| { | |
| "best_metric": 0.781834372217275, | |
| "best_model_checkpoint": "../saved_model/cino-small-v2_tusa/checkpoint-500", | |
| "epoch": 40.0, | |
| "eval_steps": 500, | |
| "global_step": 10000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.759, | |
| "eval_f1": 0.7449735449735448, | |
| "eval_loss": 0.538531482219696, | |
| "eval_precision": 0.7910112359550562, | |
| "eval_recall": 0.704, | |
| "eval_runtime": 3.843, | |
| "eval_samples_per_second": 260.21, | |
| "eval_steps_per_second": 8.327, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 2.5e-05, | |
| "loss": 0.4062, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.755, | |
| "eval_f1": 0.781834372217275, | |
| "eval_loss": 0.5724849700927734, | |
| "eval_precision": 0.7046548956661316, | |
| "eval_recall": 0.878, | |
| "eval_runtime": 3.8652, | |
| "eval_samples_per_second": 258.722, | |
| "eval_steps_per_second": 8.279, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.763, | |
| "eval_f1": 0.7718960538979788, | |
| "eval_loss": 0.6156216263771057, | |
| "eval_precision": 0.7439703153988868, | |
| "eval_recall": 0.802, | |
| "eval_runtime": 3.8674, | |
| "eval_samples_per_second": 258.57, | |
| "eval_steps_per_second": 8.274, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 5e-05, | |
| "loss": 0.1767, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.763, | |
| "eval_f1": 0.7351955307262571, | |
| "eval_loss": 0.7199212908744812, | |
| "eval_precision": 0.8329113924050633, | |
| "eval_recall": 0.658, | |
| "eval_runtime": 3.8486, | |
| "eval_samples_per_second": 259.833, | |
| "eval_steps_per_second": 8.315, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.72, | |
| "eval_f1": 0.7188755020080321, | |
| "eval_loss": 1.1636953353881836, | |
| "eval_precision": 0.7217741935483871, | |
| "eval_recall": 0.716, | |
| "eval_runtime": 3.8679, | |
| "eval_samples_per_second": 258.536, | |
| "eval_steps_per_second": 8.273, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "learning_rate": 4.722222222222222e-05, | |
| "loss": 0.0851, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.668, | |
| "eval_f1": 0.7167235494880547, | |
| "eval_loss": 1.6249794960021973, | |
| "eval_precision": 0.625, | |
| "eval_recall": 0.84, | |
| "eval_runtime": 3.8509, | |
| "eval_samples_per_second": 259.682, | |
| "eval_steps_per_second": 8.31, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.66, | |
| "eval_f1": 0.6903460837887068, | |
| "eval_loss": 1.97330904006958, | |
| "eval_precision": 0.6337792642140468, | |
| "eval_recall": 0.758, | |
| "eval_runtime": 3.8611, | |
| "eval_samples_per_second": 258.997, | |
| "eval_steps_per_second": 8.288, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "learning_rate": 4.4444444444444447e-05, | |
| "loss": 0.0374, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.672, | |
| "eval_f1": 0.6371681415929205, | |
| "eval_loss": 2.0722925662994385, | |
| "eval_precision": 0.7128712871287128, | |
| "eval_recall": 0.576, | |
| "eval_runtime": 3.8504, | |
| "eval_samples_per_second": 259.712, | |
| "eval_steps_per_second": 8.311, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.712, | |
| "eval_f1": 0.7037037037037037, | |
| "eval_loss": 2.0254833698272705, | |
| "eval_precision": 0.7245762711864406, | |
| "eval_recall": 0.684, | |
| "eval_runtime": 3.8436, | |
| "eval_samples_per_second": 260.171, | |
| "eval_steps_per_second": 8.325, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "learning_rate": 4.166666666666667e-05, | |
| "loss": 0.0196, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.72, | |
| "eval_f1": 0.7421731123388583, | |
| "eval_loss": 2.1418700218200684, | |
| "eval_precision": 0.6877133105802048, | |
| "eval_recall": 0.806, | |
| "eval_runtime": 3.8516, | |
| "eval_samples_per_second": 259.63, | |
| "eval_steps_per_second": 8.308, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_accuracy": 0.732, | |
| "eval_f1": 0.7709401709401711, | |
| "eval_loss": 2.037372350692749, | |
| "eval_precision": 0.673134328358209, | |
| "eval_recall": 0.902, | |
| "eval_runtime": 3.8641, | |
| "eval_samples_per_second": 258.791, | |
| "eval_steps_per_second": 8.281, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "learning_rate": 3.888888888888889e-05, | |
| "loss": 0.0178, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_accuracy": 0.734, | |
| "eval_f1": 0.7387033398821218, | |
| "eval_loss": 2.172621488571167, | |
| "eval_precision": 0.7258687258687259, | |
| "eval_recall": 0.752, | |
| "eval_runtime": 3.853, | |
| "eval_samples_per_second": 259.541, | |
| "eval_steps_per_second": 8.305, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_accuracy": 0.694, | |
| "eval_f1": 0.66, | |
| "eval_loss": 2.3637471199035645, | |
| "eval_precision": 0.7425, | |
| "eval_recall": 0.594, | |
| "eval_runtime": 3.8612, | |
| "eval_samples_per_second": 258.984, | |
| "eval_steps_per_second": 8.288, | |
| "step": 3250 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "learning_rate": 3.611111111111111e-05, | |
| "loss": 0.0166, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_accuracy": 0.679, | |
| "eval_f1": 0.664576802507837, | |
| "eval_loss": 2.3070590496063232, | |
| "eval_precision": 0.6958424507658644, | |
| "eval_recall": 0.636, | |
| "eval_runtime": 3.8463, | |
| "eval_samples_per_second": 259.991, | |
| "eval_steps_per_second": 8.32, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_accuracy": 0.72, | |
| "eval_f1": 0.7265625, | |
| "eval_loss": 2.2964775562286377, | |
| "eval_precision": 0.7099236641221374, | |
| "eval_recall": 0.744, | |
| "eval_runtime": 3.8729, | |
| "eval_samples_per_second": 258.207, | |
| "eval_steps_per_second": 8.263, | |
| "step": 3750 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "learning_rate": 3.3333333333333335e-05, | |
| "loss": 0.0074, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_accuracy": 0.708, | |
| "eval_f1": 0.7176015473887815, | |
| "eval_loss": 2.449232339859009, | |
| "eval_precision": 0.6947565543071161, | |
| "eval_recall": 0.742, | |
| "eval_runtime": 3.861, | |
| "eval_samples_per_second": 258.998, | |
| "eval_steps_per_second": 8.288, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_accuracy": 0.726, | |
| "eval_f1": 0.727634194831014, | |
| "eval_loss": 2.2294154167175293, | |
| "eval_precision": 0.7233201581027668, | |
| "eval_recall": 0.732, | |
| "eval_runtime": 3.8559, | |
| "eval_samples_per_second": 259.344, | |
| "eval_steps_per_second": 8.299, | |
| "step": 4250 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "learning_rate": 3.055555555555556e-05, | |
| "loss": 0.0127, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_accuracy": 0.736, | |
| "eval_f1": 0.7333333333333334, | |
| "eval_loss": 2.354280471801758, | |
| "eval_precision": 0.7408163265306122, | |
| "eval_recall": 0.726, | |
| "eval_runtime": 3.8511, | |
| "eval_samples_per_second": 259.668, | |
| "eval_steps_per_second": 8.309, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_accuracy": 0.72, | |
| "eval_f1": 0.688195991091314, | |
| "eval_loss": 2.646545886993408, | |
| "eval_precision": 0.7763819095477387, | |
| "eval_recall": 0.618, | |
| "eval_runtime": 3.8559, | |
| "eval_samples_per_second": 259.341, | |
| "eval_steps_per_second": 8.299, | |
| "step": 4750 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "learning_rate": 2.777777777777778e-05, | |
| "loss": 0.0071, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_accuracy": 0.728, | |
| "eval_f1": 0.7190082644628099, | |
| "eval_loss": 2.392427682876587, | |
| "eval_precision": 0.7435897435897436, | |
| "eval_recall": 0.696, | |
| "eval_runtime": 3.841, | |
| "eval_samples_per_second": 260.348, | |
| "eval_steps_per_second": 8.331, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "eval_accuracy": 0.712, | |
| "eval_f1": 0.7181996086105676, | |
| "eval_loss": 2.35839581489563, | |
| "eval_precision": 0.7030651340996169, | |
| "eval_recall": 0.734, | |
| "eval_runtime": 3.8576, | |
| "eval_samples_per_second": 259.231, | |
| "eval_steps_per_second": 8.295, | |
| "step": 5250 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "learning_rate": 2.5e-05, | |
| "loss": 0.0085, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_accuracy": 0.739, | |
| "eval_f1": 0.7116022099447514, | |
| "eval_loss": 1.9896154403686523, | |
| "eval_precision": 0.7950617283950617, | |
| "eval_recall": 0.644, | |
| "eval_runtime": 3.8513, | |
| "eval_samples_per_second": 259.655, | |
| "eval_steps_per_second": 8.309, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "eval_accuracy": 0.751, | |
| "eval_f1": 0.7278688524590164, | |
| "eval_loss": 2.147411584854126, | |
| "eval_precision": 0.8024096385542169, | |
| "eval_recall": 0.666, | |
| "eval_runtime": 3.9041, | |
| "eval_samples_per_second": 256.144, | |
| "eval_steps_per_second": 8.197, | |
| "step": 5750 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "learning_rate": 2.2222222222222223e-05, | |
| "loss": 0.0069, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_accuracy": 0.75, | |
| "eval_f1": 0.7306034482758621, | |
| "eval_loss": 2.172492265701294, | |
| "eval_precision": 0.7920560747663551, | |
| "eval_recall": 0.678, | |
| "eval_runtime": 3.8555, | |
| "eval_samples_per_second": 259.368, | |
| "eval_steps_per_second": 8.3, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "eval_accuracy": 0.732, | |
| "eval_f1": 0.7112068965517241, | |
| "eval_loss": 2.3147952556610107, | |
| "eval_precision": 0.7710280373831776, | |
| "eval_recall": 0.66, | |
| "eval_runtime": 3.8522, | |
| "eval_samples_per_second": 259.59, | |
| "eval_steps_per_second": 8.307, | |
| "step": 6250 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "learning_rate": 1.9444444444444445e-05, | |
| "loss": 0.0067, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "eval_accuracy": 0.726, | |
| "eval_f1": 0.7385496183206107, | |
| "eval_loss": 2.2775561809539795, | |
| "eval_precision": 0.7062043795620438, | |
| "eval_recall": 0.774, | |
| "eval_runtime": 3.8474, | |
| "eval_samples_per_second": 259.914, | |
| "eval_steps_per_second": 8.317, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "eval_accuracy": 0.718, | |
| "eval_f1": 0.699360341151386, | |
| "eval_loss": 2.492180109024048, | |
| "eval_precision": 0.7488584474885844, | |
| "eval_recall": 0.656, | |
| "eval_runtime": 3.8591, | |
| "eval_samples_per_second": 259.131, | |
| "eval_steps_per_second": 8.292, | |
| "step": 6750 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "learning_rate": 1.6666666666666667e-05, | |
| "loss": 0.0032, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "eval_accuracy": 0.726, | |
| "eval_f1": 0.7133891213389122, | |
| "eval_loss": 2.4056484699249268, | |
| "eval_precision": 0.7478070175438597, | |
| "eval_recall": 0.682, | |
| "eval_runtime": 3.8454, | |
| "eval_samples_per_second": 260.05, | |
| "eval_steps_per_second": 8.322, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 29.0, | |
| "eval_accuracy": 0.724, | |
| "eval_f1": 0.7177914110429447, | |
| "eval_loss": 2.5566771030426025, | |
| "eval_precision": 0.7343096234309623, | |
| "eval_recall": 0.702, | |
| "eval_runtime": 3.8506, | |
| "eval_samples_per_second": 259.7, | |
| "eval_steps_per_second": 8.31, | |
| "step": 7250 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "learning_rate": 1.388888888888889e-05, | |
| "loss": 0.001, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "eval_accuracy": 0.733, | |
| "eval_f1": 0.7516279069767442, | |
| "eval_loss": 2.586094617843628, | |
| "eval_precision": 0.7026086956521739, | |
| "eval_recall": 0.808, | |
| "eval_runtime": 3.8404, | |
| "eval_samples_per_second": 260.391, | |
| "eval_steps_per_second": 8.333, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 31.0, | |
| "eval_accuracy": 0.738, | |
| "eval_f1": 0.7282157676348547, | |
| "eval_loss": 2.566093683242798, | |
| "eval_precision": 0.7564655172413793, | |
| "eval_recall": 0.702, | |
| "eval_runtime": 3.8517, | |
| "eval_samples_per_second": 259.629, | |
| "eval_steps_per_second": 8.308, | |
| "step": 7750 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "learning_rate": 1.1111111111111112e-05, | |
| "loss": 0.0016, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "eval_accuracy": 0.74, | |
| "eval_f1": 0.7192224622030238, | |
| "eval_loss": 2.6535303592681885, | |
| "eval_precision": 0.7816901408450704, | |
| "eval_recall": 0.666, | |
| "eval_runtime": 3.9723, | |
| "eval_samples_per_second": 251.74, | |
| "eval_steps_per_second": 8.056, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 33.0, | |
| "eval_accuracy": 0.74, | |
| "eval_f1": 0.717391304347826, | |
| "eval_loss": 2.7205991744995117, | |
| "eval_precision": 0.7857142857142857, | |
| "eval_recall": 0.66, | |
| "eval_runtime": 3.9432, | |
| "eval_samples_per_second": 253.602, | |
| "eval_steps_per_second": 8.115, | |
| "step": 8250 | |
| }, | |
| { | |
| "epoch": 34.0, | |
| "learning_rate": 8.333333333333334e-06, | |
| "loss": 0.0014, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 34.0, | |
| "eval_accuracy": 0.725, | |
| "eval_f1": 0.7422680412371134, | |
| "eval_loss": 2.6660046577453613, | |
| "eval_precision": 0.6984126984126984, | |
| "eval_recall": 0.792, | |
| "eval_runtime": 3.9996, | |
| "eval_samples_per_second": 250.027, | |
| "eval_steps_per_second": 8.001, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 35.0, | |
| "eval_accuracy": 0.742, | |
| "eval_f1": 0.7361963190184049, | |
| "eval_loss": 2.5981457233428955, | |
| "eval_precision": 0.7531380753138075, | |
| "eval_recall": 0.72, | |
| "eval_runtime": 3.9964, | |
| "eval_samples_per_second": 250.226, | |
| "eval_steps_per_second": 8.007, | |
| "step": 8750 | |
| }, | |
| { | |
| "epoch": 36.0, | |
| "learning_rate": 5.555555555555556e-06, | |
| "loss": 0.0003, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 36.0, | |
| "eval_accuracy": 0.745, | |
| "eval_f1": 0.7384615384615385, | |
| "eval_loss": 2.6297309398651123, | |
| "eval_precision": 0.7578947368421053, | |
| "eval_recall": 0.72, | |
| "eval_runtime": 3.9747, | |
| "eval_samples_per_second": 251.591, | |
| "eval_steps_per_second": 8.051, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 37.0, | |
| "eval_accuracy": 0.738, | |
| "eval_f1": 0.7358870967741935, | |
| "eval_loss": 2.6337215900421143, | |
| "eval_precision": 0.741869918699187, | |
| "eval_recall": 0.73, | |
| "eval_runtime": 3.9752, | |
| "eval_samples_per_second": 251.558, | |
| "eval_steps_per_second": 8.05, | |
| "step": 9250 | |
| }, | |
| { | |
| "epoch": 38.0, | |
| "learning_rate": 2.777777777777778e-06, | |
| "loss": 0.0007, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 38.0, | |
| "eval_accuracy": 0.739, | |
| "eval_f1": 0.7300930713547054, | |
| "eval_loss": 2.645658493041992, | |
| "eval_precision": 0.7558886509635975, | |
| "eval_recall": 0.706, | |
| "eval_runtime": 3.9919, | |
| "eval_samples_per_second": 250.508, | |
| "eval_steps_per_second": 8.016, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 39.0, | |
| "eval_accuracy": 0.741, | |
| "eval_f1": 0.7304890738813736, | |
| "eval_loss": 2.6524713039398193, | |
| "eval_precision": 0.7613882863340564, | |
| "eval_recall": 0.702, | |
| "eval_runtime": 3.9627, | |
| "eval_samples_per_second": 252.356, | |
| "eval_steps_per_second": 8.075, | |
| "step": 9750 | |
| }, | |
| { | |
| "epoch": 40.0, | |
| "learning_rate": 0.0, | |
| "loss": 0.0, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 40.0, | |
| "eval_accuracy": 0.742, | |
| "eval_f1": 0.734020618556701, | |
| "eval_loss": 2.6484949588775635, | |
| "eval_precision": 0.7574468085106383, | |
| "eval_recall": 0.712, | |
| "eval_runtime": 3.9449, | |
| "eval_samples_per_second": 253.495, | |
| "eval_steps_per_second": 8.112, | |
| "step": 10000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 10000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 40, | |
| "save_steps": 500, | |
| "total_flos": 4.238956756992e+16, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |