matCercola18 commited on
Commit
0578bb2
·
verified ·
1 Parent(s): c00382a

results/icml_r3/X10M

Browse files
results/icml_r3/X10M/PREREG.md ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # X10M PREREG: learned-metric retrieval on the full-state MIMO family (written 2026-10-13 09:10 PDT, before any validation or test scoring of a new arm)
2
+
3
+ virgo CPU only (CUDA_VISIBLE_DEVICES empty), <= 12 workers. Code: `nonlinear/mimo_unstable/x10m.py` (reuses the X10 predictors in `x10_retr.py` and the X10 grid `x10_val.grid`); family `nonlinear/mimo_unstable/family.py` (family C, full-state gain K 2x4, verifier = one `family.verify` call per candidate).
4
+
5
+ ## Question
6
+ r3.md F4g: on full-state MIMO the generator certifies 96.1% at b = 1 against 43.3% for unit-box retrieval (+52.8 pts; 3 training seeds, 120 iid plants), and the regression twin ties the generator there (96.9). Does the margin over retrieval survive (a) a learned (ARD) metric, (b) the other X10 non-generative predictors (jitter, bootstrap, local-linear, GP-ARD posterior samples)? Nothing in the full-state family has been re-run with these predictors before.
7
+
8
+ ## Setup (X10 protocol)
9
+ - Bank: `results/icml_r2/C_mimo/bank6000.npz` (6000 plants, ONE certified gain each; the oracle SDP certified every kept plant). Because there is one stored solution per plant, `retr_all` (sols = 1) is plain kNN and the solution-ensemble KRR arms are degenerate and are NOT run. Arms: kNN with unit-box metric (D3f/C retrieval, reference) and ARD metric (GP-ARD marginal-likelihood fit on a 1500-plant bank subsample, `x10_retr.fit_ard`, bank only), jitter, bootstrap-mean, local-linear (`ll`) and local-linear with 4 bandwidths (`llbw`), GP-ARD posterior samples (`gp_ard`), each over the X10 grids (metric in {unit, ARD}).
10
+ - Validation: 160 FRESH plants (`build_bank --kind heldout --n 160 --seed 31415`, kept only if the oracle SDP certified a gain, exactly as the held-out sets; never in the bank or either test file), sampler seed 100. For each (arm, b in {1, 4, 16}) the configuration with the highest validation rate (ties: grid order) is frozen (`selected_configs.json`) BEFORE the test plants are verified.
11
+ - Test: the 120 iid plants `heldout_4242` + `heldout_9001` (the F4g set), sampler seeds 0, 1, 2 for stochastic arms. Generator = `cand_mlp_diff_s{0,1,2}.npz` (3 training seeds), regression twin = `cand_mlp_reg_s{0,1,2}.npz` (deterministic at b = 1, MC dropout beyond), re-verified here (check: generator 0.958 / 0.958 / 0.967 at b = 1 per seed, `eval_mlp.json`).
12
+
13
+ ## Tests
14
+ - Primary family, 6 tests (X10 rule): for b in {1, 4, 16}: generator minus **M1** (kNN with ARD metric; the `retr_all` ard sols=1 arm, no free hyperparameter) and generator minus **M2** (the non-generative arm among {unit-box kNN, jit, boot, ll, llbw, gp_ard} with the highest VALIDATION rate at that b). Exact two-sided McNemar per training seed (generator seed s against the comparator run with sampler seed s; deterministic comparators are the same run for every s), Holm on the worst-seed p over the 6 tests, a test is called significant only if all three per-seed tests pass after Holm; bootstrap 95% CI (10,000) over plants of the seed-averaged paired difference. The seed is the resampling unit for any across-seed statement; no pooled-over-seeds McNemar.
15
+ - Secondary (descriptive): generator minus every arm at its validation-selected configuration; the best non-generative arm on the test plants (post hoc, labelled); twin as a reference (not a comparator here).
16
+ - Falsification: the claim "the full-state margin over retrieval is not an artefact of a weak metric" is refuted at b if the generator minus M1 or M2 has a CI containing 0 or the Holm test is not significant for all seeds. Negative outcomes are reported as such.
17
+ - Limits: 120 plants, family rates are conditional on oracle solvability (as everywhere in family C); the 160 validation plants choose among ~100 configurations (winner's curse on the comparator, as in X10).
results/icml_r3/X10M/ard.npz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:816abeb9d6ed07a0685e8c3945ae7bbe5f6cf4dfd2d1a74be29467e3c26afa64
3
+ size 1318
results/icml_r3/X10M/driver.sh ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/bin/bash
2
+ # X8 seeds stage, then the X10M chain (validation plants, validation sweep, selection, test, report). CPU only, <= 12 workers, sequential.
3
+ cd /home/cercola/matte_server/DiffusionControl
4
+ export CUDA_VISIBLE_DEVICES= OMP_NUM_THREADS=1 RAYON_NUM_THREADS=1 PYTHONPATH=.
5
+ PY=myenv/bin/python
6
+ log() { echo "$(date '+%F %T') $*" >> results/icml_r3/X10M/logs/driver.log; }
7
+ mkdir -p results/icml_r3/X10M/logs
8
+ log "X8 seeds stage"; $PY -u -m nonlinear.experiments.furuta_x8 test --workers 12 --stage seeds > results/icml_r3/X8/logs/test_seeds.log 2>&1
9
+ log "X10M build val plants"; [ -f results/icml_r3/X10M/val.npz ] || $PY -u -m nonlinear.mimo_unstable.build_bank --kind heldout --n 160 --seed 31415 --workers 12 --out results/icml_r3/X10M/val.npz > results/icml_r3/X10M/logs/build_val.log 2>&1
10
+ log "X10M val sweep"; $PY -u -m nonlinear.mimo_unstable.x10m val --workers 12 > results/icml_r3/X10M/logs/val.log 2>&1
11
+ log "X10M select"; $PY -u -m nonlinear.mimo_unstable.x10m sel > results/icml_r3/X10M/logs/sel.log 2>&1
12
+ log "X10M test"; $PY -u -m nonlinear.mimo_unstable.x10m run --workers 12 > results/icml_r3/X10M/logs/test.log 2>&1
13
+ log "X10M report"; $PY -u -m nonlinear.mimo_unstable.x10m rep > results/icml_r3/X10M/logs/rep.log 2>&1
14
+ log "DONE"
results/icml_r3/X10M/logs/build_val.log ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ 160/160 0s
2
+ done results/icml_r3/X10M/val.npz
results/icml_r3/X10M/logs/driver.log ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ 2026-10-13 12:14:18 X8 seeds stage
2
+ 2026-10-13 12:30:24 X10M build val plants
3
+ 2026-10-13 12:30:24 X10M val sweep
4
+ 2026-10-13 12:34:11 X10M select
5
+ 2026-10-13 12:34:12 X10M test
6
+ 2026-10-13 12:34:47 X10M report
7
+ 2026-10-13 12:34:48 DONE
results/icml_r3/X10M/logs/rep.log ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # X10M test report (full-state MIMO, 120 iid plants, bank6000 = 6000 plants x 1 gain; validation 160 fresh plants)
2
+
3
+ test success at b = 1 / 4 / 16 (3 training/sampler seeds averaged)
4
+ arm b=1 b=4 b=16
5
+ diffusion 0.961 0.997 1.000
6
+ retr 0.433 0.700 0.833 cfg
7
+ knn_ard 0.425 0.675 0.883 cfg
8
+ jit 0.425 0.675 0.881 cfg b1:metric=ard,k=1,sigma=0.1; b4:metric=ard,k=4,sigma=0.1; b16:metric=ard,k=4,sigma=1.0
9
+ boot 0.617 0.736 0.819 cfg b1:metric=ard,kb=8; b4:metric=ard,kb=8; b16:metric=ard,kb=8
10
+ ll 0.867 0.867 0.939 cfg b1:metric=ard,k=64,lam=0.3; b4:metric=ard,k=128,lam=0.01; b16:metric=ard,k=64,lam=0.03
11
+ llbw 0.867 0.883 0.883 cfg b1:metric=ard,k=64,lam=0.3; b4:metric=ard,k=16,lam=0.01; b16:metric=ard,k=16,lam=0.01
12
+ gp_ard 0.717 0.811 0.919 cfg b1:c=0.7,eta=10.0,tau=0.5; b4:c=0.7,eta=0.3,tau=0.5; b16:c=0.7,eta=0.3,tau=0.5
13
+ twin 0.969 0.972 0.983 cfg
14
+
15
+ ## PRIMARY family (6 tests, X10 rule: exact McNemar per training seed, Holm on the worst-seed p; significant only if all 3 seeds pass), pp [95% bootstrap CI over plants]
16
+ b=1 diffusion - kNN-ARD (M1) +53.6 [+44.7,+62.8] per-seed (gen-only/comp-only, p): 65/1 p=1.82e-18; 66/2 p=1.59e-17; 66/1 p=9.22e-19 worst-seed p 1.59e-17 Holm 9.54e-17
17
+ b=1 diffusion - best validated non-generative arm (M2 = ll) +9.4 [ +4.7,+14.7] per-seed (gen-only/comp-only, p): 11/0 p=0.000977; 12/1 p=0.00342; 13/1 p=0.00183 worst-seed p 0.00342 Holm 0.00684
18
+ b=4 diffusion - kNN-ARD (M1) +32.2 [+23.9,+40.6] per-seed (gen-only/comp-only, p): 38/0 p=7.28e-12; 39/0 p=3.64e-12; 39/0 p=3.64e-12 worst-seed p 7.28e-12 Holm 3.64e-11
19
+ b=4 diffusion - best validated non-generative arm (M2 = ll) +13.1 [ +7.8,+18.9] per-seed (gen-only/comp-only, p): 15/0 p=6.1e-05; 16/0 p=3.05e-05; 16/0 p=3.05e-05 worst-seed p 6.1e-05 Holm 0.000244
20
+ b=16 diffusion - kNN-ARD (M1) +11.7 [ +6.7,+17.5] per-seed (gen-only/comp-only, p): 14/0 p=0.000122; 14/0 p=0.000122; 14/0 p=0.000122 worst-seed p 0.000122 Holm 0.000366
21
+ b=16 diffusion - best validated non-generative arm (M2 = ll) +6.1 [ +2.8,+10.0] per-seed (gen-only/comp-only, p): 7/0 p=0.0156; 7/0 p=0.0156; 8/0 p=0.00781 worst-seed p 0.0156 Holm 0.0156
22
+
23
+ ## SECONDARY: diffusion minus every arm (unadjusted)
24
+ b= 1 diffusion - retr : +52.8 [+44.2,+61.4] per-seed p 2.17e-19, 2.17e-19, 1.82e-18
25
+ b= 1 diffusion - knn_ard : +53.6 [+44.7,+62.8] per-seed p 1.82e-18, 1.59e-17, 9.22e-19
26
+ b= 1 diffusion - jit : +53.6 [+44.4,+62.2] per-seed p 1.82e-18, 1.59e-17, 9.22e-19
27
+ b= 1 diffusion - boot : +34.4 [+26.1,+43.1] per-seed p 9.09e-13, 5.89e-11, 3.08e-11
28
+ b= 1 diffusion - ll : +9.4 [ +4.7,+14.5] per-seed p 0.000977, 0.00342, 0.00183
29
+ b= 1 diffusion - llbw : +9.4 [ +4.7,+14.7] per-seed p 0.000977, 0.00342, 0.00183
30
+ b= 1 diffusion - gp_ard : +24.4 [+16.9,+31.9] per-seed p 3.73e-09, 2.98e-08, 1.54e-08
31
+ b= 1 diffusion - twin : -0.8 [ -2.2, +0.6] per-seed p 0.625, 0.5, 1
32
+ b= 4 diffusion - retr : +29.7 [+21.7,+38.1] per-seed p 5.82e-11, 2.91e-11, 2.91e-11
33
+ b= 4 diffusion - knn_ard : +32.2 [+23.9,+40.8] per-seed p 7.28e-12, 3.64e-12, 3.64e-12
34
+ b= 4 diffusion - jit : +32.2 [+24.2,+40.6] per-seed p 7.28e-12, 3.64e-12, 3.64e-12
35
+ b= 4 diffusion - boot : +26.1 [+19.2,+33.3] per-seed p 7.45e-09, 1.16e-10, 4.66e-10
36
+ b= 4 diffusion - ll : +13.1 [ +7.8,+18.6] per-seed p 6.1e-05, 3.05e-05, 3.05e-05
37
+ b= 4 diffusion - llbw : +11.4 [ +5.8,+17.5] per-seed p 0.000244, 0.000122, 0.000122
38
+ b= 4 diffusion - gp_ard : +18.6 [+12.8,+25.0] per-seed p 7.63e-06, 2.38e-07, 2.98e-08
39
+ b= 4 diffusion - twin : +2.5 [ +0.3, +5.3] per-seed p 0.625, 0.25, 0.125
40
+ b=16 diffusion - retr : +16.7 [+10.0,+23.3] per-seed p 1.91e-06, 1.91e-06, 1.91e-06
41
+ b=16 diffusion - knn_ard : +11.7 [ +6.6,+17.5] per-seed p 0.000122, 0.000122, 0.000122
42
+ b=16 diffusion - jit : +11.9 [ +7.2,+16.9] per-seed p 0.000488, 1.53e-05, 0.000122
43
+ b=16 diffusion - boot : +18.1 [+11.7,+24.7] per-seed p 1.91e-06, 4.77e-07, 2.38e-07
44
+ b=16 diffusion - ll : +6.1 [ +2.8,+10.3] per-seed p 0.0156, 0.0156, 0.00781
45
+ b=16 diffusion - llbw : +11.7 [ +6.7,+17.5] per-seed p 0.000122, 0.000122, 0.000122
46
+ b=16 diffusion - gp_ard : +8.1 [ +4.2,+12.5] per-seed p 0.00391, 0.000977, 0.00391
47
+ b=16 diffusion - twin : +1.7 [ +0.0, +3.9] per-seed p 1, 0.5, 0.25
48
+
49
+ ## Post hoc: best non-generative arm ON the test plants per b
50
+ b=1: ll 0.867; diffusion minus it +9.4 [+4.7,+14.7]
51
+ b=4: llbw 0.883; diffusion minus it +11.4 [+5.8,+17.2]
52
+ b=16: ll 0.939; diffusion minus it +6.1 [+2.8,+10.0]
results/icml_r3/X10M/logs/sel.log ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "1": {
3
+ "arm": "ll",
4
+ "val_rate": 0.925
5
+ },
6
+ "4": {
7
+ "arm": "ll",
8
+ "val_rate": 0.94375
9
+ },
10
+ "16": {
11
+ "arm": "ll",
12
+ "val_rate": 0.96875
13
+ }
14
+ } {
15
+ "1": {
16
+ "key": "retr_all:metric=ard,sols=1",
17
+ "val_rate": 0.43125
18
+ },
19
+ "4": {
20
+ "key": "retr_all:metric=ard,sols=1",
21
+ "val_rate": 0.725
22
+ },
23
+ "16": {
24
+ "key": "retr_all:metric=ard,sols=1",
25
+ "val_rate": 0.9
26
+ }
27
+ }
results/icml_r3/X10M/logs/test.log ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ candidates built 40 4s
2
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
3
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
4
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
5
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
6
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
7
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
8
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
9
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
10
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
11
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
12
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
13
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
14
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
15
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
16
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
17
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
18
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
19
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
20
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
21
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
22
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
23
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
24
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
25
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
26
+ 0 40 5s
27
+ 5 40 9s
28
+ 10 40 12s
29
+ 15 40 17s
30
+ 20 40 21s
31
+ 25 40 24s
32
+ 30 40 29s
33
+ 35 40 33s
34
+ run done 35s
results/icml_r3/X10M/logs/val.log ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
2
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
3
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
4
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
5
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
6
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
7
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
8
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
9
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
10
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
11
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
12
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
13
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
14
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
15
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
16
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
17
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
18
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
19
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
20
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
21
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
22
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
23
+ /home/cercola/matte_server/DiffusionControl/nonlinear/mimo_unstable/family.py:60: UserWarning: Solution may be inaccurate. Try another solver, adjusting the solver settings, or solve with verbose=True for more information.
24
+ cp.Problem(cp.Minimize(0), cons).solve(solver=SOLVER)
25
+ 0 172 1s {1: 0.425, 4: 0.625, 16: 0.8625}
26
+ 10 172 12s {1: 0.425, 4: 0.625, 16: 0.76875}
27
+ 20 172 22s {1: 0.4375, 4: 0.51875, 16: 0.5875}
28
+ 30 172 35s {1: 0.39375, 4: 0.5, 16: 0.58125}
29
+ 40 172 42s {1: 0.34375, 4: 0.46875, 16: 0.46875}
30
+ 50 172 50s {1: 0.43125, 4: 0.55625, 16: 0.80625}
31
+ 60 172 59s {1: 0.43125, 4: 0.725, 16: 0.85625}
32
+ 70 172 70s {1: 0.8625, 4: 0.9, 16: 0.93125}
33
+ 80 172 82s {1: 0.81875, 4: 0.94375, 16: 0.94375}
34
+ 90 172 86s {1: 0.90625, 4: 0.93125, 16: 0.93125}
35
+ 100 172 103s {1: 0.23125, 4: 0.3125, 16: 0.46875}
36
+ 110 172 119s {1: 0.65, 4: 0.6875, 16: 0.825}
37
+ 120 172 136s {1: 0.65625, 4: 0.7375, 16: 0.8625}
38
+ 130 172 152s {1: 0.8125, 4: 0.8625, 16: 0.9375}
39
+ 140 172 168s {1: 0.7625, 4: 0.825, 16: 0.8875}
40
+ 150 172 184s {1: 0.775, 4: 0.84375, 16: 0.94375}
41
+ 160 172 200s {1: 0.61875, 4: 0.6375, 16: 0.68125}
42
+ 170 172 216s {1: 0.5875, 4: 0.7125, 16: 0.8375}
43
+ val done 217s
results/icml_r3/X10M/selected_configs.json ADDED
@@ -0,0 +1,201 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "n_val": 160,
3
+ "arms": {
4
+ "retr_all": {
5
+ "1": {
6
+ "key": "retr_all:metric=ard,sols=1",
7
+ "cfg": {
8
+ "metric": "ard",
9
+ "sols": 1
10
+ },
11
+ "val_rate": 0.43125
12
+ },
13
+ "4": {
14
+ "key": "retr_all:metric=ard,sols=1",
15
+ "cfg": {
16
+ "metric": "ard",
17
+ "sols": 1
18
+ },
19
+ "val_rate": 0.725
20
+ },
21
+ "16": {
22
+ "key": "retr_all:metric=ard,sols=1",
23
+ "cfg": {
24
+ "metric": "ard",
25
+ "sols": 1
26
+ },
27
+ "val_rate": 0.9
28
+ }
29
+ },
30
+ "jit": {
31
+ "1": {
32
+ "key": "jit:k=1,metric=ard,sigma=0.1",
33
+ "cfg": {
34
+ "metric": "ard",
35
+ "k": 1,
36
+ "sigma": 0.1
37
+ },
38
+ "val_rate": 0.43125
39
+ },
40
+ "4": {
41
+ "key": "jit:k=4,metric=ard,sigma=0.1",
42
+ "cfg": {
43
+ "metric": "ard",
44
+ "k": 4,
45
+ "sigma": 0.1
46
+ },
47
+ "val_rate": 0.725
48
+ },
49
+ "16": {
50
+ "key": "jit:k=4,metric=ard,sigma=1.0",
51
+ "cfg": {
52
+ "metric": "ard",
53
+ "k": 4,
54
+ "sigma": 1.0
55
+ },
56
+ "val_rate": 0.9
57
+ }
58
+ },
59
+ "boot": {
60
+ "1": {
61
+ "key": "boot:kb=8,metric=ard",
62
+ "cfg": {
63
+ "metric": "ard",
64
+ "kb": 8
65
+ },
66
+ "val_rate": 0.5625
67
+ },
68
+ "4": {
69
+ "key": "boot:kb=8,metric=ard",
70
+ "cfg": {
71
+ "metric": "ard",
72
+ "kb": 8
73
+ },
74
+ "val_rate": 0.69375
75
+ },
76
+ "16": {
77
+ "key": "boot:kb=8,metric=ard",
78
+ "cfg": {
79
+ "metric": "ard",
80
+ "kb": 8
81
+ },
82
+ "val_rate": 0.81875
83
+ }
84
+ },
85
+ "ll": {
86
+ "1": {
87
+ "key": "ll:k=64,lam=0.3,metric=ard",
88
+ "cfg": {
89
+ "metric": "ard",
90
+ "k": 64,
91
+ "lam": 0.3
92
+ },
93
+ "val_rate": 0.925
94
+ },
95
+ "4": {
96
+ "key": "ll:k=128,lam=0.01,metric=ard",
97
+ "cfg": {
98
+ "metric": "ard",
99
+ "k": 128,
100
+ "lam": 0.01
101
+ },
102
+ "val_rate": 0.94375
103
+ },
104
+ "16": {
105
+ "key": "ll:k=64,lam=0.03,metric=ard",
106
+ "cfg": {
107
+ "metric": "ard",
108
+ "k": 64,
109
+ "lam": 0.03
110
+ },
111
+ "val_rate": 0.96875
112
+ }
113
+ },
114
+ "llbw": {
115
+ "1": {
116
+ "key": "llbw:k=64,lam=0.3,metric=ard",
117
+ "cfg": {
118
+ "metric": "ard",
119
+ "k": 64,
120
+ "lam": 0.3
121
+ },
122
+ "val_rate": 0.925
123
+ },
124
+ "4": {
125
+ "key": "llbw:k=16,lam=0.01,metric=ard",
126
+ "cfg": {
127
+ "metric": "ard",
128
+ "k": 16,
129
+ "lam": 0.01
130
+ },
131
+ "val_rate": 0.94375
132
+ },
133
+ "16": {
134
+ "key": "llbw:k=16,lam=0.01,metric=ard",
135
+ "cfg": {
136
+ "metric": "ard",
137
+ "k": 16,
138
+ "lam": 0.01
139
+ },
140
+ "val_rate": 0.94375
141
+ }
142
+ },
143
+ "gp_ard": {
144
+ "1": {
145
+ "key": "gp_ard:c=0.7,eta=10.0,tau=0.5",
146
+ "cfg": {
147
+ "c": 0.7,
148
+ "eta": 10.0,
149
+ "tau": 0.5
150
+ },
151
+ "val_rate": 0.83125
152
+ },
153
+ "4": {
154
+ "key": "gp_ard:c=0.7,eta=0.3,tau=0.5",
155
+ "cfg": {
156
+ "c": 0.7,
157
+ "eta": 0.3,
158
+ "tau": 0.5
159
+ },
160
+ "val_rate": 0.90625
161
+ },
162
+ "16": {
163
+ "key": "gp_ard:c=0.7,eta=0.3,tau=0.5",
164
+ "cfg": {
165
+ "c": 0.7,
166
+ "eta": 0.3,
167
+ "tau": 0.5
168
+ },
169
+ "val_rate": 0.9625
170
+ }
171
+ }
172
+ },
173
+ "comparator": {
174
+ "1": {
175
+ "arm": "ll",
176
+ "val_rate": 0.925
177
+ },
178
+ "4": {
179
+ "arm": "ll",
180
+ "val_rate": 0.94375
181
+ },
182
+ "16": {
183
+ "arm": "ll",
184
+ "val_rate": 0.96875
185
+ }
186
+ },
187
+ "knn_ard": {
188
+ "1": {
189
+ "key": "retr_all:metric=ard,sols=1",
190
+ "val_rate": 0.43125
191
+ },
192
+ "4": {
193
+ "key": "retr_all:metric=ard,sols=1",
194
+ "val_rate": 0.725
195
+ },
196
+ "16": {
197
+ "key": "retr_all:metric=ard,sols=1",
198
+ "val_rate": 0.9
199
+ }
200
+ }
201
+ }
results/icml_r3/X10M/test.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:40b208bbafa0f08ccd620846346f8b6812947a62ae417bd79a31342e459a2499
3
+ size 78947
results/icml_r3/X10M/test_report.txt ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # X10M test report (full-state MIMO, 120 iid plants, bank6000 = 6000 plants x 1 gain; validation 160 fresh plants)
2
+
3
+ test success at b = 1 / 4 / 16 (3 training/sampler seeds averaged)
4
+ arm b=1 b=4 b=16
5
+ diffusion 0.961 0.997 1.000
6
+ retr 0.433 0.700 0.833 cfg
7
+ knn_ard 0.425 0.675 0.883 cfg
8
+ jit 0.425 0.675 0.881 cfg b1:metric=ard,k=1,sigma=0.1; b4:metric=ard,k=4,sigma=0.1; b16:metric=ard,k=4,sigma=1.0
9
+ boot 0.617 0.736 0.819 cfg b1:metric=ard,kb=8; b4:metric=ard,kb=8; b16:metric=ard,kb=8
10
+ ll 0.867 0.867 0.939 cfg b1:metric=ard,k=64,lam=0.3; b4:metric=ard,k=128,lam=0.01; b16:metric=ard,k=64,lam=0.03
11
+ llbw 0.867 0.883 0.883 cfg b1:metric=ard,k=64,lam=0.3; b4:metric=ard,k=16,lam=0.01; b16:metric=ard,k=16,lam=0.01
12
+ gp_ard 0.717 0.811 0.919 cfg b1:c=0.7,eta=10.0,tau=0.5; b4:c=0.7,eta=0.3,tau=0.5; b16:c=0.7,eta=0.3,tau=0.5
13
+ twin 0.969 0.972 0.983 cfg
14
+
15
+ ## PRIMARY family (6 tests, X10 rule: exact McNemar per training seed, Holm on the worst-seed p; significant only if all 3 seeds pass), pp [95% bootstrap CI over plants]
16
+ b=1 diffusion - kNN-ARD (M1) +53.6 [+44.7,+62.8] per-seed (gen-only/comp-only, p): 65/1 p=1.82e-18; 66/2 p=1.59e-17; 66/1 p=9.22e-19 worst-seed p 1.59e-17 Holm 9.54e-17
17
+ b=1 diffusion - best validated non-generative arm (M2 = ll) +9.4 [ +4.7,+14.7] per-seed (gen-only/comp-only, p): 11/0 p=0.000977; 12/1 p=0.00342; 13/1 p=0.00183 worst-seed p 0.00342 Holm 0.00684
18
+ b=4 diffusion - kNN-ARD (M1) +32.2 [+23.9,+40.6] per-seed (gen-only/comp-only, p): 38/0 p=7.28e-12; 39/0 p=3.64e-12; 39/0 p=3.64e-12 worst-seed p 7.28e-12 Holm 3.64e-11
19
+ b=4 diffusion - best validated non-generative arm (M2 = ll) +13.1 [ +7.8,+18.9] per-seed (gen-only/comp-only, p): 15/0 p=6.1e-05; 16/0 p=3.05e-05; 16/0 p=3.05e-05 worst-seed p 6.1e-05 Holm 0.000244
20
+ b=16 diffusion - kNN-ARD (M1) +11.7 [ +6.7,+17.5] per-seed (gen-only/comp-only, p): 14/0 p=0.000122; 14/0 p=0.000122; 14/0 p=0.000122 worst-seed p 0.000122 Holm 0.000366
21
+ b=16 diffusion - best validated non-generative arm (M2 = ll) +6.1 [ +2.8,+10.0] per-seed (gen-only/comp-only, p): 7/0 p=0.0156; 7/0 p=0.0156; 8/0 p=0.00781 worst-seed p 0.0156 Holm 0.0156
22
+
23
+ ## SECONDARY: diffusion minus every arm (unadjusted)
24
+ b= 1 diffusion - retr : +52.8 [+44.2,+61.4] per-seed p 2.17e-19, 2.17e-19, 1.82e-18
25
+ b= 1 diffusion - knn_ard : +53.6 [+44.7,+62.8] per-seed p 1.82e-18, 1.59e-17, 9.22e-19
26
+ b= 1 diffusion - jit : +53.6 [+44.4,+62.2] per-seed p 1.82e-18, 1.59e-17, 9.22e-19
27
+ b= 1 diffusion - boot : +34.4 [+26.1,+43.1] per-seed p 9.09e-13, 5.89e-11, 3.08e-11
28
+ b= 1 diffusion - ll : +9.4 [ +4.7,+14.5] per-seed p 0.000977, 0.00342, 0.00183
29
+ b= 1 diffusion - llbw : +9.4 [ +4.7,+14.7] per-seed p 0.000977, 0.00342, 0.00183
30
+ b= 1 diffusion - gp_ard : +24.4 [+16.9,+31.9] per-seed p 3.73e-09, 2.98e-08, 1.54e-08
31
+ b= 1 diffusion - twin : -0.8 [ -2.2, +0.6] per-seed p 0.625, 0.5, 1
32
+ b= 4 diffusion - retr : +29.7 [+21.7,+38.1] per-seed p 5.82e-11, 2.91e-11, 2.91e-11
33
+ b= 4 diffusion - knn_ard : +32.2 [+23.9,+40.8] per-seed p 7.28e-12, 3.64e-12, 3.64e-12
34
+ b= 4 diffusion - jit : +32.2 [+24.2,+40.6] per-seed p 7.28e-12, 3.64e-12, 3.64e-12
35
+ b= 4 diffusion - boot : +26.1 [+19.2,+33.3] per-seed p 7.45e-09, 1.16e-10, 4.66e-10
36
+ b= 4 diffusion - ll : +13.1 [ +7.8,+18.6] per-seed p 6.1e-05, 3.05e-05, 3.05e-05
37
+ b= 4 diffusion - llbw : +11.4 [ +5.8,+17.5] per-seed p 0.000244, 0.000122, 0.000122
38
+ b= 4 diffusion - gp_ard : +18.6 [+12.8,+25.0] per-seed p 7.63e-06, 2.38e-07, 2.98e-08
39
+ b= 4 diffusion - twin : +2.5 [ +0.3, +5.3] per-seed p 0.625, 0.25, 0.125
40
+ b=16 diffusion - retr : +16.7 [+10.0,+23.3] per-seed p 1.91e-06, 1.91e-06, 1.91e-06
41
+ b=16 diffusion - knn_ard : +11.7 [ +6.6,+17.5] per-seed p 0.000122, 0.000122, 0.000122
42
+ b=16 diffusion - jit : +11.9 [ +7.2,+16.9] per-seed p 0.000488, 1.53e-05, 0.000122
43
+ b=16 diffusion - boot : +18.1 [+11.7,+24.7] per-seed p 1.91e-06, 4.77e-07, 2.38e-07
44
+ b=16 diffusion - ll : +6.1 [ +2.8,+10.3] per-seed p 0.0156, 0.0156, 0.00781
45
+ b=16 diffusion - llbw : +11.7 [ +6.7,+17.5] per-seed p 0.000122, 0.000122, 0.000122
46
+ b=16 diffusion - gp_ard : +8.1 [ +4.2,+12.5] per-seed p 0.00391, 0.000977, 0.00391
47
+ b=16 diffusion - twin : +1.7 [ +0.0, +3.9] per-seed p 1, 0.5, 0.25
48
+
49
+ ## Post hoc: best non-generative arm ON the test plants per b
50
+ b=1: ll 0.867; diffusion minus it +9.4 [+4.7,+14.7]
51
+ b=4: llbw 0.883; diffusion minus it +11.4 [+5.8,+17.2]
52
+ b=16: ll 0.939; diffusion minus it +6.1 [+2.8,+10.0]
results/icml_r3/X10M/val.json ADDED
@@ -0,0 +1,2241 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "n_val": 160,
3
+ "ard_l": [
4
+ 1.141477565178084,
5
+ 0.9193722691308082,
6
+ 2.85479311232069,
7
+ 73.71591198656911,
8
+ 0.8251647790058365,
9
+ 1.000521510421433,
10
+ 0.1389350200881672,
11
+ 0.14881183192517572,
12
+ 0.4948876672723918,
13
+ 0.5515086749273326
14
+ ],
15
+ "rows": {
16
+ "retr_all:metric=unit,sols=1": {
17
+ "arm": "retr_all",
18
+ "cfg": {
19
+ "metric": "unit",
20
+ "sols": 1
21
+ },
22
+ "rate": {
23
+ "1": 0.425,
24
+ "4": 0.625,
25
+ "16": 0.8625
26
+ }
27
+ },
28
+ "jit:k=1,metric=unit,sigma=0.1": {
29
+ "arm": "jit",
30
+ "cfg": {
31
+ "metric": "unit",
32
+ "k": 1,
33
+ "sigma": 0.1
34
+ },
35
+ "rate": {
36
+ "1": 0.425,
37
+ "4": 0.4625,
38
+ "16": 0.51875
39
+ }
40
+ },
41
+ "jit:k=1,metric=unit,sigma=0.3": {
42
+ "arm": "jit",
43
+ "cfg": {
44
+ "metric": "unit",
45
+ "k": 1,
46
+ "sigma": 0.3
47
+ },
48
+ "rate": {
49
+ "1": 0.425,
50
+ "4": 0.5375,
51
+ "16": 0.6125
52
+ }
53
+ },
54
+ "jit:k=1,metric=unit,sigma=1.0": {
55
+ "arm": "jit",
56
+ "cfg": {
57
+ "metric": "unit",
58
+ "k": 1,
59
+ "sigma": 1.0
60
+ },
61
+ "rate": {
62
+ "1": 0.425,
63
+ "4": 0.54375,
64
+ "16": 0.725
65
+ }
66
+ },
67
+ "jit:k=1,metric=unit,sigma=2.0": {
68
+ "arm": "jit",
69
+ "cfg": {
70
+ "metric": "unit",
71
+ "k": 1,
72
+ "sigma": 2.0
73
+ },
74
+ "rate": {
75
+ "1": 0.425,
76
+ "4": 0.51875,
77
+ "16": 0.7375
78
+ }
79
+ },
80
+ "jit:k=2,metric=unit,sigma=0.1": {
81
+ "arm": "jit",
82
+ "cfg": {
83
+ "metric": "unit",
84
+ "k": 2,
85
+ "sigma": 0.1
86
+ },
87
+ "rate": {
88
+ "1": 0.425,
89
+ "4": 0.5625,
90
+ "16": 0.63125
91
+ }
92
+ },
93
+ "jit:k=2,metric=unit,sigma=0.3": {
94
+ "arm": "jit",
95
+ "cfg": {
96
+ "metric": "unit",
97
+ "k": 2,
98
+ "sigma": 0.3
99
+ },
100
+ "rate": {
101
+ "1": 0.425,
102
+ "4": 0.6125,
103
+ "16": 0.7375
104
+ }
105
+ },
106
+ "jit:k=2,metric=unit,sigma=1.0": {
107
+ "arm": "jit",
108
+ "cfg": {
109
+ "metric": "unit",
110
+ "k": 2,
111
+ "sigma": 1.0
112
+ },
113
+ "rate": {
114
+ "1": 0.425,
115
+ "4": 0.63125,
116
+ "16": 0.825
117
+ }
118
+ },
119
+ "jit:k=2,metric=unit,sigma=2.0": {
120
+ "arm": "jit",
121
+ "cfg": {
122
+ "metric": "unit",
123
+ "k": 2,
124
+ "sigma": 2.0
125
+ },
126
+ "rate": {
127
+ "1": 0.425,
128
+ "4": 0.59375,
129
+ "16": 0.7125
130
+ }
131
+ },
132
+ "jit:k=4,metric=unit,sigma=0.1": {
133
+ "arm": "jit",
134
+ "cfg": {
135
+ "metric": "unit",
136
+ "k": 4,
137
+ "sigma": 0.1
138
+ },
139
+ "rate": {
140
+ "1": 0.425,
141
+ "4": 0.625,
142
+ "16": 0.675
143
+ }
144
+ },
145
+ "jit:k=4,metric=unit,sigma=0.3": {
146
+ "arm": "jit",
147
+ "cfg": {
148
+ "metric": "unit",
149
+ "k": 4,
150
+ "sigma": 0.3
151
+ },
152
+ "rate": {
153
+ "1": 0.425,
154
+ "4": 0.625,
155
+ "16": 0.76875
156
+ }
157
+ },
158
+ "jit:k=4,metric=unit,sigma=1.0": {
159
+ "arm": "jit",
160
+ "cfg": {
161
+ "metric": "unit",
162
+ "k": 4,
163
+ "sigma": 1.0
164
+ },
165
+ "rate": {
166
+ "1": 0.425,
167
+ "4": 0.625,
168
+ "16": 0.7875
169
+ }
170
+ },
171
+ "jit:k=4,metric=unit,sigma=2.0": {
172
+ "arm": "jit",
173
+ "cfg": {
174
+ "metric": "unit",
175
+ "k": 4,
176
+ "sigma": 2.0
177
+ },
178
+ "rate": {
179
+ "1": 0.425,
180
+ "4": 0.625,
181
+ "16": 0.79375
182
+ }
183
+ },
184
+ "jit:k=8,metric=unit,sigma=0.1": {
185
+ "arm": "jit",
186
+ "cfg": {
187
+ "metric": "unit",
188
+ "k": 8,
189
+ "sigma": 0.1
190
+ },
191
+ "rate": {
192
+ "1": 0.425,
193
+ "4": 0.625,
194
+ "16": 0.80625
195
+ }
196
+ },
197
+ "jit:k=8,metric=unit,sigma=0.3": {
198
+ "arm": "jit",
199
+ "cfg": {
200
+ "metric": "unit",
201
+ "k": 8,
202
+ "sigma": 0.3
203
+ },
204
+ "rate": {
205
+ "1": 0.425,
206
+ "4": 0.625,
207
+ "16": 0.81875
208
+ }
209
+ },
210
+ "jit:k=8,metric=unit,sigma=1.0": {
211
+ "arm": "jit",
212
+ "cfg": {
213
+ "metric": "unit",
214
+ "k": 8,
215
+ "sigma": 1.0
216
+ },
217
+ "rate": {
218
+ "1": 0.425,
219
+ "4": 0.625,
220
+ "16": 0.83125
221
+ }
222
+ },
223
+ "jit:k=8,metric=unit,sigma=2.0": {
224
+ "arm": "jit",
225
+ "cfg": {
226
+ "metric": "unit",
227
+ "k": 8,
228
+ "sigma": 2.0
229
+ },
230
+ "rate": {
231
+ "1": 0.425,
232
+ "4": 0.625,
233
+ "16": 0.83125
234
+ }
235
+ },
236
+ "boot:kb=4,metric=unit": {
237
+ "arm": "boot",
238
+ "cfg": {
239
+ "metric": "unit",
240
+ "kb": 4
241
+ },
242
+ "rate": {
243
+ "1": 0.46875,
244
+ "4": 0.56875,
245
+ "16": 0.64375
246
+ }
247
+ },
248
+ "boot:kb=8,metric=unit": {
249
+ "arm": "boot",
250
+ "cfg": {
251
+ "metric": "unit",
252
+ "kb": 8
253
+ },
254
+ "rate": {
255
+ "1": 0.45625,
256
+ "4": 0.575,
257
+ "16": 0.6875
258
+ }
259
+ },
260
+ "boot:kb=16,metric=unit": {
261
+ "arm": "boot",
262
+ "cfg": {
263
+ "metric": "unit",
264
+ "kb": 16
265
+ },
266
+ "rate": {
267
+ "1": 0.45,
268
+ "4": 0.5375,
269
+ "16": 0.64375
270
+ }
271
+ },
272
+ "boot:kb=32,metric=unit": {
273
+ "arm": "boot",
274
+ "cfg": {
275
+ "metric": "unit",
276
+ "kb": 32
277
+ },
278
+ "rate": {
279
+ "1": 0.4375,
280
+ "4": 0.51875,
281
+ "16": 0.5875
282
+ }
283
+ },
284
+ "boot:kb=64,metric=unit": {
285
+ "arm": "boot",
286
+ "cfg": {
287
+ "metric": "unit",
288
+ "kb": 64
289
+ },
290
+ "rate": {
291
+ "1": 0.39375,
292
+ "4": 0.44375,
293
+ "16": 0.50625
294
+ }
295
+ },
296
+ "ll:k=32,lam=0.01,metric=unit": {
297
+ "arm": "ll",
298
+ "cfg": {
299
+ "metric": "unit",
300
+ "k": 32,
301
+ "lam": 0.01
302
+ },
303
+ "rate": {
304
+ "1": 0.35,
305
+ "4": 0.525,
306
+ "16": 0.7
307
+ }
308
+ },
309
+ "ll:k=32,lam=0.03,metric=unit": {
310
+ "arm": "ll",
311
+ "cfg": {
312
+ "metric": "unit",
313
+ "k": 32,
314
+ "lam": 0.03
315
+ },
316
+ "rate": {
317
+ "1": 0.375,
318
+ "4": 0.5875,
319
+ "16": 0.70625
320
+ }
321
+ },
322
+ "ll:k=32,lam=0.3,metric=unit": {
323
+ "arm": "ll",
324
+ "cfg": {
325
+ "metric": "unit",
326
+ "k": 32,
327
+ "lam": 0.3
328
+ },
329
+ "rate": {
330
+ "1": 0.475,
331
+ "4": 0.5625,
332
+ "16": 0.6375
333
+ }
334
+ },
335
+ "ll:k=64,lam=0.01,metric=unit": {
336
+ "arm": "ll",
337
+ "cfg": {
338
+ "metric": "unit",
339
+ "k": 64,
340
+ "lam": 0.01
341
+ },
342
+ "rate": {
343
+ "1": 0.34375,
344
+ "4": 0.475,
345
+ "16": 0.65
346
+ }
347
+ },
348
+ "ll:k=64,lam=0.03,metric=unit": {
349
+ "arm": "ll",
350
+ "cfg": {
351
+ "metric": "unit",
352
+ "k": 64,
353
+ "lam": 0.03
354
+ },
355
+ "rate": {
356
+ "1": 0.35625,
357
+ "4": 0.48125,
358
+ "16": 0.6125
359
+ }
360
+ },
361
+ "ll:k=64,lam=0.3,metric=unit": {
362
+ "arm": "ll",
363
+ "cfg": {
364
+ "metric": "unit",
365
+ "k": 64,
366
+ "lam": 0.3
367
+ },
368
+ "rate": {
369
+ "1": 0.40625,
370
+ "4": 0.51875,
371
+ "16": 0.6125
372
+ }
373
+ },
374
+ "ll:k=128,lam=0.01,metric=unit": {
375
+ "arm": "ll",
376
+ "cfg": {
377
+ "metric": "unit",
378
+ "k": 128,
379
+ "lam": 0.01
380
+ },
381
+ "rate": {
382
+ "1": 0.35,
383
+ "4": 0.475,
384
+ "16": 0.6125
385
+ }
386
+ },
387
+ "ll:k=128,lam=0.03,metric=unit": {
388
+ "arm": "ll",
389
+ "cfg": {
390
+ "metric": "unit",
391
+ "k": 128,
392
+ "lam": 0.03
393
+ },
394
+ "rate": {
395
+ "1": 0.35,
396
+ "4": 0.45,
397
+ "16": 0.58125
398
+ }
399
+ },
400
+ "ll:k=128,lam=0.3,metric=unit": {
401
+ "arm": "ll",
402
+ "cfg": {
403
+ "metric": "unit",
404
+ "k": 128,
405
+ "lam": 0.3
406
+ },
407
+ "rate": {
408
+ "1": 0.39375,
409
+ "4": 0.5,
410
+ "16": 0.58125
411
+ }
412
+ },
413
+ "ll:k=256,lam=0.01,metric=unit": {
414
+ "arm": "ll",
415
+ "cfg": {
416
+ "metric": "unit",
417
+ "k": 256,
418
+ "lam": 0.01
419
+ },
420
+ "rate": {
421
+ "1": 0.3375,
422
+ "4": 0.4375,
423
+ "16": 0.55
424
+ }
425
+ },
426
+ "ll:k=256,lam=0.03,metric=unit": {
427
+ "arm": "ll",
428
+ "cfg": {
429
+ "metric": "unit",
430
+ "k": 256,
431
+ "lam": 0.03
432
+ },
433
+ "rate": {
434
+ "1": 0.34375,
435
+ "4": 0.4375,
436
+ "16": 0.575
437
+ }
438
+ },
439
+ "ll:k=256,lam=0.3,metric=unit": {
440
+ "arm": "ll",
441
+ "cfg": {
442
+ "metric": "unit",
443
+ "k": 256,
444
+ "lam": 0.3
445
+ },
446
+ "rate": {
447
+ "1": 0.35,
448
+ "4": 0.425,
449
+ "16": 0.55625
450
+ }
451
+ },
452
+ "llbw:k=16,lam=0.01,metric=unit": {
453
+ "arm": "llbw",
454
+ "cfg": {
455
+ "metric": "unit",
456
+ "k": 16,
457
+ "lam": 0.01
458
+ },
459
+ "rate": {
460
+ "1": 0.40625,
461
+ "4": 0.53125,
462
+ "16": 0.53125
463
+ }
464
+ },
465
+ "llbw:k=16,lam=0.03,metric=unit": {
466
+ "arm": "llbw",
467
+ "cfg": {
468
+ "metric": "unit",
469
+ "k": 16,
470
+ "lam": 0.03
471
+ },
472
+ "rate": {
473
+ "1": 0.45,
474
+ "4": 0.55,
475
+ "16": 0.55
476
+ }
477
+ },
478
+ "llbw:k=16,lam=0.3,metric=unit": {
479
+ "arm": "llbw",
480
+ "cfg": {
481
+ "metric": "unit",
482
+ "k": 16,
483
+ "lam": 0.3
484
+ },
485
+ "rate": {
486
+ "1": 0.5125,
487
+ "4": 0.56875,
488
+ "16": 0.56875
489
+ }
490
+ },
491
+ "llbw:k=32,lam=0.01,metric=unit": {
492
+ "arm": "llbw",
493
+ "cfg": {
494
+ "metric": "unit",
495
+ "k": 32,
496
+ "lam": 0.01
497
+ },
498
+ "rate": {
499
+ "1": 0.35,
500
+ "4": 0.475,
501
+ "16": 0.475
502
+ }
503
+ },
504
+ "llbw:k=32,lam=0.03,metric=unit": {
505
+ "arm": "llbw",
506
+ "cfg": {
507
+ "metric": "unit",
508
+ "k": 32,
509
+ "lam": 0.03
510
+ },
511
+ "rate": {
512
+ "1": 0.375,
513
+ "4": 0.49375,
514
+ "16": 0.49375
515
+ }
516
+ },
517
+ "llbw:k=32,lam=0.3,metric=unit": {
518
+ "arm": "llbw",
519
+ "cfg": {
520
+ "metric": "unit",
521
+ "k": 32,
522
+ "lam": 0.3
523
+ },
524
+ "rate": {
525
+ "1": 0.475,
526
+ "4": 0.5375,
527
+ "16": 0.5375
528
+ }
529
+ },
530
+ "llbw:k=64,lam=0.01,metric=unit": {
531
+ "arm": "llbw",
532
+ "cfg": {
533
+ "metric": "unit",
534
+ "k": 64,
535
+ "lam": 0.01
536
+ },
537
+ "rate": {
538
+ "1": 0.34375,
539
+ "4": 0.46875,
540
+ "16": 0.46875
541
+ }
542
+ },
543
+ "llbw:k=64,lam=0.03,metric=unit": {
544
+ "arm": "llbw",
545
+ "cfg": {
546
+ "metric": "unit",
547
+ "k": 64,
548
+ "lam": 0.03
549
+ },
550
+ "rate": {
551
+ "1": 0.35625,
552
+ "4": 0.48125,
553
+ "16": 0.48125
554
+ }
555
+ },
556
+ "llbw:k=64,lam=0.3,metric=unit": {
557
+ "arm": "llbw",
558
+ "cfg": {
559
+ "metric": "unit",
560
+ "k": 64,
561
+ "lam": 0.3
562
+ },
563
+ "rate": {
564
+ "1": 0.40625,
565
+ "4": 0.5125,
566
+ "16": 0.5125
567
+ }
568
+ },
569
+ "llbw:k=128,lam=0.01,metric=unit": {
570
+ "arm": "llbw",
571
+ "cfg": {
572
+ "metric": "unit",
573
+ "k": 128,
574
+ "lam": 0.01
575
+ },
576
+ "rate": {
577
+ "1": 0.35,
578
+ "4": 0.45625,
579
+ "16": 0.45625
580
+ }
581
+ },
582
+ "llbw:k=128,lam=0.03,metric=unit": {
583
+ "arm": "llbw",
584
+ "cfg": {
585
+ "metric": "unit",
586
+ "k": 128,
587
+ "lam": 0.03
588
+ },
589
+ "rate": {
590
+ "1": 0.35,
591
+ "4": 0.45625,
592
+ "16": 0.45625
593
+ }
594
+ },
595
+ "llbw:k=128,lam=0.3,metric=unit": {
596
+ "arm": "llbw",
597
+ "cfg": {
598
+ "metric": "unit",
599
+ "k": 128,
600
+ "lam": 0.3
601
+ },
602
+ "rate": {
603
+ "1": 0.39375,
604
+ "4": 0.46875,
605
+ "16": 0.46875
606
+ }
607
+ },
608
+ "retr_all:metric=ard,sols=1": {
609
+ "arm": "retr_all",
610
+ "cfg": {
611
+ "metric": "ard",
612
+ "sols": 1
613
+ },
614
+ "rate": {
615
+ "1": 0.43125,
616
+ "4": 0.725,
617
+ "16": 0.9
618
+ }
619
+ },
620
+ "jit:k=1,metric=ard,sigma=0.1": {
621
+ "arm": "jit",
622
+ "cfg": {
623
+ "metric": "ard",
624
+ "k": 1,
625
+ "sigma": 0.1
626
+ },
627
+ "rate": {
628
+ "1": 0.43125,
629
+ "4": 0.46875,
630
+ "16": 0.51875
631
+ }
632
+ },
633
+ "jit:k=1,metric=ard,sigma=0.3": {
634
+ "arm": "jit",
635
+ "cfg": {
636
+ "metric": "ard",
637
+ "k": 1,
638
+ "sigma": 0.3
639
+ },
640
+ "rate": {
641
+ "1": 0.43125,
642
+ "4": 0.5875,
643
+ "16": 0.68125
644
+ }
645
+ },
646
+ "jit:k=1,metric=ard,sigma=1.0": {
647
+ "arm": "jit",
648
+ "cfg": {
649
+ "metric": "ard",
650
+ "k": 1,
651
+ "sigma": 1.0
652
+ },
653
+ "rate": {
654
+ "1": 0.43125,
655
+ "4": 0.6,
656
+ "16": 0.80625
657
+ }
658
+ },
659
+ "jit:k=1,metric=ard,sigma=2.0": {
660
+ "arm": "jit",
661
+ "cfg": {
662
+ "metric": "ard",
663
+ "k": 1,
664
+ "sigma": 2.0
665
+ },
666
+ "rate": {
667
+ "1": 0.43125,
668
+ "4": 0.55625,
669
+ "16": 0.80625
670
+ }
671
+ },
672
+ "jit:k=2,metric=ard,sigma=0.1": {
673
+ "arm": "jit",
674
+ "cfg": {
675
+ "metric": "ard",
676
+ "k": 2,
677
+ "sigma": 0.1
678
+ },
679
+ "rate": {
680
+ "1": 0.43125,
681
+ "4": 0.58125,
682
+ "16": 0.65625
683
+ }
684
+ },
685
+ "jit:k=2,metric=ard,sigma=0.3": {
686
+ "arm": "jit",
687
+ "cfg": {
688
+ "metric": "ard",
689
+ "k": 2,
690
+ "sigma": 0.3
691
+ },
692
+ "rate": {
693
+ "1": 0.43125,
694
+ "4": 0.625,
695
+ "16": 0.775
696
+ }
697
+ },
698
+ "jit:k=2,metric=ard,sigma=1.0": {
699
+ "arm": "jit",
700
+ "cfg": {
701
+ "metric": "ard",
702
+ "k": 2,
703
+ "sigma": 1.0
704
+ },
705
+ "rate": {
706
+ "1": 0.43125,
707
+ "4": 0.69375,
708
+ "16": 0.85
709
+ }
710
+ },
711
+ "jit:k=2,metric=ard,sigma=2.0": {
712
+ "arm": "jit",
713
+ "cfg": {
714
+ "metric": "ard",
715
+ "k": 2,
716
+ "sigma": 2.0
717
+ },
718
+ "rate": {
719
+ "1": 0.43125,
720
+ "4": 0.64375,
721
+ "16": 0.825
722
+ }
723
+ },
724
+ "jit:k=4,metric=ard,sigma=0.1": {
725
+ "arm": "jit",
726
+ "cfg": {
727
+ "metric": "ard",
728
+ "k": 4,
729
+ "sigma": 0.1
730
+ },
731
+ "rate": {
732
+ "1": 0.43125,
733
+ "4": 0.725,
734
+ "16": 0.75625
735
+ }
736
+ },
737
+ "jit:k=4,metric=ard,sigma=0.3": {
738
+ "arm": "jit",
739
+ "cfg": {
740
+ "metric": "ard",
741
+ "k": 4,
742
+ "sigma": 0.3
743
+ },
744
+ "rate": {
745
+ "1": 0.43125,
746
+ "4": 0.725,
747
+ "16": 0.8375
748
+ }
749
+ },
750
+ "jit:k=4,metric=ard,sigma=1.0": {
751
+ "arm": "jit",
752
+ "cfg": {
753
+ "metric": "ard",
754
+ "k": 4,
755
+ "sigma": 1.0
756
+ },
757
+ "rate": {
758
+ "1": 0.43125,
759
+ "4": 0.725,
760
+ "16": 0.9
761
+ }
762
+ },
763
+ "jit:k=4,metric=ard,sigma=2.0": {
764
+ "arm": "jit",
765
+ "cfg": {
766
+ "metric": "ard",
767
+ "k": 4,
768
+ "sigma": 2.0
769
+ },
770
+ "rate": {
771
+ "1": 0.43125,
772
+ "4": 0.725,
773
+ "16": 0.85
774
+ }
775
+ },
776
+ "jit:k=8,metric=ard,sigma=0.1": {
777
+ "arm": "jit",
778
+ "cfg": {
779
+ "metric": "ard",
780
+ "k": 8,
781
+ "sigma": 0.1
782
+ },
783
+ "rate": {
784
+ "1": 0.43125,
785
+ "4": 0.725,
786
+ "16": 0.85
787
+ }
788
+ },
789
+ "jit:k=8,metric=ard,sigma=0.3": {
790
+ "arm": "jit",
791
+ "cfg": {
792
+ "metric": "ard",
793
+ "k": 8,
794
+ "sigma": 0.3
795
+ },
796
+ "rate": {
797
+ "1": 0.43125,
798
+ "4": 0.725,
799
+ "16": 0.85625
800
+ }
801
+ },
802
+ "jit:k=8,metric=ard,sigma=1.0": {
803
+ "arm": "jit",
804
+ "cfg": {
805
+ "metric": "ard",
806
+ "k": 8,
807
+ "sigma": 1.0
808
+ },
809
+ "rate": {
810
+ "1": 0.43125,
811
+ "4": 0.725,
812
+ "16": 0.8875
813
+ }
814
+ },
815
+ "jit:k=8,metric=ard,sigma=2.0": {
816
+ "arm": "jit",
817
+ "cfg": {
818
+ "metric": "ard",
819
+ "k": 8,
820
+ "sigma": 2.0
821
+ },
822
+ "rate": {
823
+ "1": 0.43125,
824
+ "4": 0.725,
825
+ "16": 0.89375
826
+ }
827
+ },
828
+ "boot:kb=4,metric=ard": {
829
+ "arm": "boot",
830
+ "cfg": {
831
+ "metric": "ard",
832
+ "kb": 4
833
+ },
834
+ "rate": {
835
+ "1": 0.51875,
836
+ "4": 0.6375,
837
+ "16": 0.74375
838
+ }
839
+ },
840
+ "boot:kb=8,metric=ard": {
841
+ "arm": "boot",
842
+ "cfg": {
843
+ "metric": "ard",
844
+ "kb": 8
845
+ },
846
+ "rate": {
847
+ "1": 0.5625,
848
+ "4": 0.69375,
849
+ "16": 0.81875
850
+ }
851
+ },
852
+ "boot:kb=16,metric=ard": {
853
+ "arm": "boot",
854
+ "cfg": {
855
+ "metric": "ard",
856
+ "kb": 16
857
+ },
858
+ "rate": {
859
+ "1": 0.4875,
860
+ "4": 0.65,
861
+ "16": 0.75625
862
+ }
863
+ },
864
+ "boot:kb=32,metric=ard": {
865
+ "arm": "boot",
866
+ "cfg": {
867
+ "metric": "ard",
868
+ "kb": 32
869
+ },
870
+ "rate": {
871
+ "1": 0.5125,
872
+ "4": 0.5875,
873
+ "16": 0.6875
874
+ }
875
+ },
876
+ "boot:kb=64,metric=ard": {
877
+ "arm": "boot",
878
+ "cfg": {
879
+ "metric": "ard",
880
+ "kb": 64
881
+ },
882
+ "rate": {
883
+ "1": 0.5,
884
+ "4": 0.56875,
885
+ "16": 0.65
886
+ }
887
+ },
888
+ "ll:k=32,lam=0.01,metric=ard": {
889
+ "arm": "ll",
890
+ "cfg": {
891
+ "metric": "ard",
892
+ "k": 32,
893
+ "lam": 0.01
894
+ },
895
+ "rate": {
896
+ "1": 0.86875,
897
+ "4": 0.91875,
898
+ "16": 0.9625
899
+ }
900
+ },
901
+ "ll:k=32,lam=0.03,metric=ard": {
902
+ "arm": "ll",
903
+ "cfg": {
904
+ "metric": "ard",
905
+ "k": 32,
906
+ "lam": 0.03
907
+ },
908
+ "rate": {
909
+ "1": 0.875,
910
+ "4": 0.9375,
911
+ "16": 0.95625
912
+ }
913
+ },
914
+ "ll:k=32,lam=0.3,metric=ard": {
915
+ "arm": "ll",
916
+ "cfg": {
917
+ "metric": "ard",
918
+ "k": 32,
919
+ "lam": 0.3
920
+ },
921
+ "rate": {
922
+ "1": 0.8625,
923
+ "4": 0.9,
924
+ "16": 0.93125
925
+ }
926
+ },
927
+ "ll:k=64,lam=0.01,metric=ard": {
928
+ "arm": "ll",
929
+ "cfg": {
930
+ "metric": "ard",
931
+ "k": 64,
932
+ "lam": 0.01
933
+ },
934
+ "rate": {
935
+ "1": 0.90625,
936
+ "4": 0.925,
937
+ "16": 0.9625
938
+ }
939
+ },
940
+ "ll:k=64,lam=0.03,metric=ard": {
941
+ "arm": "ll",
942
+ "cfg": {
943
+ "metric": "ard",
944
+ "k": 64,
945
+ "lam": 0.03
946
+ },
947
+ "rate": {
948
+ "1": 0.9125,
949
+ "4": 0.9375,
950
+ "16": 0.96875
951
+ }
952
+ },
953
+ "ll:k=64,lam=0.3,metric=ard": {
954
+ "arm": "ll",
955
+ "cfg": {
956
+ "metric": "ard",
957
+ "k": 64,
958
+ "lam": 0.3
959
+ },
960
+ "rate": {
961
+ "1": 0.925,
962
+ "4": 0.9375,
963
+ "16": 0.95
964
+ }
965
+ },
966
+ "ll:k=128,lam=0.01,metric=ard": {
967
+ "arm": "ll",
968
+ "cfg": {
969
+ "metric": "ard",
970
+ "k": 128,
971
+ "lam": 0.01
972
+ },
973
+ "rate": {
974
+ "1": 0.90625,
975
+ "4": 0.94375,
976
+ "16": 0.9625
977
+ }
978
+ },
979
+ "ll:k=128,lam=0.03,metric=ard": {
980
+ "arm": "ll",
981
+ "cfg": {
982
+ "metric": "ard",
983
+ "k": 128,
984
+ "lam": 0.03
985
+ },
986
+ "rate": {
987
+ "1": 0.90625,
988
+ "4": 0.925,
989
+ "16": 0.9625
990
+ }
991
+ },
992
+ "ll:k=128,lam=0.3,metric=ard": {
993
+ "arm": "ll",
994
+ "cfg": {
995
+ "metric": "ard",
996
+ "k": 128,
997
+ "lam": 0.3
998
+ },
999
+ "rate": {
1000
+ "1": 0.9125,
1001
+ "4": 0.9375,
1002
+ "16": 0.95625
1003
+ }
1004
+ },
1005
+ "ll:k=256,lam=0.01,metric=ard": {
1006
+ "arm": "ll",
1007
+ "cfg": {
1008
+ "metric": "ard",
1009
+ "k": 256,
1010
+ "lam": 0.01
1011
+ },
1012
+ "rate": {
1013
+ "1": 0.89375,
1014
+ "4": 0.91875,
1015
+ "16": 0.9375
1016
+ }
1017
+ },
1018
+ "ll:k=256,lam=0.03,metric=ard": {
1019
+ "arm": "ll",
1020
+ "cfg": {
1021
+ "metric": "ard",
1022
+ "k": 256,
1023
+ "lam": 0.03
1024
+ },
1025
+ "rate": {
1026
+ "1": 0.89375,
1027
+ "4": 0.9375,
1028
+ "16": 0.95
1029
+ }
1030
+ },
1031
+ "ll:k=256,lam=0.3,metric=ard": {
1032
+ "arm": "ll",
1033
+ "cfg": {
1034
+ "metric": "ard",
1035
+ "k": 256,
1036
+ "lam": 0.3
1037
+ },
1038
+ "rate": {
1039
+ "1": 0.9,
1040
+ "4": 0.9125,
1041
+ "16": 0.94375
1042
+ }
1043
+ },
1044
+ "llbw:k=16,lam=0.01,metric=ard": {
1045
+ "arm": "llbw",
1046
+ "cfg": {
1047
+ "metric": "ard",
1048
+ "k": 16,
1049
+ "lam": 0.01
1050
+ },
1051
+ "rate": {
1052
+ "1": 0.81875,
1053
+ "4": 0.94375,
1054
+ "16": 0.94375
1055
+ }
1056
+ },
1057
+ "llbw:k=16,lam=0.03,metric=ard": {
1058
+ "arm": "llbw",
1059
+ "cfg": {
1060
+ "metric": "ard",
1061
+ "k": 16,
1062
+ "lam": 0.03
1063
+ },
1064
+ "rate": {
1065
+ "1": 0.8125,
1066
+ "4": 0.9375,
1067
+ "16": 0.9375
1068
+ }
1069
+ },
1070
+ "llbw:k=16,lam=0.3,metric=ard": {
1071
+ "arm": "llbw",
1072
+ "cfg": {
1073
+ "metric": "ard",
1074
+ "k": 16,
1075
+ "lam": 0.3
1076
+ },
1077
+ "rate": {
1078
+ "1": 0.70625,
1079
+ "4": 0.93125,
1080
+ "16": 0.93125
1081
+ }
1082
+ },
1083
+ "llbw:k=32,lam=0.01,metric=ard": {
1084
+ "arm": "llbw",
1085
+ "cfg": {
1086
+ "metric": "ard",
1087
+ "k": 32,
1088
+ "lam": 0.01
1089
+ },
1090
+ "rate": {
1091
+ "1": 0.86875,
1092
+ "4": 0.925,
1093
+ "16": 0.925
1094
+ }
1095
+ },
1096
+ "llbw:k=32,lam=0.03,metric=ard": {
1097
+ "arm": "llbw",
1098
+ "cfg": {
1099
+ "metric": "ard",
1100
+ "k": 32,
1101
+ "lam": 0.03
1102
+ },
1103
+ "rate": {
1104
+ "1": 0.875,
1105
+ "4": 0.925,
1106
+ "16": 0.925
1107
+ }
1108
+ },
1109
+ "llbw:k=32,lam=0.3,metric=ard": {
1110
+ "arm": "llbw",
1111
+ "cfg": {
1112
+ "metric": "ard",
1113
+ "k": 32,
1114
+ "lam": 0.3
1115
+ },
1116
+ "rate": {
1117
+ "1": 0.8625,
1118
+ "4": 0.9375,
1119
+ "16": 0.9375
1120
+ }
1121
+ },
1122
+ "llbw:k=64,lam=0.01,metric=ard": {
1123
+ "arm": "llbw",
1124
+ "cfg": {
1125
+ "metric": "ard",
1126
+ "k": 64,
1127
+ "lam": 0.01
1128
+ },
1129
+ "rate": {
1130
+ "1": 0.90625,
1131
+ "4": 0.93125,
1132
+ "16": 0.93125
1133
+ }
1134
+ },
1135
+ "llbw:k=64,lam=0.03,metric=ard": {
1136
+ "arm": "llbw",
1137
+ "cfg": {
1138
+ "metric": "ard",
1139
+ "k": 64,
1140
+ "lam": 0.03
1141
+ },
1142
+ "rate": {
1143
+ "1": 0.9125,
1144
+ "4": 0.93125,
1145
+ "16": 0.93125
1146
+ }
1147
+ },
1148
+ "llbw:k=64,lam=0.3,metric=ard": {
1149
+ "arm": "llbw",
1150
+ "cfg": {
1151
+ "metric": "ard",
1152
+ "k": 64,
1153
+ "lam": 0.3
1154
+ },
1155
+ "rate": {
1156
+ "1": 0.925,
1157
+ "4": 0.9375,
1158
+ "16": 0.9375
1159
+ }
1160
+ },
1161
+ "llbw:k=128,lam=0.01,metric=ard": {
1162
+ "arm": "llbw",
1163
+ "cfg": {
1164
+ "metric": "ard",
1165
+ "k": 128,
1166
+ "lam": 0.01
1167
+ },
1168
+ "rate": {
1169
+ "1": 0.90625,
1170
+ "4": 0.925,
1171
+ "16": 0.925
1172
+ }
1173
+ },
1174
+ "llbw:k=128,lam=0.03,metric=ard": {
1175
+ "arm": "llbw",
1176
+ "cfg": {
1177
+ "metric": "ard",
1178
+ "k": 128,
1179
+ "lam": 0.03
1180
+ },
1181
+ "rate": {
1182
+ "1": 0.90625,
1183
+ "4": 0.93125,
1184
+ "16": 0.93125
1185
+ }
1186
+ },
1187
+ "llbw:k=128,lam=0.3,metric=ard": {
1188
+ "arm": "llbw",
1189
+ "cfg": {
1190
+ "metric": "ard",
1191
+ "k": 128,
1192
+ "lam": 0.3
1193
+ },
1194
+ "rate": {
1195
+ "1": 0.9125,
1196
+ "4": 0.93125,
1197
+ "16": 0.93125
1198
+ }
1199
+ },
1200
+ "gp_ard:c=0.35,eta=0.3,tau=0.5": {
1201
+ "arm": "gp_ard",
1202
+ "cfg": {
1203
+ "c": 0.35,
1204
+ "eta": 0.3,
1205
+ "tau": 0.5
1206
+ },
1207
+ "rate": {
1208
+ "1": 0.23125,
1209
+ "4": 0.28125,
1210
+ "16": 0.53125
1211
+ }
1212
+ },
1213
+ "gp_ard:c=0.35,eta=0.3,tau=1.0": {
1214
+ "arm": "gp_ard",
1215
+ "cfg": {
1216
+ "c": 0.35,
1217
+ "eta": 0.3,
1218
+ "tau": 1.0
1219
+ },
1220
+ "rate": {
1221
+ "1": 0.23125,
1222
+ "4": 0.29375,
1223
+ "16": 0.50625
1224
+ }
1225
+ },
1226
+ "gp_ard:c=0.35,eta=0.3,tau=2.0": {
1227
+ "arm": "gp_ard",
1228
+ "cfg": {
1229
+ "c": 0.35,
1230
+ "eta": 0.3,
1231
+ "tau": 2.0
1232
+ },
1233
+ "rate": {
1234
+ "1": 0.23125,
1235
+ "4": 0.29375,
1236
+ "16": 0.5
1237
+ }
1238
+ },
1239
+ "gp_ard:c=0.35,eta=0.3,tau=3.0": {
1240
+ "arm": "gp_ard",
1241
+ "cfg": {
1242
+ "c": 0.35,
1243
+ "eta": 0.3,
1244
+ "tau": 3.0
1245
+ },
1246
+ "rate": {
1247
+ "1": 0.23125,
1248
+ "4": 0.29375,
1249
+ "16": 0.59375
1250
+ }
1251
+ },
1252
+ "gp_ard:c=0.35,eta=1.0,tau=0.5": {
1253
+ "arm": "gp_ard",
1254
+ "cfg": {
1255
+ "c": 0.35,
1256
+ "eta": 1.0,
1257
+ "tau": 0.5
1258
+ },
1259
+ "rate": {
1260
+ "1": 0.23125,
1261
+ "4": 0.3,
1262
+ "16": 0.5375
1263
+ }
1264
+ },
1265
+ "gp_ard:c=0.35,eta=1.0,tau=1.0": {
1266
+ "arm": "gp_ard",
1267
+ "cfg": {
1268
+ "c": 0.35,
1269
+ "eta": 1.0,
1270
+ "tau": 1.0
1271
+ },
1272
+ "rate": {
1273
+ "1": 0.23125,
1274
+ "4": 0.2875,
1275
+ "16": 0.48125
1276
+ }
1277
+ },
1278
+ "gp_ard:c=0.35,eta=1.0,tau=2.0": {
1279
+ "arm": "gp_ard",
1280
+ "cfg": {
1281
+ "c": 0.35,
1282
+ "eta": 1.0,
1283
+ "tau": 2.0
1284
+ },
1285
+ "rate": {
1286
+ "1": 0.23125,
1287
+ "4": 0.30625,
1288
+ "16": 0.53125
1289
+ }
1290
+ },
1291
+ "gp_ard:c=0.35,eta=1.0,tau=3.0": {
1292
+ "arm": "gp_ard",
1293
+ "cfg": {
1294
+ "c": 0.35,
1295
+ "eta": 1.0,
1296
+ "tau": 3.0
1297
+ },
1298
+ "rate": {
1299
+ "1": 0.23125,
1300
+ "4": 0.3,
1301
+ "16": 0.525
1302
+ }
1303
+ },
1304
+ "gp_ard:c=0.35,eta=3.0,tau=0.5": {
1305
+ "arm": "gp_ard",
1306
+ "cfg": {
1307
+ "c": 0.35,
1308
+ "eta": 3.0,
1309
+ "tau": 0.5
1310
+ },
1311
+ "rate": {
1312
+ "1": 0.23125,
1313
+ "4": 0.3125,
1314
+ "16": 0.46875
1315
+ }
1316
+ },
1317
+ "gp_ard:c=0.35,eta=3.0,tau=1.0": {
1318
+ "arm": "gp_ard",
1319
+ "cfg": {
1320
+ "c": 0.35,
1321
+ "eta": 3.0,
1322
+ "tau": 1.0
1323
+ },
1324
+ "rate": {
1325
+ "1": 0.23125,
1326
+ "4": 0.2625,
1327
+ "16": 0.50625
1328
+ }
1329
+ },
1330
+ "gp_ard:c=0.35,eta=3.0,tau=2.0": {
1331
+ "arm": "gp_ard",
1332
+ "cfg": {
1333
+ "c": 0.35,
1334
+ "eta": 3.0,
1335
+ "tau": 2.0
1336
+ },
1337
+ "rate": {
1338
+ "1": 0.23125,
1339
+ "4": 0.30625,
1340
+ "16": 0.525
1341
+ }
1342
+ },
1343
+ "gp_ard:c=0.35,eta=3.0,tau=3.0": {
1344
+ "arm": "gp_ard",
1345
+ "cfg": {
1346
+ "c": 0.35,
1347
+ "eta": 3.0,
1348
+ "tau": 3.0
1349
+ },
1350
+ "rate": {
1351
+ "1": 0.23125,
1352
+ "4": 0.3125,
1353
+ "16": 0.53125
1354
+ }
1355
+ },
1356
+ "gp_ard:c=0.35,eta=10.0,tau=0.5": {
1357
+ "arm": "gp_ard",
1358
+ "cfg": {
1359
+ "c": 0.35,
1360
+ "eta": 10.0,
1361
+ "tau": 0.5
1362
+ },
1363
+ "rate": {
1364
+ "1": 0.23125,
1365
+ "4": 0.3,
1366
+ "16": 0.525
1367
+ }
1368
+ },
1369
+ "gp_ard:c=0.35,eta=10.0,tau=1.0": {
1370
+ "arm": "gp_ard",
1371
+ "cfg": {
1372
+ "c": 0.35,
1373
+ "eta": 10.0,
1374
+ "tau": 1.0
1375
+ },
1376
+ "rate": {
1377
+ "1": 0.23125,
1378
+ "4": 0.2875,
1379
+ "16": 0.49375
1380
+ }
1381
+ },
1382
+ "gp_ard:c=0.35,eta=10.0,tau=2.0": {
1383
+ "arm": "gp_ard",
1384
+ "cfg": {
1385
+ "c": 0.35,
1386
+ "eta": 10.0,
1387
+ "tau": 2.0
1388
+ },
1389
+ "rate": {
1390
+ "1": 0.23125,
1391
+ "4": 0.28125,
1392
+ "16": 0.475
1393
+ }
1394
+ },
1395
+ "gp_ard:c=0.35,eta=10.0,tau=3.0": {
1396
+ "arm": "gp_ard",
1397
+ "cfg": {
1398
+ "c": 0.35,
1399
+ "eta": 10.0,
1400
+ "tau": 3.0
1401
+ },
1402
+ "rate": {
1403
+ "1": 0.23125,
1404
+ "4": 0.31875,
1405
+ "16": 0.49375
1406
+ }
1407
+ },
1408
+ "gp_ard:c=0.5,eta=0.3,tau=0.5": {
1409
+ "arm": "gp_ard",
1410
+ "cfg": {
1411
+ "c": 0.5,
1412
+ "eta": 0.3,
1413
+ "tau": 0.5
1414
+ },
1415
+ "rate": {
1416
+ "1": 0.65,
1417
+ "4": 0.725,
1418
+ "16": 0.875
1419
+ }
1420
+ },
1421
+ "gp_ard:c=0.5,eta=0.3,tau=1.0": {
1422
+ "arm": "gp_ard",
1423
+ "cfg": {
1424
+ "c": 0.5,
1425
+ "eta": 0.3,
1426
+ "tau": 1.0
1427
+ },
1428
+ "rate": {
1429
+ "1": 0.65,
1430
+ "4": 0.70625,
1431
+ "16": 0.83125
1432
+ }
1433
+ },
1434
+ "gp_ard:c=0.5,eta=0.3,tau=2.0": {
1435
+ "arm": "gp_ard",
1436
+ "cfg": {
1437
+ "c": 0.5,
1438
+ "eta": 0.3,
1439
+ "tau": 2.0
1440
+ },
1441
+ "rate": {
1442
+ "1": 0.65,
1443
+ "4": 0.6875,
1444
+ "16": 0.825
1445
+ }
1446
+ },
1447
+ "gp_ard:c=0.5,eta=0.3,tau=3.0": {
1448
+ "arm": "gp_ard",
1449
+ "cfg": {
1450
+ "c": 0.5,
1451
+ "eta": 0.3,
1452
+ "tau": 3.0
1453
+ },
1454
+ "rate": {
1455
+ "1": 0.65,
1456
+ "4": 0.7,
1457
+ "16": 0.84375
1458
+ }
1459
+ },
1460
+ "gp_ard:c=0.5,eta=1.0,tau=0.5": {
1461
+ "arm": "gp_ard",
1462
+ "cfg": {
1463
+ "c": 0.5,
1464
+ "eta": 1.0,
1465
+ "tau": 0.5
1466
+ },
1467
+ "rate": {
1468
+ "1": 0.65,
1469
+ "4": 0.7375,
1470
+ "16": 0.825
1471
+ }
1472
+ },
1473
+ "gp_ard:c=0.5,eta=1.0,tau=1.0": {
1474
+ "arm": "gp_ard",
1475
+ "cfg": {
1476
+ "c": 0.5,
1477
+ "eta": 1.0,
1478
+ "tau": 1.0
1479
+ },
1480
+ "rate": {
1481
+ "1": 0.65,
1482
+ "4": 0.7,
1483
+ "16": 0.81875
1484
+ }
1485
+ },
1486
+ "gp_ard:c=0.5,eta=1.0,tau=2.0": {
1487
+ "arm": "gp_ard",
1488
+ "cfg": {
1489
+ "c": 0.5,
1490
+ "eta": 1.0,
1491
+ "tau": 2.0
1492
+ },
1493
+ "rate": {
1494
+ "1": 0.65,
1495
+ "4": 0.71875,
1496
+ "16": 0.825
1497
+ }
1498
+ },
1499
+ "gp_ard:c=0.5,eta=1.0,tau=3.0": {
1500
+ "arm": "gp_ard",
1501
+ "cfg": {
1502
+ "c": 0.5,
1503
+ "eta": 1.0,
1504
+ "tau": 3.0
1505
+ },
1506
+ "rate": {
1507
+ "1": 0.65,
1508
+ "4": 0.6875,
1509
+ "16": 0.8375
1510
+ }
1511
+ },
1512
+ "gp_ard:c=0.5,eta=3.0,tau=0.5": {
1513
+ "arm": "gp_ard",
1514
+ "cfg": {
1515
+ "c": 0.5,
1516
+ "eta": 3.0,
1517
+ "tau": 0.5
1518
+ },
1519
+ "rate": {
1520
+ "1": 0.65,
1521
+ "4": 0.73125,
1522
+ "16": 0.86875
1523
+ }
1524
+ },
1525
+ "gp_ard:c=0.5,eta=3.0,tau=1.0": {
1526
+ "arm": "gp_ard",
1527
+ "cfg": {
1528
+ "c": 0.5,
1529
+ "eta": 3.0,
1530
+ "tau": 1.0
1531
+ },
1532
+ "rate": {
1533
+ "1": 0.65,
1534
+ "4": 0.7,
1535
+ "16": 0.86875
1536
+ }
1537
+ },
1538
+ "gp_ard:c=0.5,eta=3.0,tau=2.0": {
1539
+ "arm": "gp_ard",
1540
+ "cfg": {
1541
+ "c": 0.5,
1542
+ "eta": 3.0,
1543
+ "tau": 2.0
1544
+ },
1545
+ "rate": {
1546
+ "1": 0.65,
1547
+ "4": 0.66875,
1548
+ "16": 0.7875
1549
+ }
1550
+ },
1551
+ "gp_ard:c=0.5,eta=3.0,tau=3.0": {
1552
+ "arm": "gp_ard",
1553
+ "cfg": {
1554
+ "c": 0.5,
1555
+ "eta": 3.0,
1556
+ "tau": 3.0
1557
+ },
1558
+ "rate": {
1559
+ "1": 0.65,
1560
+ "4": 0.70625,
1561
+ "16": 0.8125
1562
+ }
1563
+ },
1564
+ "gp_ard:c=0.5,eta=10.0,tau=0.5": {
1565
+ "arm": "gp_ard",
1566
+ "cfg": {
1567
+ "c": 0.5,
1568
+ "eta": 10.0,
1569
+ "tau": 0.5
1570
+ },
1571
+ "rate": {
1572
+ "1": 0.65625,
1573
+ "4": 0.7375,
1574
+ "16": 0.8625
1575
+ }
1576
+ },
1577
+ "gp_ard:c=0.5,eta=10.0,tau=1.0": {
1578
+ "arm": "gp_ard",
1579
+ "cfg": {
1580
+ "c": 0.5,
1581
+ "eta": 10.0,
1582
+ "tau": 1.0
1583
+ },
1584
+ "rate": {
1585
+ "1": 0.65625,
1586
+ "4": 0.725,
1587
+ "16": 0.84375
1588
+ }
1589
+ },
1590
+ "gp_ard:c=0.5,eta=10.0,tau=2.0": {
1591
+ "arm": "gp_ard",
1592
+ "cfg": {
1593
+ "c": 0.5,
1594
+ "eta": 10.0,
1595
+ "tau": 2.0
1596
+ },
1597
+ "rate": {
1598
+ "1": 0.65625,
1599
+ "4": 0.69375,
1600
+ "16": 0.83125
1601
+ }
1602
+ },
1603
+ "gp_ard:c=0.5,eta=10.0,tau=3.0": {
1604
+ "arm": "gp_ard",
1605
+ "cfg": {
1606
+ "c": 0.5,
1607
+ "eta": 10.0,
1608
+ "tau": 3.0
1609
+ },
1610
+ "rate": {
1611
+ "1": 0.65625,
1612
+ "4": 0.69375,
1613
+ "16": 0.81875
1614
+ }
1615
+ },
1616
+ "gp_ard:c=0.7,eta=0.3,tau=0.5": {
1617
+ "arm": "gp_ard",
1618
+ "cfg": {
1619
+ "c": 0.7,
1620
+ "eta": 0.3,
1621
+ "tau": 0.5
1622
+ },
1623
+ "rate": {
1624
+ "1": 0.8125,
1625
+ "4": 0.90625,
1626
+ "16": 0.9625
1627
+ }
1628
+ },
1629
+ "gp_ard:c=0.7,eta=0.3,tau=1.0": {
1630
+ "arm": "gp_ard",
1631
+ "cfg": {
1632
+ "c": 0.7,
1633
+ "eta": 0.3,
1634
+ "tau": 1.0
1635
+ },
1636
+ "rate": {
1637
+ "1": 0.8125,
1638
+ "4": 0.84375,
1639
+ "16": 0.9625
1640
+ }
1641
+ },
1642
+ "gp_ard:c=0.7,eta=0.3,tau=2.0": {
1643
+ "arm": "gp_ard",
1644
+ "cfg": {
1645
+ "c": 0.7,
1646
+ "eta": 0.3,
1647
+ "tau": 2.0
1648
+ },
1649
+ "rate": {
1650
+ "1": 0.8125,
1651
+ "4": 0.84375,
1652
+ "16": 0.91875
1653
+ }
1654
+ },
1655
+ "gp_ard:c=0.7,eta=0.3,tau=3.0": {
1656
+ "arm": "gp_ard",
1657
+ "cfg": {
1658
+ "c": 0.7,
1659
+ "eta": 0.3,
1660
+ "tau": 3.0
1661
+ },
1662
+ "rate": {
1663
+ "1": 0.8125,
1664
+ "4": 0.85625,
1665
+ "16": 0.93125
1666
+ }
1667
+ },
1668
+ "gp_ard:c=0.7,eta=1.0,tau=0.5": {
1669
+ "arm": "gp_ard",
1670
+ "cfg": {
1671
+ "c": 0.7,
1672
+ "eta": 1.0,
1673
+ "tau": 0.5
1674
+ },
1675
+ "rate": {
1676
+ "1": 0.8125,
1677
+ "4": 0.9,
1678
+ "16": 0.9625
1679
+ }
1680
+ },
1681
+ "gp_ard:c=0.7,eta=1.0,tau=1.0": {
1682
+ "arm": "gp_ard",
1683
+ "cfg": {
1684
+ "c": 0.7,
1685
+ "eta": 1.0,
1686
+ "tau": 1.0
1687
+ },
1688
+ "rate": {
1689
+ "1": 0.8125,
1690
+ "4": 0.89375,
1691
+ "16": 0.95625
1692
+ }
1693
+ },
1694
+ "gp_ard:c=0.7,eta=1.0,tau=2.0": {
1695
+ "arm": "gp_ard",
1696
+ "cfg": {
1697
+ "c": 0.7,
1698
+ "eta": 1.0,
1699
+ "tau": 2.0
1700
+ },
1701
+ "rate": {
1702
+ "1": 0.8125,
1703
+ "4": 0.8625,
1704
+ "16": 0.9375
1705
+ }
1706
+ },
1707
+ "gp_ard:c=0.7,eta=1.0,tau=3.0": {
1708
+ "arm": "gp_ard",
1709
+ "cfg": {
1710
+ "c": 0.7,
1711
+ "eta": 1.0,
1712
+ "tau": 3.0
1713
+ },
1714
+ "rate": {
1715
+ "1": 0.8125,
1716
+ "4": 0.8375,
1717
+ "16": 0.90625
1718
+ }
1719
+ },
1720
+ "gp_ard:c=0.7,eta=3.0,tau=0.5": {
1721
+ "arm": "gp_ard",
1722
+ "cfg": {
1723
+ "c": 0.7,
1724
+ "eta": 3.0,
1725
+ "tau": 0.5
1726
+ },
1727
+ "rate": {
1728
+ "1": 0.81875,
1729
+ "4": 0.88125,
1730
+ "16": 0.95625
1731
+ }
1732
+ },
1733
+ "gp_ard:c=0.7,eta=3.0,tau=1.0": {
1734
+ "arm": "gp_ard",
1735
+ "cfg": {
1736
+ "c": 0.7,
1737
+ "eta": 3.0,
1738
+ "tau": 1.0
1739
+ },
1740
+ "rate": {
1741
+ "1": 0.81875,
1742
+ "4": 0.85625,
1743
+ "16": 0.9375
1744
+ }
1745
+ },
1746
+ "gp_ard:c=0.7,eta=3.0,tau=2.0": {
1747
+ "arm": "gp_ard",
1748
+ "cfg": {
1749
+ "c": 0.7,
1750
+ "eta": 3.0,
1751
+ "tau": 2.0
1752
+ },
1753
+ "rate": {
1754
+ "1": 0.81875,
1755
+ "4": 0.84375,
1756
+ "16": 0.925
1757
+ }
1758
+ },
1759
+ "gp_ard:c=0.7,eta=3.0,tau=3.0": {
1760
+ "arm": "gp_ard",
1761
+ "cfg": {
1762
+ "c": 0.7,
1763
+ "eta": 3.0,
1764
+ "tau": 3.0
1765
+ },
1766
+ "rate": {
1767
+ "1": 0.81875,
1768
+ "4": 0.85625,
1769
+ "16": 0.925
1770
+ }
1771
+ },
1772
+ "gp_ard:c=0.7,eta=10.0,tau=0.5": {
1773
+ "arm": "gp_ard",
1774
+ "cfg": {
1775
+ "c": 0.7,
1776
+ "eta": 10.0,
1777
+ "tau": 0.5
1778
+ },
1779
+ "rate": {
1780
+ "1": 0.83125,
1781
+ "4": 0.88125,
1782
+ "16": 0.95625
1783
+ }
1784
+ },
1785
+ "gp_ard:c=0.7,eta=10.0,tau=1.0": {
1786
+ "arm": "gp_ard",
1787
+ "cfg": {
1788
+ "c": 0.7,
1789
+ "eta": 10.0,
1790
+ "tau": 1.0
1791
+ },
1792
+ "rate": {
1793
+ "1": 0.83125,
1794
+ "4": 0.8625,
1795
+ "16": 0.95
1796
+ }
1797
+ },
1798
+ "gp_ard:c=0.7,eta=10.0,tau=2.0": {
1799
+ "arm": "gp_ard",
1800
+ "cfg": {
1801
+ "c": 0.7,
1802
+ "eta": 10.0,
1803
+ "tau": 2.0
1804
+ },
1805
+ "rate": {
1806
+ "1": 0.83125,
1807
+ "4": 0.85625,
1808
+ "16": 0.9375
1809
+ }
1810
+ },
1811
+ "gp_ard:c=0.7,eta=10.0,tau=3.0": {
1812
+ "arm": "gp_ard",
1813
+ "cfg": {
1814
+ "c": 0.7,
1815
+ "eta": 10.0,
1816
+ "tau": 3.0
1817
+ },
1818
+ "rate": {
1819
+ "1": 0.83125,
1820
+ "4": 0.8625,
1821
+ "16": 0.9375
1822
+ }
1823
+ },
1824
+ "gp_ard:c=1.0,eta=0.3,tau=0.5": {
1825
+ "arm": "gp_ard",
1826
+ "cfg": {
1827
+ "c": 1.0,
1828
+ "eta": 0.3,
1829
+ "tau": 0.5
1830
+ },
1831
+ "rate": {
1832
+ "1": 0.7625,
1833
+ "4": 0.825,
1834
+ "16": 0.8875
1835
+ }
1836
+ },
1837
+ "gp_ard:c=1.0,eta=0.3,tau=1.0": {
1838
+ "arm": "gp_ard",
1839
+ "cfg": {
1840
+ "c": 1.0,
1841
+ "eta": 0.3,
1842
+ "tau": 1.0
1843
+ },
1844
+ "rate": {
1845
+ "1": 0.7625,
1846
+ "4": 0.85625,
1847
+ "16": 0.925
1848
+ }
1849
+ },
1850
+ "gp_ard:c=1.0,eta=0.3,tau=2.0": {
1851
+ "arm": "gp_ard",
1852
+ "cfg": {
1853
+ "c": 1.0,
1854
+ "eta": 0.3,
1855
+ "tau": 2.0
1856
+ },
1857
+ "rate": {
1858
+ "1": 0.7625,
1859
+ "4": 0.85625,
1860
+ "16": 0.9375
1861
+ }
1862
+ },
1863
+ "gp_ard:c=1.0,eta=0.3,tau=3.0": {
1864
+ "arm": "gp_ard",
1865
+ "cfg": {
1866
+ "c": 1.0,
1867
+ "eta": 0.3,
1868
+ "tau": 3.0
1869
+ },
1870
+ "rate": {
1871
+ "1": 0.7625,
1872
+ "4": 0.84375,
1873
+ "16": 0.93125
1874
+ }
1875
+ },
1876
+ "gp_ard:c=1.0,eta=1.0,tau=0.5": {
1877
+ "arm": "gp_ard",
1878
+ "cfg": {
1879
+ "c": 1.0,
1880
+ "eta": 1.0,
1881
+ "tau": 0.5
1882
+ },
1883
+ "rate": {
1884
+ "1": 0.76875,
1885
+ "4": 0.84375,
1886
+ "16": 0.8875
1887
+ }
1888
+ },
1889
+ "gp_ard:c=1.0,eta=1.0,tau=1.0": {
1890
+ "arm": "gp_ard",
1891
+ "cfg": {
1892
+ "c": 1.0,
1893
+ "eta": 1.0,
1894
+ "tau": 1.0
1895
+ },
1896
+ "rate": {
1897
+ "1": 0.76875,
1898
+ "4": 0.8625,
1899
+ "16": 0.94375
1900
+ }
1901
+ },
1902
+ "gp_ard:c=1.0,eta=1.0,tau=2.0": {
1903
+ "arm": "gp_ard",
1904
+ "cfg": {
1905
+ "c": 1.0,
1906
+ "eta": 1.0,
1907
+ "tau": 2.0
1908
+ },
1909
+ "rate": {
1910
+ "1": 0.76875,
1911
+ "4": 0.8375,
1912
+ "16": 0.9375
1913
+ }
1914
+ },
1915
+ "gp_ard:c=1.0,eta=1.0,tau=3.0": {
1916
+ "arm": "gp_ard",
1917
+ "cfg": {
1918
+ "c": 1.0,
1919
+ "eta": 1.0,
1920
+ "tau": 3.0
1921
+ },
1922
+ "rate": {
1923
+ "1": 0.76875,
1924
+ "4": 0.84375,
1925
+ "16": 0.94375
1926
+ }
1927
+ },
1928
+ "gp_ard:c=1.0,eta=3.0,tau=0.5": {
1929
+ "arm": "gp_ard",
1930
+ "cfg": {
1931
+ "c": 1.0,
1932
+ "eta": 3.0,
1933
+ "tau": 0.5
1934
+ },
1935
+ "rate": {
1936
+ "1": 0.775,
1937
+ "4": 0.85625,
1938
+ "16": 0.9
1939
+ }
1940
+ },
1941
+ "gp_ard:c=1.0,eta=3.0,tau=1.0": {
1942
+ "arm": "gp_ard",
1943
+ "cfg": {
1944
+ "c": 1.0,
1945
+ "eta": 3.0,
1946
+ "tau": 1.0
1947
+ },
1948
+ "rate": {
1949
+ "1": 0.775,
1950
+ "4": 0.85,
1951
+ "16": 0.95
1952
+ }
1953
+ },
1954
+ "gp_ard:c=1.0,eta=3.0,tau=2.0": {
1955
+ "arm": "gp_ard",
1956
+ "cfg": {
1957
+ "c": 1.0,
1958
+ "eta": 3.0,
1959
+ "tau": 2.0
1960
+ },
1961
+ "rate": {
1962
+ "1": 0.775,
1963
+ "4": 0.84375,
1964
+ "16": 0.94375
1965
+ }
1966
+ },
1967
+ "gp_ard:c=1.0,eta=3.0,tau=3.0": {
1968
+ "arm": "gp_ard",
1969
+ "cfg": {
1970
+ "c": 1.0,
1971
+ "eta": 3.0,
1972
+ "tau": 3.0
1973
+ },
1974
+ "rate": {
1975
+ "1": 0.775,
1976
+ "4": 0.84375,
1977
+ "16": 0.9625
1978
+ }
1979
+ },
1980
+ "gp_ard:c=1.0,eta=10.0,tau=0.5": {
1981
+ "arm": "gp_ard",
1982
+ "cfg": {
1983
+ "c": 1.0,
1984
+ "eta": 10.0,
1985
+ "tau": 0.5
1986
+ },
1987
+ "rate": {
1988
+ "1": 0.7875,
1989
+ "4": 0.86875,
1990
+ "16": 0.93125
1991
+ }
1992
+ },
1993
+ "gp_ard:c=1.0,eta=10.0,tau=1.0": {
1994
+ "arm": "gp_ard",
1995
+ "cfg": {
1996
+ "c": 1.0,
1997
+ "eta": 10.0,
1998
+ "tau": 1.0
1999
+ },
2000
+ "rate": {
2001
+ "1": 0.7875,
2002
+ "4": 0.89375,
2003
+ "16": 0.9375
2004
+ }
2005
+ },
2006
+ "gp_ard:c=1.0,eta=10.0,tau=2.0": {
2007
+ "arm": "gp_ard",
2008
+ "cfg": {
2009
+ "c": 1.0,
2010
+ "eta": 10.0,
2011
+ "tau": 2.0
2012
+ },
2013
+ "rate": {
2014
+ "1": 0.7875,
2015
+ "4": 0.825,
2016
+ "16": 0.93125
2017
+ }
2018
+ },
2019
+ "gp_ard:c=1.0,eta=10.0,tau=3.0": {
2020
+ "arm": "gp_ard",
2021
+ "cfg": {
2022
+ "c": 1.0,
2023
+ "eta": 10.0,
2024
+ "tau": 3.0
2025
+ },
2026
+ "rate": {
2027
+ "1": 0.7875,
2028
+ "4": 0.875,
2029
+ "16": 0.95625
2030
+ }
2031
+ },
2032
+ "gp_ard:c=2.0,eta=0.3,tau=0.5": {
2033
+ "arm": "gp_ard",
2034
+ "cfg": {
2035
+ "c": 2.0,
2036
+ "eta": 0.3,
2037
+ "tau": 0.5
2038
+ },
2039
+ "rate": {
2040
+ "1": 0.59375,
2041
+ "4": 0.625,
2042
+ "16": 0.6375
2043
+ }
2044
+ },
2045
+ "gp_ard:c=2.0,eta=0.3,tau=1.0": {
2046
+ "arm": "gp_ard",
2047
+ "cfg": {
2048
+ "c": 2.0,
2049
+ "eta": 0.3,
2050
+ "tau": 1.0
2051
+ },
2052
+ "rate": {
2053
+ "1": 0.59375,
2054
+ "4": 0.63125,
2055
+ "16": 0.6625
2056
+ }
2057
+ },
2058
+ "gp_ard:c=2.0,eta=0.3,tau=2.0": {
2059
+ "arm": "gp_ard",
2060
+ "cfg": {
2061
+ "c": 2.0,
2062
+ "eta": 0.3,
2063
+ "tau": 2.0
2064
+ },
2065
+ "rate": {
2066
+ "1": 0.59375,
2067
+ "4": 0.66875,
2068
+ "16": 0.725
2069
+ }
2070
+ },
2071
+ "gp_ard:c=2.0,eta=0.3,tau=3.0": {
2072
+ "arm": "gp_ard",
2073
+ "cfg": {
2074
+ "c": 2.0,
2075
+ "eta": 0.3,
2076
+ "tau": 3.0
2077
+ },
2078
+ "rate": {
2079
+ "1": 0.59375,
2080
+ "4": 0.7125,
2081
+ "16": 0.7625
2082
+ }
2083
+ },
2084
+ "gp_ard:c=2.0,eta=1.0,tau=0.5": {
2085
+ "arm": "gp_ard",
2086
+ "cfg": {
2087
+ "c": 2.0,
2088
+ "eta": 1.0,
2089
+ "tau": 0.5
2090
+ },
2091
+ "rate": {
2092
+ "1": 0.61875,
2093
+ "4": 0.6375,
2094
+ "16": 0.68125
2095
+ }
2096
+ },
2097
+ "gp_ard:c=2.0,eta=1.0,tau=1.0": {
2098
+ "arm": "gp_ard",
2099
+ "cfg": {
2100
+ "c": 2.0,
2101
+ "eta": 1.0,
2102
+ "tau": 1.0
2103
+ },
2104
+ "rate": {
2105
+ "1": 0.61875,
2106
+ "4": 0.675,
2107
+ "16": 0.7125
2108
+ }
2109
+ },
2110
+ "gp_ard:c=2.0,eta=1.0,tau=2.0": {
2111
+ "arm": "gp_ard",
2112
+ "cfg": {
2113
+ "c": 2.0,
2114
+ "eta": 1.0,
2115
+ "tau": 2.0
2116
+ },
2117
+ "rate": {
2118
+ "1": 0.61875,
2119
+ "4": 0.68125,
2120
+ "16": 0.7625
2121
+ }
2122
+ },
2123
+ "gp_ard:c=2.0,eta=1.0,tau=3.0": {
2124
+ "arm": "gp_ard",
2125
+ "cfg": {
2126
+ "c": 2.0,
2127
+ "eta": 1.0,
2128
+ "tau": 3.0
2129
+ },
2130
+ "rate": {
2131
+ "1": 0.61875,
2132
+ "4": 0.71875,
2133
+ "16": 0.81875
2134
+ }
2135
+ },
2136
+ "gp_ard:c=2.0,eta=3.0,tau=0.5": {
2137
+ "arm": "gp_ard",
2138
+ "cfg": {
2139
+ "c": 2.0,
2140
+ "eta": 3.0,
2141
+ "tau": 0.5
2142
+ },
2143
+ "rate": {
2144
+ "1": 0.63125,
2145
+ "4": 0.65625,
2146
+ "16": 0.68125
2147
+ }
2148
+ },
2149
+ "gp_ard:c=2.0,eta=3.0,tau=1.0": {
2150
+ "arm": "gp_ard",
2151
+ "cfg": {
2152
+ "c": 2.0,
2153
+ "eta": 3.0,
2154
+ "tau": 1.0
2155
+ },
2156
+ "rate": {
2157
+ "1": 0.63125,
2158
+ "4": 0.675,
2159
+ "16": 0.725
2160
+ }
2161
+ },
2162
+ "gp_ard:c=2.0,eta=3.0,tau=2.0": {
2163
+ "arm": "gp_ard",
2164
+ "cfg": {
2165
+ "c": 2.0,
2166
+ "eta": 3.0,
2167
+ "tau": 2.0
2168
+ },
2169
+ "rate": {
2170
+ "1": 0.63125,
2171
+ "4": 0.73125,
2172
+ "16": 0.8125
2173
+ }
2174
+ },
2175
+ "gp_ard:c=2.0,eta=3.0,tau=3.0": {
2176
+ "arm": "gp_ard",
2177
+ "cfg": {
2178
+ "c": 2.0,
2179
+ "eta": 3.0,
2180
+ "tau": 3.0
2181
+ },
2182
+ "rate": {
2183
+ "1": 0.63125,
2184
+ "4": 0.70625,
2185
+ "16": 0.83125
2186
+ }
2187
+ },
2188
+ "gp_ard:c=2.0,eta=10.0,tau=0.5": {
2189
+ "arm": "gp_ard",
2190
+ "cfg": {
2191
+ "c": 2.0,
2192
+ "eta": 10.0,
2193
+ "tau": 0.5
2194
+ },
2195
+ "rate": {
2196
+ "1": 0.5875,
2197
+ "4": 0.61875,
2198
+ "16": 0.66875
2199
+ }
2200
+ },
2201
+ "gp_ard:c=2.0,eta=10.0,tau=1.0": {
2202
+ "arm": "gp_ard",
2203
+ "cfg": {
2204
+ "c": 2.0,
2205
+ "eta": 10.0,
2206
+ "tau": 1.0
2207
+ },
2208
+ "rate": {
2209
+ "1": 0.5875,
2210
+ "4": 0.64375,
2211
+ "16": 0.7875
2212
+ }
2213
+ },
2214
+ "gp_ard:c=2.0,eta=10.0,tau=2.0": {
2215
+ "arm": "gp_ard",
2216
+ "cfg": {
2217
+ "c": 2.0,
2218
+ "eta": 10.0,
2219
+ "tau": 2.0
2220
+ },
2221
+ "rate": {
2222
+ "1": 0.5875,
2223
+ "4": 0.7125,
2224
+ "16": 0.8375
2225
+ }
2226
+ },
2227
+ "gp_ard:c=2.0,eta=10.0,tau=3.0": {
2228
+ "arm": "gp_ard",
2229
+ "cfg": {
2230
+ "c": 2.0,
2231
+ "eta": 10.0,
2232
+ "tau": 3.0
2233
+ },
2234
+ "rate": {
2235
+ "1": 0.5875,
2236
+ "4": 0.7375,
2237
+ "16": 0.88125
2238
+ }
2239
+ }
2240
+ }
2241
+ }
results/icml_r3/X10M/val.npz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5dbbc8a6060646963c1bb10d7ad1a5cca295b28b44d1286769e56bd6235e82b
3
+ size 25054