Local Config:
dataset = adult
n_estimators = 1
max_depth_t = None
seed = 0
Using dataset adult, training set size is 10 with 19 attributes.
accuracy_train= 1.0 accuracy_test= 0.7657718120805369
RF parsing done!
Model creation done!
Starting CP-SAT solver v9.8.3296
Parameters: random_seed: 0 max_time_in_seconds: 18000 log_search_progress: true num_workers: 8
Initial satisfaction model '': (model_fingerprint: 0x388a71457e5c97d2)
#Variables: 190
- 190 Booleans in [0,1]
#kLinear0: 10
#kLinearN: 10 (#terms: 60)
Starting presolve at 0.00s
INFEASIBLE: 'proven during initial copy of constraint #10:
linear {
domain: [1, 1]
}
With current variable domains:
'
Presolve summary:
- 0 affine relations were detected.
Problem closed by presolve.
CpSolverResponse summary:
status: INFEASIBLE
objective: NA
best_bound: NA
integers: 0
booleans: 0
conflicts: 0
branches: 0
propagations: 0
integer_propagations: 0
restarts: 0
lp_iterations: 0
walltime: 0.000337
usertime: 0.000337
deterministic_time: 0
gap_integral: 0
Traceback (most recent call last):
File "DRAFT/data_extraction_from_rf_experiments.py", line 265, in <module>
dict_res = extractor.fit(
^^^^^^^^^^^^^^
File "DRAFT/DRAFT.py", line 299, in fit
self.perform_reconstruction_v1_CP_SAT(
File "DRAFT/DRAFT.py", line 474, in perform_reconstruction_v1_CP_SAT
raise RuntimeError(
RuntimeError: Infeasible model: the reconstruction problem has no solution. Please make sure the provided one-hot encoding constraints are correct. Else, report this issue to the developers.
I have not touched the one-hot encoding and it works for the other 2 datasets.
Additionally, as a general comment, model-building time in your experiments is part of the solve_duration_time
, maybe this should be considered as a separate timer as it doesn't reflect the solver's ability to solve the problem but the speed of the API (which can of course vary depending on the implementation and the language used).