Created
June 23, 2011 07:54
-
-
Save danbri/1042113 to your computer and use it in GitHub Desktop.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| a few test runs | |
| in mode 0, first run was with 0.8, | |
| Final result. Obj=305652, TRAIN RMSE= 0.7663 VALIDATION RMSE= 1.2225. | |
| then 0.9 train/test ratio | |
| Final result. Obj=356242, TRAIN RMSE= 0.7842 VALIDATION RMSE= 1.1875. | |
| then mode 1 at .9 ratio | |
| Final result. Obj=821376, TRAIN RMSE= 1.2244 VALIDATION RMSE= 1.2988. | |
| INFO: pmf.cpp(main:1285): PMF/ALS/SVD++/SGD Code written By Danny Bickson, CMU | |
| Send bug reports and comments to [email protected] | |
| WARNING: pmf.cpp(main:1287): Code compiled with GL_NO_MULT_EDGES flag - this mode does not support multiple edges between user and movie in different times | |
| Setting run mode ALS_MATRIX | |
| INFO: pmf.cpp(main:1335): ALS_MATRIX starting | |
| loading data file 2010beebzbin | |
| Loading 2010beebzbin TRAINING | |
| Matrix size is: USERS 19500 MOVIES 12777 TIME BINS 1 | |
| Creating 958503 edges (observed ratings)... | |
| .....loading data file 2010beebzbine | |
| Loading 2010beebzbine VALIDATION | |
| Matrix size is: USERS 19500 MOVIES 12777 TIME BINS 1 | |
| Creating 239626 edges (observed ratings)... | |
| ..loading data file 2010beebzbint | |
| Loading 2010beebzbint TEST | |
| skipping file | |
| setting regularization weight to 0.065 | |
| ALS_MATRIX for matrix (19500, 12777, 1):958503. D=20 | |
| pU=0.065, pV=0.065, pT=1, muT=1, D=20 | |
| complete. Obj=3.21913e+07, TRAIN RMSE=8.1957 VALIDATION RMSE=8.1947. | |
| max iterations = 10 | |
| step = 1 | |
| max_iterations = 10 | |
| INFO: asynchronous_engine.hpp(run:94): Worker 0 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 2 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 1 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 3 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 4 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 5 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 6 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 7 started. | |
| Entering last iter with 1 | |
| 2.32317) Iter ALS_MATRIX 1 Obj=8.55617e+06, TRAIN RMSE=4.1873 VALIDATION RMSE=2.3568. | |
| Entering last iter with 2 | |
| 4.69663) Iter ALS_MATRIX 2 Obj=4.54246e+06, TRAIN RMSE=3.0502 VALIDATION RMSE=1.4099. | |
| Entering last iter with 3 | |
| 7.05022) Iter ALS_MATRIX 3 Obj=1.03557e+06, TRAIN RMSE=1.4276 VALIDATION RMSE=1.3620. | |
| Entering last iter with 4 | |
| 9.71574) Iter ALS_MATRIX 4 Obj=672046, TRAIN RMSE=1.1433 VALIDATION RMSE=1.3318. | |
| Entering last iter with 5 | |
| 11.7241) Iter ALS_MATRIX 5 Obj=529501, TRAIN RMSE=1.0111 VALIDATION RMSE=1.2939. | |
| Entering last iter with 6 | |
| 14.0961) Iter ALS_MATRIX 6 Obj=445449, TRAIN RMSE=0.9261 VALIDATION RMSE=1.2766. | |
| Entering last iter with 7 | |
| 16.4428) Iter ALS_MATRIX 7 Obj=390646, TRAIN RMSE=0.8667 VALIDATION RMSE=1.2591. | |
| Entering last iter with 8 | |
| 18.6453) Iter ALS_MATRIX 8 Obj=356157, TRAIN RMSE=0.8273 VALIDATION RMSE=1.2463. | |
| Entering last iter with 9 | |
| 21.0598) Iter ALS_MATRIX 9 Obj=328914, TRAIN RMSE=0.7953 VALIDATION RMSE=1.2369. | |
| Entering last iter with 10 | |
| 23.2363) Iter ALS_MATRIX 10 Obj=310411, TRAIN RMSE=0.7727 VALIDATION RMSE=1.2231. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 1 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 6 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 2 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 7 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 3 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 0 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 4 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 5 finished. | |
| Final result. Obj=305652, TRAIN RMSE= 0.7663 VALIDATION RMSE= 1.2225. | |
| Finished in 23.355002 | |
| Performance counters are: 0) EDGE_TRAVERSAL, 30.6182 | |
| Performance counters are: 2) CALC_RMSE_Q, 0.008503 | |
| Performance counters are: 3) ALS_LEAST_SQUARES, 51.7636 | |
| Performance counters are: 6) CALC_OBJ, 0.106283 | |
| === REPORT FOR core() === | |
| [Numeric] | |
| ncpus: 8 | |
| [Other] | |
| affinities: false | |
| compile_flags: | |
| engine: async | |
| scheduler: round_robin | |
| schedyield: true | |
| scope: edge | |
| === REPORT FOR engine() === | |
| [Numeric] | |
| num_edges: 958503 | |
| num_syncs: 0 | |
| num_vertices: 32277 | |
| updatecount: 322770 | |
| [Timings] | |
| runtime: 23.3 s | |
| [Other] | |
| termination_reason: task depletion (natural) | |
| [Numeric] | |
| updatecount_vector: 322770 (count: 8, min: 37506, max: 43344, avg: 40346.2) | |
| updatecount_vector.values: 38304,43344,43344,37506,40320,40320,40320,39312, | |
| and again with 0.9 ie more training data, | |
| TellyClub:danbri danbri$ cat go.sh | |
| #!/bin/sh | |
| pmf 2010beebzbin 0 --ncpus=8 --float=true --lambda=0.065 --scheduler="round_robin(max_iterations=10)" | |
| TellyClub:danbri danbri$ ./go.sh | |
| INFO: pmf.cpp(main:1285): PMF/ALS/SVD++/SGD Code written By Danny Bickson, CMU | |
| Send bug reports and comments to [email protected] | |
| WARNING: pmf.cpp(main:1287): Code compiled with GL_NO_MULT_EDGES flag - this mode does not support multiple edges between user and movie in different times | |
| Setting run mode ALS_MATRIX | |
| INFO: pmf.cpp(main:1335): ALS_MATRIX starting | |
| loading data file 2010beebzbin | |
| Loading 2010beebzbin TRAINING | |
| Matrix size is: USERS 19500 MOVIES 12777 TIME BINS 1 | |
| Creating 1078316 edges (observed ratings)... | |
| ......loading data file 2010beebzbine | |
| Loading 2010beebzbine VALIDATION | |
| Matrix size is: USERS 19500 MOVIES 12777 TIME BINS 1 | |
| Creating 119813 edges (observed ratings)... | |
| .loading data file 2010beebzbint | |
| Loading 2010beebzbint TEST | |
| skipping file | |
| setting regularization weight to 0.065 | |
| ALS_MATRIX for matrix (19500, 12777, 1):1078316. D=20 | |
| pU=0.065, pV=0.065, pT=1, muT=1, D=20 | |
| complete. Obj=3.62124e+07, TRAIN RMSE=8.1954 VALIDATION RMSE=8.1965. | |
| max iterations = 10 | |
| step = 1 | |
| max_iterations = 10 | |
| INFO: asynchronous_engine.hpp(run:94): Worker 0 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 1 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 2 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 3 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 4 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 5 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 6 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 7 started. | |
| Entering last iter with 1 | |
| 2.46273) Iter ALS_MATRIX 1 Obj=9.599e+06, TRAIN RMSE=4.1853 VALIDATION RMSE=2.4512. | |
| Entering last iter with 2 | |
| 5.0279) Iter ALS_MATRIX 2 Obj=5.14698e+06, TRAIN RMSE=3.0638 VALIDATION RMSE=1.4456. | |
| Entering last iter with 3 | |
| 7.68911) Iter ALS_MATRIX 3 Obj=1.1557e+06, TRAIN RMSE=1.4253 VALIDATION RMSE=1.3210. | |
| Entering last iter with 4 | |
| 10.4047) Iter ALS_MATRIX 4 Obj=781178, TRAIN RMSE=1.1661 VALIDATION RMSE=1.2854. | |
| Entering last iter with 5 | |
| 12.998) Iter ALS_MATRIX 5 Obj=610431, TRAIN RMSE=1.0282 VALIDATION RMSE=1.2554. | |
| Entering last iter with 6 | |
| 15.5638) Iter ALS_MATRIX 6 Obj=515583, TRAIN RMSE=0.9438 VALIDATION RMSE=1.2351. | |
| Entering last iter with 7 | |
| 18.2514) Iter ALS_MATRIX 7 Obj=452888, TRAIN RMSE=0.8840 VALIDATION RMSE=1.2211. | |
| Entering last iter with 8 | |
| 20.9628) Iter ALS_MATRIX 8 Obj=411440, TRAIN RMSE=0.8426 VALIDATION RMSE=1.2082. | |
| Entering last iter with 9 | |
| 23.7272) Iter ALS_MATRIX 9 Obj=379939, TRAIN RMSE=0.8099 VALIDATION RMSE=1.1997. | |
| Entering last iter with 10 | |
| 26.1217) Iter ALS_MATRIX 10 Obj=359368, TRAIN RMSE=0.7879 VALIDATION RMSE=1.1921. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 3 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 6 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 1 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 2 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 5 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 4 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 0 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 7 finished. | |
| Final result. Obj=356242, TRAIN RMSE= 0.7842 VALIDATION RMSE= 1.1875. | |
| Finished in 26.337415 | |
| Performance counters are: 0) EDGE_TRAVERSAL, 38.0618 | |
| Performance counters are: 2) CALC_RMSE_Q, 0.007834 | |
| Performance counters are: 3) ALS_LEAST_SQUARES, 62.4937 | |
| Performance counters are: 6) CALC_OBJ, 0.080001 | |
| === REPORT FOR core() === | |
| [Numeric] | |
| ncpus: 8 | |
| [Other] | |
| affinities: false | |
| compile_flags: | |
| engine: async | |
| scheduler: round_robin | |
| schedyield: true | |
| scope: edge | |
| === REPORT FOR engine() === | |
| [Numeric] | |
| num_edges: 1.07832e+06 | |
| num_syncs: 0 | |
| num_vertices: 32277 | |
| updatecount: 322770 | |
| [Timings] | |
| runtime: 26.3 s | |
| [Other] | |
| termination_reason: task depletion (natural) | |
| [Numeric] | |
| updatecount_vector: 322770 (count: 8, min: 38304, max: 43344, avg: 40346.2) | |
| updatecount_vector.values: 39312,43344,40320,38304,41328,40320,40320,39522, | |
| and again in another mode, 1 | |
| TellyClub:danbri danbri$ sh go.sh | |
| INFO: pmf.cpp(main:1285): PMF/ALS/SVD++/SGD Code written By Danny Bickson, CMU | |
| Send bug reports and comments to [email protected] | |
| WARNING: pmf.cpp(main:1287): Code compiled with GL_NO_MULT_EDGES flag - this mode does not support multiple edges between user and movie in different times | |
| Setting run mode SVD++ | |
| INFO: pmf.cpp(main:1335): SVD++ starting | |
| ERROR: pmf.cpp(main:1358): Can not run required algorithm without GL_SVD_PP flag. Please define flag on pmf.h and recompile | |
| TellyClub:danbri danbri$ nano go.sh | |
| TellyClub:danbri danbri$ sh go.sh | |
| INFO: pmf.cpp(main:1285): PMF/ALS/SVD++/SGD Code written By Danny Bickson, CMU | |
| Send bug reports and comments to [email protected] | |
| WARNING: pmf.cpp(main:1287): Code compiled with GL_NO_MULT_EDGES flag - this mode does not support multiple edges between user and movie in different times | |
| Setting run mode BPTF_MATRIX | |
| INFO: pmf.cpp(main:1335): BPTF_MATRIX starting | |
| loading data file 2010beebzbin | |
| Loading 2010beebzbin TRAINING | |
| Matrix size is: USERS 19500 MOVIES 12777 TIME BINS 1 | |
| Creating 1078316 edges (observed ratings)... | |
| ......loading data file 2010beebzbine | |
| Loading 2010beebzbine VALIDATION | |
| Matrix size is: USERS 19500 MOVIES 12777 TIME BINS 1 | |
| Creating 119813 edges (observed ratings)... | |
| .loading data file 2010beebzbint | |
| Loading 2010beebzbint TEST | |
| skipping file | |
| setting regularization weight to 0.065 | |
| BPTF_MATRIX for matrix (19500, 12777, 1):1078316. D=20 | |
| pU=0.065, pV=0.065, pT=1, muT=1, D=20 | |
| nuAlpha=1, Walpha=1, mu=0, muT=1, nu=20, beta=1, W=1, WT=1 BURN_IN=10 | |
| complete. Obj=3.62124e+07, TRAIN RMSE=8.1954 VALIDATION RMSE=8.1965. | |
| sampled alpha is 1.00191 | |
| max iterations = 10 | |
| step = 1 | |
| max_iterations = 10 | |
| INFO: asynchronous_engine.hpp(run:94): Worker 0 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 1 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 2 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 3 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 5 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 6 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 7 started. | |
| INFO: asynchronous_engine.hpp(run:94): Worker 4 started. | |
| Entering last iter with 1 | |
| 3.68712) Iter BPTF_MATRIX 1 Obj=3.00105e+07, TRAIN RMSE=7.4606 VALIDATION RMSE=7.5836. | |
| sampled alpha is 0.017951 | |
| Entering last iter with 2 | |
| 7.45824) Iter BPTF_MATRIX 2 Obj=2.75293e+07, TRAIN RMSE=7.1453 VALIDATION RMSE=5.1945. | |
| sampled alpha is 0.0195931 | |
| Entering last iter with 3 | |
| 11.3544) Iter BPTF_MATRIX 3 Obj=9.87986e+06, TRAIN RMSE=4.2790 VALIDATION RMSE=2.2782. | |
| sampled alpha is 0.0546689 | |
| Entering last iter with 4 | |
| 15.4137) Iter BPTF_MATRIX 4 Obj=1.98435e+06, TRAIN RMSE=1.9130 VALIDATION RMSE=1.5121. | |
| sampled alpha is 0.272499 | |
| Entering last iter with 5 | |
| 19.2782) Iter BPTF_MATRIX 5 Obj=1.01961e+06, TRAIN RMSE=1.3667 VALIDATION RMSE=1.3328. | |
| sampled alpha is 0.536882 | |
| Entering last iter with 6 | |
| 23.1699) Iter BPTF_MATRIX 6 Obj=851561, TRAIN RMSE=1.2472 VALIDATION RMSE=1.3053. | |
| sampled alpha is 0.642909 | |
| Entering last iter with 7 | |
| 27.2852) Iter BPTF_MATRIX 7 Obj=826420, TRAIN RMSE=1.2282 VALIDATION RMSE=1.2982. | |
| sampled alpha is 0.664127 | |
| Entering last iter with 8 | |
| 31.8299) Iter BPTF_MATRIX 8 Obj=821087, TRAIN RMSE=1.2242 VALIDATION RMSE=1.2988. | |
| sampled alpha is 0.668049 | |
| Entering last iter with 9 | |
| 35.7247) Iter BPTF_MATRIX 9 Obj=821367, TRAIN RMSE=1.2243 VALIDATION RMSE=1.2974. | |
| Finished burn-in period. starting to aggregate samples | |
| sampled alpha is 0.666922 | |
| Entering last iter with 10 | |
| 39.4087) Iter BPTF_MATRIX 10 Obj=821436, TRAIN RMSE=1.2244 VALIDATION RMSE=1.2987. | |
| sampled alpha is 0.667033 | |
| INFO: asynchronous_engine.hpp(run:102): Worker 4 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 3 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 7 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 1 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 5 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 6 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 2 finished. | |
| INFO: asynchronous_engine.hpp(run:102): Worker 0 finished. | |
| Final result. Obj=821376, TRAIN RMSE= 1.2244 VALIDATION RMSE= 1.2988. | |
| Finished in 39.663386 | |
| Performance counters are: 0) EDGE_TRAVERSAL, 44.4005 | |
| Performance counters are: 1) BPTF_SAMPLE_STEP, 1.73233 | |
| Performance counters are: 2) CALC_RMSE_Q, 0.005454 | |
| Performance counters are: 6) CALC_OBJ, 0.099321 | |
| Performance counters are: 7) BPTF_MVN_RNDEX, 30.5902 | |
| Performance counters are: 8) BPTF_LEAST_SQUARES2, 14.2635 | |
| === REPORT FOR core() === | |
| [Numeric] | |
| ncpus: 8 | |
| [Other] | |
| affinities: false | |
| compile_flags: | |
| engine: async | |
| scheduler: round_robin | |
| schedyield: true | |
| scope: edge | |
| === REPORT FOR engine() === | |
| [Numeric] | |
| num_edges: 1.07832e+06 | |
| num_syncs: 0 | |
| num_vertices: 32277 | |
| updatecount: 322770 | |
| [Timings] | |
| runtime: 39.6 s | |
| [Other] | |
| termination_reason: task depletion (natural) | |
| [Numeric] | |
| updatecount_vector: 322770 (count: 8, min: 39312, max: 41328, avg: 40346.2) | |
| updatecount_vector.values: 40320,40320,39522,41328,41328,39312,39312,41328, | |
| TellyClub:danbri danbri$ cat go.sh | |
| #!/bin/sh | |
| pmf 2010beebzbin 1 --ncpus=8 --float=true --lambda=0.065 --scheduler="round_robin(max_iterations=10)" | |
| # 0 = Matrix factorization using alternating least squares | |
| # 1 = Matrix factorization using MCMC procedure | |
| # 2 = Tensor factorization using MCMC procedure, single edge exist between user and movies | |
| # 3 = Tensor factorization, using MCMC procedure with support for multiple edges between user and movies in different times | |
| # 4 = Tensor factorization using alternating least squars | |
| # 5 = SVD++ algorithm |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment