| { | |
| "best_metric": 0.8517706990242004, | |
| "best_model_checkpoint": "./output/training_results/C017_random_sample_llama3-8b-base_instruct_20240504_182259/checkpoint-40", | |
| "epoch": 4.0, | |
| "eval_steps": 20, | |
| "global_step": 192, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.020833333333333332, | |
| "grad_norm": 0.0, | |
| "learning_rate": 0.0, | |
| "loss": 0.9877, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.10416666666666667, | |
| "grad_norm": 13.88322004479745, | |
| "learning_rate": 1.5e-06, | |
| "loss": 0.9606, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.20833333333333334, | |
| "grad_norm": 5.819857619989807, | |
| "learning_rate": 5.25e-06, | |
| "loss": 0.9201, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3125, | |
| "grad_norm": 5.213327490718728, | |
| "learning_rate": 9e-06, | |
| "loss": 0.8431, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.4166666666666667, | |
| "grad_norm": 4.886645368516583, | |
| "learning_rate": 1.275e-05, | |
| "loss": 0.8222, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.4166666666666667, | |
| "eval_loss": 0.8593380451202393, | |
| "eval_runtime": 1.9967, | |
| "eval_samples_per_second": 170.28, | |
| "eval_steps_per_second": 1.502, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.5208333333333334, | |
| "grad_norm": 4.291987293679674, | |
| "learning_rate": 1.3195176200175283e-05, | |
| "loss": 0.8289, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.625, | |
| "grad_norm": 4.031254003708488, | |
| "learning_rate": 9.515676612044427e-06, | |
| "loss": 0.8983, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.7291666666666666, | |
| "grad_norm": 3.73543629054737, | |
| "learning_rate": 6.797580677308734e-06, | |
| "loss": 0.8188, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.8333333333333334, | |
| "grad_norm": 4.212558198275198, | |
| "learning_rate": 4.808575415542887e-06, | |
| "loss": 0.8014, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.8333333333333334, | |
| "eval_loss": 0.8517706990242004, | |
| "eval_runtime": 1.9645, | |
| "eval_samples_per_second": 173.07, | |
| "eval_steps_per_second": 1.527, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.9375, | |
| "grad_norm": 3.963009491354867, | |
| "learning_rate": 3.3676619069852654e-06, | |
| "loss": 0.8782, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 1.0416666666666667, | |
| "grad_norm": 3.785430208940067, | |
| "learning_rate": 2.334947896124909e-06, | |
| "loss": 0.7259, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.1458333333333333, | |
| "grad_norm": 2.9958758287554184, | |
| "learning_rate": 1.603233215095547e-06, | |
| "loss": 0.4621, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 3.5920078433753297, | |
| "learning_rate": 1.0911174606561334e-06, | |
| "loss": 0.4422, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "eval_loss": 0.8722255825996399, | |
| "eval_runtime": 1.9622, | |
| "eval_samples_per_second": 173.278, | |
| "eval_steps_per_second": 1.529, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.3541666666666667, | |
| "grad_norm": 3.9286205240212144, | |
| "learning_rate": 7.373930741131784e-07, | |
| "loss": 0.4117, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 1.4583333333333333, | |
| "grad_norm": 4.373192522719682, | |
| "learning_rate": 4.965174334325768e-07, | |
| "loss": 0.4422, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.5625, | |
| "grad_norm": 5.66550951351165, | |
| "learning_rate": 3.349849877937343e-07, | |
| "loss": 0.4432, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 1.6666666666666665, | |
| "grad_norm": 3.7845006550086153, | |
| "learning_rate": 2.2844505627726646e-07, | |
| "loss": 0.4551, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.6666666666666665, | |
| "eval_loss": 0.8555447459220886, | |
| "eval_runtime": 1.9656, | |
| "eval_samples_per_second": 172.972, | |
| "eval_steps_per_second": 1.526, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.7708333333333335, | |
| "grad_norm": 3.572304742292393, | |
| "learning_rate": 1.594328760942437e-07, | |
| "loss": 0.4222, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 1.875, | |
| "grad_norm": 3.488459009987249, | |
| "learning_rate": 1.156010161291434e-07, | |
| "loss": 0.4309, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.9791666666666665, | |
| "grad_norm": 3.900536398698168, | |
| "learning_rate": 8.835555547373544e-08, | |
| "loss": 0.4793, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 2.0833333333333335, | |
| "grad_norm": 3.413641170228889, | |
| "learning_rate": 7.181664349277562e-08, | |
| "loss": 0.3806, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.0833333333333335, | |
| "eval_loss": 0.8529651165008545, | |
| "eval_runtime": 1.9631, | |
| "eval_samples_per_second": 173.199, | |
| "eval_steps_per_second": 1.528, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.1875, | |
| "grad_norm": 3.3820950755420904, | |
| "learning_rate": 6.203637972657601e-08, | |
| "loss": 0.3745, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 2.2916666666666665, | |
| "grad_norm": 3.247185633907683, | |
| "learning_rate": 5.6418543066491835e-08, | |
| "loss": 0.409, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 2.3958333333333335, | |
| "grad_norm": 3.3611611000807198, | |
| "learning_rate": 5.329471712759216e-08, | |
| "loss": 0.3605, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 4.406858841387693, | |
| "learning_rate": 5.161995210302015e-08, | |
| "loss": 0.4011, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "eval_loss": 0.8576686978340149, | |
| "eval_runtime": 1.9674, | |
| "eval_samples_per_second": 172.821, | |
| "eval_steps_per_second": 1.525, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.6041666666666665, | |
| "grad_norm": 3.4544518361326575, | |
| "learning_rate": 5.075841465580837e-08, | |
| "loss": 0.3872, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 2.7083333333333335, | |
| "grad_norm": 3.733645772230528, | |
| "learning_rate": 5.033564114946932e-08, | |
| "loss": 0.3853, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.8125, | |
| "grad_norm": 3.40895522907687, | |
| "learning_rate": 5.013915282607116e-08, | |
| "loss": 0.3832, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 2.9166666666666665, | |
| "grad_norm": 3.2871773154531923, | |
| "learning_rate": 5.005343402153039e-08, | |
| "loss": 0.37, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.9166666666666665, | |
| "eval_loss": 0.8621854782104492, | |
| "eval_runtime": 1.9654, | |
| "eval_samples_per_second": 172.991, | |
| "eval_steps_per_second": 1.526, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 3.0208333333333335, | |
| "grad_norm": 3.105947874400041, | |
| "learning_rate": 5.001872829857116e-08, | |
| "loss": 0.3799, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 3.125, | |
| "grad_norm": 3.27679756840579, | |
| "learning_rate": 5.000587713853837e-08, | |
| "loss": 0.3821, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 3.2291666666666665, | |
| "grad_norm": 3.2917595668052524, | |
| "learning_rate": 5.0001608748597456e-08, | |
| "loss": 0.3799, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 3.3333333333333335, | |
| "grad_norm": 4.375036973754038, | |
| "learning_rate": 5.0000370319656156e-08, | |
| "loss": 0.3626, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 3.3333333333333335, | |
| "eval_loss": 0.8658773303031921, | |
| "eval_runtime": 1.9678, | |
| "eval_samples_per_second": 172.779, | |
| "eval_steps_per_second": 1.525, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 3.4375, | |
| "grad_norm": 3.315514407544336, | |
| "learning_rate": 5.0000067945715855e-08, | |
| "loss": 0.3646, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 3.5416666666666665, | |
| "grad_norm": 3.562684382929414, | |
| "learning_rate": 5.0000009144677036e-08, | |
| "loss": 0.347, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 3.6458333333333335, | |
| "grad_norm": 3.3664300561373466, | |
| "learning_rate": 5.0000000785521776e-08, | |
| "loss": 0.3734, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 3.75, | |
| "grad_norm": 3.499144173589757, | |
| "learning_rate": 5.000000003317662e-08, | |
| "loss": 0.3708, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 3.75, | |
| "eval_loss": 0.8686714768409729, | |
| "eval_runtime": 1.965, | |
| "eval_samples_per_second": 173.025, | |
| "eval_steps_per_second": 1.527, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 3.8541666666666665, | |
| "grad_norm": 3.7255205262291726, | |
| "learning_rate": 5.000000000038355e-08, | |
| "loss": 0.3746, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 3.9583333333333335, | |
| "grad_norm": 3.1641205572409237, | |
| "learning_rate": 5.000000000000018e-08, | |
| "loss": 0.3734, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 192, | |
| "total_flos": 5360344104960.0, | |
| "train_loss": 0.5152150879924496, | |
| "train_runtime": 1052.2097, | |
| "train_samples_per_second": 11.61, | |
| "train_steps_per_second": 0.182 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 192, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 20, | |
| "total_flos": 5360344104960.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |