| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9972941631233089, | |
| "eval_steps": 500, | |
| "global_step": 2580, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.003865481252415926, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.9935575312459736e-05, | |
| "loss": 0.0, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.007730962504831852, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.987115062491947e-05, | |
| "loss": 0.0, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.011596443757247778, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.98067259373792e-05, | |
| "loss": 0.0, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.015461925009663703, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.974230124983894e-05, | |
| "loss": 0.0, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.019327406262079627, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.967787656229868e-05, | |
| "loss": 0.0, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.023192887514495556, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.961345187475841e-05, | |
| "loss": 0.0, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.02705836876691148, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.9549027187218144e-05, | |
| "loss": 0.0, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.030923850019327407, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.948460249967788e-05, | |
| "loss": 0.0, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.03478933127174333, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.942017781213761e-05, | |
| "loss": 0.0, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.038654812524159254, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.935575312459735e-05, | |
| "loss": 0.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.04252029377657519, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.9291328437057085e-05, | |
| "loss": 0.0, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.04638577502899111, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.922690374951682e-05, | |
| "loss": 0.0, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.05025125628140704, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.916247906197655e-05, | |
| "loss": 0.0, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.05411673753382296, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.9098054374436285e-05, | |
| "loss": 0.0, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.05798221878623889, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.903362968689602e-05, | |
| "loss": 0.0, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.061847700038654814, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.896920499935575e-05, | |
| "loss": 0.0, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.06571318129107073, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.890478031181549e-05, | |
| "loss": 0.0, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.06957866254348666, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.8840355624275226e-05, | |
| "loss": 0.0, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.07344414379590258, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.877593093673496e-05, | |
| "loss": 0.0, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.07730962504831851, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.8711506249194693e-05, | |
| "loss": 0.0, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.08117510630073445, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.864708156165443e-05, | |
| "loss": 0.0, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.08504058755315037, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.858265687411416e-05, | |
| "loss": 0.0, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.0889060688055663, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.85182321865739e-05, | |
| "loss": 0.0, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.09277155005798222, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.8453807499033634e-05, | |
| "loss": 0.0, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.09663703131039815, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.838938281149337e-05, | |
| "loss": 0.0, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.10050251256281408, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.83249581239531e-05, | |
| "loss": 0.0, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.10436799381523, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.8260533436412835e-05, | |
| "loss": 0.0, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.10823347506764593, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.819610874887257e-05, | |
| "loss": 0.0, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.11209895632006185, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.81316840613323e-05, | |
| "loss": 0.0, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.11596443757247778, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.806725937379204e-05, | |
| "loss": 0.0, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.1198299188248937, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.8002834686251776e-05, | |
| "loss": 0.0, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.12369540007730963, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.793840999871151e-05, | |
| "loss": 0.0, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.12756088132972554, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.787398531117124e-05, | |
| "loss": 0.0, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.13142636258214146, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.7809560623630976e-05, | |
| "loss": 0.0, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.1352918438345574, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.774513593609071e-05, | |
| "loss": 0.0, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.13915732508697332, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.768071124855045e-05, | |
| "loss": 0.0, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.14302280633938924, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.7616286561010184e-05, | |
| "loss": 0.0, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.14688828759180517, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.755186187346992e-05, | |
| "loss": 0.0, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.1507537688442211, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.748743718592965e-05, | |
| "loss": 0.0, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.15461925009663702, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.7423012498389384e-05, | |
| "loss": 0.0, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.15848473134905297, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.735858781084912e-05, | |
| "loss": 0.0, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.1623502126014689, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.729416312330885e-05, | |
| "loss": 0.0, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.16621569385388482, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.722973843576859e-05, | |
| "loss": 0.0, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.17008117510630075, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.7165313748228325e-05, | |
| "loss": 0.0, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.17394665635871667, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.710088906068806e-05, | |
| "loss": 0.0, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.1778121376111326, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.703646437314779e-05, | |
| "loss": 0.0, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.18167761886354852, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.6972039685607526e-05, | |
| "loss": 0.0, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.18554310011596445, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.690761499806726e-05, | |
| "loss": 0.0, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.18940858136838037, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.6843190310527e-05, | |
| "loss": 0.0, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.1932740626207963, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.677876562298673e-05, | |
| "loss": 0.0, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.19713954387321223, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.671434093544647e-05, | |
| "loss": 0.0, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.20100502512562815, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.66499162479062e-05, | |
| "loss": 0.0, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.20487050637804408, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.6585491560365934e-05, | |
| "loss": 0.0, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.20873598763046, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.652106687282567e-05, | |
| "loss": 0.0, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.21260146888287593, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.64566421852854e-05, | |
| "loss": 0.0, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.21646695013529185, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.639221749774514e-05, | |
| "loss": 0.0, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.22033243138770778, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.6327792810204875e-05, | |
| "loss": 0.0, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.2241979126401237, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.626336812266461e-05, | |
| "loss": 0.0, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.22806339389253963, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.619894343512434e-05, | |
| "loss": 0.0, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.23192887514495555, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.6134518747584075e-05, | |
| "loss": 0.0, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.23579435639737148, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.607009406004381e-05, | |
| "loss": 0.0, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.2396598376497874, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.600566937250354e-05, | |
| "loss": 0.0, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.24352531890220333, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.594124468496328e-05, | |
| "loss": 0.0, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.24739080015461926, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.5876819997423016e-05, | |
| "loss": 0.0, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.25125628140703515, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.581239530988275e-05, | |
| "loss": 0.0, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.2551217626594511, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.5747970622342483e-05, | |
| "loss": 0.0, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.258987243911867, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.568354593480222e-05, | |
| "loss": 0.0, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.26285272516428293, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.561912124726195e-05, | |
| "loss": 0.0, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.26671820641669886, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.555469655972169e-05, | |
| "loss": 0.0, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.2705836876691148, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.5490271872181424e-05, | |
| "loss": 0.0, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.2744491689215307, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.542584718464116e-05, | |
| "loss": 0.0, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.27831465017394663, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.536142249710089e-05, | |
| "loss": 0.0, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.28218013142636256, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.5296997809560625e-05, | |
| "loss": 0.0, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.2860456126787785, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.523257312202036e-05, | |
| "loss": 0.0, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.2899110939311944, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.516814843448009e-05, | |
| "loss": 0.0, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.29377657518361033, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.510372374693983e-05, | |
| "loss": 0.0, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.29764205643602626, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.5039299059399566e-05, | |
| "loss": 0.0, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.3015075376884422, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.49748743718593e-05, | |
| "loss": 0.0, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.3053730189408581, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.491044968431903e-05, | |
| "loss": 0.0, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.30923850019327404, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.4846024996778767e-05, | |
| "loss": 0.0, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.31310398144569, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.47816003092385e-05, | |
| "loss": 0.0, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.31696946269810594, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.471717562169824e-05, | |
| "loss": 0.0, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.32083494395052187, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.4652750934157974e-05, | |
| "loss": 0.0, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.3247004252029378, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.458832624661771e-05, | |
| "loss": 0.0, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.3285659064553537, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.452390155907744e-05, | |
| "loss": 0.0, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.33243138770776964, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.4459476871537175e-05, | |
| "loss": 0.0, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.33629686896018557, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.439505218399691e-05, | |
| "loss": 0.0, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.3401623502126015, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.433062749645664e-05, | |
| "loss": 0.0, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.3440278314650174, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.426620280891638e-05, | |
| "loss": 0.0, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.34789331271743335, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.4201778121376115e-05, | |
| "loss": 0.0, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.35175879396984927, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.413735343383585e-05, | |
| "loss": 0.0, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.3556242752222652, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.407292874629558e-05, | |
| "loss": 0.0, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.3594897564746811, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.4008504058755316e-05, | |
| "loss": 0.0, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.36335523772709705, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.394407937121505e-05, | |
| "loss": 0.0, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.367220718979513, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.387965468367479e-05, | |
| "loss": 0.0, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.3710862002319289, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.3815229996134523e-05, | |
| "loss": 0.0, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 0.3749516814843448, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.375080530859426e-05, | |
| "loss": 0.0, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 0.37881716273676075, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.368638062105399e-05, | |
| "loss": 0.0, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 0.3826826439891767, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.3621955933513724e-05, | |
| "loss": 0.0, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 0.3865481252415926, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.355753124597346e-05, | |
| "loss": 0.0, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.3904136064940085, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.349310655843319e-05, | |
| "loss": 0.0, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 0.39427908774642445, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.342868187089293e-05, | |
| "loss": 0.0, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 0.3981445689988404, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.3364257183352665e-05, | |
| "loss": 0.0, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 0.4020100502512563, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.32998324958124e-05, | |
| "loss": 0.0, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 0.4058755315036722, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.323540780827213e-05, | |
| "loss": 0.0, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 0.40974101275608815, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.3170983120731866e-05, | |
| "loss": 0.0, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 0.4136064940085041, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.31065584331916e-05, | |
| "loss": 0.0, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 0.41747197526092, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.304213374565134e-05, | |
| "loss": 0.0, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 0.42133745651333593, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.297770905811107e-05, | |
| "loss": 0.0, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 0.42520293776575185, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.2913284370570806e-05, | |
| "loss": 0.0, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.4290684190181678, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.284885968303054e-05, | |
| "loss": 0.0, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 0.4329339002705837, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.2784434995490274e-05, | |
| "loss": 0.0, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 0.43679938152299963, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.272001030795001e-05, | |
| "loss": 0.0, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 0.44066486277541556, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.265558562040974e-05, | |
| "loss": 0.0, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 0.4445303440278315, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.259116093286948e-05, | |
| "loss": 0.0, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 0.4483958252802474, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.2526736245329214e-05, | |
| "loss": 0.0, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 0.45226130653266333, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.246231155778895e-05, | |
| "loss": 0.0, | |
| "step": 1170 | |
| }, | |
| { | |
| "epoch": 0.45612678778507926, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.239788687024868e-05, | |
| "loss": 0.0, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 0.4599922690374952, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.2333462182708415e-05, | |
| "loss": 0.0, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 0.4638577502899111, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.226903749516815e-05, | |
| "loss": 0.0, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.46772323154232703, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.220461280762789e-05, | |
| "loss": 0.0, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 0.47158871279474296, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.214018812008762e-05, | |
| "loss": 0.0, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 0.4754541940471589, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.2075763432547356e-05, | |
| "loss": 0.0, | |
| "step": 1230 | |
| }, | |
| { | |
| "epoch": 0.4793196752995748, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.201133874500709e-05, | |
| "loss": 0.0, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 0.48318515655199074, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.194691405746682e-05, | |
| "loss": 0.0, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 0.48705063780440666, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.1882489369926557e-05, | |
| "loss": 0.0, | |
| "step": 1260 | |
| }, | |
| { | |
| "epoch": 0.4909161190568226, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.181806468238629e-05, | |
| "loss": 0.0, | |
| "step": 1270 | |
| }, | |
| { | |
| "epoch": 0.4947816003092385, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.175363999484603e-05, | |
| "loss": 0.0, | |
| "step": 1280 | |
| }, | |
| { | |
| "epoch": 0.49864708156165444, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.1689215307305764e-05, | |
| "loss": 0.0, | |
| "step": 1290 | |
| }, | |
| { | |
| "epoch": 0.5025125628140703, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.16247906197655e-05, | |
| "loss": 0.0, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.5063780440664862, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.156036593222523e-05, | |
| "loss": 0.0, | |
| "step": 1310 | |
| }, | |
| { | |
| "epoch": 0.5102435253189022, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.1495941244684965e-05, | |
| "loss": 0.0, | |
| "step": 1320 | |
| }, | |
| { | |
| "epoch": 0.5141090065713181, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.14315165571447e-05, | |
| "loss": 0.0, | |
| "step": 1330 | |
| }, | |
| { | |
| "epoch": 0.517974487823734, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.136709186960444e-05, | |
| "loss": 0.0, | |
| "step": 1340 | |
| }, | |
| { | |
| "epoch": 0.5218399690761499, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.130266718206417e-05, | |
| "loss": 0.0, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 0.5257054503285659, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.1238242494523905e-05, | |
| "loss": 0.0, | |
| "step": 1360 | |
| }, | |
| { | |
| "epoch": 0.5295709315809818, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.117381780698364e-05, | |
| "loss": 0.0, | |
| "step": 1370 | |
| }, | |
| { | |
| "epoch": 0.5334364128333977, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.110939311944337e-05, | |
| "loss": 0.0, | |
| "step": 1380 | |
| }, | |
| { | |
| "epoch": 0.5373018940858136, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.1044968431903106e-05, | |
| "loss": 0.0, | |
| "step": 1390 | |
| }, | |
| { | |
| "epoch": 0.5411673753382296, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.098054374436284e-05, | |
| "loss": 0.0, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.5450328565906455, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.091611905682258e-05, | |
| "loss": 0.0, | |
| "step": 1410 | |
| }, | |
| { | |
| "epoch": 0.5488983378430614, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.0851694369282313e-05, | |
| "loss": 0.0, | |
| "step": 1420 | |
| }, | |
| { | |
| "epoch": 0.5527638190954773, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.078726968174205e-05, | |
| "loss": 0.0, | |
| "step": 1430 | |
| }, | |
| { | |
| "epoch": 0.5566293003478933, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.072284499420178e-05, | |
| "loss": 0.0, | |
| "step": 1440 | |
| }, | |
| { | |
| "epoch": 0.5604947816003092, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.0658420306661514e-05, | |
| "loss": 0.0, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 0.5643602628527251, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.059399561912125e-05, | |
| "loss": 0.0, | |
| "step": 1460 | |
| }, | |
| { | |
| "epoch": 0.568225744105141, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.052957093158099e-05, | |
| "loss": 0.0, | |
| "step": 1470 | |
| }, | |
| { | |
| "epoch": 0.572091225357557, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.046514624404072e-05, | |
| "loss": 0.0, | |
| "step": 1480 | |
| }, | |
| { | |
| "epoch": 0.5759567066099729, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.0400721556500455e-05, | |
| "loss": 0.0, | |
| "step": 1490 | |
| }, | |
| { | |
| "epoch": 0.5798221878623888, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.033629686896019e-05, | |
| "loss": 0.0, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.5836876691148047, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.027187218141992e-05, | |
| "loss": 0.0, | |
| "step": 1510 | |
| }, | |
| { | |
| "epoch": 0.5875531503672207, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.0207447493879656e-05, | |
| "loss": 0.0, | |
| "step": 1520 | |
| }, | |
| { | |
| "epoch": 0.5914186316196366, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.014302280633939e-05, | |
| "loss": 0.0, | |
| "step": 1530 | |
| }, | |
| { | |
| "epoch": 0.5952841128720525, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.007859811879913e-05, | |
| "loss": 0.0, | |
| "step": 1540 | |
| }, | |
| { | |
| "epoch": 0.5991495941244684, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.001417343125886e-05, | |
| "loss": 0.0, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 0.6030150753768844, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.9949748743718597e-05, | |
| "loss": 0.0, | |
| "step": 1560 | |
| }, | |
| { | |
| "epoch": 0.6068805566293003, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.988532405617833e-05, | |
| "loss": 0.0, | |
| "step": 1570 | |
| }, | |
| { | |
| "epoch": 0.6107460378817162, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.9820899368638064e-05, | |
| "loss": 0.0, | |
| "step": 1580 | |
| }, | |
| { | |
| "epoch": 0.6146115191341321, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.97564746810978e-05, | |
| "loss": 0.0, | |
| "step": 1590 | |
| }, | |
| { | |
| "epoch": 0.6184770003865481, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.969204999355754e-05, | |
| "loss": 0.0, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.622342481638964, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.962762530601727e-05, | |
| "loss": 0.0, | |
| "step": 1610 | |
| }, | |
| { | |
| "epoch": 0.62620796289138, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.9563200618477005e-05, | |
| "loss": 0.0, | |
| "step": 1620 | |
| }, | |
| { | |
| "epoch": 0.630073444143796, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.949877593093674e-05, | |
| "loss": 0.0, | |
| "step": 1630 | |
| }, | |
| { | |
| "epoch": 0.6339389253962119, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.943435124339647e-05, | |
| "loss": 0.0, | |
| "step": 1640 | |
| }, | |
| { | |
| "epoch": 0.6378044066486278, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.9369926555856205e-05, | |
| "loss": 0.0, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 0.6416698879010437, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.930550186831594e-05, | |
| "loss": 0.0, | |
| "step": 1660 | |
| }, | |
| { | |
| "epoch": 0.6455353691534597, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.924107718077568e-05, | |
| "loss": 0.0, | |
| "step": 1670 | |
| }, | |
| { | |
| "epoch": 0.6494008504058756, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.917665249323541e-05, | |
| "loss": 0.0, | |
| "step": 1680 | |
| }, | |
| { | |
| "epoch": 0.6532663316582915, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.9112227805695146e-05, | |
| "loss": 0.0, | |
| "step": 1690 | |
| }, | |
| { | |
| "epoch": 0.6571318129107074, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.904780311815488e-05, | |
| "loss": 0.0, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.6609972941631234, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.898337843061461e-05, | |
| "loss": 0.0, | |
| "step": 1710 | |
| }, | |
| { | |
| "epoch": 0.6648627754155393, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.891895374307435e-05, | |
| "loss": 0.0, | |
| "step": 1720 | |
| }, | |
| { | |
| "epoch": 0.6687282566679552, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.885452905553408e-05, | |
| "loss": 0.0, | |
| "step": 1730 | |
| }, | |
| { | |
| "epoch": 0.6725937379203711, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.879010436799382e-05, | |
| "loss": 0.0, | |
| "step": 1740 | |
| }, | |
| { | |
| "epoch": 0.6764592191727871, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.8725679680453554e-05, | |
| "loss": 0.0, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 0.680324700425203, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.866125499291329e-05, | |
| "loss": 0.0, | |
| "step": 1760 | |
| }, | |
| { | |
| "epoch": 0.6841901816776189, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.859683030537302e-05, | |
| "loss": 0.0, | |
| "step": 1770 | |
| }, | |
| { | |
| "epoch": 0.6880556629300348, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.8532405617832755e-05, | |
| "loss": 0.0, | |
| "step": 1780 | |
| }, | |
| { | |
| "epoch": 0.6919211441824508, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.846798093029249e-05, | |
| "loss": 0.0, | |
| "step": 1790 | |
| }, | |
| { | |
| "epoch": 0.6957866254348667, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.840355624275223e-05, | |
| "loss": 0.0, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.6996521066872826, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.833913155521196e-05, | |
| "loss": 0.0, | |
| "step": 1810 | |
| }, | |
| { | |
| "epoch": 0.7035175879396985, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.8274706867671696e-05, | |
| "loss": 0.0, | |
| "step": 1820 | |
| }, | |
| { | |
| "epoch": 0.7073830691921145, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.821028218013143e-05, | |
| "loss": 0.0, | |
| "step": 1830 | |
| }, | |
| { | |
| "epoch": 0.7112485504445304, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.814585749259116e-05, | |
| "loss": 0.0, | |
| "step": 1840 | |
| }, | |
| { | |
| "epoch": 0.7151140316969463, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.8081432805050896e-05, | |
| "loss": 0.0, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 0.7189795129493622, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.801700811751063e-05, | |
| "loss": 0.0, | |
| "step": 1860 | |
| }, | |
| { | |
| "epoch": 0.7228449942017782, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.795258342997037e-05, | |
| "loss": 0.0, | |
| "step": 1870 | |
| }, | |
| { | |
| "epoch": 0.7267104754541941, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.7888158742430104e-05, | |
| "loss": 0.0, | |
| "step": 1880 | |
| }, | |
| { | |
| "epoch": 0.73057595670661, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.782373405488984e-05, | |
| "loss": 0.0, | |
| "step": 1890 | |
| }, | |
| { | |
| "epoch": 0.734441437959026, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.775930936734957e-05, | |
| "loss": 0.0, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 0.7383069192114419, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.7694884679809304e-05, | |
| "loss": 0.0, | |
| "step": 1910 | |
| }, | |
| { | |
| "epoch": 0.7421724004638578, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.763045999226904e-05, | |
| "loss": 0.0, | |
| "step": 1920 | |
| }, | |
| { | |
| "epoch": 0.7460378817162737, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.756603530472878e-05, | |
| "loss": 0.0, | |
| "step": 1930 | |
| }, | |
| { | |
| "epoch": 0.7499033629686896, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.750161061718851e-05, | |
| "loss": 0.0, | |
| "step": 1940 | |
| }, | |
| { | |
| "epoch": 0.7537688442211056, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.7437185929648245e-05, | |
| "loss": 0.0, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 0.7576343254735215, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.737276124210798e-05, | |
| "loss": 0.0, | |
| "step": 1960 | |
| }, | |
| { | |
| "epoch": 0.7614998067259374, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.730833655456771e-05, | |
| "loss": 0.0, | |
| "step": 1970 | |
| }, | |
| { | |
| "epoch": 0.7653652879783533, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.7243911867027446e-05, | |
| "loss": 0.0, | |
| "step": 1980 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.717948717948718e-05, | |
| "loss": 0.0, | |
| "step": 1990 | |
| }, | |
| { | |
| "epoch": 0.7730962504831852, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.711506249194692e-05, | |
| "loss": 0.0, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.7769617317356011, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.705063780440665e-05, | |
| "loss": 0.0, | |
| "step": 2010 | |
| }, | |
| { | |
| "epoch": 0.780827212988017, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.6986213116866387e-05, | |
| "loss": 0.0, | |
| "step": 2020 | |
| }, | |
| { | |
| "epoch": 0.784692694240433, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.692178842932612e-05, | |
| "loss": 0.0, | |
| "step": 2030 | |
| }, | |
| { | |
| "epoch": 0.7885581754928489, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.6857363741785854e-05, | |
| "loss": 0.0, | |
| "step": 2040 | |
| }, | |
| { | |
| "epoch": 0.7924236567452648, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.679293905424559e-05, | |
| "loss": 0.0, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 0.7962891379976808, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.672851436670533e-05, | |
| "loss": 0.0, | |
| "step": 2060 | |
| }, | |
| { | |
| "epoch": 0.8001546192500967, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.666408967916506e-05, | |
| "loss": 0.0, | |
| "step": 2070 | |
| }, | |
| { | |
| "epoch": 0.8040201005025126, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.6599664991624795e-05, | |
| "loss": 0.0, | |
| "step": 2080 | |
| }, | |
| { | |
| "epoch": 0.8078855817549285, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.653524030408453e-05, | |
| "loss": 0.0, | |
| "step": 2090 | |
| }, | |
| { | |
| "epoch": 0.8117510630073445, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.647081561654426e-05, | |
| "loss": 0.0, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 0.8156165442597604, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.6406390929003995e-05, | |
| "loss": 0.0, | |
| "step": 2110 | |
| }, | |
| { | |
| "epoch": 0.8194820255121763, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.634196624146373e-05, | |
| "loss": 0.0, | |
| "step": 2120 | |
| }, | |
| { | |
| "epoch": 0.8233475067645922, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.627754155392347e-05, | |
| "loss": 0.0, | |
| "step": 2130 | |
| }, | |
| { | |
| "epoch": 0.8272129880170082, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.62131168663832e-05, | |
| "loss": 0.0, | |
| "step": 2140 | |
| }, | |
| { | |
| "epoch": 0.8310784692694241, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.6148692178842936e-05, | |
| "loss": 0.0, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 0.83494395052184, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.608426749130267e-05, | |
| "loss": 0.0, | |
| "step": 2160 | |
| }, | |
| { | |
| "epoch": 0.8388094317742559, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.60198428037624e-05, | |
| "loss": 0.0, | |
| "step": 2170 | |
| }, | |
| { | |
| "epoch": 0.8426749130266719, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.595541811622214e-05, | |
| "loss": 0.0, | |
| "step": 2180 | |
| }, | |
| { | |
| "epoch": 0.8465403942790878, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.589099342868188e-05, | |
| "loss": 0.0, | |
| "step": 2190 | |
| }, | |
| { | |
| "epoch": 0.8504058755315037, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.582656874114161e-05, | |
| "loss": 0.0, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 0.8542713567839196, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.5762144053601344e-05, | |
| "loss": 0.0, | |
| "step": 2210 | |
| }, | |
| { | |
| "epoch": 0.8581368380363356, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.569771936606108e-05, | |
| "loss": 0.0, | |
| "step": 2220 | |
| }, | |
| { | |
| "epoch": 0.8620023192887515, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.563329467852081e-05, | |
| "loss": 0.0, | |
| "step": 2230 | |
| }, | |
| { | |
| "epoch": 0.8658678005411674, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.5568869990980545e-05, | |
| "loss": 0.0, | |
| "step": 2240 | |
| }, | |
| { | |
| "epoch": 0.8697332817935833, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.550444530344028e-05, | |
| "loss": 0.0, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 0.8735987630459993, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.544002061590002e-05, | |
| "loss": 0.0, | |
| "step": 2260 | |
| }, | |
| { | |
| "epoch": 0.8774642442984152, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.537559592835975e-05, | |
| "loss": 0.0, | |
| "step": 2270 | |
| }, | |
| { | |
| "epoch": 0.8813297255508311, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.5311171240819486e-05, | |
| "loss": 0.0, | |
| "step": 2280 | |
| }, | |
| { | |
| "epoch": 0.885195206803247, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.524674655327922e-05, | |
| "loss": 0.0, | |
| "step": 2290 | |
| }, | |
| { | |
| "epoch": 0.889060688055663, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.518232186573895e-05, | |
| "loss": 0.0, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 0.8929261693080789, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.5117897178198686e-05, | |
| "loss": 0.0, | |
| "step": 2310 | |
| }, | |
| { | |
| "epoch": 0.8967916505604948, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.5053472490658427e-05, | |
| "loss": 0.0, | |
| "step": 2320 | |
| }, | |
| { | |
| "epoch": 0.9006571318129107, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.498904780311816e-05, | |
| "loss": 0.0, | |
| "step": 2330 | |
| }, | |
| { | |
| "epoch": 0.9045226130653267, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.4924623115577894e-05, | |
| "loss": 0.0, | |
| "step": 2340 | |
| }, | |
| { | |
| "epoch": 0.9083880943177426, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.486019842803763e-05, | |
| "loss": 0.0, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 0.9122535755701585, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.479577374049736e-05, | |
| "loss": 0.0, | |
| "step": 2360 | |
| }, | |
| { | |
| "epoch": 0.9161190568225744, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.4731349052957094e-05, | |
| "loss": 0.0, | |
| "step": 2370 | |
| }, | |
| { | |
| "epoch": 0.9199845380749904, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.466692436541683e-05, | |
| "loss": 0.0, | |
| "step": 2380 | |
| }, | |
| { | |
| "epoch": 0.9238500193274063, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.460249967787657e-05, | |
| "loss": 0.0, | |
| "step": 2390 | |
| }, | |
| { | |
| "epoch": 0.9277155005798222, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.45380749903363e-05, | |
| "loss": 0.0, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 0.9315809818322381, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.4473650302796035e-05, | |
| "loss": 0.0, | |
| "step": 2410 | |
| }, | |
| { | |
| "epoch": 0.9354464630846541, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.440922561525577e-05, | |
| "loss": 0.0, | |
| "step": 2420 | |
| }, | |
| { | |
| "epoch": 0.93931194433707, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.43448009277155e-05, | |
| "loss": 0.0, | |
| "step": 2430 | |
| }, | |
| { | |
| "epoch": 0.9431774255894859, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.4280376240175236e-05, | |
| "loss": 0.0, | |
| "step": 2440 | |
| }, | |
| { | |
| "epoch": 0.9470429068419018, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.4215951552634976e-05, | |
| "loss": 0.0, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 0.9509083880943178, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.415152686509471e-05, | |
| "loss": 0.0, | |
| "step": 2460 | |
| }, | |
| { | |
| "epoch": 0.9547738693467337, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.408710217755444e-05, | |
| "loss": 0.0, | |
| "step": 2470 | |
| }, | |
| { | |
| "epoch": 0.9586393505991496, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.402267749001418e-05, | |
| "loss": 0.0, | |
| "step": 2480 | |
| }, | |
| { | |
| "epoch": 0.9625048318515655, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.395825280247391e-05, | |
| "loss": 0.0, | |
| "step": 2490 | |
| }, | |
| { | |
| "epoch": 0.9663703131039815, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.3893828114933644e-05, | |
| "loss": 0.0, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.9702357943563974, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.382940342739338e-05, | |
| "loss": 0.0, | |
| "step": 2510 | |
| }, | |
| { | |
| "epoch": 0.9741012756088133, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.376497873985312e-05, | |
| "loss": 0.0, | |
| "step": 2520 | |
| }, | |
| { | |
| "epoch": 0.9779667568612292, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.370055405231285e-05, | |
| "loss": 0.0, | |
| "step": 2530 | |
| }, | |
| { | |
| "epoch": 0.9818322381136452, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.3636129364772585e-05, | |
| "loss": 0.0, | |
| "step": 2540 | |
| }, | |
| { | |
| "epoch": 0.9856977193660611, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.357170467723232e-05, | |
| "loss": 0.0, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 0.989563200618477, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.350727998969205e-05, | |
| "loss": 0.0, | |
| "step": 2560 | |
| }, | |
| { | |
| "epoch": 0.993428681870893, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.3442855302151785e-05, | |
| "loss": 0.0, | |
| "step": 2570 | |
| }, | |
| { | |
| "epoch": 0.9972941631233089, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.3378430614611526e-05, | |
| "loss": 0.0, | |
| "step": 2580 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 7761, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 10, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |