diff options
| author | Yuren Hao <yurenh2@illinois.edu> | 2026-07-05 03:53:57 -0500 |
|---|---|---|
| committer | Yuren Hao <yurenh2@illinois.edu> | 2026-07-05 03:53:57 -0500 |
| commit | cbecb171b1af77fe6510fd60f1dfa4c7938a20d6 (patch) | |
| tree | adbe7b5ad4514930f510267963213afd21b9ebc4 /ep_run/holo_fast_gate.log | |
| parent | 1cea113ef78d2703b024a088703a06ac5d235c5c (diff) | |
holofast: exact halved-jvp AEP track — 1.55x nudged phase, gradient-gate passed
holo_a_track computed the doubled-batch jvp/vjp on [v0; -v0] at a shared
anchor zbar — exact antisymmetric redundancy (phase deviations from the
common mode are exact negatives). holo_a_track_fast computes at batch B and
mirrors: single-eval parity 6e-7 (exact); trajectory-level 45% divergence
SHARED with the original's own FD noise floor (1e-6 state noise -> 49%
self-divergence — the 2r=0.04 finite difference amplifies fp noise; training
averages it via pema/momentum). Ship gate: cos(EP,BPTT) orig vs fast
indistinguishable (0.907/0.912, 0.853/0.853, 0.918/0.918). Timing 6.43->4.16s
on the T2=40 nudged phase (contended GPU, relative). --holofast flag,
default off; queued ablation arms deliberately stay on orig for fidelity.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
Diffstat (limited to 'ep_run/holo_fast_gate.log')
| -rw-r--r-- | ep_run/holo_fast_gate.log | 7 |
1 files changed, 7 insertions, 0 deletions
diff --git a/ep_run/holo_fast_gate.log b/ep_run/holo_fast_gate.log new file mode 100644 index 0000000..a057dd4 --- /dev/null +++ b/ep_run/holo_fast_gate.log @@ -0,0 +1,7 @@ +batch cos_orig cos_fast res +/home/yurenh2/miniconda3/lib/python3.13/site-packages/torch/autograd/graph.py:865: UserWarning: Attempting to run cuBLAS, but there was no current CUDA context! Attempting to set the primary context... (Triggered internally at /pytorch/aten/src/ATen/cuda/CublasHandlePool.cpp:330.) + return Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + 0 0.9068 0.9117 9.47e-03 + 1 0.8533 0.8525 4.00e-03 + 2 0.9177 0.9178 4.00e-03 +HOLO_FAST_GATE_DONE |