diff options
Diffstat (limited to 'results/snapshot_no_outln_v1/run_s123.log')
| -rw-r--r-- | results/snapshot_no_outln_v1/run_s123.log | 51 |
1 files changed, 51 insertions, 0 deletions
diff --git a/results/snapshot_no_outln_v1/run_s123.log b/results/snapshot_no_outln_v1/run_s123.log new file mode 100644 index 0000000..a946d6d --- /dev/null +++ b/results/snapshot_no_outln_v1/run_s123.log @@ -0,0 +1,51 @@ +NO-OUT_LN VARIANT: depth=4, d_hidden=256, epochs=100, seed=123 + +=== BP training (NO out_ln) === + [BP-noLN] Ep 0: ||h_L||=8.791e+00 ||g||=5.394e-04 acc=0.1016 + [BP-noLN] Ep 1: ||h_L||=3.456e+01 ||g||=1.382e-04 acc=0.4355 + [BP-noLN] Ep 5: ||h_L||=5.193e+01 ||g||=9.721e-05 acc=0.5049 + [BP-noLN] Ep 10: ||h_L||=6.115e+01 ||g||=8.778e-05 acc=0.5430 + [BP-noLN] Ep 15: ||h_L||=6.668e+01 ||g||=8.387e-05 acc=0.5586 + [BP-noLN] Ep 20: ||h_L||=6.880e+01 ||g||=8.997e-05 acc=0.5176 + [BP-noLN] Ep 25: ||h_L||=7.092e+01 ||g||=8.184e-05 acc=0.5850 + [BP-noLN] Ep 30: ||h_L||=7.285e+01 ||g||=7.960e-05 acc=0.5693 + [BP-noLN] Ep 35: ||h_L||=7.522e+01 ||g||=8.247e-05 acc=0.5869 + [BP-noLN] Ep 40: ||h_L||=7.551e+01 ||g||=8.186e-05 acc=0.6201 + [BP-noLN] Ep 45: ||h_L||=7.421e+01 ||g||=8.565e-05 acc=0.5918 + [BP-noLN] Ep 50: ||h_L||=7.473e+01 ||g||=9.094e-05 acc=0.5840 + [BP-noLN] Ep 55: ||h_L||=7.556e+01 ||g||=8.896e-05 acc=0.6143 + [BP-noLN] Ep 60: ||h_L||=7.545e+01 ||g||=9.015e-05 acc=0.6396 + [BP-noLN] Ep 65: ||h_L||=7.509e+01 ||g||=9.009e-05 acc=0.6455 + [BP-noLN] Ep 70: ||h_L||=7.466e+01 ||g||=9.170e-05 acc=0.6230 + [BP-noLN] Ep 75: ||h_L||=7.410e+01 ||g||=1.030e-04 acc=0.6455 + [BP-noLN] Ep 80: ||h_L||=7.313e+01 ||g||=1.043e-04 acc=0.6309 + [BP-noLN] Ep 85: ||h_L||=7.275e+01 ||g||=1.035e-04 acc=0.6396 + [BP-noLN] Ep 90: ||h_L||=7.262e+01 ||g||=1.079e-04 acc=0.6445 + [BP-noLN] Ep 95: ||h_L||=7.256e+01 ||g||=1.087e-04 acc=0.6377 + [BP-noLN] Ep 100: ||h_L||=7.248e+01 ||g||=1.082e-04 acc=0.6396 + +=== DFA training (NO out_ln) === + [DFA-noLN] Ep 0: ||h_L||=8.791e+00 ||g||=5.394e-04 acc=0.1016 + [DFA-noLN] Ep 1: ||h_L||=1.231e+03 ||g||=6.461e-04 acc=0.2129 γ=0.0130 + [DFA-noLN] Ep 5: ||h_L||=1.290e+04 ||g||=7.423e-04 acc=0.2080 γ=-0.0053 + [DFA-noLN] Ep 10: ||h_L||=6.970e+04 ||g||=7.399e-04 acc=0.2031 γ=-0.0038 + [DFA-noLN] Ep 15: ||h_L||=2.144e+05 ||g||=7.383e-04 acc=0.1807 γ=0.0158 + [DFA-noLN] Ep 20: ||h_L||=5.289e+05 ||g||=7.379e-04 acc=0.2295 γ=0.0205 + [DFA-noLN] Ep 25: ||h_L||=1.047e+06 ||g||=7.582e-04 acc=0.1729 γ=-0.0056 + [DFA-noLN] Ep 30: ||h_L||=1.809e+06 ||g||=7.676e-04 acc=0.1934 γ=0.0171 + [DFA-noLN] Ep 35: ||h_L||=2.741e+06 ||g||=7.757e-04 acc=0.1504 γ=0.0181 + [DFA-noLN] Ep 40: ||h_L||=3.868e+06 ||g||=7.547e-04 acc=0.1377 γ=0.0036 + [DFA-noLN] Ep 45: ||h_L||=5.087e+06 ||g||=7.677e-04 acc=0.2246 γ=0.0148 + [DFA-noLN] Ep 50: ||h_L||=6.378e+06 ||g||=7.632e-04 acc=0.2002 γ=-0.0010 + [DFA-noLN] Ep 55: ||h_L||=7.652e+06 ||g||=7.547e-04 acc=0.0908 γ=0.0198 + [DFA-noLN] Ep 60: ||h_L||=8.786e+06 ||g||=7.196e-04 acc=0.1670 γ=0.0255 + [DFA-noLN] Ep 65: ||h_L||=9.795e+06 ||g||=7.598e-04 acc=0.1699 γ=0.0191 + [DFA-noLN] Ep 70: ||h_L||=1.063e+07 ||g||=7.313e-04 acc=0.2422 γ=0.0235 + [DFA-noLN] Ep 75: ||h_L||=1.127e+07 ||g||=7.316e-04 acc=0.2490 γ=0.0445 + [DFA-noLN] Ep 80: ||h_L||=1.173e+07 ||g||=7.425e-04 acc=0.1904 γ=0.0283 + [DFA-noLN] Ep 85: ||h_L||=1.203e+07 ||g||=7.233e-04 acc=0.2500 γ=0.0402 + [DFA-noLN] Ep 90: ||h_L||=1.216e+07 ||g||=7.064e-04 acc=0.3096 γ=0.0335 + [DFA-noLN] Ep 95: ||h_L||=1.218e+07 ||g||=7.087e-04 acc=0.2920 γ=0.0295 + [DFA-noLN] Ep 100: ||h_L||=1.218e+07 ||g||=7.086e-04 acc=0.3135 γ=0.0273 + +Saved results/snapshot_no_outln_v1/snapshot_noLN_s123.json |
