| { | |
| "best_metric": 0.33645617961883545, | |
| "best_model_checkpoint": "mikhail_panzo/zlm_b128_le4_s12000/checkpoint-3500", | |
| "epoch": 5.863874345549738, | |
| "eval_steps": 500, | |
| "global_step": 3500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.08376963350785341, | |
| "grad_norm": 2.469545841217041, | |
| "learning_rate": 2.4500000000000003e-06, | |
| "loss": 1.0402, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.16753926701570682, | |
| "grad_norm": 3.3642780780792236, | |
| "learning_rate": 4.950000000000001e-06, | |
| "loss": 0.8472, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2513089005235602, | |
| "grad_norm": 1.8375087976455688, | |
| "learning_rate": 7.45e-06, | |
| "loss": 0.7331, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33507853403141363, | |
| "grad_norm": 3.700824737548828, | |
| "learning_rate": 9.950000000000001e-06, | |
| "loss": 0.6472, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.418848167539267, | |
| "grad_norm": 1.7632888555526733, | |
| "learning_rate": 1.2450000000000001e-05, | |
| "loss": 0.6112, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.5026178010471204, | |
| "grad_norm": 3.1861908435821533, | |
| "learning_rate": 1.4950000000000001e-05, | |
| "loss": 0.5865, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.5863874345549738, | |
| "grad_norm": 2.2845046520233154, | |
| "learning_rate": 1.745e-05, | |
| "loss": 0.5682, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.6701570680628273, | |
| "grad_norm": 2.079210042953491, | |
| "learning_rate": 1.995e-05, | |
| "loss": 0.5465, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.7539267015706806, | |
| "grad_norm": 3.669891119003296, | |
| "learning_rate": 2.245e-05, | |
| "loss": 0.5302, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.837696335078534, | |
| "grad_norm": 2.4679417610168457, | |
| "learning_rate": 2.495e-05, | |
| "loss": 0.5137, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.837696335078534, | |
| "eval_loss": 0.4513999819755554, | |
| "eval_runtime": 277.1074, | |
| "eval_samples_per_second": 30.634, | |
| "eval_steps_per_second": 3.832, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.9214659685863874, | |
| "grad_norm": 2.703871726989746, | |
| "learning_rate": 2.7450000000000003e-05, | |
| "loss": 0.5071, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.0052356020942408, | |
| "grad_norm": 3.8184635639190674, | |
| "learning_rate": 2.995e-05, | |
| "loss": 0.4971, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.0890052356020943, | |
| "grad_norm": 2.2857866287231445, | |
| "learning_rate": 3.245e-05, | |
| "loss": 0.4955, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 1.1727748691099475, | |
| "grad_norm": 3.5974085330963135, | |
| "learning_rate": 3.495e-05, | |
| "loss": 0.4935, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.256544502617801, | |
| "grad_norm": 1.720818281173706, | |
| "learning_rate": 3.745e-05, | |
| "loss": 0.4763, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 1.3403141361256545, | |
| "grad_norm": 3.080139636993408, | |
| "learning_rate": 3.995e-05, | |
| "loss": 0.4735, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.4240837696335078, | |
| "grad_norm": 4.877579212188721, | |
| "learning_rate": 4.245e-05, | |
| "loss": 0.4654, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 1.5078534031413613, | |
| "grad_norm": 3.383965253829956, | |
| "learning_rate": 4.495e-05, | |
| "loss": 0.4628, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.5916230366492146, | |
| "grad_norm": 3.3636982440948486, | |
| "learning_rate": 4.745e-05, | |
| "loss": 0.4541, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 1.675392670157068, | |
| "grad_norm": 1.666568398475647, | |
| "learning_rate": 4.995e-05, | |
| "loss": 0.4565, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.675392670157068, | |
| "eval_loss": 0.4093586802482605, | |
| "eval_runtime": 269.4005, | |
| "eval_samples_per_second": 31.511, | |
| "eval_steps_per_second": 3.942, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.7591623036649215, | |
| "grad_norm": 1.6522510051727295, | |
| "learning_rate": 5.245e-05, | |
| "loss": 0.4541, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 1.8429319371727748, | |
| "grad_norm": 1.6531606912612915, | |
| "learning_rate": 5.495e-05, | |
| "loss": 0.4448, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.9267015706806283, | |
| "grad_norm": 3.3253750801086426, | |
| "learning_rate": 5.745e-05, | |
| "loss": 0.4346, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 2.0104712041884816, | |
| "grad_norm": 2.0393073558807373, | |
| "learning_rate": 5.995000000000001e-05, | |
| "loss": 0.4314, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 2.094240837696335, | |
| "grad_norm": 1.906546950340271, | |
| "learning_rate": 6.245000000000001e-05, | |
| "loss": 0.4327, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 2.1780104712041886, | |
| "grad_norm": 1.7925021648406982, | |
| "learning_rate": 6.494999999999999e-05, | |
| "loss": 0.4285, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 2.261780104712042, | |
| "grad_norm": 2.5238988399505615, | |
| "learning_rate": 6.745e-05, | |
| "loss": 0.4251, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 2.345549738219895, | |
| "grad_norm": 2.53450345993042, | |
| "learning_rate": 6.995e-05, | |
| "loss": 0.4284, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 2.4293193717277486, | |
| "grad_norm": 1.9077616930007935, | |
| "learning_rate": 7.245000000000001e-05, | |
| "loss": 0.4244, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 2.513089005235602, | |
| "grad_norm": 1.5720113515853882, | |
| "learning_rate": 7.495e-05, | |
| "loss": 0.4171, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.513089005235602, | |
| "eval_loss": 0.384281188249588, | |
| "eval_runtime": 274.8763, | |
| "eval_samples_per_second": 30.883, | |
| "eval_steps_per_second": 3.864, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.5968586387434556, | |
| "grad_norm": 2.272549867630005, | |
| "learning_rate": 7.745e-05, | |
| "loss": 0.4175, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 2.680628272251309, | |
| "grad_norm": 1.3440821170806885, | |
| "learning_rate": 7.995e-05, | |
| "loss": 0.4134, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 2.7643979057591626, | |
| "grad_norm": 1.99045729637146, | |
| "learning_rate": 8.245e-05, | |
| "loss": 0.4117, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 2.8481675392670156, | |
| "grad_norm": 1.3599165678024292, | |
| "learning_rate": 8.495e-05, | |
| "loss": 0.4101, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 2.931937172774869, | |
| "grad_norm": 2.3722105026245117, | |
| "learning_rate": 8.745000000000001e-05, | |
| "loss": 0.413, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 3.0157068062827226, | |
| "grad_norm": 1.5704238414764404, | |
| "learning_rate": 8.995e-05, | |
| "loss": 0.4086, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 3.099476439790576, | |
| "grad_norm": 3.272968053817749, | |
| "learning_rate": 9.245e-05, | |
| "loss": 0.4057, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 3.183246073298429, | |
| "grad_norm": 2.658064126968384, | |
| "learning_rate": 9.495e-05, | |
| "loss": 0.4071, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 3.2670157068062826, | |
| "grad_norm": 1.688971757888794, | |
| "learning_rate": 9.745000000000001e-05, | |
| "loss": 0.3984, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 3.350785340314136, | |
| "grad_norm": 1.4563082456588745, | |
| "learning_rate": 9.995e-05, | |
| "loss": 0.404, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.350785340314136, | |
| "eval_loss": 0.36596959829330444, | |
| "eval_runtime": 268.3638, | |
| "eval_samples_per_second": 31.632, | |
| "eval_steps_per_second": 3.957, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.4345549738219896, | |
| "grad_norm": 1.6672757863998413, | |
| "learning_rate": 9.951e-05, | |
| "loss": 0.4032, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 3.518324607329843, | |
| "grad_norm": 2.0084328651428223, | |
| "learning_rate": 9.901e-05, | |
| "loss": 0.4003, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 3.6020942408376966, | |
| "grad_norm": 1.8289756774902344, | |
| "learning_rate": 9.851e-05, | |
| "loss": 0.3986, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 3.6858638743455496, | |
| "grad_norm": 1.791348934173584, | |
| "learning_rate": 9.801e-05, | |
| "loss": 0.3928, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 3.769633507853403, | |
| "grad_norm": 1.9079582691192627, | |
| "learning_rate": 9.751e-05, | |
| "loss": 0.3899, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 3.8534031413612566, | |
| "grad_norm": 7.154463768005371, | |
| "learning_rate": 9.701e-05, | |
| "loss": 0.3894, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 3.93717277486911, | |
| "grad_norm": 1.9638899564743042, | |
| "learning_rate": 9.651e-05, | |
| "loss": 0.399, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 4.020942408376963, | |
| "grad_norm": 2.147676706314087, | |
| "learning_rate": 9.601e-05, | |
| "loss": 0.3902, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 4.104712041884817, | |
| "grad_norm": 2.282815456390381, | |
| "learning_rate": 9.551e-05, | |
| "loss": 0.3906, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 4.18848167539267, | |
| "grad_norm": 1.2255228757858276, | |
| "learning_rate": 9.501e-05, | |
| "loss": 0.3851, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 4.18848167539267, | |
| "eval_loss": 0.354680597782135, | |
| "eval_runtime": 269.5668, | |
| "eval_samples_per_second": 31.491, | |
| "eval_steps_per_second": 3.94, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 4.272251308900524, | |
| "grad_norm": 1.2088522911071777, | |
| "learning_rate": 9.451000000000002e-05, | |
| "loss": 0.3891, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 4.356020942408377, | |
| "grad_norm": 1.3467260599136353, | |
| "learning_rate": 9.402000000000001e-05, | |
| "loss": 0.3854, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 4.439790575916231, | |
| "grad_norm": 1.2227791547775269, | |
| "learning_rate": 9.352000000000001e-05, | |
| "loss": 0.3818, | |
| "step": 2650 | |
| }, | |
| { | |
| "epoch": 4.523560209424084, | |
| "grad_norm": 1.0672763586044312, | |
| "learning_rate": 9.302e-05, | |
| "loss": 0.389, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 4.607329842931938, | |
| "grad_norm": 1.6198370456695557, | |
| "learning_rate": 9.252e-05, | |
| "loss": 0.3811, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 4.69109947643979, | |
| "grad_norm": 1.296200156211853, | |
| "learning_rate": 9.202e-05, | |
| "loss": 0.382, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 4.774869109947644, | |
| "grad_norm": 1.533375859260559, | |
| "learning_rate": 9.152e-05, | |
| "loss": 0.381, | |
| "step": 2850 | |
| }, | |
| { | |
| "epoch": 4.858638743455497, | |
| "grad_norm": 2.161743640899658, | |
| "learning_rate": 9.102e-05, | |
| "loss": 0.3829, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 4.942408376963351, | |
| "grad_norm": 1.2308694124221802, | |
| "learning_rate": 9.052000000000001e-05, | |
| "loss": 0.3806, | |
| "step": 2950 | |
| }, | |
| { | |
| "epoch": 5.026178010471204, | |
| "grad_norm": 1.2448184490203857, | |
| "learning_rate": 9.002000000000001e-05, | |
| "loss": 0.3808, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 5.026178010471204, | |
| "eval_loss": 0.3451012670993805, | |
| "eval_runtime": 274.5252, | |
| "eval_samples_per_second": 30.922, | |
| "eval_steps_per_second": 3.868, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 5.109947643979058, | |
| "grad_norm": 1.4961705207824707, | |
| "learning_rate": 8.952000000000001e-05, | |
| "loss": 0.3755, | |
| "step": 3050 | |
| }, | |
| { | |
| "epoch": 5.193717277486911, | |
| "grad_norm": 1.2276802062988281, | |
| "learning_rate": 8.902e-05, | |
| "loss": 0.3771, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 5.277486910994765, | |
| "grad_norm": 1.3186664581298828, | |
| "learning_rate": 8.852e-05, | |
| "loss": 0.3766, | |
| "step": 3150 | |
| }, | |
| { | |
| "epoch": 5.361256544502618, | |
| "grad_norm": 1.4056960344314575, | |
| "learning_rate": 8.802e-05, | |
| "loss": 0.3761, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 5.445026178010472, | |
| "grad_norm": 1.4609332084655762, | |
| "learning_rate": 8.752e-05, | |
| "loss": 0.377, | |
| "step": 3250 | |
| }, | |
| { | |
| "epoch": 5.528795811518324, | |
| "grad_norm": 1.2348676919937134, | |
| "learning_rate": 8.702e-05, | |
| "loss": 0.3761, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 5.612565445026178, | |
| "grad_norm": 1.2290446758270264, | |
| "learning_rate": 8.652e-05, | |
| "loss": 0.3735, | |
| "step": 3350 | |
| }, | |
| { | |
| "epoch": 5.696335078534031, | |
| "grad_norm": 1.0205050706863403, | |
| "learning_rate": 8.602e-05, | |
| "loss": 0.3709, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 5.780104712041885, | |
| "grad_norm": 1.83589768409729, | |
| "learning_rate": 8.552e-05, | |
| "loss": 0.3709, | |
| "step": 3450 | |
| }, | |
| { | |
| "epoch": 5.863874345549738, | |
| "grad_norm": 1.8419536352157593, | |
| "learning_rate": 8.502e-05, | |
| "loss": 0.3704, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 5.863874345549738, | |
| "eval_loss": 0.33645617961883545, | |
| "eval_runtime": 267.7957, | |
| "eval_samples_per_second": 31.7, | |
| "eval_steps_per_second": 3.966, | |
| "step": 3500 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 12000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 21, | |
| "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": 6.292785237299827e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |