experiment. It predicts efficiency and Vout from the graph-based topology and duty inputs used by this project.

This is a custom PyTorch checkpoint, not a Transformers checkpoint. Load it with shared_energy.gtn_surrogate.GTNSurrogateModel from this repository.

Selected validation result

• Best epoch: 662
• Best validation mean RSE: 0.091561
• Validation RSE, efficiency: 0.060639
• Validation RSE, Vout: 0.122483
• Validation MSE: 261.503524

Test evaluation was intentionally skipped, so the listed metrics are validation metrics.

Training setup

{
  "completed_epochs": 700,
  "gradient_clip_norm": 1.0,
  "loop_learning_rate": 0.0003,
  "loss": "rse",
  "mode": "independent_gtn_loop_ab",
  "phase": "loop",
  "warmup_epochs": 450,
  "warmup_learning_rate": 0.001
}

The B arm resumes the common 450-epoch no-loop warmup, enables loop attention from epoch 451, reduces the loop-phase learning rate to 3e-4, and applies global gradient-norm clipping of 1.0. The checkpoint is restored to the best validation epoch (662), rather than the final training epoch (700).

Split summary:

{
  "test_row_count": 4493,
  "test_topology_count": 979,
  "train_row_count": 15635,
  "train_topology_count": 3425,
  "val_row_count": 2247,
  "val_topology_count": 489
}

Files

• gtn_surrogate_model.pt: selected inference checkpoint containing model_state_dict, model_config, target names, and best-validation metadata.
• manifest.json: training and validation summary.
• README.md: this model card.

final_training_state.pt, periodic training_states/, and batch diagnostics are deliberately excluded. They are resumption/debug artifacts, not inference weights. Training history and validation predictions are uploaded only when their explicit flags are supplied.

Loading

from huggingface_hub import hf_hub_download
import torch
from shared_energy.gtn_surrogate import GTNSurrogateModel

checkpoint_path = hf_hub_download(
    repo_id="tjwjdgns011119/gtn_surrogate_lamagic_5comp_v1", filename="gtn_surrogate_model.pt"
)
payload = torch.load(checkpoint_path, map_location="cpu")
model = GTNSurrogateModel(**payload["model_config"])
model.load_state_dict(payload["model_state_dict"], strict=True)
model.eval()

The checkpoint is for research workflows in this repository. Its outputs are surrogate estimates, not direct SPICE-verified measurements.