✓ Initialized. View run at 
https://modal.com/apps/seth-45699/main/ap-L1S2KGTqxMKOoeFaRYjhEc
✓ Created objects.
├── 🔨 Created mount 
│   /private/tmp/claude-501/-Users-sethstafford-dev-research-sutro/aabe6221-ffef
│   -4b0f-bafe-aa419c95eb0e/scratchpad/c6/run_modal.py
├── 🔨 Created mount 
│   /private/tmp/claude-501/-Users-sethstafford-dev-research-sutro/aabe6221-ffef
│   -4b0f-bafe-aa419c95eb0e/scratchpad/c6/mnist.py
└── 🔨 Created function remote_score.

==========
== CUDA ==
==========

CUDA Version 13.3.0

Container image Copyright (c) 2016-2023, NVIDIA CORPORATION & AFFILIATES. All rights reserved.



--- run 1:
triton_smoke.py:fast_mlp at 5.40% error: 15 timed calls, each in a fresh process
  NVIDIA A100 80GB PCIe, torch 2.12.0+cu130, sandbox on
  call  1/15: 50.218 ms, scorer's clock 53.462 ms
  call  2/15: 50.168 ms, scorer's clock 53.111 ms
  call  3/15: 50.135 ms, scorer's clock 50.360 ms
  call  4/15: 50.231 ms, scorer's clock 53.830 ms
  call  5/15: 50.154 ms, scorer's clock 50.209 ms
  call  6/15: 50.157 ms, scorer's clock 53.732 ms
  call  7/15: 50.235 ms, scorer's clock 49.997 ms
  call  8/15: 50.108 ms, scorer's clock 53.673 ms
  call  9/15: 50.121 ms, scorer's clock 50.451 ms
  call 10/15: 50.054 ms, scorer's clock 53.876 ms
  call 11/15: 50.116 ms, scorer's clock 51.103 ms
  call 12/15: 50.091 ms, scorer's clock 53.779 ms
  call 13/15: 50.161 ms, scorer's clock 53.819 ms
  call 14/15: 50.223 ms, scorer's clock 50.903 ms
  call 15/15: 50.719 ms, scorer's clock 56.184 ms
MNIST 95.28% (104,810/110,000), 52.133 ms/call; hold-out (kmnist) 92.00%, 52.440 ms/call (floored by the scorer's clock); score 52.440 ms
Stopping app - local entrypoint completed.
✓ App completed. View run at 
https://modal.com/apps/seth-45699/main/ap-L1S2KGTqxMKOoeFaRYjhEc
EXIT 0