| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.868411617067049, | |
| "eval_steps": 2000, | |
| "global_step": 8000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 2.6817848682403564, | |
| "learning_rate": 1e-06, | |
| "loss": 1.077, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 3.5721559524536133, | |
| "learning_rate": 9.898989898989898e-07, | |
| "loss": 0.8347, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 4.235135078430176, | |
| "learning_rate": 9.7989898989899e-07, | |
| "loss": 0.6862, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 2.8351337909698486, | |
| "learning_rate": 9.697979797979798e-07, | |
| "loss": 0.666, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 4.0803093910217285, | |
| "learning_rate": 9.598989898989899e-07, | |
| "loss": 0.6536, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 3.8616509437561035, | |
| "learning_rate": 9.497979797979798e-07, | |
| "loss": 0.6579, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 3.376267194747925, | |
| "learning_rate": 9.396969696969696e-07, | |
| "loss": 0.6223, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 3.7751758098602295, | |
| "learning_rate": 9.295959595959596e-07, | |
| "loss": 0.6141, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 3.669135093688965, | |
| "learning_rate": 9.194949494949495e-07, | |
| "loss": 0.627, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 3.837573528289795, | |
| "learning_rate": 9.093939393939394e-07, | |
| "loss": 0.6309, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 3.461555242538452, | |
| "learning_rate": 8.992929292929292e-07, | |
| "loss": 0.6138, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 2.907550573348999, | |
| "learning_rate": 8.891919191919191e-07, | |
| "loss": 0.6367, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 5.007347106933594, | |
| "learning_rate": 8.790909090909091e-07, | |
| "loss": 0.6444, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 4.709310054779053, | |
| "learning_rate": 8.68989898989899e-07, | |
| "loss": 0.6418, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 2.976135730743408, | |
| "learning_rate": 8.588888888888888e-07, | |
| "loss": 0.6572, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 2.4408750534057617, | |
| "learning_rate": 8.487878787878787e-07, | |
| "loss": 0.6347, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 4.144632816314697, | |
| "learning_rate": 8.386868686868687e-07, | |
| "loss": 0.6001, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 3.0854594707489014, | |
| "learning_rate": 8.285858585858585e-07, | |
| "loss": 0.6446, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 4.574347496032715, | |
| "learning_rate": 8.185858585858586e-07, | |
| "loss": 0.6206, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 4.453900337219238, | |
| "learning_rate": 8.084848484848484e-07, | |
| "loss": 0.6555, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "eval_loss": 0.7467872500419617, | |
| "eval_runtime": 199.2545, | |
| "eval_samples_per_second": 5.019, | |
| "eval_steps_per_second": 1.255, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 4.515957832336426, | |
| "learning_rate": 7.983838383838384e-07, | |
| "loss": 0.6001, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 5.094139099121094, | |
| "learning_rate": 7.882828282828282e-07, | |
| "loss": 0.5989, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 3.6258816719055176, | |
| "learning_rate": 7.781818181818182e-07, | |
| "loss": 0.5893, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 3.713240385055542, | |
| "learning_rate": 7.68080808080808e-07, | |
| "loss": 0.5944, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 2.879484176635742, | |
| "learning_rate": 7.579797979797979e-07, | |
| "loss": 0.6098, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 6.193981170654297, | |
| "learning_rate": 7.478787878787879e-07, | |
| "loss": 0.6232, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 5.713097095489502, | |
| "learning_rate": 7.377777777777777e-07, | |
| "loss": 0.5912, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 4.410184383392334, | |
| "learning_rate": 7.276767676767677e-07, | |
| "loss": 0.6023, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "grad_norm": 5.159379005432129, | |
| "learning_rate": 7.175757575757575e-07, | |
| "loss": 0.6148, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "grad_norm": 4.885188102722168, | |
| "learning_rate": 7.074747474747474e-07, | |
| "loss": 0.5715, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "grad_norm": 2.8274130821228027, | |
| "learning_rate": 6.973737373737374e-07, | |
| "loss": 0.5843, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 4.313437461853027, | |
| "learning_rate": 6.872727272727273e-07, | |
| "loss": 0.5781, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "grad_norm": 3.335306167602539, | |
| "learning_rate": 6.771717171717171e-07, | |
| "loss": 0.5875, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "grad_norm": 3.799851417541504, | |
| "learning_rate": 6.67070707070707e-07, | |
| "loss": 0.5903, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 5.783286094665527, | |
| "learning_rate": 6.56969696969697e-07, | |
| "loss": 0.591, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "grad_norm": 2.4970409870147705, | |
| "learning_rate": 6.468686868686868e-07, | |
| "loss": 0.576, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "grad_norm": 3.698167562484741, | |
| "learning_rate": 6.367676767676767e-07, | |
| "loss": 0.5971, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "grad_norm": 5.005897521972656, | |
| "learning_rate": 6.266666666666667e-07, | |
| "loss": 0.5847, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "grad_norm": 4.98276424407959, | |
| "learning_rate": 6.165656565656565e-07, | |
| "loss": 0.6004, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "grad_norm": 4.1894402503967285, | |
| "learning_rate": 6.064646464646465e-07, | |
| "loss": 0.594, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "eval_loss": 0.7255015969276428, | |
| "eval_runtime": 204.1006, | |
| "eval_samples_per_second": 4.9, | |
| "eval_steps_per_second": 1.225, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "grad_norm": 4.014637470245361, | |
| "learning_rate": 5.963636363636363e-07, | |
| "loss": 0.6014, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "grad_norm": 4.055298805236816, | |
| "learning_rate": 5.862626262626262e-07, | |
| "loss": 0.575, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "grad_norm": 3.810657262802124, | |
| "learning_rate": 5.761616161616162e-07, | |
| "loss": 0.5879, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "grad_norm": 2.8757894039154053, | |
| "learning_rate": 5.660606060606061e-07, | |
| "loss": 0.5864, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "grad_norm": 5.22218132019043, | |
| "learning_rate": 5.559595959595959e-07, | |
| "loss": 0.5881, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "grad_norm": 4.59492826461792, | |
| "learning_rate": 5.458585858585858e-07, | |
| "loss": 0.5628, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 4.455421447753906, | |
| "learning_rate": 5.357575757575758e-07, | |
| "loss": 0.6008, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "grad_norm": 4.579429626464844, | |
| "learning_rate": 5.256565656565657e-07, | |
| "loss": 0.5957, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "grad_norm": 5.034527778625488, | |
| "learning_rate": 5.155555555555555e-07, | |
| "loss": 0.556, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "grad_norm": 6.71533203125, | |
| "learning_rate": 5.054545454545454e-07, | |
| "loss": 0.5811, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.83, | |
| "grad_norm": 3.7940480709075928, | |
| "learning_rate": 4.953535353535353e-07, | |
| "loss": 0.5907, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "grad_norm": 4.488739013671875, | |
| "learning_rate": 4.852525252525253e-07, | |
| "loss": 0.5975, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "grad_norm": 4.032153129577637, | |
| "learning_rate": 4.7515151515151514e-07, | |
| "loss": 0.5771, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "grad_norm": 4.4360504150390625, | |
| "learning_rate": 4.65050505050505e-07, | |
| "loss": 0.5625, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "grad_norm": 4.770937919616699, | |
| "learning_rate": 4.549494949494949e-07, | |
| "loss": 0.5647, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.01, | |
| "grad_norm": 4.940979957580566, | |
| "learning_rate": 4.448484848484848e-07, | |
| "loss": 0.5871, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 2.04, | |
| "grad_norm": 3.6415045261383057, | |
| "learning_rate": 4.3474747474747475e-07, | |
| "loss": 0.5615, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "grad_norm": 5.734036922454834, | |
| "learning_rate": 4.246464646464646e-07, | |
| "loss": 0.5491, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 2.12, | |
| "grad_norm": 7.926692008972168, | |
| "learning_rate": 4.1454545454545453e-07, | |
| "loss": 0.5602, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "grad_norm": 3.576627492904663, | |
| "learning_rate": 4.044444444444444e-07, | |
| "loss": 0.596, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "eval_loss": 0.7305542826652527, | |
| "eval_runtime": 199.7984, | |
| "eval_samples_per_second": 5.005, | |
| "eval_steps_per_second": 1.251, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.19, | |
| "grad_norm": 5.2710347175598145, | |
| "learning_rate": 3.943434343434343e-07, | |
| "loss": 0.558, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "grad_norm": 6.788084030151367, | |
| "learning_rate": 3.8424242424242423e-07, | |
| "loss": 0.5531, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "grad_norm": 5.253055572509766, | |
| "learning_rate": 3.7414141414141414e-07, | |
| "loss": 0.5767, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "grad_norm": 5.5620341300964355, | |
| "learning_rate": 3.6414141414141413e-07, | |
| "loss": 0.5362, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 2.33, | |
| "grad_norm": 5.042788505554199, | |
| "learning_rate": 3.5404040404040405e-07, | |
| "loss": 0.5863, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "grad_norm": 4.5857415199279785, | |
| "learning_rate": 3.439393939393939e-07, | |
| "loss": 0.5759, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "grad_norm": 5.151732444763184, | |
| "learning_rate": 3.3383838383838383e-07, | |
| "loss": 0.5481, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "grad_norm": 4.4784393310546875, | |
| "learning_rate": 3.237373737373737e-07, | |
| "loss": 0.5686, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "grad_norm": 5.631476879119873, | |
| "learning_rate": 3.1363636363636366e-07, | |
| "loss": 0.5803, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 2.51, | |
| "grad_norm": 6.012638092041016, | |
| "learning_rate": 3.035353535353535e-07, | |
| "loss": 0.5937, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "grad_norm": 5.126097202301025, | |
| "learning_rate": 2.9343434343434344e-07, | |
| "loss": 0.57, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "grad_norm": 6.094301700592041, | |
| "learning_rate": 2.833333333333333e-07, | |
| "loss": 0.5468, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "grad_norm": 5.276275634765625, | |
| "learning_rate": 2.732323232323232e-07, | |
| "loss": 0.5411, | |
| "step": 7300 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "grad_norm": 5.292468547821045, | |
| "learning_rate": 2.631313131313131e-07, | |
| "loss": 0.5579, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "grad_norm": 5.443357944488525, | |
| "learning_rate": 2.5303030303030305e-07, | |
| "loss": 0.5844, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 2.72, | |
| "grad_norm": 5.82540225982666, | |
| "learning_rate": 2.429292929292929e-07, | |
| "loss": 0.5693, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 2.76, | |
| "grad_norm": 4.729764938354492, | |
| "learning_rate": 2.3282828282828283e-07, | |
| "loss": 0.5705, | |
| "step": 7700 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "grad_norm": 5.492878437042236, | |
| "learning_rate": 2.2272727272727272e-07, | |
| "loss": 0.5694, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "grad_norm": 5.059893608093262, | |
| "learning_rate": 2.1262626262626264e-07, | |
| "loss": 0.5543, | |
| "step": 7900 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "grad_norm": 5.176052093505859, | |
| "learning_rate": 2.0252525252525253e-07, | |
| "loss": 0.5411, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "eval_loss": 0.725495457649231, | |
| "eval_runtime": 564.5599, | |
| "eval_samples_per_second": 1.771, | |
| "eval_steps_per_second": 0.443, | |
| "step": 8000 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 10000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 2000, | |
| "total_flos": 7.541625072742564e+17, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |