tiny-gpt2 / trainer_state.json
taufeeque's picture
Update to 1M parameter model
9763874
raw
history blame
11.9 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 14.013452914798206,
"global_step": 50000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.14,
"learning_rate": 9.900000000000001e-05,
"loss": 9.7624,
"step": 500
},
{
"epoch": 0.28,
"learning_rate": 9.8e-05,
"loss": 8.2156,
"step": 1000
},
{
"epoch": 0.42,
"learning_rate": 9.7e-05,
"loss": 7.5205,
"step": 1500
},
{
"epoch": 0.56,
"learning_rate": 9.6e-05,
"loss": 7.3362,
"step": 2000
},
{
"epoch": 0.7,
"learning_rate": 9.5e-05,
"loss": 7.1343,
"step": 2500
},
{
"epoch": 0.84,
"learning_rate": 9.4e-05,
"loss": 6.9313,
"step": 3000
},
{
"epoch": 0.98,
"learning_rate": 9.300000000000001e-05,
"loss": 6.7408,
"step": 3500
},
{
"epoch": 1.12,
"learning_rate": 9.200000000000001e-05,
"loss": 6.5858,
"step": 4000
},
{
"epoch": 1.26,
"learning_rate": 9.1e-05,
"loss": 6.4674,
"step": 4500
},
{
"epoch": 1.4,
"learning_rate": 9e-05,
"loss": 6.3797,
"step": 5000
},
{
"epoch": 1.54,
"learning_rate": 8.900000000000001e-05,
"loss": 6.3089,
"step": 5500
},
{
"epoch": 1.68,
"learning_rate": 8.800000000000001e-05,
"loss": 6.2535,
"step": 6000
},
{
"epoch": 1.82,
"learning_rate": 8.7e-05,
"loss": 6.2054,
"step": 6500
},
{
"epoch": 1.96,
"learning_rate": 8.6e-05,
"loss": 6.1582,
"step": 7000
},
{
"epoch": 2.1,
"learning_rate": 8.5e-05,
"loss": 6.1192,
"step": 7500
},
{
"epoch": 2.24,
"learning_rate": 8.4e-05,
"loss": 6.0872,
"step": 8000
},
{
"epoch": 2.38,
"learning_rate": 8.3e-05,
"loss": 6.0494,
"step": 8500
},
{
"epoch": 2.52,
"learning_rate": 8.2e-05,
"loss": 6.0233,
"step": 9000
},
{
"epoch": 2.66,
"learning_rate": 8.1e-05,
"loss": 5.993,
"step": 9500
},
{
"epoch": 2.8,
"learning_rate": 8e-05,
"loss": 5.9731,
"step": 10000
},
{
"epoch": 2.94,
"learning_rate": 7.900000000000001e-05,
"loss": 5.9495,
"step": 10500
},
{
"epoch": 3.08,
"learning_rate": 7.800000000000001e-05,
"loss": 5.9287,
"step": 11000
},
{
"epoch": 3.22,
"learning_rate": 7.7e-05,
"loss": 5.9077,
"step": 11500
},
{
"epoch": 3.36,
"learning_rate": 7.6e-05,
"loss": 5.8884,
"step": 12000
},
{
"epoch": 3.5,
"learning_rate": 7.500000000000001e-05,
"loss": 5.8712,
"step": 12500
},
{
"epoch": 3.64,
"learning_rate": 7.4e-05,
"loss": 5.8562,
"step": 13000
},
{
"epoch": 3.78,
"learning_rate": 7.3e-05,
"loss": 5.836,
"step": 13500
},
{
"epoch": 3.92,
"learning_rate": 7.2e-05,
"loss": 5.8282,
"step": 14000
},
{
"epoch": 4.06,
"learning_rate": 7.1e-05,
"loss": 5.8108,
"step": 14500
},
{
"epoch": 4.2,
"learning_rate": 7e-05,
"loss": 5.7997,
"step": 15000
},
{
"epoch": 4.34,
"learning_rate": 6.9e-05,
"loss": 5.7869,
"step": 15500
},
{
"epoch": 4.48,
"learning_rate": 6.800000000000001e-05,
"loss": 5.7755,
"step": 16000
},
{
"epoch": 4.62,
"learning_rate": 6.7e-05,
"loss": 5.7686,
"step": 16500
},
{
"epoch": 4.76,
"learning_rate": 6.6e-05,
"loss": 5.7561,
"step": 17000
},
{
"epoch": 4.9,
"learning_rate": 6.500000000000001e-05,
"loss": 5.7481,
"step": 17500
},
{
"epoch": 5.04,
"learning_rate": 6.400000000000001e-05,
"loss": 5.7397,
"step": 18000
},
{
"epoch": 5.18,
"learning_rate": 6.3e-05,
"loss": 5.7285,
"step": 18500
},
{
"epoch": 5.33,
"learning_rate": 6.2e-05,
"loss": 5.7194,
"step": 19000
},
{
"epoch": 5.47,
"learning_rate": 6.1e-05,
"loss": 5.7117,
"step": 19500
},
{
"epoch": 5.61,
"learning_rate": 6e-05,
"loss": 5.7066,
"step": 20000
},
{
"epoch": 5.75,
"learning_rate": 5.9e-05,
"loss": 5.6968,
"step": 20500
},
{
"epoch": 5.89,
"learning_rate": 5.8e-05,
"loss": 5.6889,
"step": 21000
},
{
"epoch": 6.03,
"learning_rate": 5.6999999999999996e-05,
"loss": 5.681,
"step": 21500
},
{
"epoch": 6.17,
"learning_rate": 5.6000000000000006e-05,
"loss": 5.6788,
"step": 22000
},
{
"epoch": 6.31,
"learning_rate": 5.500000000000001e-05,
"loss": 5.6668,
"step": 22500
},
{
"epoch": 6.45,
"learning_rate": 5.4000000000000005e-05,
"loss": 5.6649,
"step": 23000
},
{
"epoch": 6.59,
"learning_rate": 5.300000000000001e-05,
"loss": 5.6574,
"step": 23500
},
{
"epoch": 6.73,
"learning_rate": 5.2000000000000004e-05,
"loss": 5.6499,
"step": 24000
},
{
"epoch": 6.87,
"learning_rate": 5.1000000000000006e-05,
"loss": 5.6465,
"step": 24500
},
{
"epoch": 7.01,
"learning_rate": 5e-05,
"loss": 5.6419,
"step": 25000
},
{
"epoch": 7.15,
"learning_rate": 4.9e-05,
"loss": 5.6358,
"step": 25500
},
{
"epoch": 7.29,
"learning_rate": 4.8e-05,
"loss": 5.6312,
"step": 26000
},
{
"epoch": 7.43,
"learning_rate": 4.7e-05,
"loss": 5.6293,
"step": 26500
},
{
"epoch": 7.57,
"learning_rate": 4.600000000000001e-05,
"loss": 5.6218,
"step": 27000
},
{
"epoch": 7.71,
"learning_rate": 4.5e-05,
"loss": 5.618,
"step": 27500
},
{
"epoch": 7.85,
"learning_rate": 4.4000000000000006e-05,
"loss": 5.613,
"step": 28000
},
{
"epoch": 7.99,
"learning_rate": 4.3e-05,
"loss": 5.6101,
"step": 28500
},
{
"epoch": 8.13,
"learning_rate": 4.2e-05,
"loss": 5.6048,
"step": 29000
},
{
"epoch": 8.27,
"learning_rate": 4.1e-05,
"loss": 5.6,
"step": 29500
},
{
"epoch": 8.41,
"learning_rate": 4e-05,
"loss": 5.5976,
"step": 30000
},
{
"epoch": 8.55,
"learning_rate": 3.9000000000000006e-05,
"loss": 5.5936,
"step": 30500
},
{
"epoch": 8.69,
"learning_rate": 3.8e-05,
"loss": 5.5931,
"step": 31000
},
{
"epoch": 8.83,
"learning_rate": 3.7e-05,
"loss": 5.5898,
"step": 31500
},
{
"epoch": 8.97,
"learning_rate": 3.6e-05,
"loss": 5.5888,
"step": 32000
},
{
"epoch": 9.11,
"learning_rate": 3.5e-05,
"loss": 5.582,
"step": 32500
},
{
"epoch": 9.25,
"learning_rate": 3.4000000000000007e-05,
"loss": 5.5817,
"step": 33000
},
{
"epoch": 9.39,
"learning_rate": 3.3e-05,
"loss": 5.5726,
"step": 33500
},
{
"epoch": 9.53,
"learning_rate": 3.2000000000000005e-05,
"loss": 5.5778,
"step": 34000
},
{
"epoch": 9.67,
"learning_rate": 3.1e-05,
"loss": 5.5718,
"step": 34500
},
{
"epoch": 9.81,
"learning_rate": 3e-05,
"loss": 5.569,
"step": 35000
},
{
"epoch": 9.95,
"learning_rate": 2.9e-05,
"loss": 5.5703,
"step": 35500
},
{
"epoch": 10.09,
"learning_rate": 2.8000000000000003e-05,
"loss": 5.5668,
"step": 36000
},
{
"epoch": 10.23,
"learning_rate": 2.7000000000000002e-05,
"loss": 5.561,
"step": 36500
},
{
"epoch": 10.37,
"learning_rate": 2.6000000000000002e-05,
"loss": 5.5554,
"step": 37000
},
{
"epoch": 10.51,
"learning_rate": 2.5e-05,
"loss": 5.5621,
"step": 37500
},
{
"epoch": 10.65,
"learning_rate": 2.4e-05,
"loss": 5.558,
"step": 38000
},
{
"epoch": 10.79,
"learning_rate": 2.3000000000000003e-05,
"loss": 5.5557,
"step": 38500
},
{
"epoch": 10.93,
"learning_rate": 2.2000000000000003e-05,
"loss": 5.5572,
"step": 39000
},
{
"epoch": 11.07,
"learning_rate": 2.1e-05,
"loss": 5.5519,
"step": 39500
},
{
"epoch": 11.21,
"learning_rate": 2e-05,
"loss": 5.5502,
"step": 40000
},
{
"epoch": 11.35,
"learning_rate": 1.9e-05,
"loss": 5.5436,
"step": 40500
},
{
"epoch": 11.49,
"learning_rate": 1.8e-05,
"loss": 5.5482,
"step": 41000
},
{
"epoch": 11.63,
"learning_rate": 1.7000000000000003e-05,
"loss": 5.5524,
"step": 41500
},
{
"epoch": 11.77,
"learning_rate": 1.6000000000000003e-05,
"loss": 5.5449,
"step": 42000
},
{
"epoch": 11.91,
"learning_rate": 1.5e-05,
"loss": 5.5443,
"step": 42500
},
{
"epoch": 12.05,
"learning_rate": 1.4000000000000001e-05,
"loss": 5.5466,
"step": 43000
},
{
"epoch": 12.19,
"learning_rate": 1.3000000000000001e-05,
"loss": 5.5388,
"step": 43500
},
{
"epoch": 12.33,
"learning_rate": 1.2e-05,
"loss": 5.5397,
"step": 44000
},
{
"epoch": 12.47,
"learning_rate": 1.1000000000000001e-05,
"loss": 5.5385,
"step": 44500
},
{
"epoch": 12.61,
"learning_rate": 1e-05,
"loss": 5.5437,
"step": 45000
},
{
"epoch": 12.75,
"learning_rate": 9e-06,
"loss": 5.541,
"step": 45500
},
{
"epoch": 12.89,
"learning_rate": 8.000000000000001e-06,
"loss": 5.5395,
"step": 46000
},
{
"epoch": 13.03,
"learning_rate": 7.000000000000001e-06,
"loss": 5.5357,
"step": 46500
},
{
"epoch": 13.17,
"learning_rate": 6e-06,
"loss": 5.5351,
"step": 47000
},
{
"epoch": 13.31,
"learning_rate": 5e-06,
"loss": 5.5353,
"step": 47500
},
{
"epoch": 13.45,
"learning_rate": 4.000000000000001e-06,
"loss": 5.5371,
"step": 48000
},
{
"epoch": 13.59,
"learning_rate": 3e-06,
"loss": 5.5343,
"step": 48500
},
{
"epoch": 13.73,
"learning_rate": 2.0000000000000003e-06,
"loss": 5.5371,
"step": 49000
},
{
"epoch": 13.87,
"learning_rate": 1.0000000000000002e-06,
"loss": 5.538,
"step": 49500
},
{
"epoch": 14.01,
"learning_rate": 0.0,
"loss": 5.5386,
"step": 50000
},
{
"epoch": 14.01,
"step": 50000,
"total_flos": 336513375928320.0,
"train_loss": 5.854185776367188,
"train_runtime": 10672.3435,
"train_samples_per_second": 149.92,
"train_steps_per_second": 4.685
}
],
"max_steps": 50000,
"num_train_epochs": 15,
"total_flos": 336513375928320.0,
"trial_name": null,
"trial_params": null
}