[bootstrap] device=cuda  skip_val=False

── unet @ 100% ────────────────────────────────
  ✓ copied checkpoint → unet_100_best.pth
  ✓ parsed 50 epochs from training_logs.txt
  · running val pass with new metrics code… done in 7.4s  dice=0.9376  iou=0.8825  miou=0.9372  pixel_acc=0.9924
  ✓ wrote log → unet_100.json

── segformer_b0 @ 100% ────────────────────────────────
  ✓ copied checkpoint → segformer_b0_100_best.pth
  ✓ parsed 50 epochs from training_logs.txt
  · running val pass with new metrics code… [transformers] You passed `num_labels=1` which is incompatible to the `id2label` map of length `1000`.
Warning: You are sending unauthenticated requests to the HF Hub. Please set a HF_TOKEN to enable higher rate limits and faster downloads.

Loading weights:   0%|          | 0/192 [00:00<?, ?it/s]
Loading weights: 100%|██████████| 192/192 [00:00<00:00, 10568.19it/s]
[transformers] SegformerForSemanticSegmentation LOAD REPORT from: nvidia/mit-b0
Key                                           | Status     | 
----------------------------------------------+------------+-
classifier.bias                               | UNEXPECTED | 
classifier.weight                             | UNEXPECTED | 
decode_head.linear_c.{0, 1, 2, 3}.proj.bias   | MISSING    | 
decode_head.batch_norm.num_batches_tracked    | MISSING    | 
decode_head.classifier.weight                 | MISSING    | 
decode_head.classifier.bias                   | MISSING    | 
decode_head.batch_norm.running_mean           | MISSING    | 
decode_head.batch_norm.running_var            | MISSING    | 
decode_head.linear_c.{0, 1, 2, 3}.proj.weight | MISSING    | 
decode_head.batch_norm.weight                 | MISSING    | 
decode_head.batch_norm.bias                   | MISSING    | 
decode_head.linear_fuse.weight                | MISSING    | 

Notes:
- UNEXPECTED:	can be ignored when loading from different task/architecture; not ok if you expect identical arch.
- MISSING:	those params were newly initialized because missing from the checkpoint. Consider training on your downstream task.
done in 3.7s  dice=0.9293  iou=0.8679  miou=0.9294  pixel_acc=0.9914
  ✓ wrote log → segformer_b0_100.json

[bootstrap] done.