jung-geun/PSO
1
0
Fork 0
mirror of https://github.com/jung-geun/PSO.git synced 2026-09-20 14:11:48 +09:00
Code Issues Packages Projects Releases Wiki Activity

feat: modernize PSO and add convergence research

Browse Source 
Migrate the package and examples to the tensor-native PyTorch implementation, add benchmark evidence, and add the guarded post-training convergence protocol with TensorBoard progress monitoring and hash-verified recovery.

Constraint: Preserve one-shot official-test sealing and auditable research artifacts

Rejected: Commit local .omc runs and downloaded datasets | multi-gigabyte runtime state is machine-local

Confidence: high

Scope-risk: broad

Not-tested: Production CUDA run on pieroot-server
This commit is contained in:
jung-geun
2026-09-07 22:03:25 +09:00
parent 9745bb7ad4
commit 813433000a
120 changed files with 198148 additions and 4745 deletions
Download Patch File Download Diff File Expand all files Collapse all files
benchmark_results/pso_v6_heavy_autoresearch.csv
+9
View File
@@ -0,0 +1,9 @@
workload,split,seeds,accuracy_mean,accuracy_sd,nll_mean,nll_sd,accuracy_gain_vs_baseline_pp,relative_nll_reduction_vs_baseline_pct,state_ratio,ratio,projection_mode,projection_seed
mnist_compact,development_seeds_101_103,101-103,49.053333,0.417652,1.525704,0.01494,-0.1,-0.501617,0.4918032786885246,0.5,fixed,1800044939
mnist_compact,confirmation_seeds_111_113,111-113,48.466667,2.515002,1.550212,0.046101,-0.686666,-2.116016,0.4918032786885246,0.5,fixed,1800044939
mnist_wide,development_seeds_101_103,101-103,48.346667,3.214052,1.677712,0.072138,4.486667,2.52995,0.5,0.5,explicit,592157828
mnist_wide,confirmation_seeds_111_113,111-113,45.723333,1.05633,1.734905,0.034461,1.863333,-0.792792,0.5,0.5,explicit,592157828
fashion_compact,development_seeds_101_103,101-103,50.796667,3.095228,1.385324,0.035646,3.793334,8.330571,0.4918032786885246,0.5,fixed,1363313651
fashion_compact,confirmation_seeds_111_113,111-113,53.196667,2.025669,1.328168,0.078488,6.193334,12.112688,0.4918032786885246,0.5,fixed,1363313651
fashion_wide,development_seeds_101_103,101-103,48.263333,3.71862,1.531421,0.069986,1.953333,-0.371883,0.5,0.5,fixed,189641451
fashion_wide,confirmation_seeds_111_113,111-113,49.476667,4.250251,1.513876,0.067096,3.166667,0.778045,0.5,0.5,fixed,189641451
1 workload split seeds accuracy_mean accuracy_sd nll_mean nll_sd accuracy_gain_vs_baseline_pp relative_nll_reduction_vs_baseline_pct state_ratio ratio projection_mode projection_seed
2 mnist_compact development_seeds_101_103 101-103 49.053333 0.417652 1.525704 0.01494 -0.1 -0.501617 0.4918032786885246 0.5 fixed 1800044939
3 mnist_compact confirmation_seeds_111_113 111-113 48.466667 2.515002 1.550212 0.046101 -0.686666 -2.116016 0.4918032786885246 0.5 fixed 1800044939
4 mnist_wide development_seeds_101_103 101-103 48.346667 3.214052 1.677712 0.072138 4.486667 2.52995 0.5 0.5 explicit 592157828
5 mnist_wide confirmation_seeds_111_113 111-113 45.723333 1.05633 1.734905 0.034461 1.863333 -0.792792 0.5 0.5 explicit 592157828
6 fashion_compact development_seeds_101_103 101-103 50.796667 3.095228 1.385324 0.035646 3.793334 8.330571 0.4918032786885246 0.5 fixed 1363313651
7 fashion_compact confirmation_seeds_111_113 111-113 53.196667 2.025669 1.328168 0.078488 6.193334 12.112688 0.4918032786885246 0.5 fixed 1363313651
8 fashion_wide development_seeds_101_103 101-103 48.263333 3.71862 1.531421 0.069986 1.953333 -0.371883 0.5 0.5 fixed 189641451
9 fashion_wide confirmation_seeds_111_113 111-113 49.476667 4.250251 1.513876 0.067096 3.166667 0.778045 0.5 0.5 fixed 189641451