| { | |
| "best_metric": 0.5248246192932129, | |
| "best_model_checkpoint": "./vit-beta2-0.99/checkpoint-5778", | |
| "epoch": 28.0, | |
| "eval_steps": 500, | |
| "global_step": 8988, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 27.36143684387207, | |
| "learning_rate": 1.8291979226774382e-05, | |
| "loss": 1.7217, | |
| "step": 321 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7035367545076283, | |
| "eval_f1": 0.6527006206875058, | |
| "eval_loss": 1.0190271139144897, | |
| "eval_precision": 0.6635371340134015, | |
| "eval_recall": 0.7035367545076283, | |
| "eval_runtime": 22.7678, | |
| "eval_samples_per_second": 126.67, | |
| "eval_steps_per_second": 15.856, | |
| "step": 321 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 32.75101089477539, | |
| "learning_rate": 3.681477207155222e-05, | |
| "loss": 1.1622, | |
| "step": 642 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.705617198335645, | |
| "eval_f1": 0.7191659456414141, | |
| "eval_loss": 0.7385604977607727, | |
| "eval_precision": 0.7575809849286596, | |
| "eval_recall": 0.705617198335645, | |
| "eval_runtime": 23.0256, | |
| "eval_samples_per_second": 125.252, | |
| "eval_steps_per_second": 15.678, | |
| "step": 642 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 7.780852317810059, | |
| "learning_rate": 5.5337564916330066e-05, | |
| "loss": 1.0368, | |
| "step": 963 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7517337031900139, | |
| "eval_f1": 0.7214294760102591, | |
| "eval_loss": 0.655035138130188, | |
| "eval_precision": 0.7563510144231254, | |
| "eval_recall": 0.7517337031900139, | |
| "eval_runtime": 22.8211, | |
| "eval_samples_per_second": 126.374, | |
| "eval_steps_per_second": 15.819, | |
| "step": 963 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 30.053781509399414, | |
| "learning_rate": 7.386035776110792e-05, | |
| "loss": 0.9653, | |
| "step": 1284 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.7843273231622746, | |
| "eval_f1": 0.7863050211732793, | |
| "eval_loss": 0.5640625953674316, | |
| "eval_precision": 0.7948170600442105, | |
| "eval_recall": 0.7843273231622746, | |
| "eval_runtime": 22.4916, | |
| "eval_samples_per_second": 128.226, | |
| "eval_steps_per_second": 16.05, | |
| "step": 1284 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 10.72180461883545, | |
| "learning_rate": 9.238315060588575e-05, | |
| "loss": 0.9272, | |
| "step": 1605 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.6768377253814147, | |
| "eval_f1": 0.703514416314774, | |
| "eval_loss": 0.7956904172897339, | |
| "eval_precision": 0.7959127541379581, | |
| "eval_recall": 0.6768377253814147, | |
| "eval_runtime": 22.4157, | |
| "eval_samples_per_second": 128.66, | |
| "eval_steps_per_second": 16.105, | |
| "step": 1605 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 9.665838241577148, | |
| "learning_rate": 9.984715255878176e-05, | |
| "loss": 0.9878, | |
| "step": 1926 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.7871012482662968, | |
| "eval_f1": 0.7903728309886872, | |
| "eval_loss": 0.580937922000885, | |
| "eval_precision": 0.8061861183645468, | |
| "eval_recall": 0.7871012482662968, | |
| "eval_runtime": 22.8489, | |
| "eval_samples_per_second": 126.221, | |
| "eval_steps_per_second": 15.799, | |
| "step": 1926 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 7.413944244384766, | |
| "learning_rate": 9.889061131437471e-05, | |
| "loss": 0.872, | |
| "step": 2247 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.7215672676837726, | |
| "eval_f1": 0.7441866963098906, | |
| "eval_loss": 0.681545078754425, | |
| "eval_precision": 0.8081398492651465, | |
| "eval_recall": 0.7215672676837726, | |
| "eval_runtime": 22.5223, | |
| "eval_samples_per_second": 128.051, | |
| "eval_steps_per_second": 16.029, | |
| "step": 2247 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 3.1963610649108887, | |
| "learning_rate": 9.707265436104638e-05, | |
| "loss": 0.7998, | |
| "step": 2568 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.7558945908460472, | |
| "eval_f1": 0.7722745826002211, | |
| "eval_loss": 0.6104105114936829, | |
| "eval_precision": 0.8143413614427701, | |
| "eval_recall": 0.7558945908460472, | |
| "eval_runtime": 22.1997, | |
| "eval_samples_per_second": 129.912, | |
| "eval_steps_per_second": 16.261, | |
| "step": 2568 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 5.708278656005859, | |
| "learning_rate": 9.44253127296151e-05, | |
| "loss": 0.733, | |
| "step": 2889 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.8148404993065187, | |
| "eval_f1": 0.817195928987888, | |
| "eval_loss": 0.5296399593353271, | |
| "eval_precision": 0.8254366092776788, | |
| "eval_recall": 0.8148404993065187, | |
| "eval_runtime": 22.2658, | |
| "eval_samples_per_second": 129.526, | |
| "eval_steps_per_second": 16.213, | |
| "step": 2889 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 8.459081649780273, | |
| "learning_rate": 9.099523058358976e-05, | |
| "loss": 0.6957, | |
| "step": 3210 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.7995839112343966, | |
| "eval_f1": 0.8052016823215509, | |
| "eval_loss": 0.579708456993103, | |
| "eval_precision": 0.8322303167903897, | |
| "eval_recall": 0.7995839112343966, | |
| "eval_runtime": 22.874, | |
| "eval_samples_per_second": 126.082, | |
| "eval_steps_per_second": 15.782, | |
| "step": 3210 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "grad_norm": 23.16131591796875, | |
| "learning_rate": 8.684284338417735e-05, | |
| "loss": 0.6271, | |
| "step": 3531 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_accuracy": 0.7933425797503467, | |
| "eval_f1": 0.8057886771155975, | |
| "eval_loss": 0.5925618410110474, | |
| "eval_precision": 0.8342721256554735, | |
| "eval_recall": 0.7933425797503467, | |
| "eval_runtime": 22.4086, | |
| "eval_samples_per_second": 128.7, | |
| "eval_steps_per_second": 16.11, | |
| "step": 3531 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 13.86120319366455, | |
| "learning_rate": 8.204131306302357e-05, | |
| "loss": 0.5614, | |
| "step": 3852 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_accuracy": 0.7919556171983356, | |
| "eval_f1": 0.8060480658501534, | |
| "eval_loss": 0.58785080909729, | |
| "eval_precision": 0.8384030380760077, | |
| "eval_recall": 0.7919556171983356, | |
| "eval_runtime": 22.7711, | |
| "eval_samples_per_second": 126.652, | |
| "eval_steps_per_second": 15.853, | |
| "step": 3852 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "grad_norm": 14.25642204284668, | |
| "learning_rate": 7.667523896413962e-05, | |
| "loss": 0.4576, | |
| "step": 4173 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_accuracy": 0.8138002773925104, | |
| "eval_f1": 0.802756356097153, | |
| "eval_loss": 0.6664562225341797, | |
| "eval_precision": 0.8312195418319966, | |
| "eval_recall": 0.8138002773925104, | |
| "eval_runtime": 22.7626, | |
| "eval_samples_per_second": 126.699, | |
| "eval_steps_per_second": 15.859, | |
| "step": 4173 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "grad_norm": 2.4907939434051514, | |
| "learning_rate": 7.083916726724684e-05, | |
| "loss": 0.4645, | |
| "step": 4494 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_accuracy": 0.8294036061026352, | |
| "eval_f1": 0.8329373166180268, | |
| "eval_loss": 0.5514557957649231, | |
| "eval_precision": 0.8470000675264909, | |
| "eval_recall": 0.8294036061026352, | |
| "eval_runtime": 22.6903, | |
| "eval_samples_per_second": 127.103, | |
| "eval_steps_per_second": 15.91, | |
| "step": 4494 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "grad_norm": 6.062983512878418, | |
| "learning_rate": 6.463592515537568e-05, | |
| "loss": 0.3913, | |
| "step": 4815 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_accuracy": 0.8224687933425797, | |
| "eval_f1": 0.8288240804348163, | |
| "eval_loss": 0.5473943948745728, | |
| "eval_precision": 0.8465516026472886, | |
| "eval_recall": 0.8224687933425797, | |
| "eval_runtime": 22.9153, | |
| "eval_samples_per_second": 125.855, | |
| "eval_steps_per_second": 15.754, | |
| "step": 4815 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 8.832720756530762, | |
| "learning_rate": 5.8174809077430184e-05, | |
| "loss": 0.3693, | |
| "step": 5136 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_accuracy": 0.823509015256588, | |
| "eval_f1": 0.8308490392951902, | |
| "eval_loss": 0.5769267082214355, | |
| "eval_precision": 0.8463977879369676, | |
| "eval_recall": 0.823509015256588, | |
| "eval_runtime": 22.4364, | |
| "eval_samples_per_second": 128.541, | |
| "eval_steps_per_second": 16.09, | |
| "step": 5136 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "grad_norm": 3.554053783416748, | |
| "learning_rate": 5.156965902716534e-05, | |
| "loss": 0.2794, | |
| "step": 5457 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_accuracy": 0.8509015256588072, | |
| "eval_f1": 0.851625286103631, | |
| "eval_loss": 0.5327965617179871, | |
| "eval_precision": 0.8571269483137886, | |
| "eval_recall": 0.8509015256588072, | |
| "eval_runtime": 22.5058, | |
| "eval_samples_per_second": 128.145, | |
| "eval_steps_per_second": 16.04, | |
| "step": 5457 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "grad_norm": 12.536526679992676, | |
| "learning_rate": 4.493685276832998e-05, | |
| "loss": 0.2677, | |
| "step": 5778 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_accuracy": 0.8623439667128987, | |
| "eval_f1": 0.8596257980493167, | |
| "eval_loss": 0.5248246192932129, | |
| "eval_precision": 0.8583558881059454, | |
| "eval_recall": 0.8623439667128987, | |
| "eval_runtime": 23.0138, | |
| "eval_samples_per_second": 125.316, | |
| "eval_steps_per_second": 15.686, | |
| "step": 5778 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "grad_norm": 0.8805328011512756, | |
| "learning_rate": 3.839325534621579e-05, | |
| "loss": 0.2104, | |
| "step": 6099 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_accuracy": 0.8432732316227461, | |
| "eval_f1": 0.8473236873027821, | |
| "eval_loss": 0.6283748745918274, | |
| "eval_precision": 0.8572165883311328, | |
| "eval_recall": 0.8432732316227461, | |
| "eval_runtime": 22.0808, | |
| "eval_samples_per_second": 130.611, | |
| "eval_steps_per_second": 16.349, | |
| "step": 6099 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "grad_norm": 0.04168611392378807, | |
| "learning_rate": 3.2073473743477955e-05, | |
| "loss": 0.2459, | |
| "step": 6420 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_accuracy": 0.8543689320388349, | |
| "eval_f1": 0.8554941368452752, | |
| "eval_loss": 0.6137195229530334, | |
| "eval_precision": 0.8595513589850039, | |
| "eval_recall": 0.8543689320388349, | |
| "eval_runtime": 22.5224, | |
| "eval_samples_per_second": 128.05, | |
| "eval_steps_per_second": 16.028, | |
| "step": 6420 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "grad_norm": 3.362337350845337, | |
| "learning_rate": 2.604941738980618e-05, | |
| "loss": 0.1769, | |
| "step": 6741 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "eval_accuracy": 0.8637309292649098, | |
| "eval_f1": 0.8566147037629545, | |
| "eval_loss": 0.5959635376930237, | |
| "eval_precision": 0.8572769111277326, | |
| "eval_recall": 0.8637309292649098, | |
| "eval_runtime": 22.6796, | |
| "eval_samples_per_second": 127.162, | |
| "eval_steps_per_second": 15.917, | |
| "step": 6741 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "grad_norm": 32.62522506713867, | |
| "learning_rate": 2.0447352243517255e-05, | |
| "loss": 0.1294, | |
| "step": 7062 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_accuracy": 0.8699722607489597, | |
| "eval_f1": 0.868659902198456, | |
| "eval_loss": 0.5843542814254761, | |
| "eval_precision": 0.8687454531579298, | |
| "eval_recall": 0.8699722607489597, | |
| "eval_runtime": 22.7408, | |
| "eval_samples_per_second": 126.82, | |
| "eval_steps_per_second": 15.875, | |
| "step": 7062 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "grad_norm": 4.362905502319336, | |
| "learning_rate": 1.536598246865575e-05, | |
| "loss": 0.1597, | |
| "step": 7383 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "eval_accuracy": 0.866504854368932, | |
| "eval_f1": 0.8589306843328246, | |
| "eval_loss": 0.6580309867858887, | |
| "eval_precision": 0.8603887819342586, | |
| "eval_recall": 0.866504854368932, | |
| "eval_runtime": 22.8065, | |
| "eval_samples_per_second": 126.455, | |
| "eval_steps_per_second": 15.829, | |
| "step": 7383 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "grad_norm": 1.3405486345291138, | |
| "learning_rate": 1.0894837969414489e-05, | |
| "loss": 0.1227, | |
| "step": 7704 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_accuracy": 0.8730929264909847, | |
| "eval_f1": 0.8712364081669894, | |
| "eval_loss": 0.6225888729095459, | |
| "eval_precision": 0.8720005451749013, | |
| "eval_recall": 0.8730929264909847, | |
| "eval_runtime": 22.4336, | |
| "eval_samples_per_second": 128.557, | |
| "eval_steps_per_second": 16.092, | |
| "step": 7704 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "grad_norm": 0.08982106298208237, | |
| "learning_rate": 7.112696940726155e-06, | |
| "loss": 0.1054, | |
| "step": 8025 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "eval_accuracy": 0.8751733703190014, | |
| "eval_f1": 0.8720900906443607, | |
| "eval_loss": 0.6197877526283264, | |
| "eval_precision": 0.8727615417504975, | |
| "eval_recall": 0.8751733703190014, | |
| "eval_runtime": 22.3517, | |
| "eval_samples_per_second": 129.028, | |
| "eval_steps_per_second": 16.151, | |
| "step": 8025 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "grad_norm": 16.338762283325195, | |
| "learning_rate": 4.094394131694684e-06, | |
| "loss": 0.0945, | |
| "step": 8346 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "eval_accuracy": 0.8793342579750347, | |
| "eval_f1": 0.8764130000059128, | |
| "eval_loss": 0.6049804091453552, | |
| "eval_precision": 0.8757375120219936, | |
| "eval_recall": 0.8793342579750347, | |
| "eval_runtime": 22.9615, | |
| "eval_samples_per_second": 125.601, | |
| "eval_steps_per_second": 15.722, | |
| "step": 8346 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "grad_norm": 0.07402696460485458, | |
| "learning_rate": 1.8742732027931087e-06, | |
| "loss": 0.1242, | |
| "step": 8667 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "eval_accuracy": 0.8828016643550625, | |
| "eval_f1": 0.8797538930980352, | |
| "eval_loss": 0.6077802181243896, | |
| "eval_precision": 0.8788497875866979, | |
| "eval_recall": 0.8828016643550625, | |
| "eval_runtime": 22.9003, | |
| "eval_samples_per_second": 125.937, | |
| "eval_steps_per_second": 15.764, | |
| "step": 8667 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "grad_norm": 0.06073630228638649, | |
| "learning_rate": 5.020912943263345e-07, | |
| "loss": 0.0819, | |
| "step": 8988 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "eval_accuracy": 0.8796809986130375, | |
| "eval_f1": 0.8755693874564652, | |
| "eval_loss": 0.6189650893211365, | |
| "eval_precision": 0.8748178599277232, | |
| "eval_recall": 0.8796809986130375, | |
| "eval_runtime": 23.1664, | |
| "eval_samples_per_second": 124.49, | |
| "eval_steps_per_second": 15.583, | |
| "step": 8988 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "step": 8988, | |
| "total_flos": 1.1127108458244538e+19, | |
| "train_loss": 0.5275238134301817, | |
| "train_runtime": 3329.2548, | |
| "train_samples_per_second": 154.028, | |
| "train_steps_per_second": 9.642 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 32100, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 100, | |
| "save_steps": 500, | |
| "total_flos": 1.1127108458244538e+19, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |