| { | |
| "best_metric": 0.34451597929000854, | |
| "best_model_checkpoint": "model_output/e2e_opentable_5_way__19684-shot__seed-42__gpt2/checkpoint-550", | |
| "epoch": 3.571428571428571, | |
| "global_step": 550, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.9783362218370885e-05, | |
| "loss": 6.8575, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.956672443674177e-05, | |
| "loss": 6.173, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "eval_distillation_accuracy_counterfactual": 0.43894304771677783, | |
| "eval_distillation_accuracy_factual": 0.8099025141097999, | |
| "eval_distillation_f1_counterfactual": 0.41910630617913947, | |
| "eval_distillation_f1_factual": 0.7867273237866255, | |
| "eval_groundtruth_accuracy_counterfactual": 0.39148281169830684, | |
| "eval_groundtruth_f1_counterfactual": 0.37160171109948154, | |
| "eval_groundtruth_f1_factual": 0.6498734087641715, | |
| "eval_icace_cosine": 0.5216067433357239, | |
| "eval_icace_l2": 0.6332615613937378, | |
| "eval_icace_normdiff": 0.4962432086467743, | |
| "eval_loss": 6.244416236877441, | |
| "eval_runtime": 6.3453, | |
| "eval_samples_per_second": 614.313, | |
| "eval_steps_per_second": 4.886, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 4.935008665511265e-05, | |
| "loss": 6.0611, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 4.913344887348354e-05, | |
| "loss": 5.6355, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 4.891681109185442e-05, | |
| "loss": 5.4416, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "eval_distillation_accuracy_counterfactual": 0.4910210364289379, | |
| "eval_distillation_accuracy_factual": 0.8419702411493073, | |
| "eval_distillation_f1_counterfactual": 0.41441222306682873, | |
| "eval_distillation_f1_factual": 0.8223199536105706, | |
| "eval_groundtruth_accuracy_counterfactual": 0.4486916367367881, | |
| "eval_groundtruth_f1_counterfactual": 0.37593058398927565, | |
| "eval_groundtruth_f1_factual": 0.6500439526437162, | |
| "eval_icace_cosine": 0.48116251826286316, | |
| "eval_icace_l2": 0.5798823237419128, | |
| "eval_icace_normdiff": 0.46706676483154297, | |
| "eval_loss": 5.620999336242676, | |
| "eval_runtime": 6.0381, | |
| "eval_samples_per_second": 645.568, | |
| "eval_steps_per_second": 5.134, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 4.8700173310225307e-05, | |
| "loss": 5.3368, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 4.848353552859619e-05, | |
| "loss": 5.21, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "eval_distillation_accuracy_counterfactual": 0.5736275012827091, | |
| "eval_distillation_accuracy_factual": 0.8542842483324782, | |
| "eval_distillation_f1_counterfactual": 0.5108131897133708, | |
| "eval_distillation_f1_factual": 0.8349821571428248, | |
| "eval_groundtruth_accuracy_counterfactual": 0.5374551051821447, | |
| "eval_groundtruth_f1_counterfactual": 0.47925180036361503, | |
| "eval_groundtruth_f1_factual": 0.6521939459011195, | |
| "eval_icace_cosine": 0.429977685213089, | |
| "eval_icace_l2": 0.5049712061882019, | |
| "eval_icace_normdiff": 0.37107908725738525, | |
| "eval_loss": 4.268450736999512, | |
| "eval_runtime": 9.0497, | |
| "eval_samples_per_second": 430.732, | |
| "eval_steps_per_second": 3.426, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "learning_rate": 4.826689774696707e-05, | |
| "loss": 4.6547, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "learning_rate": 4.8050259965337955e-05, | |
| "loss": 4.0985, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "learning_rate": 4.7833622183708845e-05, | |
| "loss": 3.9061, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "eval_distillation_accuracy_counterfactual": 0.5938943047716778, | |
| "eval_distillation_accuracy_factual": 0.8627501282709081, | |
| "eval_distillation_f1_counterfactual": 0.5569296055565827, | |
| "eval_distillation_f1_factual": 0.8462828359256841, | |
| "eval_groundtruth_accuracy_counterfactual": 0.5677270395074397, | |
| "eval_groundtruth_f1_counterfactual": 0.5296349124253801, | |
| "eval_groundtruth_f1_factual": 0.6586176111497359, | |
| "eval_icace_cosine": 0.39845770597457886, | |
| "eval_icace_l2": 0.4688405990600586, | |
| "eval_icace_normdiff": 0.28705334663391113, | |
| "eval_loss": 3.954227924346924, | |
| "eval_runtime": 6.0652, | |
| "eval_samples_per_second": 642.681, | |
| "eval_steps_per_second": 5.111, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "learning_rate": 4.761698440207972e-05, | |
| "loss": 3.7083, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "learning_rate": 4.740034662045061e-05, | |
| "loss": 3.5838, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "eval_distillation_accuracy_counterfactual": 0.6244227809132888, | |
| "eval_distillation_accuracy_factual": 0.8471010774756286, | |
| "eval_distillation_f1_counterfactual": 0.5612464478394067, | |
| "eval_distillation_f1_factual": 0.8168933248114726, | |
| "eval_groundtruth_accuracy_counterfactual": 0.581836839404823, | |
| "eval_groundtruth_f1_counterfactual": 0.5188128402781068, | |
| "eval_groundtruth_f1_factual": 0.6448810803991499, | |
| "eval_icace_cosine": 0.385066956281662, | |
| "eval_icace_l2": 0.4372788965702057, | |
| "eval_icace_normdiff": 0.2642575800418854, | |
| "eval_loss": 3.669494867324829, | |
| "eval_runtime": 6.0794, | |
| "eval_samples_per_second": 641.177, | |
| "eval_steps_per_second": 5.099, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "learning_rate": 4.7183708838821494e-05, | |
| "loss": 3.4503, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 4.6967071057192376e-05, | |
| "loss": 3.2902, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "learning_rate": 4.675043327556326e-05, | |
| "loss": 3.1492, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "eval_distillation_accuracy_counterfactual": 0.642380708055413, | |
| "eval_distillation_accuracy_factual": 0.8673678809645973, | |
| "eval_distillation_f1_counterfactual": 0.60503663340424, | |
| "eval_distillation_f1_factual": 0.8479942560980248, | |
| "eval_groundtruth_accuracy_counterfactual": 0.5872242175474602, | |
| "eval_groundtruth_f1_counterfactual": 0.5475602868329816, | |
| "eval_groundtruth_f1_factual": 0.6552421371317594, | |
| "eval_icace_cosine": 0.381485253572464, | |
| "eval_icace_l2": 0.4177369773387909, | |
| "eval_icace_normdiff": 0.2561808228492737, | |
| "eval_loss": 3.5663232803344727, | |
| "eval_runtime": 8.5489, | |
| "eval_samples_per_second": 455.967, | |
| "eval_steps_per_second": 3.626, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 4.653379549393415e-05, | |
| "loss": 3.01, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "learning_rate": 4.6317157712305025e-05, | |
| "loss": 2.966, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.27, | |
| "eval_distillation_accuracy_counterfactual": 0.6436634171369934, | |
| "eval_distillation_accuracy_factual": 0.8550538737814264, | |
| "eval_distillation_f1_counterfactual": 0.606514121797564, | |
| "eval_distillation_f1_factual": 0.8415474619556556, | |
| "eval_groundtruth_accuracy_counterfactual": 0.5995382247306311, | |
| "eval_groundtruth_f1_counterfactual": 0.5608025209119987, | |
| "eval_groundtruth_f1_factual": 0.6735103649630443, | |
| "eval_icace_cosine": 0.37723952531814575, | |
| "eval_icace_l2": 0.41544613242149353, | |
| "eval_icace_normdiff": 0.2528684437274933, | |
| "eval_loss": 3.5517373085021973, | |
| "eval_runtime": 6.2733, | |
| "eval_samples_per_second": 621.367, | |
| "eval_steps_per_second": 4.942, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 4.6100519930675915e-05, | |
| "loss": 3.0867, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "learning_rate": 4.58838821490468e-05, | |
| "loss": 3.006, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "learning_rate": 4.566724436741768e-05, | |
| "loss": 2.9638, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "eval_distillation_accuracy_counterfactual": 0.6575166752180606, | |
| "eval_distillation_accuracy_factual": 0.8527449974345819, | |
| "eval_distillation_f1_counterfactual": 0.613799464505768, | |
| "eval_distillation_f1_factual": 0.8339444902298532, | |
| "eval_groundtruth_accuracy_counterfactual": 0.6023601847101078, | |
| "eval_groundtruth_f1_counterfactual": 0.553976724654705, | |
| "eval_groundtruth_f1_factual": 0.6607569496254541, | |
| "eval_icace_cosine": 0.3669813871383667, | |
| "eval_icace_l2": 0.4078792929649353, | |
| "eval_icace_normdiff": 0.24327021837234497, | |
| "eval_loss": 3.4745211601257324, | |
| "eval_runtime": 6.2941, | |
| "eval_samples_per_second": 619.306, | |
| "eval_steps_per_second": 4.925, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.73, | |
| "learning_rate": 4.5450606585788563e-05, | |
| "loss": 2.8851, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "learning_rate": 4.5233968804159446e-05, | |
| "loss": 3.0178, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 2.92, | |
| "eval_distillation_accuracy_counterfactual": 0.6372498717290919, | |
| "eval_distillation_accuracy_factual": 0.8527449974345819, | |
| "eval_distillation_f1_counterfactual": 0.588239631522627, | |
| "eval_distillation_f1_factual": 0.8336195132025683, | |
| "eval_groundtruth_accuracy_counterfactual": 0.6067213955874807, | |
| "eval_groundtruth_f1_counterfactual": 0.5579137249617195, | |
| "eval_groundtruth_f1_factual": 0.6659281894600338, | |
| "eval_icace_cosine": 0.36822977662086487, | |
| "eval_icace_l2": 0.41515976190567017, | |
| "eval_icace_normdiff": 0.2492736279964447, | |
| "eval_loss": 3.339986562728882, | |
| "eval_runtime": 8.6524, | |
| "eval_samples_per_second": 450.51, | |
| "eval_steps_per_second": 3.583, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "learning_rate": 4.501733102253033e-05, | |
| "loss": 2.8144, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 3.12, | |
| "learning_rate": 4.480069324090121e-05, | |
| "loss": 2.6693, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 3.25, | |
| "learning_rate": 4.45840554592721e-05, | |
| "loss": 2.6417, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.25, | |
| "eval_distillation_accuracy_counterfactual": 0.6503335043612108, | |
| "eval_distillation_accuracy_factual": 0.8619805028219599, | |
| "eval_distillation_f1_counterfactual": 0.6114857659963661, | |
| "eval_distillation_f1_factual": 0.8435060729033399, | |
| "eval_groundtruth_accuracy_counterfactual": 0.6062083119548486, | |
| "eval_groundtruth_f1_counterfactual": 0.5663665924537191, | |
| "eval_groundtruth_f1_factual": 0.6765095780155879, | |
| "eval_icace_cosine": 0.37013715505599976, | |
| "eval_icace_l2": 0.40843912959098816, | |
| "eval_icace_normdiff": 0.23886314034461975, | |
| "eval_loss": 3.3956122398376465, | |
| "eval_runtime": 6.3102, | |
| "eval_samples_per_second": 617.733, | |
| "eval_steps_per_second": 4.913, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.38, | |
| "learning_rate": 4.436741767764298e-05, | |
| "loss": 2.5693, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 3.51, | |
| "learning_rate": 4.415077989601387e-05, | |
| "loss": 2.6584, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 3.57, | |
| "eval_distillation_accuracy_counterfactual": 0.6477680861980503, | |
| "eval_distillation_accuracy_factual": 0.8527449974345819, | |
| "eval_distillation_f1_counterfactual": 0.6050503103766632, | |
| "eval_distillation_f1_factual": 0.8352243795490729, | |
| "eval_groundtruth_accuracy_counterfactual": 0.5974858902001026, | |
| "eval_groundtruth_f1_counterfactual": 0.5539329237527222, | |
| "eval_groundtruth_f1_factual": 0.6650957003091867, | |
| "eval_icace_cosine": 0.34451597929000854, | |
| "eval_icace_l2": 0.4020812511444092, | |
| "eval_icace_normdiff": 0.2356698215007782, | |
| "eval_loss": 3.3548502922058105, | |
| "eval_runtime": 6.1488, | |
| "eval_samples_per_second": 633.947, | |
| "eval_steps_per_second": 5.042, | |
| "step": 550 | |
| } | |
| ], | |
| "max_steps": 4616, | |
| "num_train_epochs": 30, | |
| "total_flos": 4625353826374656.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |