| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 34.99853587115666, | |
| "eval_steps": 500, | |
| "global_step": 11952, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.29282576866764276, | |
| "grad_norm": 0.2411503791809082, | |
| "learning_rate": 0.0001, | |
| "loss": 2.0009, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.5856515373352855, | |
| "grad_norm": 0.22286373376846313, | |
| "learning_rate": 0.0001, | |
| "loss": 1.9538, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.8784773060029283, | |
| "grad_norm": 0.24079065024852753, | |
| "learning_rate": 0.0001, | |
| "loss": 1.9571, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9985358711566618, | |
| "eval_accuracy": 0.45379487179487177, | |
| "eval_loss": 3.9512319564819336, | |
| "eval_runtime": 5.0647, | |
| "eval_samples_per_second": 98.722, | |
| "eval_steps_per_second": 12.439, | |
| "step": 341 | |
| }, | |
| { | |
| "epoch": 0.9985358711566618, | |
| "eval_bleu": 0.28767194245703276, | |
| "eval_exact_match": 0.03, | |
| "eval_prefix_exact_match": 0.042, | |
| "step": 341 | |
| }, | |
| { | |
| "epoch": 1.171303074670571, | |
| "grad_norm": 0.33546799421310425, | |
| "learning_rate": 0.0001, | |
| "loss": 1.9255, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.4641288433382138, | |
| "grad_norm": 0.3515666723251343, | |
| "learning_rate": 0.0001, | |
| "loss": 1.8684, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.7569546120058566, | |
| "grad_norm": 0.35733896493911743, | |
| "learning_rate": 0.0001, | |
| "loss": 1.8819, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.4482564102564103, | |
| "eval_loss": 4.112804412841797, | |
| "eval_runtime": 5.0848, | |
| "eval_samples_per_second": 98.333, | |
| "eval_steps_per_second": 12.39, | |
| "step": 683 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_bleu": 0.24521038067633488, | |
| "eval_exact_match": 0.02, | |
| "eval_prefix_exact_match": 0.044, | |
| "step": 683 | |
| }, | |
| { | |
| "epoch": 2.049780380673499, | |
| "grad_norm": 0.3784458339214325, | |
| "learning_rate": 0.0001, | |
| "loss": 1.8613, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.342606149341142, | |
| "grad_norm": 0.46702849864959717, | |
| "learning_rate": 0.0001, | |
| "loss": 1.743, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.6354319180087846, | |
| "grad_norm": 0.5156407356262207, | |
| "learning_rate": 0.0001, | |
| "loss": 1.7653, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.9282576866764276, | |
| "grad_norm": 0.4755038619041443, | |
| "learning_rate": 0.0001, | |
| "loss": 1.7702, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.998535871156662, | |
| "eval_accuracy": 0.446051282051282, | |
| "eval_loss": 4.327652454376221, | |
| "eval_runtime": 5.1641, | |
| "eval_samples_per_second": 96.822, | |
| "eval_steps_per_second": 12.2, | |
| "step": 1024 | |
| }, | |
| { | |
| "epoch": 2.998535871156662, | |
| "eval_bleu": 0.2598440457703149, | |
| "eval_exact_match": 0.018, | |
| "eval_prefix_exact_match": 0.034, | |
| "step": 1024 | |
| }, | |
| { | |
| "epoch": 3.22108345534407, | |
| "grad_norm": 0.5288150906562805, | |
| "learning_rate": 0.0001, | |
| "loss": 1.6325, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 3.513909224011713, | |
| "grad_norm": 0.5784010887145996, | |
| "learning_rate": 0.0001, | |
| "loss": 1.5896, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.8067349926793557, | |
| "grad_norm": 0.6458429098129272, | |
| "learning_rate": 0.0001, | |
| "loss": 1.6163, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.44235897435897437, | |
| "eval_loss": 4.584905624389648, | |
| "eval_runtime": 4.9243, | |
| "eval_samples_per_second": 101.538, | |
| "eval_steps_per_second": 12.794, | |
| "step": 1366 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_bleu": 0.27967222841565076, | |
| "eval_exact_match": 0.01, | |
| "eval_prefix_exact_match": 0.038, | |
| "step": 1366 | |
| }, | |
| { | |
| "epoch": 4.099560761346998, | |
| "grad_norm": 0.732841432094574, | |
| "learning_rate": 0.0001, | |
| "loss": 1.5441, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 4.392386530014641, | |
| "grad_norm": 0.7372816205024719, | |
| "learning_rate": 0.0001, | |
| "loss": 1.4142, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 4.685212298682284, | |
| "grad_norm": 0.8235049247741699, | |
| "learning_rate": 0.0001, | |
| "loss": 1.4198, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.978038067349927, | |
| "grad_norm": 0.7765743732452393, | |
| "learning_rate": 0.0001, | |
| "loss": 1.4427, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 4.998535871156662, | |
| "eval_accuracy": 0.43861538461538463, | |
| "eval_loss": 4.850330829620361, | |
| "eval_runtime": 4.9395, | |
| "eval_samples_per_second": 101.224, | |
| "eval_steps_per_second": 12.754, | |
| "step": 1707 | |
| }, | |
| { | |
| "epoch": 4.998535871156662, | |
| "eval_bleu": 0.2170008640509292, | |
| "eval_exact_match": 0.006, | |
| "eval_prefix_exact_match": 0.032, | |
| "step": 1707 | |
| }, | |
| { | |
| "epoch": 5.270863836017569, | |
| "grad_norm": 0.9119400978088379, | |
| "learning_rate": 0.0001, | |
| "loss": 1.2469, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 5.563689604685212, | |
| "grad_norm": 0.8966185450553894, | |
| "learning_rate": 0.0001, | |
| "loss": 1.2294, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 5.856515373352855, | |
| "grad_norm": 0.9629669189453125, | |
| "learning_rate": 0.0001, | |
| "loss": 1.2498, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.4348717948717949, | |
| "eval_loss": 5.09260892868042, | |
| "eval_runtime": 4.9572, | |
| "eval_samples_per_second": 100.863, | |
| "eval_steps_per_second": 12.709, | |
| "step": 2049 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_bleu": 0.20196479624851718, | |
| "eval_exact_match": 0.002, | |
| "eval_prefix_exact_match": 0.03, | |
| "step": 2049 | |
| }, | |
| { | |
| "epoch": 6.149341142020498, | |
| "grad_norm": 1.0808535814285278, | |
| "learning_rate": 0.0001, | |
| "loss": 1.1238, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 6.44216691068814, | |
| "grad_norm": 1.1639196872711182, | |
| "learning_rate": 0.0001, | |
| "loss": 1.0301, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 6.734992679355783, | |
| "grad_norm": 1.0662291049957275, | |
| "learning_rate": 0.0001, | |
| "loss": 1.0655, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 6.998535871156662, | |
| "eval_accuracy": 0.4326153846153846, | |
| "eval_loss": 5.270845890045166, | |
| "eval_runtime": 5.089, | |
| "eval_samples_per_second": 98.251, | |
| "eval_steps_per_second": 12.38, | |
| "step": 2390 | |
| }, | |
| { | |
| "epoch": 6.998535871156662, | |
| "eval_bleu": 0.17780932376070374, | |
| "eval_exact_match": 0.004, | |
| "eval_prefix_exact_match": 0.024, | |
| "step": 2390 | |
| }, | |
| { | |
| "epoch": 7.027818448023426, | |
| "grad_norm": 1.6462695598602295, | |
| "learning_rate": 0.0001, | |
| "loss": 1.0358, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 7.320644216691068, | |
| "grad_norm": 1.2542779445648193, | |
| "learning_rate": 0.0001, | |
| "loss": 0.8413, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 7.613469985358711, | |
| "grad_norm": 1.334762454032898, | |
| "learning_rate": 0.0001, | |
| "loss": 0.8654, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 7.906295754026354, | |
| "grad_norm": 1.2401787042617798, | |
| "learning_rate": 0.0001, | |
| "loss": 0.8733, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.4316923076923077, | |
| "eval_loss": 5.402440547943115, | |
| "eval_runtime": 4.9383, | |
| "eval_samples_per_second": 101.25, | |
| "eval_steps_per_second": 12.758, | |
| "step": 2732 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_bleu": 0.21418356891762502, | |
| "eval_exact_match": 0.002, | |
| "eval_prefix_exact_match": 0.026, | |
| "step": 2732 | |
| }, | |
| { | |
| "epoch": 8.199121522693996, | |
| "grad_norm": 1.255996584892273, | |
| "learning_rate": 0.0001, | |
| "loss": 0.7302, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 8.49194729136164, | |
| "grad_norm": 1.2836525440216064, | |
| "learning_rate": 0.0001, | |
| "loss": 0.6875, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 8.784773060029282, | |
| "grad_norm": 1.375715732574463, | |
| "learning_rate": 0.0001, | |
| "loss": 0.7219, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 8.998535871156662, | |
| "eval_accuracy": 0.42938461538461536, | |
| "eval_loss": 5.534832000732422, | |
| "eval_runtime": 5.2363, | |
| "eval_samples_per_second": 95.487, | |
| "eval_steps_per_second": 12.031, | |
| "step": 3073 | |
| }, | |
| { | |
| "epoch": 8.998535871156662, | |
| "eval_bleu": 0.20459866130314297, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.024, | |
| "step": 3073 | |
| }, | |
| { | |
| "epoch": 9.077598828696924, | |
| "grad_norm": 1.481998324394226, | |
| "learning_rate": 0.0001, | |
| "loss": 0.6722, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 9.370424597364568, | |
| "grad_norm": 1.4086748361587524, | |
| "learning_rate": 0.0001, | |
| "loss": 0.5489, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 9.66325036603221, | |
| "grad_norm": 1.4381046295166016, | |
| "learning_rate": 0.0001, | |
| "loss": 0.565, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 9.956076134699854, | |
| "grad_norm": 1.7364269495010376, | |
| "learning_rate": 0.0001, | |
| "loss": 0.5932, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.42605128205128207, | |
| "eval_loss": 5.768980503082275, | |
| "eval_runtime": 5.0821, | |
| "eval_samples_per_second": 98.384, | |
| "eval_steps_per_second": 12.396, | |
| "step": 3415 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_bleu": 0.19515580420261575, | |
| "eval_exact_match": 0.002, | |
| "eval_prefix_exact_match": 0.026, | |
| "step": 3415 | |
| }, | |
| { | |
| "epoch": 10.248901903367496, | |
| "grad_norm": 1.599753737449646, | |
| "learning_rate": 0.0001, | |
| "loss": 0.4581, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 10.541727672035138, | |
| "grad_norm": 1.5595277547836304, | |
| "learning_rate": 0.0001, | |
| "loss": 0.4519, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 10.834553440702782, | |
| "grad_norm": 1.6201436519622803, | |
| "learning_rate": 0.0001, | |
| "loss": 0.4719, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 10.998535871156662, | |
| "eval_accuracy": 0.42538461538461536, | |
| "eval_loss": 5.89434289932251, | |
| "eval_runtime": 5.0717, | |
| "eval_samples_per_second": 98.586, | |
| "eval_steps_per_second": 12.422, | |
| "step": 3756 | |
| }, | |
| { | |
| "epoch": 10.998535871156662, | |
| "eval_bleu": 0.2074466521823303, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.02, | |
| "step": 3756 | |
| }, | |
| { | |
| "epoch": 11.127379209370424, | |
| "grad_norm": 1.6209241151809692, | |
| "learning_rate": 0.0001, | |
| "loss": 0.4263, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 11.420204978038067, | |
| "grad_norm": 1.5148394107818604, | |
| "learning_rate": 0.0001, | |
| "loss": 0.3625, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 11.71303074670571, | |
| "grad_norm": 1.6862399578094482, | |
| "learning_rate": 0.0001, | |
| "loss": 0.3838, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_accuracy": 0.4247179487179487, | |
| "eval_loss": 6.019149303436279, | |
| "eval_runtime": 5.3028, | |
| "eval_samples_per_second": 94.291, | |
| "eval_steps_per_second": 11.881, | |
| "step": 4098 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_bleu": 0.16189430124724263, | |
| "eval_exact_match": 0.002, | |
| "eval_prefix_exact_match": 0.028, | |
| "step": 4098 | |
| }, | |
| { | |
| "epoch": 12.005856515373353, | |
| "grad_norm": 1.3296407461166382, | |
| "learning_rate": 0.0001, | |
| "loss": 0.3952, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 12.298682284040996, | |
| "grad_norm": 1.3439353704452515, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2901, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 12.591508052708638, | |
| "grad_norm": 1.5329118967056274, | |
| "learning_rate": 0.0001, | |
| "loss": 0.3114, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 12.88433382137628, | |
| "grad_norm": 1.649775743484497, | |
| "learning_rate": 0.0001, | |
| "loss": 0.329, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 12.998535871156662, | |
| "eval_accuracy": 0.42456410256410254, | |
| "eval_loss": 6.104373455047607, | |
| "eval_runtime": 5.9944, | |
| "eval_samples_per_second": 83.411, | |
| "eval_steps_per_second": 10.51, | |
| "step": 4439 | |
| }, | |
| { | |
| "epoch": 12.998535871156662, | |
| "eval_bleu": 0.17170793542335236, | |
| "eval_exact_match": 0.002, | |
| "eval_prefix_exact_match": 0.028, | |
| "step": 4439 | |
| }, | |
| { | |
| "epoch": 13.177159590043924, | |
| "grad_norm": 1.5262281894683838, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2758, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 13.469985358711567, | |
| "grad_norm": 1.5962131023406982, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2594, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 13.762811127379209, | |
| "grad_norm": 1.4258382320404053, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2742, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_accuracy": 0.4216410256410256, | |
| "eval_loss": 6.146524906158447, | |
| "eval_runtime": 4.9576, | |
| "eval_samples_per_second": 100.855, | |
| "eval_steps_per_second": 12.708, | |
| "step": 4781 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_bleu": 0.20004000747911055, | |
| "eval_exact_match": 0.002, | |
| "eval_prefix_exact_match": 0.026, | |
| "step": 4781 | |
| }, | |
| { | |
| "epoch": 14.055636896046853, | |
| "grad_norm": 1.388717532157898, | |
| "learning_rate": 0.0001, | |
| "loss": 0.269, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 14.348462664714495, | |
| "grad_norm": 1.2370668649673462, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2226, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 14.641288433382137, | |
| "grad_norm": 1.3778610229492188, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2334, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 14.93411420204978, | |
| "grad_norm": 1.6371277570724487, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2432, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 14.998535871156662, | |
| "eval_accuracy": 0.4226666666666667, | |
| "eval_loss": 6.3254475593566895, | |
| "eval_runtime": 5.2196, | |
| "eval_samples_per_second": 95.794, | |
| "eval_steps_per_second": 12.07, | |
| "step": 5122 | |
| }, | |
| { | |
| "epoch": 14.998535871156662, | |
| "eval_bleu": 0.16083749853030369, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.024, | |
| "step": 5122 | |
| }, | |
| { | |
| "epoch": 15.226939970717423, | |
| "grad_norm": 1.2761411666870117, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2066, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 15.519765739385067, | |
| "grad_norm": 1.6878156661987305, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2044, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 15.812591508052709, | |
| "grad_norm": 1.4964863061904907, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2158, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_accuracy": 0.4227692307692308, | |
| "eval_loss": 6.441042423248291, | |
| "eval_runtime": 4.9987, | |
| "eval_samples_per_second": 100.025, | |
| "eval_steps_per_second": 12.603, | |
| "step": 5464 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_bleu": 0.1349628182313855, | |
| "eval_exact_match": 0.002, | |
| "eval_prefix_exact_match": 0.026, | |
| "step": 5464 | |
| }, | |
| { | |
| "epoch": 16.105417276720353, | |
| "grad_norm": 0.9965784549713135, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2038, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 16.398243045387993, | |
| "grad_norm": 1.2694511413574219, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1857, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 16.691068814055637, | |
| "grad_norm": 1.436286211013794, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1944, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 16.98389458272328, | |
| "grad_norm": 1.526081919670105, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2013, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 16.998535871156662, | |
| "eval_accuracy": 0.4214871794871795, | |
| "eval_loss": 6.39243745803833, | |
| "eval_runtime": 5.1033, | |
| "eval_samples_per_second": 97.975, | |
| "eval_steps_per_second": 12.345, | |
| "step": 5805 | |
| }, | |
| { | |
| "epoch": 16.998535871156662, | |
| "eval_bleu": 0.18366001556911046, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.018, | |
| "step": 5805 | |
| }, | |
| { | |
| "epoch": 17.27672035139092, | |
| "grad_norm": 1.1292760372161865, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1694, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 17.569546120058565, | |
| "grad_norm": 1.440766453742981, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1782, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 17.86237188872621, | |
| "grad_norm": 1.2640533447265625, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1851, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_accuracy": 0.42005128205128206, | |
| "eval_loss": 6.5217366218566895, | |
| "eval_runtime": 5.0655, | |
| "eval_samples_per_second": 98.707, | |
| "eval_steps_per_second": 12.437, | |
| "step": 6147 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_bleu": 0.14616910828479238, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.026, | |
| "step": 6147 | |
| }, | |
| { | |
| "epoch": 18.15519765739385, | |
| "grad_norm": 1.430617094039917, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1717, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 18.448023426061493, | |
| "grad_norm": 1.1885381937026978, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1656, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 18.740849194729137, | |
| "grad_norm": 1.64545738697052, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1721, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 18.998535871156662, | |
| "eval_accuracy": 0.4208717948717949, | |
| "eval_loss": 6.557307720184326, | |
| "eval_runtime": 5.092, | |
| "eval_samples_per_second": 98.193, | |
| "eval_steps_per_second": 12.372, | |
| "step": 6488 | |
| }, | |
| { | |
| "epoch": 18.998535871156662, | |
| "eval_bleu": 0.16916503510298256, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.02, | |
| "step": 6488 | |
| }, | |
| { | |
| "epoch": 19.03367496339678, | |
| "grad_norm": 1.2168452739715576, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1743, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 19.32650073206442, | |
| "grad_norm": 1.2320917844772339, | |
| "learning_rate": 0.0001, | |
| "loss": 0.152, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 19.619326500732065, | |
| "grad_norm": 1.1541388034820557, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1619, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 19.91215226939971, | |
| "grad_norm": 1.2821346521377563, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1676, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_accuracy": 0.42138461538461536, | |
| "eval_loss": 6.566097736358643, | |
| "eval_runtime": 5.0056, | |
| "eval_samples_per_second": 99.888, | |
| "eval_steps_per_second": 12.586, | |
| "step": 6830 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_bleu": 0.21213643830050385, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.026, | |
| "step": 6830 | |
| }, | |
| { | |
| "epoch": 20.20497803806735, | |
| "grad_norm": 1.1950445175170898, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1533, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 20.497803806734993, | |
| "grad_norm": 1.253625750541687, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1522, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 20.790629575402637, | |
| "grad_norm": 1.193354606628418, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1579, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 20.998535871156662, | |
| "eval_accuracy": 0.42133333333333334, | |
| "eval_loss": 6.566288948059082, | |
| "eval_runtime": 5.2039, | |
| "eval_samples_per_second": 96.082, | |
| "eval_steps_per_second": 12.106, | |
| "step": 7171 | |
| }, | |
| { | |
| "epoch": 20.998535871156662, | |
| "eval_bleu": 0.16340256740662695, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.022, | |
| "step": 7171 | |
| }, | |
| { | |
| "epoch": 21.083455344070277, | |
| "grad_norm": 1.0565904378890991, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1575, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 21.37628111273792, | |
| "grad_norm": 1.1256818771362305, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1459, | |
| "step": 7300 | |
| }, | |
| { | |
| "epoch": 21.669106881405565, | |
| "grad_norm": 1.4385403394699097, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1511, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 21.961932650073205, | |
| "grad_norm": 1.1152088642120361, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1575, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_accuracy": 0.42015384615384616, | |
| "eval_loss": 6.625944137573242, | |
| "eval_runtime": 5.0735, | |
| "eval_samples_per_second": 98.552, | |
| "eval_steps_per_second": 12.418, | |
| "step": 7513 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_bleu": 0.19792591482411717, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.022, | |
| "step": 7513 | |
| }, | |
| { | |
| "epoch": 22.25475841874085, | |
| "grad_norm": 1.1440153121948242, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1414, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 22.547584187408493, | |
| "grad_norm": 1.1086493730545044, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1435, | |
| "step": 7700 | |
| }, | |
| { | |
| "epoch": 22.840409956076133, | |
| "grad_norm": 1.2570273876190186, | |
| "learning_rate": 0.0001, | |
| "loss": 0.15, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 22.998535871156662, | |
| "eval_accuracy": 0.42143589743589743, | |
| "eval_loss": 6.595519065856934, | |
| "eval_runtime": 5.2943, | |
| "eval_samples_per_second": 94.442, | |
| "eval_steps_per_second": 11.9, | |
| "step": 7854 | |
| }, | |
| { | |
| "epoch": 22.998535871156662, | |
| "eval_bleu": 0.17402208428859392, | |
| "eval_exact_match": 0.002, | |
| "eval_prefix_exact_match": 0.028, | |
| "step": 7854 | |
| }, | |
| { | |
| "epoch": 23.133235724743777, | |
| "grad_norm": 1.070988655090332, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1455, | |
| "step": 7900 | |
| }, | |
| { | |
| "epoch": 23.42606149341142, | |
| "grad_norm": 1.021813154220581, | |
| "learning_rate": 0.0001, | |
| "loss": 0.138, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 23.718887262079065, | |
| "grad_norm": 1.2333232164382935, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1427, | |
| "step": 8100 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_accuracy": 0.4216410256410256, | |
| "eval_loss": 6.62971830368042, | |
| "eval_runtime": 4.9115, | |
| "eval_samples_per_second": 101.802, | |
| "eval_steps_per_second": 12.827, | |
| "step": 8196 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_bleu": 0.1531918581052099, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.028, | |
| "step": 8196 | |
| }, | |
| { | |
| "epoch": 24.011713030746705, | |
| "grad_norm": 0.9755038022994995, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1488, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 24.30453879941435, | |
| "grad_norm": 1.2587764263153076, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1328, | |
| "step": 8300 | |
| }, | |
| { | |
| "epoch": 24.597364568081993, | |
| "grad_norm": 0.9521226286888123, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1384, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 24.890190336749633, | |
| "grad_norm": 1.2158619165420532, | |
| "learning_rate": 0.0001, | |
| "loss": 0.145, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 24.998535871156662, | |
| "eval_accuracy": 0.4226666666666667, | |
| "eval_loss": 6.57574462890625, | |
| "eval_runtime": 5.0443, | |
| "eval_samples_per_second": 99.122, | |
| "eval_steps_per_second": 12.489, | |
| "step": 8537 | |
| }, | |
| { | |
| "epoch": 24.998535871156662, | |
| "eval_bleu": 0.17454308300545518, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.022, | |
| "step": 8537 | |
| }, | |
| { | |
| "epoch": 25.183016105417277, | |
| "grad_norm": 1.0446624755859375, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1357, | |
| "step": 8600 | |
| }, | |
| { | |
| "epoch": 25.47584187408492, | |
| "grad_norm": 1.07607901096344, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1341, | |
| "step": 8700 | |
| }, | |
| { | |
| "epoch": 25.76866764275256, | |
| "grad_norm": 1.0376451015472412, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1393, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "eval_accuracy": 0.42133333333333334, | |
| "eval_loss": 6.567473411560059, | |
| "eval_runtime": 5.3766, | |
| "eval_samples_per_second": 92.995, | |
| "eval_steps_per_second": 11.717, | |
| "step": 8879 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "eval_bleu": 0.183512215739427, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.026, | |
| "step": 8879 | |
| }, | |
| { | |
| "epoch": 26.061493411420205, | |
| "grad_norm": 0.8689504861831665, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1415, | |
| "step": 8900 | |
| }, | |
| { | |
| "epoch": 26.35431918008785, | |
| "grad_norm": 0.8533344864845276, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1292, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 26.64714494875549, | |
| "grad_norm": 1.0425468683242798, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1348, | |
| "step": 9100 | |
| }, | |
| { | |
| "epoch": 26.939970717423133, | |
| "grad_norm": 1.2276948690414429, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1405, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 26.998535871156662, | |
| "eval_accuracy": 0.42133333333333334, | |
| "eval_loss": 6.664970874786377, | |
| "eval_runtime": 5.6177, | |
| "eval_samples_per_second": 89.004, | |
| "eval_steps_per_second": 11.215, | |
| "step": 9220 | |
| }, | |
| { | |
| "epoch": 26.998535871156662, | |
| "eval_bleu": 0.19671240169357163, | |
| "eval_exact_match": 0.002, | |
| "eval_prefix_exact_match": 0.026, | |
| "step": 9220 | |
| }, | |
| { | |
| "epoch": 27.232796486090777, | |
| "grad_norm": 1.0451931953430176, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1282, | |
| "step": 9300 | |
| }, | |
| { | |
| "epoch": 27.525622254758417, | |
| "grad_norm": 1.0023565292358398, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1307, | |
| "step": 9400 | |
| }, | |
| { | |
| "epoch": 27.81844802342606, | |
| "grad_norm": 0.9287552833557129, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1365, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "eval_accuracy": 0.420974358974359, | |
| "eval_loss": 6.642728328704834, | |
| "eval_runtime": 4.9204, | |
| "eval_samples_per_second": 101.618, | |
| "eval_steps_per_second": 12.804, | |
| "step": 9562 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "eval_bleu": 0.17149559728372338, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.028, | |
| "step": 9562 | |
| }, | |
| { | |
| "epoch": 28.111273792093705, | |
| "grad_norm": 0.7279195785522461, | |
| "learning_rate": 0.0001, | |
| "loss": 0.133, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 28.404099560761345, | |
| "grad_norm": 0.9348853230476379, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1268, | |
| "step": 9700 | |
| }, | |
| { | |
| "epoch": 28.69692532942899, | |
| "grad_norm": 0.9402393102645874, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1323, | |
| "step": 9800 | |
| }, | |
| { | |
| "epoch": 28.989751098096633, | |
| "grad_norm": 0.9316591620445251, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1372, | |
| "step": 9900 | |
| }, | |
| { | |
| "epoch": 28.998535871156662, | |
| "eval_accuracy": 0.4208717948717949, | |
| "eval_loss": 6.548105239868164, | |
| "eval_runtime": 4.8211, | |
| "eval_samples_per_second": 103.71, | |
| "eval_steps_per_second": 13.067, | |
| "step": 9903 | |
| }, | |
| { | |
| "epoch": 28.998535871156662, | |
| "eval_bleu": 0.16884007365872924, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.02, | |
| "step": 9903 | |
| }, | |
| { | |
| "epoch": 29.282576866764277, | |
| "grad_norm": 0.9301176071166992, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1234, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 29.575402635431917, | |
| "grad_norm": 0.8069124221801758, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1287, | |
| "step": 10100 | |
| }, | |
| { | |
| "epoch": 29.86822840409956, | |
| "grad_norm": 1.0093735456466675, | |
| "learning_rate": 0.0001, | |
| "loss": 0.134, | |
| "step": 10200 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "eval_accuracy": 0.41994871794871796, | |
| "eval_loss": 6.661746025085449, | |
| "eval_runtime": 4.8287, | |
| "eval_samples_per_second": 103.548, | |
| "eval_steps_per_second": 13.047, | |
| "step": 10245 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "eval_bleu": 0.19612712097933296, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.028, | |
| "step": 10245 | |
| }, | |
| { | |
| "epoch": 30.161054172767205, | |
| "grad_norm": 0.965752899646759, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1268, | |
| "step": 10300 | |
| }, | |
| { | |
| "epoch": 30.453879941434845, | |
| "grad_norm": 0.9576678276062012, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1241, | |
| "step": 10400 | |
| }, | |
| { | |
| "epoch": 30.74670571010249, | |
| "grad_norm": 0.8335288763046265, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1287, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 30.998535871156662, | |
| "eval_accuracy": 0.4206666666666667, | |
| "eval_loss": 6.624051094055176, | |
| "eval_runtime": 5.0567, | |
| "eval_samples_per_second": 98.879, | |
| "eval_steps_per_second": 12.459, | |
| "step": 10586 | |
| }, | |
| { | |
| "epoch": 30.998535871156662, | |
| "eval_bleu": 0.17063063536886222, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.026, | |
| "step": 10586 | |
| }, | |
| { | |
| "epoch": 31.039531478770133, | |
| "grad_norm": 0.731296718120575, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1327, | |
| "step": 10600 | |
| }, | |
| { | |
| "epoch": 31.332357247437773, | |
| "grad_norm": 0.9108264446258545, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1204, | |
| "step": 10700 | |
| }, | |
| { | |
| "epoch": 31.625183016105417, | |
| "grad_norm": 0.9855822324752808, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1259, | |
| "step": 10800 | |
| }, | |
| { | |
| "epoch": 31.91800878477306, | |
| "grad_norm": 1.1106101274490356, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1305, | |
| "step": 10900 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "eval_accuracy": 0.41994871794871796, | |
| "eval_loss": 6.609373092651367, | |
| "eval_runtime": 5.0029, | |
| "eval_samples_per_second": 99.941, | |
| "eval_steps_per_second": 12.593, | |
| "step": 10928 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "eval_bleu": 0.1600123189131272, | |
| "eval_exact_match": 0.002, | |
| "eval_prefix_exact_match": 0.03, | |
| "step": 10928 | |
| }, | |
| { | |
| "epoch": 32.210834553440705, | |
| "grad_norm": 0.8511720895767212, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1216, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 32.503660322108345, | |
| "grad_norm": 1.0120123624801636, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1225, | |
| "step": 11100 | |
| }, | |
| { | |
| "epoch": 32.796486090775986, | |
| "grad_norm": 0.9636625051498413, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1274, | |
| "step": 11200 | |
| }, | |
| { | |
| "epoch": 32.99853587115666, | |
| "eval_accuracy": 0.4165128205128205, | |
| "eval_loss": 6.682253360748291, | |
| "eval_runtime": 5.2415, | |
| "eval_samples_per_second": 95.392, | |
| "eval_steps_per_second": 12.019, | |
| "step": 11269 | |
| }, | |
| { | |
| "epoch": 32.99853587115666, | |
| "eval_bleu": 0.21084378912778606, | |
| "eval_exact_match": 0.002, | |
| "eval_prefix_exact_match": 0.022, | |
| "step": 11269 | |
| }, | |
| { | |
| "epoch": 33.08931185944363, | |
| "grad_norm": 0.991923451423645, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1271, | |
| "step": 11300 | |
| }, | |
| { | |
| "epoch": 33.38213762811127, | |
| "grad_norm": 0.8520097136497498, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1195, | |
| "step": 11400 | |
| }, | |
| { | |
| "epoch": 33.674963396778914, | |
| "grad_norm": 1.1480833292007446, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1242, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 33.96778916544656, | |
| "grad_norm": 0.8655959367752075, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1296, | |
| "step": 11600 | |
| }, | |
| { | |
| "epoch": 34.0, | |
| "eval_accuracy": 0.4195384615384615, | |
| "eval_loss": 6.621041774749756, | |
| "eval_runtime": 5.1712, | |
| "eval_samples_per_second": 96.689, | |
| "eval_steps_per_second": 12.183, | |
| "step": 11611 | |
| }, | |
| { | |
| "epoch": 34.0, | |
| "eval_bleu": 0.2088991994544206, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.02, | |
| "step": 11611 | |
| }, | |
| { | |
| "epoch": 34.2606149341142, | |
| "grad_norm": 0.7514725923538208, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1199, | |
| "step": 11700 | |
| }, | |
| { | |
| "epoch": 34.55344070278184, | |
| "grad_norm": 0.9057220220565796, | |
| "learning_rate": 0.0001, | |
| "loss": 0.121, | |
| "step": 11800 | |
| }, | |
| { | |
| "epoch": 34.84626647144949, | |
| "grad_norm": 0.8322867155075073, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1271, | |
| "step": 11900 | |
| }, | |
| { | |
| "epoch": 34.99853587115666, | |
| "eval_accuracy": 0.41846153846153844, | |
| "eval_loss": 6.704195976257324, | |
| "eval_runtime": 5.1841, | |
| "eval_samples_per_second": 96.448, | |
| "eval_steps_per_second": 12.153, | |
| "step": 11952 | |
| }, | |
| { | |
| "epoch": 34.99853587115666, | |
| "eval_bleu": 0.17311282679501735, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.022, | |
| "step": 11952 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 17050, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.0263778553086607e+18, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |