3085 lines
84 KiB
JSON
3085 lines
84 KiB
JSON
|
{
|
||
|
|
"best_global_step": null,
|
||
|
|
"best_metric": null,
|
||
|
|
"best_model_checkpoint": null,
|
||
|
|
"epoch": 4.0,
|
||
|
|
"eval_steps": 100,
|
||
|
|
"global_step": 2500,
|
||
|
|
"is_hyper_param_search": false,
|
||
|
|
"is_local_process_zero": true,
|
||
|
|
"is_world_process_zero": true,
|
||
|
|
"log_history": [
|
||
|
|
{
|
||
|
|
"entropy": 0.6421742392703891,
|
||
|
|
"epoch": 0.016,
|
||
|
|
"grad_norm": 4.4375,
|
||
|
|
"learning_rate": 1.8e-07,
|
||
|
|
"loss": 0.953,
|
||
|
|
"mean_token_accuracy": 0.792544937506318,
|
||
|
|
"num_tokens": 282849.0,
|
||
|
|
"step": 10
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6719903893768787,
|
||
|
|
"epoch": 0.032,
|
||
|
|
"grad_norm": 4.25,
|
||
|
|
"learning_rate": 3.8e-07,
|
||
|
|
"loss": 0.9758,
|
||
|
|
"mean_token_accuracy": 0.7857426963746548,
|
||
|
|
"num_tokens": 561726.0,
|
||
|
|
"step": 20
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6603780068457127,
|
||
|
|
"epoch": 0.048,
|
||
|
|
"grad_norm": 4.40625,
|
||
|
|
"learning_rate": 5.800000000000001e-07,
|
||
|
|
"loss": 0.9582,
|
||
|
|
"mean_token_accuracy": 0.7894285101443529,
|
||
|
|
"num_tokens": 842092.0,
|
||
|
|
"step": 30
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6775440879166126,
|
||
|
|
"epoch": 0.064,
|
||
|
|
"grad_norm": 3.890625,
|
||
|
|
"learning_rate": 7.8e-07,
|
||
|
|
"loss": 0.9776,
|
||
|
|
"mean_token_accuracy": 0.785004175081849,
|
||
|
|
"num_tokens": 1133349.0,
|
||
|
|
"step": 40
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6926743468269706,
|
||
|
|
"epoch": 0.08,
|
||
|
|
"grad_norm": 4.6875,
|
||
|
|
"learning_rate": 9.800000000000001e-07,
|
||
|
|
"loss": 0.9497,
|
||
|
|
"mean_token_accuracy": 0.7896979171782732,
|
||
|
|
"num_tokens": 1422340.0,
|
||
|
|
"step": 50
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7043748194351792,
|
||
|
|
"epoch": 0.096,
|
||
|
|
"grad_norm": 3.484375,
|
||
|
|
"learning_rate": 1.1800000000000001e-06,
|
||
|
|
"loss": 0.9217,
|
||
|
|
"mean_token_accuracy": 0.7949878334999084,
|
||
|
|
"num_tokens": 1711821.0,
|
||
|
|
"step": 60
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7551965553313493,
|
||
|
|
"epoch": 0.112,
|
||
|
|
"grad_norm": 2.71875,
|
||
|
|
"learning_rate": 1.3800000000000001e-06,
|
||
|
|
"loss": 0.931,
|
||
|
|
"mean_token_accuracy": 0.7912098452448845,
|
||
|
|
"num_tokens": 1991855.0,
|
||
|
|
"step": 70
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7913430579006672,
|
||
|
|
"epoch": 0.128,
|
||
|
|
"grad_norm": 2.359375,
|
||
|
|
"learning_rate": 1.5800000000000001e-06,
|
||
|
|
"loss": 0.9172,
|
||
|
|
"mean_token_accuracy": 0.7935858864337206,
|
||
|
|
"num_tokens": 2288720.0,
|
||
|
|
"step": 80
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7881425807252527,
|
||
|
|
"epoch": 0.144,
|
||
|
|
"grad_norm": 2.125,
|
||
|
|
"learning_rate": 1.7800000000000001e-06,
|
||
|
|
"loss": 0.917,
|
||
|
|
"mean_token_accuracy": 0.7932404045015573,
|
||
|
|
"num_tokens": 2564352.0,
|
||
|
|
"step": 90
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7627748234197498,
|
||
|
|
"epoch": 0.16,
|
||
|
|
"grad_norm": 2.125,
|
||
|
|
"learning_rate": 1.98e-06,
|
||
|
|
"loss": 0.9014,
|
||
|
|
"mean_token_accuracy": 0.7970093633979559,
|
||
|
|
"num_tokens": 2841876.0,
|
||
|
|
"step": 100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.16,
|
||
|
|
"eval_math_entropy": 0.7972113332748413,
|
||
|
|
"eval_math_loss": 0.9417253136634827,
|
||
|
|
"eval_math_mean_token_accuracy": 0.7907100429534912,
|
||
|
|
"eval_math_num_tokens": 2841876.0,
|
||
|
|
"eval_math_runtime": 35.8728,
|
||
|
|
"eval_math_samples_per_second": 13.938,
|
||
|
|
"eval_math_steps_per_second": 3.485,
|
||
|
|
"step": 100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.16,
|
||
|
|
"eval_physics_entropy": 0.7183168392181396,
|
||
|
|
"eval_physics_loss": 0.7405859231948853,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.822186062335968,
|
||
|
|
"eval_physics_num_tokens": 2841876.0,
|
||
|
|
"eval_physics_runtime": 41.4814,
|
||
|
|
"eval_physics_samples_per_second": 12.054,
|
||
|
|
"eval_physics_steps_per_second": 3.013,
|
||
|
|
"step": 100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7770459342747926,
|
||
|
|
"epoch": 0.176,
|
||
|
|
"grad_norm": 1.9765625,
|
||
|
|
"learning_rate": 2.1800000000000003e-06,
|
||
|
|
"loss": 0.8952,
|
||
|
|
"mean_token_accuracy": 0.7969978783279658,
|
||
|
|
"num_tokens": 3129237.0,
|
||
|
|
"step": 110
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7958783155307174,
|
||
|
|
"epoch": 0.192,
|
||
|
|
"grad_norm": 2.3125,
|
||
|
|
"learning_rate": 2.38e-06,
|
||
|
|
"loss": 0.9119,
|
||
|
|
"mean_token_accuracy": 0.793396720290184,
|
||
|
|
"num_tokens": 3414875.0,
|
||
|
|
"step": 120
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7911858700215817,
|
||
|
|
"epoch": 0.208,
|
||
|
|
"grad_norm": 2.4375,
|
||
|
|
"learning_rate": 2.5800000000000003e-06,
|
||
|
|
"loss": 0.8979,
|
||
|
|
"mean_token_accuracy": 0.7956417176872492,
|
||
|
|
"num_tokens": 3695623.0,
|
||
|
|
"step": 130
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7871318949386478,
|
||
|
|
"epoch": 0.224,
|
||
|
|
"grad_norm": 2.09375,
|
||
|
|
"learning_rate": 2.7800000000000005e-06,
|
||
|
|
"loss": 0.8882,
|
||
|
|
"mean_token_accuracy": 0.7958744909614325,
|
||
|
|
"num_tokens": 3977666.0,
|
||
|
|
"step": 140
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7629930060356855,
|
||
|
|
"epoch": 0.24,
|
||
|
|
"grad_norm": 1.9453125,
|
||
|
|
"learning_rate": 2.9800000000000003e-06,
|
||
|
|
"loss": 0.8654,
|
||
|
|
"mean_token_accuracy": 0.8001766365021468,
|
||
|
|
"num_tokens": 4269287.0,
|
||
|
|
"step": 150
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7563591374084353,
|
||
|
|
"epoch": 0.256,
|
||
|
|
"grad_norm": 2.015625,
|
||
|
|
"learning_rate": 3.1800000000000005e-06,
|
||
|
|
"loss": 0.846,
|
||
|
|
"mean_token_accuracy": 0.8012917254120111,
|
||
|
|
"num_tokens": 4558632.0,
|
||
|
|
"step": 160
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7706880155950785,
|
||
|
|
"epoch": 0.272,
|
||
|
|
"grad_norm": 1.859375,
|
||
|
|
"learning_rate": 3.3800000000000007e-06,
|
||
|
|
"loss": 0.8589,
|
||
|
|
"mean_token_accuracy": 0.8001386649906636,
|
||
|
|
"num_tokens": 4839270.0,
|
||
|
|
"step": 170
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7551738245412707,
|
||
|
|
"epoch": 0.288,
|
||
|
|
"grad_norm": 1.609375,
|
||
|
|
"learning_rate": 3.58e-06,
|
||
|
|
"loss": 0.8468,
|
||
|
|
"mean_token_accuracy": 0.8015156831592322,
|
||
|
|
"num_tokens": 5125447.0,
|
||
|
|
"step": 180
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7806578669697046,
|
||
|
|
"epoch": 0.304,
|
||
|
|
"grad_norm": 1.7734375,
|
||
|
|
"learning_rate": 3.7800000000000002e-06,
|
||
|
|
"loss": 0.8405,
|
||
|
|
"mean_token_accuracy": 0.7998461078852415,
|
||
|
|
"num_tokens": 5415925.0,
|
||
|
|
"step": 190
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7316340124234557,
|
||
|
|
"epoch": 0.32,
|
||
|
|
"grad_norm": 1.6953125,
|
||
|
|
"learning_rate": 3.980000000000001e-06,
|
||
|
|
"loss": 0.7942,
|
||
|
|
"mean_token_accuracy": 0.8076768793165684,
|
||
|
|
"num_tokens": 5710178.0,
|
||
|
|
"step": 200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.32,
|
||
|
|
"eval_math_entropy": 0.7726770024299622,
|
||
|
|
"eval_math_loss": 0.84871506690979,
|
||
|
|
"eval_math_mean_token_accuracy": 0.7979933972358704,
|
||
|
|
"eval_math_num_tokens": 5710178.0,
|
||
|
|
"eval_math_runtime": 35.765,
|
||
|
|
"eval_math_samples_per_second": 13.98,
|
||
|
|
"eval_math_steps_per_second": 3.495,
|
||
|
|
"step": 200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.32,
|
||
|
|
"eval_physics_entropy": 0.6965276994705201,
|
||
|
|
"eval_physics_loss": 0.7036627531051636,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.820941798210144,
|
||
|
|
"eval_physics_num_tokens": 5710178.0,
|
||
|
|
"eval_physics_runtime": 41.4597,
|
||
|
|
"eval_physics_samples_per_second": 12.06,
|
||
|
|
"eval_physics_steps_per_second": 3.015,
|
||
|
|
"step": 200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7737893080338836,
|
||
|
|
"epoch": 0.336,
|
||
|
|
"grad_norm": 1.828125,
|
||
|
|
"learning_rate": 4.18e-06,
|
||
|
|
"loss": 0.8126,
|
||
|
|
"mean_token_accuracy": 0.800426610186696,
|
||
|
|
"num_tokens": 5993843.0,
|
||
|
|
"step": 210
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7482354704290628,
|
||
|
|
"epoch": 0.352,
|
||
|
|
"grad_norm": 1.9609375,
|
||
|
|
"learning_rate": 4.38e-06,
|
||
|
|
"loss": 0.7874,
|
||
|
|
"mean_token_accuracy": 0.8061643201857805,
|
||
|
|
"num_tokens": 6279137.0,
|
||
|
|
"step": 220
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7608217550441623,
|
||
|
|
"epoch": 0.368,
|
||
|
|
"grad_norm": 1.65625,
|
||
|
|
"learning_rate": 4.58e-06,
|
||
|
|
"loss": 0.7742,
|
||
|
|
"mean_token_accuracy": 0.803000221401453,
|
||
|
|
"num_tokens": 6562387.0,
|
||
|
|
"step": 230
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.748550821095705,
|
||
|
|
"epoch": 0.384,
|
||
|
|
"grad_norm": 1.84375,
|
||
|
|
"learning_rate": 4.78e-06,
|
||
|
|
"loss": 0.7565,
|
||
|
|
"mean_token_accuracy": 0.8081529751420021,
|
||
|
|
"num_tokens": 6851520.0,
|
||
|
|
"step": 240
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7808488950133323,
|
||
|
|
"epoch": 0.4,
|
||
|
|
"grad_norm": 1.75,
|
||
|
|
"learning_rate": 4.980000000000001e-06,
|
||
|
|
"loss": 0.7953,
|
||
|
|
"mean_token_accuracy": 0.7999043371528387,
|
||
|
|
"num_tokens": 7138932.0,
|
||
|
|
"step": 250
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7674930268898607,
|
||
|
|
"epoch": 0.416,
|
||
|
|
"grad_norm": 1.6796875,
|
||
|
|
"learning_rate": 5.18e-06,
|
||
|
|
"loss": 0.7597,
|
||
|
|
"mean_token_accuracy": 0.8044768057763576,
|
||
|
|
"num_tokens": 7420720.0,
|
||
|
|
"step": 260
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7496220650151372,
|
||
|
|
"epoch": 0.432,
|
||
|
|
"grad_norm": 1.5390625,
|
||
|
|
"learning_rate": 5.380000000000001e-06,
|
||
|
|
"loss": 0.7468,
|
||
|
|
"mean_token_accuracy": 0.8073872365057468,
|
||
|
|
"num_tokens": 7695518.0,
|
||
|
|
"step": 270
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7439631385728717,
|
||
|
|
"epoch": 0.448,
|
||
|
|
"grad_norm": 2.21875,
|
||
|
|
"learning_rate": 5.580000000000001e-06,
|
||
|
|
"loss": 0.7339,
|
||
|
|
"mean_token_accuracy": 0.8076559219509363,
|
||
|
|
"num_tokens": 7983543.0,
|
||
|
|
"step": 280
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7457249015569687,
|
||
|
|
"epoch": 0.464,
|
||
|
|
"grad_norm": 2.65625,
|
||
|
|
"learning_rate": 5.78e-06,
|
||
|
|
"loss": 0.741,
|
||
|
|
"mean_token_accuracy": 0.8087894931435585,
|
||
|
|
"num_tokens": 8271787.0,
|
||
|
|
"step": 290
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7761816097423434,
|
||
|
|
"epoch": 0.48,
|
||
|
|
"grad_norm": 1.78125,
|
||
|
|
"learning_rate": 5.98e-06,
|
||
|
|
"loss": 0.7548,
|
||
|
|
"mean_token_accuracy": 0.8063571337610483,
|
||
|
|
"num_tokens": 8556170.0,
|
||
|
|
"step": 300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.48,
|
||
|
|
"eval_math_entropy": 0.7653279647827148,
|
||
|
|
"eval_math_loss": 0.7636919021606445,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8056883907318115,
|
||
|
|
"eval_math_num_tokens": 8556170.0,
|
||
|
|
"eval_math_runtime": 35.8013,
|
||
|
|
"eval_math_samples_per_second": 13.966,
|
||
|
|
"eval_math_steps_per_second": 3.491,
|
||
|
|
"step": 300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.48,
|
||
|
|
"eval_physics_entropy": 0.6890090732574463,
|
||
|
|
"eval_physics_loss": 0.6574941277503967,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8227382025718689,
|
||
|
|
"eval_physics_num_tokens": 8556170.0,
|
||
|
|
"eval_physics_runtime": 41.4883,
|
||
|
|
"eval_physics_samples_per_second": 12.052,
|
||
|
|
"eval_physics_steps_per_second": 3.013,
|
||
|
|
"step": 300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7349074123427272,
|
||
|
|
"epoch": 0.496,
|
||
|
|
"grad_norm": 1.828125,
|
||
|
|
"learning_rate": 6.18e-06,
|
||
|
|
"loss": 0.7172,
|
||
|
|
"mean_token_accuracy": 0.8132212206721305,
|
||
|
|
"num_tokens": 8840590.0,
|
||
|
|
"step": 310
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7443849379196763,
|
||
|
|
"epoch": 0.512,
|
||
|
|
"grad_norm": 1.6640625,
|
||
|
|
"learning_rate": 6.380000000000001e-06,
|
||
|
|
"loss": 0.7181,
|
||
|
|
"mean_token_accuracy": 0.812743404135108,
|
||
|
|
"num_tokens": 9131903.0,
|
||
|
|
"step": 320
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7170431386679411,
|
||
|
|
"epoch": 0.528,
|
||
|
|
"grad_norm": 1.7109375,
|
||
|
|
"learning_rate": 6.5800000000000005e-06,
|
||
|
|
"loss": 0.7045,
|
||
|
|
"mean_token_accuracy": 0.8160003747791051,
|
||
|
|
"num_tokens": 9429519.0,
|
||
|
|
"step": 330
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7328394843265414,
|
||
|
|
"epoch": 0.544,
|
||
|
|
"grad_norm": 2.234375,
|
||
|
|
"learning_rate": 6.780000000000001e-06,
|
||
|
|
"loss": 0.722,
|
||
|
|
"mean_token_accuracy": 0.8143766384571791,
|
||
|
|
"num_tokens": 9718247.0,
|
||
|
|
"step": 340
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7620531113818287,
|
||
|
|
"epoch": 0.56,
|
||
|
|
"grad_norm": 2.453125,
|
||
|
|
"learning_rate": 6.98e-06,
|
||
|
|
"loss": 0.7395,
|
||
|
|
"mean_token_accuracy": 0.8070962373167276,
|
||
|
|
"num_tokens": 9991100.0,
|
||
|
|
"step": 350
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7331368897110224,
|
||
|
|
"epoch": 0.576,
|
||
|
|
"grad_norm": 1.8515625,
|
||
|
|
"learning_rate": 7.180000000000001e-06,
|
||
|
|
"loss": 0.7173,
|
||
|
|
"mean_token_accuracy": 0.8115411166101694,
|
||
|
|
"num_tokens": 10276090.0,
|
||
|
|
"step": 360
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7342592483386398,
|
||
|
|
"epoch": 0.592,
|
||
|
|
"grad_norm": 1.546875,
|
||
|
|
"learning_rate": 7.3800000000000005e-06,
|
||
|
|
"loss": 0.7141,
|
||
|
|
"mean_token_accuracy": 0.8124774299561978,
|
||
|
|
"num_tokens": 10566762.0,
|
||
|
|
"step": 370
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7440765531733632,
|
||
|
|
"epoch": 0.608,
|
||
|
|
"grad_norm": 1.59375,
|
||
|
|
"learning_rate": 7.58e-06,
|
||
|
|
"loss": 0.7237,
|
||
|
|
"mean_token_accuracy": 0.8122259072959424,
|
||
|
|
"num_tokens": 10857002.0,
|
||
|
|
"step": 380
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7216648861765862,
|
||
|
|
"epoch": 0.624,
|
||
|
|
"grad_norm": 2.25,
|
||
|
|
"learning_rate": 7.78e-06,
|
||
|
|
"loss": 0.7057,
|
||
|
|
"mean_token_accuracy": 0.8139940846711398,
|
||
|
|
"num_tokens": 11147326.0,
|
||
|
|
"step": 390
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7119627617299557,
|
||
|
|
"epoch": 0.64,
|
||
|
|
"grad_norm": 1.5234375,
|
||
|
|
"learning_rate": 7.980000000000002e-06,
|
||
|
|
"loss": 0.6947,
|
||
|
|
"mean_token_accuracy": 0.8168974690139293,
|
||
|
|
"num_tokens": 11437710.0,
|
||
|
|
"step": 400
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.64,
|
||
|
|
"eval_math_entropy": 0.7429183626174927,
|
||
|
|
"eval_math_loss": 0.7330864071846008,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8096566972732544,
|
||
|
|
"eval_math_num_tokens": 11437710.0,
|
||
|
|
"eval_math_runtime": 35.7944,
|
||
|
|
"eval_math_samples_per_second": 13.969,
|
||
|
|
"eval_math_steps_per_second": 3.492,
|
||
|
|
"step": 400
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.64,
|
||
|
|
"eval_physics_entropy": 0.6820782516002655,
|
||
|
|
"eval_physics_loss": 0.6511842012405396,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8221303143501282,
|
||
|
|
"eval_physics_num_tokens": 11437710.0,
|
||
|
|
"eval_physics_runtime": 41.4584,
|
||
|
|
"eval_physics_samples_per_second": 12.06,
|
||
|
|
"eval_physics_steps_per_second": 3.015,
|
||
|
|
"step": 400
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.712121813185513,
|
||
|
|
"epoch": 0.656,
|
||
|
|
"grad_norm": 1.40625,
|
||
|
|
"learning_rate": 8.18e-06,
|
||
|
|
"loss": 0.6914,
|
||
|
|
"mean_token_accuracy": 0.8179911982268095,
|
||
|
|
"num_tokens": 11727268.0,
|
||
|
|
"step": 410
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7159143410623073,
|
||
|
|
"epoch": 0.672,
|
||
|
|
"grad_norm": 1.65625,
|
||
|
|
"learning_rate": 8.380000000000001e-06,
|
||
|
|
"loss": 0.7016,
|
||
|
|
"mean_token_accuracy": 0.813546035438776,
|
||
|
|
"num_tokens": 12020478.0,
|
||
|
|
"step": 420
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7149255678057671,
|
||
|
|
"epoch": 0.688,
|
||
|
|
"grad_norm": 1.5078125,
|
||
|
|
"learning_rate": 8.580000000000001e-06,
|
||
|
|
"loss": 0.6908,
|
||
|
|
"mean_token_accuracy": 0.8162215895950794,
|
||
|
|
"num_tokens": 12300380.0,
|
||
|
|
"step": 430
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7053035136312247,
|
||
|
|
"epoch": 0.704,
|
||
|
|
"grad_norm": 1.78125,
|
||
|
|
"learning_rate": 8.78e-06,
|
||
|
|
"loss": 0.6839,
|
||
|
|
"mean_token_accuracy": 0.8172036681324244,
|
||
|
|
"num_tokens": 12591366.0,
|
||
|
|
"step": 440
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7228720126673579,
|
||
|
|
"epoch": 0.72,
|
||
|
|
"grad_norm": 1.7421875,
|
||
|
|
"learning_rate": 8.98e-06,
|
||
|
|
"loss": 0.6945,
|
||
|
|
"mean_token_accuracy": 0.8145164046436548,
|
||
|
|
"num_tokens": 12869618.0,
|
||
|
|
"step": 450
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7052439017221331,
|
||
|
|
"epoch": 0.736,
|
||
|
|
"grad_norm": 1.75,
|
||
|
|
"learning_rate": 9.180000000000002e-06,
|
||
|
|
"loss": 0.6866,
|
||
|
|
"mean_token_accuracy": 0.8176059562712907,
|
||
|
|
"num_tokens": 13156311.0,
|
||
|
|
"step": 460
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.692553037032485,
|
||
|
|
"epoch": 0.752,
|
||
|
|
"grad_norm": 1.5625,
|
||
|
|
"learning_rate": 9.38e-06,
|
||
|
|
"loss": 0.6644,
|
||
|
|
"mean_token_accuracy": 0.821098442748189,
|
||
|
|
"num_tokens": 13438911.0,
|
||
|
|
"step": 470
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6714967370033265,
|
||
|
|
"epoch": 0.768,
|
||
|
|
"grad_norm": 1.3203125,
|
||
|
|
"learning_rate": 9.58e-06,
|
||
|
|
"loss": 0.6595,
|
||
|
|
"mean_token_accuracy": 0.825907539948821,
|
||
|
|
"num_tokens": 13735397.0,
|
||
|
|
"step": 480
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7148594724014401,
|
||
|
|
"epoch": 0.784,
|
||
|
|
"grad_norm": 1.7890625,
|
||
|
|
"learning_rate": 9.780000000000001e-06,
|
||
|
|
"loss": 0.6954,
|
||
|
|
"mean_token_accuracy": 0.8156244982033968,
|
||
|
|
"num_tokens": 14024497.0,
|
||
|
|
"step": 490
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7110015576705336,
|
||
|
|
"epoch": 0.8,
|
||
|
|
"grad_norm": 1.6328125,
|
||
|
|
"learning_rate": 9.980000000000001e-06,
|
||
|
|
"loss": 0.69,
|
||
|
|
"mean_token_accuracy": 0.8156023971736431,
|
||
|
|
"num_tokens": 14316427.0,
|
||
|
|
"step": 500
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.8,
|
||
|
|
"eval_math_entropy": 0.7275117125511169,
|
||
|
|
"eval_math_loss": 0.7106672525405884,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8132652158737183,
|
||
|
|
"eval_math_num_tokens": 14316427.0,
|
||
|
|
"eval_math_runtime": 35.7797,
|
||
|
|
"eval_math_samples_per_second": 13.974,
|
||
|
|
"eval_math_steps_per_second": 3.494,
|
||
|
|
"step": 500
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.8,
|
||
|
|
"eval_physics_entropy": 0.6752480027675629,
|
||
|
|
"eval_physics_loss": 0.6474179029464722,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8222041974067688,
|
||
|
|
"eval_physics_num_tokens": 14316427.0,
|
||
|
|
"eval_physics_runtime": 41.1592,
|
||
|
|
"eval_physics_samples_per_second": 12.148,
|
||
|
|
"eval_physics_steps_per_second": 3.037,
|
||
|
|
"step": 500
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7031571570783853,
|
||
|
|
"epoch": 0.816,
|
||
|
|
"grad_norm": 1.7421875,
|
||
|
|
"learning_rate": 1.018e-05,
|
||
|
|
"loss": 0.6756,
|
||
|
|
"mean_token_accuracy": 0.8171336345374585,
|
||
|
|
"num_tokens": 14607136.0,
|
||
|
|
"step": 510
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.693414007127285,
|
||
|
|
"epoch": 0.832,
|
||
|
|
"grad_norm": 1.5859375,
|
||
|
|
"learning_rate": 1.038e-05,
|
||
|
|
"loss": 0.6706,
|
||
|
|
"mean_token_accuracy": 0.8211553748697042,
|
||
|
|
"num_tokens": 14894084.0,
|
||
|
|
"step": 520
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7212638273835182,
|
||
|
|
"epoch": 0.848,
|
||
|
|
"grad_norm": 1.59375,
|
||
|
|
"learning_rate": 1.0580000000000002e-05,
|
||
|
|
"loss": 0.6965,
|
||
|
|
"mean_token_accuracy": 0.8145442601293326,
|
||
|
|
"num_tokens": 15177039.0,
|
||
|
|
"step": 530
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7037423193454743,
|
||
|
|
"epoch": 0.864,
|
||
|
|
"grad_norm": 1.4296875,
|
||
|
|
"learning_rate": 1.0780000000000002e-05,
|
||
|
|
"loss": 0.6824,
|
||
|
|
"mean_token_accuracy": 0.816459146887064,
|
||
|
|
"num_tokens": 15463393.0,
|
||
|
|
"step": 540
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7084541182965041,
|
||
|
|
"epoch": 0.88,
|
||
|
|
"grad_norm": 1.5546875,
|
||
|
|
"learning_rate": 1.0980000000000002e-05,
|
||
|
|
"loss": 0.6861,
|
||
|
|
"mean_token_accuracy": 0.8153413504362106,
|
||
|
|
"num_tokens": 15747943.0,
|
||
|
|
"step": 550
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7014794830232859,
|
||
|
|
"epoch": 0.896,
|
||
|
|
"grad_norm": 1.5078125,
|
||
|
|
"learning_rate": 1.1180000000000001e-05,
|
||
|
|
"loss": 0.6844,
|
||
|
|
"mean_token_accuracy": 0.8173291139304638,
|
||
|
|
"num_tokens": 16027931.0,
|
||
|
|
"step": 560
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6914853570982814,
|
||
|
|
"epoch": 0.912,
|
||
|
|
"grad_norm": 1.546875,
|
||
|
|
"learning_rate": 1.138e-05,
|
||
|
|
"loss": 0.6643,
|
||
|
|
"mean_token_accuracy": 0.8196637153625488,
|
||
|
|
"num_tokens": 16317127.0,
|
||
|
|
"step": 570
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6866581376641989,
|
||
|
|
"epoch": 0.928,
|
||
|
|
"grad_norm": 1.46875,
|
||
|
|
"learning_rate": 1.1580000000000001e-05,
|
||
|
|
"loss": 0.661,
|
||
|
|
"mean_token_accuracy": 0.8199340119957924,
|
||
|
|
"num_tokens": 16598757.0,
|
||
|
|
"step": 580
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7026519874110818,
|
||
|
|
"epoch": 0.944,
|
||
|
|
"grad_norm": 7.0625,
|
||
|
|
"learning_rate": 1.178e-05,
|
||
|
|
"loss": 0.6838,
|
||
|
|
"mean_token_accuracy": 0.8174264524132013,
|
||
|
|
"num_tokens": 16882236.0,
|
||
|
|
"step": 590
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7053765447810292,
|
||
|
|
"epoch": 0.96,
|
||
|
|
"grad_norm": 1.515625,
|
||
|
|
"learning_rate": 1.198e-05,
|
||
|
|
"loss": 0.6832,
|
||
|
|
"mean_token_accuracy": 0.8171489015221596,
|
||
|
|
"num_tokens": 17169495.0,
|
||
|
|
"step": 600
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.96,
|
||
|
|
"eval_math_entropy": 0.7025088634490967,
|
||
|
|
"eval_math_loss": 0.6926642656326294,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8157670006752015,
|
||
|
|
"eval_math_num_tokens": 17169495.0,
|
||
|
|
"eval_math_runtime": 35.7558,
|
||
|
|
"eval_math_samples_per_second": 13.984,
|
||
|
|
"eval_math_steps_per_second": 3.496,
|
||
|
|
"step": 600
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.96,
|
||
|
|
"eval_physics_entropy": 0.6707950072288513,
|
||
|
|
"eval_physics_loss": 0.6470632553100586,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8212249627113343,
|
||
|
|
"eval_physics_num_tokens": 17169495.0,
|
||
|
|
"eval_physics_runtime": 41.7459,
|
||
|
|
"eval_physics_samples_per_second": 11.977,
|
||
|
|
"eval_physics_steps_per_second": 2.994,
|
||
|
|
"step": 600
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.684873866662383,
|
||
|
|
"epoch": 0.976,
|
||
|
|
"grad_norm": 1.8515625,
|
||
|
|
"learning_rate": 1.218e-05,
|
||
|
|
"loss": 0.6671,
|
||
|
|
"mean_token_accuracy": 0.8196386702358722,
|
||
|
|
"num_tokens": 17454147.0,
|
||
|
|
"step": 610
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.7006405493244529,
|
||
|
|
"epoch": 0.992,
|
||
|
|
"grad_norm": 1.5390625,
|
||
|
|
"learning_rate": 1.2380000000000002e-05,
|
||
|
|
"loss": 0.672,
|
||
|
|
"mean_token_accuracy": 0.8165230687707663,
|
||
|
|
"num_tokens": 17742596.0,
|
||
|
|
"step": 620
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6851861752569676,
|
||
|
|
"epoch": 1.008,
|
||
|
|
"grad_norm": 1.3828125,
|
||
|
|
"learning_rate": 1.2580000000000002e-05,
|
||
|
|
"loss": 0.6635,
|
||
|
|
"mean_token_accuracy": 0.8197994366288185,
|
||
|
|
"num_tokens": 18033540.0,
|
||
|
|
"step": 630
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6624577900394797,
|
||
|
|
"epoch": 1.024,
|
||
|
|
"grad_norm": 1.5,
|
||
|
|
"learning_rate": 1.2780000000000001e-05,
|
||
|
|
"loss": 0.6405,
|
||
|
|
"mean_token_accuracy": 0.823651035502553,
|
||
|
|
"num_tokens": 18316346.0,
|
||
|
|
"step": 640
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6590202970430254,
|
||
|
|
"epoch": 1.04,
|
||
|
|
"grad_norm": 1.6484375,
|
||
|
|
"learning_rate": 1.2980000000000001e-05,
|
||
|
|
"loss": 0.6392,
|
||
|
|
"mean_token_accuracy": 0.8249083694070578,
|
||
|
|
"num_tokens": 18595230.0,
|
||
|
|
"step": 650
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6720681777223945,
|
||
|
|
"epoch": 1.056,
|
||
|
|
"grad_norm": 1.5625,
|
||
|
|
"learning_rate": 1.3180000000000001e-05,
|
||
|
|
"loss": 0.6481,
|
||
|
|
"mean_token_accuracy": 0.8219482194632292,
|
||
|
|
"num_tokens": 18885547.0,
|
||
|
|
"step": 660
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.666364430822432,
|
||
|
|
"epoch": 1.072,
|
||
|
|
"grad_norm": 1.5625,
|
||
|
|
"learning_rate": 1.3380000000000002e-05,
|
||
|
|
"loss": 0.6453,
|
||
|
|
"mean_token_accuracy": 0.821395493671298,
|
||
|
|
"num_tokens": 19169083.0,
|
||
|
|
"step": 670
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6612174535170198,
|
||
|
|
"epoch": 1.088,
|
||
|
|
"grad_norm": 1.515625,
|
||
|
|
"learning_rate": 1.3580000000000002e-05,
|
||
|
|
"loss": 0.6328,
|
||
|
|
"mean_token_accuracy": 0.823925781622529,
|
||
|
|
"num_tokens": 19461465.0,
|
||
|
|
"step": 680
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6591472556814552,
|
||
|
|
"epoch": 1.104,
|
||
|
|
"grad_norm": 1.3125,
|
||
|
|
"learning_rate": 1.378e-05,
|
||
|
|
"loss": 0.6414,
|
||
|
|
"mean_token_accuracy": 0.8235542017966508,
|
||
|
|
"num_tokens": 19755027.0,
|
||
|
|
"step": 690
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6446556953713298,
|
||
|
|
"epoch": 1.12,
|
||
|
|
"grad_norm": 1.5390625,
|
||
|
|
"learning_rate": 1.398e-05,
|
||
|
|
"loss": 0.6235,
|
||
|
|
"mean_token_accuracy": 0.8280277159065008,
|
||
|
|
"num_tokens": 20043293.0,
|
||
|
|
"step": 700
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.12,
|
||
|
|
"eval_math_entropy": 0.6619428229331971,
|
||
|
|
"eval_math_loss": 0.6814246773719788,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8176272907257081,
|
||
|
|
"eval_math_num_tokens": 20043293.0,
|
||
|
|
"eval_math_runtime": 35.7663,
|
||
|
|
"eval_math_samples_per_second": 13.98,
|
||
|
|
"eval_math_steps_per_second": 3.495,
|
||
|
|
"step": 700
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.12,
|
||
|
|
"eval_physics_entropy": 0.6304002203941346,
|
||
|
|
"eval_physics_loss": 0.6498542428016663,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8209347081184387,
|
||
|
|
"eval_physics_num_tokens": 20043293.0,
|
||
|
|
"eval_physics_runtime": 41.4297,
|
||
|
|
"eval_physics_samples_per_second": 12.069,
|
||
|
|
"eval_physics_steps_per_second": 3.017,
|
||
|
|
"step": 700
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6416833782568574,
|
||
|
|
"epoch": 1.1360000000000001,
|
||
|
|
"grad_norm": 1.46875,
|
||
|
|
"learning_rate": 1.418e-05,
|
||
|
|
"loss": 0.6298,
|
||
|
|
"mean_token_accuracy": 0.8263315539807081,
|
||
|
|
"num_tokens": 20323010.0,
|
||
|
|
"step": 710
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6587465461343527,
|
||
|
|
"epoch": 1.152,
|
||
|
|
"grad_norm": 1.578125,
|
||
|
|
"learning_rate": 1.4380000000000001e-05,
|
||
|
|
"loss": 0.6376,
|
||
|
|
"mean_token_accuracy": 0.8255084995180368,
|
||
|
|
"num_tokens": 20613266.0,
|
||
|
|
"step": 720
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6476276513189078,
|
||
|
|
"epoch": 1.168,
|
||
|
|
"grad_norm": 1.40625,
|
||
|
|
"learning_rate": 1.4580000000000001e-05,
|
||
|
|
"loss": 0.6316,
|
||
|
|
"mean_token_accuracy": 0.8264493472874165,
|
||
|
|
"num_tokens": 20910404.0,
|
||
|
|
"step": 730
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6569468161091209,
|
||
|
|
"epoch": 1.184,
|
||
|
|
"grad_norm": 1.546875,
|
||
|
|
"learning_rate": 1.478e-05,
|
||
|
|
"loss": 0.6336,
|
||
|
|
"mean_token_accuracy": 0.8247423082590103,
|
||
|
|
"num_tokens": 21200197.0,
|
||
|
|
"step": 740
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6399336410686374,
|
||
|
|
"epoch": 1.2,
|
||
|
|
"grad_norm": 1.40625,
|
||
|
|
"learning_rate": 1.498e-05,
|
||
|
|
"loss": 0.622,
|
||
|
|
"mean_token_accuracy": 0.8271243523806333,
|
||
|
|
"num_tokens": 21496441.0,
|
||
|
|
"step": 750
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6582916142418981,
|
||
|
|
"epoch": 1.216,
|
||
|
|
"grad_norm": 1.609375,
|
||
|
|
"learning_rate": 1.5180000000000002e-05,
|
||
|
|
"loss": 0.6353,
|
||
|
|
"mean_token_accuracy": 0.8234689589589834,
|
||
|
|
"num_tokens": 21778030.0,
|
||
|
|
"step": 760
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.637734330072999,
|
||
|
|
"epoch": 1.232,
|
||
|
|
"grad_norm": 1.5703125,
|
||
|
|
"learning_rate": 1.5380000000000002e-05,
|
||
|
|
"loss": 0.6159,
|
||
|
|
"mean_token_accuracy": 0.8288460947573185,
|
||
|
|
"num_tokens": 22060851.0,
|
||
|
|
"step": 770
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.636210915632546,
|
||
|
|
"epoch": 1.248,
|
||
|
|
"grad_norm": 1.5078125,
|
||
|
|
"learning_rate": 1.5580000000000003e-05,
|
||
|
|
"loss": 0.6165,
|
||
|
|
"mean_token_accuracy": 0.82775485701859,
|
||
|
|
"num_tokens": 22348738.0,
|
||
|
|
"step": 780
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6564677763730288,
|
||
|
|
"epoch": 1.264,
|
||
|
|
"grad_norm": 1.3671875,
|
||
|
|
"learning_rate": 1.578e-05,
|
||
|
|
"loss": 0.6349,
|
||
|
|
"mean_token_accuracy": 0.8237839676439762,
|
||
|
|
"num_tokens": 22635039.0,
|
||
|
|
"step": 790
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6571361353620887,
|
||
|
|
"epoch": 1.28,
|
||
|
|
"grad_norm": 1.421875,
|
||
|
|
"learning_rate": 1.5980000000000003e-05,
|
||
|
|
"loss": 0.6366,
|
||
|
|
"mean_token_accuracy": 0.8244204126298428,
|
||
|
|
"num_tokens": 22915190.0,
|
||
|
|
"step": 800
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.28,
|
||
|
|
"eval_math_entropy": 0.6791466112136841,
|
||
|
|
"eval_math_loss": 0.669770359992981,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8196616005897522,
|
||
|
|
"eval_math_num_tokens": 22915190.0,
|
||
|
|
"eval_math_runtime": 35.7574,
|
||
|
|
"eval_math_samples_per_second": 13.983,
|
||
|
|
"eval_math_steps_per_second": 3.496,
|
||
|
|
"step": 800
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.28,
|
||
|
|
"eval_physics_entropy": 0.6547831742763519,
|
||
|
|
"eval_physics_loss": 0.650165319442749,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.82009166097641,
|
||
|
|
"eval_physics_num_tokens": 22915190.0,
|
||
|
|
"eval_physics_runtime": 41.4354,
|
||
|
|
"eval_physics_samples_per_second": 12.067,
|
||
|
|
"eval_physics_steps_per_second": 3.017,
|
||
|
|
"step": 800
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6429510954767466,
|
||
|
|
"epoch": 1.296,
|
||
|
|
"grad_norm": 1.40625,
|
||
|
|
"learning_rate": 1.618e-05,
|
||
|
|
"loss": 0.6172,
|
||
|
|
"mean_token_accuracy": 0.8271825641393662,
|
||
|
|
"num_tokens": 23194947.0,
|
||
|
|
"step": 810
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6526322312653065,
|
||
|
|
"epoch": 1.312,
|
||
|
|
"grad_norm": 1.5390625,
|
||
|
|
"learning_rate": 1.638e-05,
|
||
|
|
"loss": 0.6414,
|
||
|
|
"mean_token_accuracy": 0.8243110898882151,
|
||
|
|
"num_tokens": 23483496.0,
|
||
|
|
"step": 820
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6679502466693521,
|
||
|
|
"epoch": 1.328,
|
||
|
|
"grad_norm": 1.6171875,
|
||
|
|
"learning_rate": 1.658e-05,
|
||
|
|
"loss": 0.6441,
|
||
|
|
"mean_token_accuracy": 0.8223706830292941,
|
||
|
|
"num_tokens": 23762589.0,
|
||
|
|
"step": 830
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6403571154922247,
|
||
|
|
"epoch": 1.3439999999999999,
|
||
|
|
"grad_norm": 1.4921875,
|
||
|
|
"learning_rate": 1.6780000000000002e-05,
|
||
|
|
"loss": 0.6222,
|
||
|
|
"mean_token_accuracy": 0.8268204662948847,
|
||
|
|
"num_tokens": 24041497.0,
|
||
|
|
"step": 840
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6487369466573,
|
||
|
|
"epoch": 1.3599999999999999,
|
||
|
|
"grad_norm": 1.203125,
|
||
|
|
"learning_rate": 1.698e-05,
|
||
|
|
"loss": 0.6282,
|
||
|
|
"mean_token_accuracy": 0.8273422639817,
|
||
|
|
"num_tokens": 24330591.0,
|
||
|
|
"step": 850
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.688577077165246,
|
||
|
|
"epoch": 1.376,
|
||
|
|
"grad_norm": 1.4453125,
|
||
|
|
"learning_rate": 1.718e-05,
|
||
|
|
"loss": 0.6651,
|
||
|
|
"mean_token_accuracy": 0.8164406325668097,
|
||
|
|
"num_tokens": 24611420.0,
|
||
|
|
"step": 860
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.643337251059711,
|
||
|
|
"epoch": 1.392,
|
||
|
|
"grad_norm": 1.4375,
|
||
|
|
"learning_rate": 1.7380000000000003e-05,
|
||
|
|
"loss": 0.6209,
|
||
|
|
"mean_token_accuracy": 0.825714922696352,
|
||
|
|
"num_tokens": 24894694.0,
|
||
|
|
"step": 870
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6371515756472945,
|
||
|
|
"epoch": 1.408,
|
||
|
|
"grad_norm": 1.453125,
|
||
|
|
"learning_rate": 1.758e-05,
|
||
|
|
"loss": 0.6156,
|
||
|
|
"mean_token_accuracy": 0.8278428696095943,
|
||
|
|
"num_tokens": 25184946.0,
|
||
|
|
"step": 880
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6312845427542925,
|
||
|
|
"epoch": 1.424,
|
||
|
|
"grad_norm": 1.25,
|
||
|
|
"learning_rate": 1.7780000000000003e-05,
|
||
|
|
"loss": 0.615,
|
||
|
|
"mean_token_accuracy": 0.8290112767368555,
|
||
|
|
"num_tokens": 25474746.0,
|
||
|
|
"step": 890
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6522130774334073,
|
||
|
|
"epoch": 1.44,
|
||
|
|
"grad_norm": 1.3671875,
|
||
|
|
"learning_rate": 1.798e-05,
|
||
|
|
"loss": 0.6295,
|
||
|
|
"mean_token_accuracy": 0.8248222030699253,
|
||
|
|
"num_tokens": 25758550.0,
|
||
|
|
"step": 900
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.44,
|
||
|
|
"eval_math_entropy": 0.6579540917873382,
|
||
|
|
"eval_math_loss": 0.6607276797294617,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8207882242202759,
|
||
|
|
"eval_math_num_tokens": 25758550.0,
|
||
|
|
"eval_math_runtime": 35.7442,
|
||
|
|
"eval_math_samples_per_second": 13.988,
|
||
|
|
"eval_math_steps_per_second": 3.497,
|
||
|
|
"step": 900
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.44,
|
||
|
|
"eval_physics_entropy": 0.6424285123348236,
|
||
|
|
"eval_physics_loss": 0.651524543762207,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8197560949325562,
|
||
|
|
"eval_physics_num_tokens": 25758550.0,
|
||
|
|
"eval_physics_runtime": 41.435,
|
||
|
|
"eval_physics_samples_per_second": 12.067,
|
||
|
|
"eval_physics_steps_per_second": 3.017,
|
||
|
|
"step": 900
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6359886834397912,
|
||
|
|
"epoch": 1.456,
|
||
|
|
"grad_norm": 1.3203125,
|
||
|
|
"learning_rate": 1.8180000000000002e-05,
|
||
|
|
"loss": 0.6177,
|
||
|
|
"mean_token_accuracy": 0.8268891926854849,
|
||
|
|
"num_tokens": 26039862.0,
|
||
|
|
"step": 910
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6429944572970271,
|
||
|
|
"epoch": 1.472,
|
||
|
|
"grad_norm": 1.28125,
|
||
|
|
"learning_rate": 1.8380000000000004e-05,
|
||
|
|
"loss": 0.6266,
|
||
|
|
"mean_token_accuracy": 0.8256016962230206,
|
||
|
|
"num_tokens": 26318666.0,
|
||
|
|
"step": 920
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.666933044232428,
|
||
|
|
"epoch": 1.488,
|
||
|
|
"grad_norm": 1.265625,
|
||
|
|
"learning_rate": 1.858e-05,
|
||
|
|
"loss": 0.6455,
|
||
|
|
"mean_token_accuracy": 0.8214054699987173,
|
||
|
|
"num_tokens": 26594867.0,
|
||
|
|
"step": 930
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6266062622889876,
|
||
|
|
"epoch": 1.504,
|
||
|
|
"grad_norm": 1.5,
|
||
|
|
"learning_rate": 1.878e-05,
|
||
|
|
"loss": 0.6092,
|
||
|
|
"mean_token_accuracy": 0.8303067740052938,
|
||
|
|
"num_tokens": 26887233.0,
|
||
|
|
"step": 940
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.659474827721715,
|
||
|
|
"epoch": 1.52,
|
||
|
|
"grad_norm": 1.140625,
|
||
|
|
"learning_rate": 1.898e-05,
|
||
|
|
"loss": 0.6344,
|
||
|
|
"mean_token_accuracy": 0.8236033659428358,
|
||
|
|
"num_tokens": 27177448.0,
|
||
|
|
"step": 950
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6371477670967579,
|
||
|
|
"epoch": 1.536,
|
||
|
|
"grad_norm": 1.1484375,
|
||
|
|
"learning_rate": 1.918e-05,
|
||
|
|
"loss": 0.6245,
|
||
|
|
"mean_token_accuracy": 0.8263927459716797,
|
||
|
|
"num_tokens": 27467857.0,
|
||
|
|
"step": 960
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6363119596615434,
|
||
|
|
"epoch": 1.552,
|
||
|
|
"grad_norm": 1.265625,
|
||
|
|
"learning_rate": 1.938e-05,
|
||
|
|
"loss": 0.6172,
|
||
|
|
"mean_token_accuracy": 0.8277775753289461,
|
||
|
|
"num_tokens": 27755138.0,
|
||
|
|
"step": 970
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.625205036252737,
|
||
|
|
"epoch": 1.568,
|
||
|
|
"grad_norm": 1.15625,
|
||
|
|
"learning_rate": 1.9580000000000002e-05,
|
||
|
|
"loss": 0.6047,
|
||
|
|
"mean_token_accuracy": 0.8324479281902313,
|
||
|
|
"num_tokens": 28049363.0,
|
||
|
|
"step": 980
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6341174798086285,
|
||
|
|
"epoch": 1.584,
|
||
|
|
"grad_norm": 1.375,
|
||
|
|
"learning_rate": 1.978e-05,
|
||
|
|
"loss": 0.6153,
|
||
|
|
"mean_token_accuracy": 0.8282931160181761,
|
||
|
|
"num_tokens": 28337890.0,
|
||
|
|
"step": 990
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6248919799923897,
|
||
|
|
"epoch": 1.6,
|
||
|
|
"grad_norm": 1.3359375,
|
||
|
|
"learning_rate": 1.9980000000000002e-05,
|
||
|
|
"loss": 0.6112,
|
||
|
|
"mean_token_accuracy": 0.827129240334034,
|
||
|
|
"num_tokens": 28626068.0,
|
||
|
|
"step": 1000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.6,
|
||
|
|
"eval_math_entropy": 0.6310200927257538,
|
||
|
|
"eval_math_loss": 0.651292085647583,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8225455293655396,
|
||
|
|
"eval_math_num_tokens": 28626068.0,
|
||
|
|
"eval_math_runtime": 35.7544,
|
||
|
|
"eval_math_samples_per_second": 13.984,
|
||
|
|
"eval_math_steps_per_second": 3.496,
|
||
|
|
"step": 1000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.6,
|
||
|
|
"eval_physics_entropy": 0.6265229551792145,
|
||
|
|
"eval_physics_loss": 0.6552274227142334,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8185877275466918,
|
||
|
|
"eval_physics_num_tokens": 28626068.0,
|
||
|
|
"eval_physics_runtime": 41.1356,
|
||
|
|
"eval_physics_samples_per_second": 12.155,
|
||
|
|
"eval_physics_steps_per_second": 3.039,
|
||
|
|
"step": 1000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6397603360936046,
|
||
|
|
"epoch": 1.616,
|
||
|
|
"grad_norm": 1.25,
|
||
|
|
"learning_rate": 1.9980000000000002e-05,
|
||
|
|
"loss": 0.6266,
|
||
|
|
"mean_token_accuracy": 0.8261015623807907,
|
||
|
|
"num_tokens": 28913208.0,
|
||
|
|
"step": 1010
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6517996810376644,
|
||
|
|
"epoch": 1.6320000000000001,
|
||
|
|
"grad_norm": 1.265625,
|
||
|
|
"learning_rate": 1.995777777777778e-05,
|
||
|
|
"loss": 0.6284,
|
||
|
|
"mean_token_accuracy": 0.8228654317557812,
|
||
|
|
"num_tokens": 29188917.0,
|
||
|
|
"step": 1020
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6239277983084321,
|
||
|
|
"epoch": 1.6480000000000001,
|
||
|
|
"grad_norm": 1.2421875,
|
||
|
|
"learning_rate": 1.9935555555555557e-05,
|
||
|
|
"loss": 0.6068,
|
||
|
|
"mean_token_accuracy": 0.8318834796547889,
|
||
|
|
"num_tokens": 29477292.0,
|
||
|
|
"step": 1030
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6378339532762766,
|
||
|
|
"epoch": 1.6640000000000001,
|
||
|
|
"grad_norm": 1.2109375,
|
||
|
|
"learning_rate": 1.9913333333333335e-05,
|
||
|
|
"loss": 0.6129,
|
||
|
|
"mean_token_accuracy": 0.8274775009602309,
|
||
|
|
"num_tokens": 29772480.0,
|
||
|
|
"step": 1040
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.64110322073102,
|
||
|
|
"epoch": 1.6800000000000002,
|
||
|
|
"grad_norm": 1.15625,
|
||
|
|
"learning_rate": 1.9891111111111112e-05,
|
||
|
|
"loss": 0.6244,
|
||
|
|
"mean_token_accuracy": 0.8272217728197575,
|
||
|
|
"num_tokens": 30061007.0,
|
||
|
|
"step": 1050
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6449864653870463,
|
||
|
|
"epoch": 1.696,
|
||
|
|
"grad_norm": 1.1875,
|
||
|
|
"learning_rate": 1.986888888888889e-05,
|
||
|
|
"loss": 0.6243,
|
||
|
|
"mean_token_accuracy": 0.8258783508092165,
|
||
|
|
"num_tokens": 30344415.0,
|
||
|
|
"step": 1060
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6212725475430488,
|
||
|
|
"epoch": 1.712,
|
||
|
|
"grad_norm": 1.1640625,
|
||
|
|
"learning_rate": 1.9846666666666668e-05,
|
||
|
|
"loss": 0.6076,
|
||
|
|
"mean_token_accuracy": 0.8308158092200756,
|
||
|
|
"num_tokens": 30640891.0,
|
||
|
|
"step": 1070
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.618477713689208,
|
||
|
|
"epoch": 1.728,
|
||
|
|
"grad_norm": 1.2109375,
|
||
|
|
"learning_rate": 1.9824444444444445e-05,
|
||
|
|
"loss": 0.5981,
|
||
|
|
"mean_token_accuracy": 0.8317991364747286,
|
||
|
|
"num_tokens": 30921295.0,
|
||
|
|
"step": 1080
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6253316847607493,
|
||
|
|
"epoch": 1.744,
|
||
|
|
"grad_norm": 1.3125,
|
||
|
|
"learning_rate": 1.9802222222222226e-05,
|
||
|
|
"loss": 0.608,
|
||
|
|
"mean_token_accuracy": 0.8302719250321389,
|
||
|
|
"num_tokens": 31206375.0,
|
||
|
|
"step": 1090
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6203237606212497,
|
||
|
|
"epoch": 1.76,
|
||
|
|
"grad_norm": 1.140625,
|
||
|
|
"learning_rate": 1.978e-05,
|
||
|
|
"loss": 0.6016,
|
||
|
|
"mean_token_accuracy": 0.829212823137641,
|
||
|
|
"num_tokens": 31491856.0,
|
||
|
|
"step": 1100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.76,
|
||
|
|
"eval_math_entropy": 0.6470205163955689,
|
||
|
|
"eval_math_loss": 0.6425853371620178,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8245493578910827,
|
||
|
|
"eval_math_num_tokens": 31491856.0,
|
||
|
|
"eval_math_runtime": 35.7795,
|
||
|
|
"eval_math_samples_per_second": 13.974,
|
||
|
|
"eval_math_steps_per_second": 3.494,
|
||
|
|
"step": 1100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.76,
|
||
|
|
"eval_physics_entropy": 0.6618543210029602,
|
||
|
|
"eval_physics_loss": 0.6551723480224609,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8183454179763794,
|
||
|
|
"eval_physics_num_tokens": 31491856.0,
|
||
|
|
"eval_physics_runtime": 41.448,
|
||
|
|
"eval_physics_samples_per_second": 12.063,
|
||
|
|
"eval_physics_steps_per_second": 3.016,
|
||
|
|
"step": 1100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6285459484905005,
|
||
|
|
"epoch": 1.776,
|
||
|
|
"grad_norm": 1.203125,
|
||
|
|
"learning_rate": 1.975777777777778e-05,
|
||
|
|
"loss": 0.608,
|
||
|
|
"mean_token_accuracy": 0.8288080658763647,
|
||
|
|
"num_tokens": 31782829.0,
|
||
|
|
"step": 1110
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6362737115472555,
|
||
|
|
"epoch": 1.792,
|
||
|
|
"grad_norm": 1.1484375,
|
||
|
|
"learning_rate": 1.9735555555555556e-05,
|
||
|
|
"loss": 0.6168,
|
||
|
|
"mean_token_accuracy": 0.8268269985914231,
|
||
|
|
"num_tokens": 32070187.0,
|
||
|
|
"step": 1120
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6320221189409494,
|
||
|
|
"epoch": 1.808,
|
||
|
|
"grad_norm": 1.203125,
|
||
|
|
"learning_rate": 1.9713333333333337e-05,
|
||
|
|
"loss": 0.6148,
|
||
|
|
"mean_token_accuracy": 0.8273419640958309,
|
||
|
|
"num_tokens": 32355072.0,
|
||
|
|
"step": 1130
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6012085109949112,
|
||
|
|
"epoch": 1.8239999999999998,
|
||
|
|
"grad_norm": 1.0859375,
|
||
|
|
"learning_rate": 1.969111111111111e-05,
|
||
|
|
"loss": 0.5802,
|
||
|
|
"mean_token_accuracy": 0.8352209012955427,
|
||
|
|
"num_tokens": 32634013.0,
|
||
|
|
"step": 1140
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6171128546819091,
|
||
|
|
"epoch": 1.8399999999999999,
|
||
|
|
"grad_norm": 1.328125,
|
||
|
|
"learning_rate": 1.9668888888888892e-05,
|
||
|
|
"loss": 0.6044,
|
||
|
|
"mean_token_accuracy": 0.8311630170792341,
|
||
|
|
"num_tokens": 32917222.0,
|
||
|
|
"step": 1150
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6315351065248251,
|
||
|
|
"epoch": 1.8559999999999999,
|
||
|
|
"grad_norm": 1.171875,
|
||
|
|
"learning_rate": 1.9646666666666666e-05,
|
||
|
|
"loss": 0.6066,
|
||
|
|
"mean_token_accuracy": 0.8292891420423985,
|
||
|
|
"num_tokens": 33207147.0,
|
||
|
|
"step": 1160
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6153346884995698,
|
||
|
|
"epoch": 1.8719999999999999,
|
||
|
|
"grad_norm": 1.234375,
|
||
|
|
"learning_rate": 1.9624444444444447e-05,
|
||
|
|
"loss": 0.5954,
|
||
|
|
"mean_token_accuracy": 0.8320576064288616,
|
||
|
|
"num_tokens": 33496572.0,
|
||
|
|
"step": 1170
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.624875632673502,
|
||
|
|
"epoch": 1.888,
|
||
|
|
"grad_norm": 1.203125,
|
||
|
|
"learning_rate": 1.9602222222222225e-05,
|
||
|
|
"loss": 0.6059,
|
||
|
|
"mean_token_accuracy": 0.8299812860786915,
|
||
|
|
"num_tokens": 33775788.0,
|
||
|
|
"step": 1180
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6256145890802145,
|
||
|
|
"epoch": 1.904,
|
||
|
|
"grad_norm": 1.1640625,
|
||
|
|
"learning_rate": 1.9580000000000002e-05,
|
||
|
|
"loss": 0.6084,
|
||
|
|
"mean_token_accuracy": 0.8293462209403515,
|
||
|
|
"num_tokens": 34052322.0,
|
||
|
|
"step": 1190
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6057112148031593,
|
||
|
|
"epoch": 1.92,
|
||
|
|
"grad_norm": 1.109375,
|
||
|
|
"learning_rate": 1.955777777777778e-05,
|
||
|
|
"loss": 0.5888,
|
||
|
|
"mean_token_accuracy": 0.8329556312412023,
|
||
|
|
"num_tokens": 34331468.0,
|
||
|
|
"step": 1200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.92,
|
||
|
|
"eval_math_entropy": 0.6444816522598267,
|
||
|
|
"eval_math_loss": 0.6337655782699585,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8254525041580201,
|
||
|
|
"eval_math_num_tokens": 34331468.0,
|
||
|
|
"eval_math_runtime": 35.7615,
|
||
|
|
"eval_math_samples_per_second": 13.982,
|
||
|
|
"eval_math_steps_per_second": 3.495,
|
||
|
|
"step": 1200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.92,
|
||
|
|
"eval_physics_entropy": 0.6525136830806733,
|
||
|
|
"eval_physics_loss": 0.6549710035324097,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8188423981666565,
|
||
|
|
"eval_physics_num_tokens": 34331468.0,
|
||
|
|
"eval_physics_runtime": 41.437,
|
||
|
|
"eval_physics_samples_per_second": 12.067,
|
||
|
|
"eval_physics_steps_per_second": 3.017,
|
||
|
|
"step": 1200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.622246976941824,
|
||
|
|
"epoch": 1.936,
|
||
|
|
"grad_norm": 1.2109375,
|
||
|
|
"learning_rate": 1.9535555555555557e-05,
|
||
|
|
"loss": 0.6052,
|
||
|
|
"mean_token_accuracy": 0.8293883047997952,
|
||
|
|
"num_tokens": 34612743.0,
|
||
|
|
"step": 1210
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6145633114501834,
|
||
|
|
"epoch": 1.952,
|
||
|
|
"grad_norm": 1.078125,
|
||
|
|
"learning_rate": 1.9513333333333335e-05,
|
||
|
|
"loss": 0.5899,
|
||
|
|
"mean_token_accuracy": 0.8315575309097767,
|
||
|
|
"num_tokens": 34891115.0,
|
||
|
|
"step": 1220
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6348631424829364,
|
||
|
|
"epoch": 1.968,
|
||
|
|
"grad_norm": 3.625,
|
||
|
|
"learning_rate": 1.9491111111111113e-05,
|
||
|
|
"loss": 0.6247,
|
||
|
|
"mean_token_accuracy": 0.8266950808465481,
|
||
|
|
"num_tokens": 35179505.0,
|
||
|
|
"step": 1230
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6054144280031324,
|
||
|
|
"epoch": 1.984,
|
||
|
|
"grad_norm": 1.0703125,
|
||
|
|
"learning_rate": 1.946888888888889e-05,
|
||
|
|
"loss": 0.5878,
|
||
|
|
"mean_token_accuracy": 0.8339846853166819,
|
||
|
|
"num_tokens": 35472596.0,
|
||
|
|
"step": 1240
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.614653673581779,
|
||
|
|
"epoch": 2.0,
|
||
|
|
"grad_norm": 1.109375,
|
||
|
|
"learning_rate": 1.9446666666666668e-05,
|
||
|
|
"loss": 0.5966,
|
||
|
|
"mean_token_accuracy": 0.8319407891482115,
|
||
|
|
"num_tokens": 35772848.0,
|
||
|
|
"step": 1250
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5460906764492393,
|
||
|
|
"epoch": 2.016,
|
||
|
|
"grad_norm": 1.3515625,
|
||
|
|
"learning_rate": 1.9424444444444446e-05,
|
||
|
|
"loss": 0.5207,
|
||
|
|
"mean_token_accuracy": 0.844809464737773,
|
||
|
|
"num_tokens": 36060330.0,
|
||
|
|
"step": 1260
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5267174040898681,
|
||
|
|
"epoch": 2.032,
|
||
|
|
"grad_norm": 1.3671875,
|
||
|
|
"learning_rate": 1.9402222222222223e-05,
|
||
|
|
"loss": 0.5098,
|
||
|
|
"mean_token_accuracy": 0.8475256130099297,
|
||
|
|
"num_tokens": 36343894.0,
|
||
|
|
"step": 1270
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5175346418283879,
|
||
|
|
"epoch": 2.048,
|
||
|
|
"grad_norm": 1.3671875,
|
||
|
|
"learning_rate": 1.938e-05,
|
||
|
|
"loss": 0.501,
|
||
|
|
"mean_token_accuracy": 0.8504213903099298,
|
||
|
|
"num_tokens": 36636410.0,
|
||
|
|
"step": 1280
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5137943865731358,
|
||
|
|
"epoch": 2.064,
|
||
|
|
"grad_norm": 1.28125,
|
||
|
|
"learning_rate": 1.935777777777778e-05,
|
||
|
|
"loss": 0.4962,
|
||
|
|
"mean_token_accuracy": 0.8507547531276941,
|
||
|
|
"num_tokens": 36925382.0,
|
||
|
|
"step": 1290
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5208146266639233,
|
||
|
|
"epoch": 2.08,
|
||
|
|
"grad_norm": 1.375,
|
||
|
|
"learning_rate": 1.9335555555555556e-05,
|
||
|
|
"loss": 0.4954,
|
||
|
|
"mean_token_accuracy": 0.849713422730565,
|
||
|
|
"num_tokens": 37208672.0,
|
||
|
|
"step": 1300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.08,
|
||
|
|
"eval_math_entropy": 0.5398842208385467,
|
||
|
|
"eval_math_loss": 0.6511635780334473,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8244616360664367,
|
||
|
|
"eval_math_num_tokens": 37208672.0,
|
||
|
|
"eval_math_runtime": 35.7792,
|
||
|
|
"eval_math_samples_per_second": 13.975,
|
||
|
|
"eval_math_steps_per_second": 3.494,
|
||
|
|
"step": 1300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.08,
|
||
|
|
"eval_physics_entropy": 0.5216713752746582,
|
||
|
|
"eval_physics_loss": 0.6896871328353882,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8157849154472351,
|
||
|
|
"eval_physics_num_tokens": 37208672.0,
|
||
|
|
"eval_physics_runtime": 41.4274,
|
||
|
|
"eval_physics_samples_per_second": 12.069,
|
||
|
|
"eval_physics_steps_per_second": 3.017,
|
||
|
|
"step": 1300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5262768002226949,
|
||
|
|
"epoch": 2.096,
|
||
|
|
"grad_norm": 1.359375,
|
||
|
|
"learning_rate": 1.9313333333333334e-05,
|
||
|
|
"loss": 0.5071,
|
||
|
|
"mean_token_accuracy": 0.8485719960182905,
|
||
|
|
"num_tokens": 37491954.0,
|
||
|
|
"step": 1310
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5144134562462568,
|
||
|
|
"epoch": 2.112,
|
||
|
|
"grad_norm": 1.390625,
|
||
|
|
"learning_rate": 1.9291111111111115e-05,
|
||
|
|
"loss": 0.4932,
|
||
|
|
"mean_token_accuracy": 0.8491108123213053,
|
||
|
|
"num_tokens": 37787033.0,
|
||
|
|
"step": 1320
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5125698197633028,
|
||
|
|
"epoch": 2.128,
|
||
|
|
"grad_norm": 1.390625,
|
||
|
|
"learning_rate": 1.926888888888889e-05,
|
||
|
|
"loss": 0.4936,
|
||
|
|
"mean_token_accuracy": 0.8515845112502575,
|
||
|
|
"num_tokens": 38068955.0,
|
||
|
|
"step": 1330
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5147779244929552,
|
||
|
|
"epoch": 2.144,
|
||
|
|
"grad_norm": 1.359375,
|
||
|
|
"learning_rate": 1.924666666666667e-05,
|
||
|
|
"loss": 0.4998,
|
||
|
|
"mean_token_accuracy": 0.8497466515749693,
|
||
|
|
"num_tokens": 38359914.0,
|
||
|
|
"step": 1340
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5150290129706263,
|
||
|
|
"epoch": 2.16,
|
||
|
|
"grad_norm": 1.28125,
|
||
|
|
"learning_rate": 1.9224444444444444e-05,
|
||
|
|
"loss": 0.4976,
|
||
|
|
"mean_token_accuracy": 0.8506554007530213,
|
||
|
|
"num_tokens": 38649684.0,
|
||
|
|
"step": 1350
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.49578714426606896,
|
||
|
|
"epoch": 2.176,
|
||
|
|
"grad_norm": 1.375,
|
||
|
|
"learning_rate": 1.9202222222222225e-05,
|
||
|
|
"loss": 0.4768,
|
||
|
|
"mean_token_accuracy": 0.8545470163226128,
|
||
|
|
"num_tokens": 38938155.0,
|
||
|
|
"step": 1360
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5279102198779583,
|
||
|
|
"epoch": 2.192,
|
||
|
|
"grad_norm": 1.2578125,
|
||
|
|
"learning_rate": 1.918e-05,
|
||
|
|
"loss": 0.5109,
|
||
|
|
"mean_token_accuracy": 0.8461268071085215,
|
||
|
|
"num_tokens": 39228800.0,
|
||
|
|
"step": 1370
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.520401804894209,
|
||
|
|
"epoch": 2.208,
|
||
|
|
"grad_norm": 1.4296875,
|
||
|
|
"learning_rate": 1.915777777777778e-05,
|
||
|
|
"loss": 0.5095,
|
||
|
|
"mean_token_accuracy": 0.8481683298945427,
|
||
|
|
"num_tokens": 39513118.0,
|
||
|
|
"step": 1380
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.517220887541771,
|
||
|
|
"epoch": 2.224,
|
||
|
|
"grad_norm": 1.3828125,
|
||
|
|
"learning_rate": 1.9135555555555555e-05,
|
||
|
|
"loss": 0.4947,
|
||
|
|
"mean_token_accuracy": 0.8505345996469259,
|
||
|
|
"num_tokens": 39804212.0,
|
||
|
|
"step": 1390
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5203588601201773,
|
||
|
|
"epoch": 2.24,
|
||
|
|
"grad_norm": 1.359375,
|
||
|
|
"learning_rate": 1.9113333333333336e-05,
|
||
|
|
"loss": 0.5071,
|
||
|
|
"mean_token_accuracy": 0.8482481308281422,
|
||
|
|
"num_tokens": 40092012.0,
|
||
|
|
"step": 1400
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.24,
|
||
|
|
"eval_math_entropy": 0.5498721191883087,
|
||
|
|
"eval_math_loss": 0.6494505405426025,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8248468618392945,
|
||
|
|
"eval_math_num_tokens": 40092012.0,
|
||
|
|
"eval_math_runtime": 35.7606,
|
||
|
|
"eval_math_samples_per_second": 13.982,
|
||
|
|
"eval_math_steps_per_second": 3.495,
|
||
|
|
"step": 1400
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.24,
|
||
|
|
"eval_physics_entropy": 0.5240244953632355,
|
||
|
|
"eval_physics_loss": 0.6975257992744446,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8148801970481873,
|
||
|
|
"eval_physics_num_tokens": 40092012.0,
|
||
|
|
"eval_physics_runtime": 41.7753,
|
||
|
|
"eval_physics_samples_per_second": 11.969,
|
||
|
|
"eval_physics_steps_per_second": 2.992,
|
||
|
|
"step": 1400
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5205017613247037,
|
||
|
|
"epoch": 2.2560000000000002,
|
||
|
|
"grad_norm": 1.3671875,
|
||
|
|
"learning_rate": 1.9091111111111113e-05,
|
||
|
|
"loss": 0.5042,
|
||
|
|
"mean_token_accuracy": 0.8483030397444964,
|
||
|
|
"num_tokens": 40380069.0,
|
||
|
|
"step": 1410
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5286959325894713,
|
||
|
|
"epoch": 2.2720000000000002,
|
||
|
|
"grad_norm": 1.5625,
|
||
|
|
"learning_rate": 1.906888888888889e-05,
|
||
|
|
"loss": 0.5089,
|
||
|
|
"mean_token_accuracy": 0.8469363551586866,
|
||
|
|
"num_tokens": 40655903.0,
|
||
|
|
"step": 1420
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5245801825076342,
|
||
|
|
"epoch": 2.288,
|
||
|
|
"grad_norm": 1.421875,
|
||
|
|
"learning_rate": 1.904666666666667e-05,
|
||
|
|
"loss": 0.505,
|
||
|
|
"mean_token_accuracy": 0.8467433623969555,
|
||
|
|
"num_tokens": 40944791.0,
|
||
|
|
"step": 1430
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5089570162817836,
|
||
|
|
"epoch": 2.304,
|
||
|
|
"grad_norm": 1.453125,
|
||
|
|
"learning_rate": 1.9024444444444446e-05,
|
||
|
|
"loss": 0.4927,
|
||
|
|
"mean_token_accuracy": 0.851346742734313,
|
||
|
|
"num_tokens": 41227146.0,
|
||
|
|
"step": 1440
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5277796785347164,
|
||
|
|
"epoch": 2.32,
|
||
|
|
"grad_norm": 1.4296875,
|
||
|
|
"learning_rate": 1.9002222222222224e-05,
|
||
|
|
"loss": 0.5143,
|
||
|
|
"mean_token_accuracy": 0.8476031333208084,
|
||
|
|
"num_tokens": 41509322.0,
|
||
|
|
"step": 1450
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5237356763333082,
|
||
|
|
"epoch": 2.336,
|
||
|
|
"grad_norm": 1.3984375,
|
||
|
|
"learning_rate": 1.898e-05,
|
||
|
|
"loss": 0.5015,
|
||
|
|
"mean_token_accuracy": 0.8484678287059069,
|
||
|
|
"num_tokens": 41792799.0,
|
||
|
|
"step": 1460
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5347706735134125,
|
||
|
|
"epoch": 2.352,
|
||
|
|
"grad_norm": 1.3828125,
|
||
|
|
"learning_rate": 1.895777777777778e-05,
|
||
|
|
"loss": 0.5169,
|
||
|
|
"mean_token_accuracy": 0.8446083668619394,
|
||
|
|
"num_tokens": 42083080.0,
|
||
|
|
"step": 1470
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5238756524398923,
|
||
|
|
"epoch": 2.368,
|
||
|
|
"grad_norm": 1.40625,
|
||
|
|
"learning_rate": 1.8935555555555556e-05,
|
||
|
|
"loss": 0.5074,
|
||
|
|
"mean_token_accuracy": 0.8478495571762323,
|
||
|
|
"num_tokens": 42368442.0,
|
||
|
|
"step": 1480
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5398916637524962,
|
||
|
|
"epoch": 2.384,
|
||
|
|
"grad_norm": 1.7734375,
|
||
|
|
"learning_rate": 1.8913333333333334e-05,
|
||
|
|
"loss": 0.5247,
|
||
|
|
"mean_token_accuracy": 0.8446283016353846,
|
||
|
|
"num_tokens": 42663926.0,
|
||
|
|
"step": 1490
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5313681181520223,
|
||
|
|
"epoch": 2.4,
|
||
|
|
"grad_norm": 1.6015625,
|
||
|
|
"learning_rate": 1.8891111111111115e-05,
|
||
|
|
"loss": 0.5162,
|
||
|
|
"mean_token_accuracy": 0.8469122361391783,
|
||
|
|
"num_tokens": 42946351.0,
|
||
|
|
"step": 1500
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.4,
|
||
|
|
"eval_math_entropy": 0.5470261330604553,
|
||
|
|
"eval_math_loss": 0.6441261172294617,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8258598875999451,
|
||
|
|
"eval_math_num_tokens": 42946351.0,
|
||
|
|
"eval_math_runtime": 35.7282,
|
||
|
|
"eval_math_samples_per_second": 13.995,
|
||
|
|
"eval_math_steps_per_second": 3.499,
|
||
|
|
"step": 1500
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.4,
|
||
|
|
"eval_physics_entropy": 0.5238613975048065,
|
||
|
|
"eval_physics_loss": 0.6977487802505493,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.81443390417099,
|
||
|
|
"eval_physics_num_tokens": 42946351.0,
|
||
|
|
"eval_physics_runtime": 41.457,
|
||
|
|
"eval_physics_samples_per_second": 12.061,
|
||
|
|
"eval_physics_steps_per_second": 3.015,
|
||
|
|
"step": 1500
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5377832099795341,
|
||
|
|
"epoch": 2.416,
|
||
|
|
"grad_norm": 1.4609375,
|
||
|
|
"learning_rate": 1.886888888888889e-05,
|
||
|
|
"loss": 0.5215,
|
||
|
|
"mean_token_accuracy": 0.8441088248044253,
|
||
|
|
"num_tokens": 43227782.0,
|
||
|
|
"step": 1510
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5219327821396291,
|
||
|
|
"epoch": 2.432,
|
||
|
|
"grad_norm": 1.3671875,
|
||
|
|
"learning_rate": 1.884666666666667e-05,
|
||
|
|
"loss": 0.5065,
|
||
|
|
"mean_token_accuracy": 0.8481958381831646,
|
||
|
|
"num_tokens": 43511939.0,
|
||
|
|
"step": 1520
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5146825276315212,
|
||
|
|
"epoch": 2.448,
|
||
|
|
"grad_norm": 1.53125,
|
||
|
|
"learning_rate": 1.8824444444444445e-05,
|
||
|
|
"loss": 0.4976,
|
||
|
|
"mean_token_accuracy": 0.8503648530691862,
|
||
|
|
"num_tokens": 43811005.0,
|
||
|
|
"step": 1530
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5279206298291683,
|
||
|
|
"epoch": 2.464,
|
||
|
|
"grad_norm": 1.484375,
|
||
|
|
"learning_rate": 1.8802222222222226e-05,
|
||
|
|
"loss": 0.5124,
|
||
|
|
"mean_token_accuracy": 0.8466742359101772,
|
||
|
|
"num_tokens": 44106223.0,
|
||
|
|
"step": 1540
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.548205547593534,
|
||
|
|
"epoch": 2.48,
|
||
|
|
"grad_norm": 1.515625,
|
||
|
|
"learning_rate": 1.878e-05,
|
||
|
|
"loss": 0.5272,
|
||
|
|
"mean_token_accuracy": 0.8421384803950787,
|
||
|
|
"num_tokens": 44379095.0,
|
||
|
|
"step": 1550
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5311242727562785,
|
||
|
|
"epoch": 2.496,
|
||
|
|
"grad_norm": 1.3515625,
|
||
|
|
"learning_rate": 1.875777777777778e-05,
|
||
|
|
"loss": 0.5146,
|
||
|
|
"mean_token_accuracy": 0.8455508589744568,
|
||
|
|
"num_tokens": 44670316.0,
|
||
|
|
"step": 1560
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5254273321479559,
|
||
|
|
"epoch": 2.512,
|
||
|
|
"grad_norm": 1.4765625,
|
||
|
|
"learning_rate": 1.873555555555556e-05,
|
||
|
|
"loss": 0.5098,
|
||
|
|
"mean_token_accuracy": 0.8476663187146187,
|
||
|
|
"num_tokens": 44952877.0,
|
||
|
|
"step": 1570
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5215749599039554,
|
||
|
|
"epoch": 2.528,
|
||
|
|
"grad_norm": 1.390625,
|
||
|
|
"learning_rate": 1.8713333333333336e-05,
|
||
|
|
"loss": 0.5036,
|
||
|
|
"mean_token_accuracy": 0.8488380990922451,
|
||
|
|
"num_tokens": 45237653.0,
|
||
|
|
"step": 1580
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5184456476941705,
|
||
|
|
"epoch": 2.544,
|
||
|
|
"grad_norm": 1.46875,
|
||
|
|
"learning_rate": 1.8691111111111114e-05,
|
||
|
|
"loss": 0.5051,
|
||
|
|
"mean_token_accuracy": 0.8494234204292297,
|
||
|
|
"num_tokens": 45524866.0,
|
||
|
|
"step": 1590
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5188256312161684,
|
||
|
|
"epoch": 2.56,
|
||
|
|
"grad_norm": 1.3359375,
|
||
|
|
"learning_rate": 1.866888888888889e-05,
|
||
|
|
"loss": 0.4994,
|
||
|
|
"mean_token_accuracy": 0.8480058163404465,
|
||
|
|
"num_tokens": 45803783.0,
|
||
|
|
"step": 1600
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.56,
|
||
|
|
"eval_math_entropy": 0.5415253615379334,
|
||
|
|
"eval_math_loss": 0.6402944326400757,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8268161854743957,
|
||
|
|
"eval_math_num_tokens": 45803783.0,
|
||
|
|
"eval_math_runtime": 35.8351,
|
||
|
|
"eval_math_samples_per_second": 13.953,
|
||
|
|
"eval_math_steps_per_second": 3.488,
|
||
|
|
"step": 1600
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.56,
|
||
|
|
"eval_physics_entropy": 0.5146132915019989,
|
||
|
|
"eval_physics_loss": 0.7024555802345276,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8143429985046386,
|
||
|
|
"eval_physics_num_tokens": 45803783.0,
|
||
|
|
"eval_physics_runtime": 41.4942,
|
||
|
|
"eval_physics_samples_per_second": 12.05,
|
||
|
|
"eval_physics_steps_per_second": 3.012,
|
||
|
|
"step": 1600
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5206789819523692,
|
||
|
|
"epoch": 2.576,
|
||
|
|
"grad_norm": 1.40625,
|
||
|
|
"learning_rate": 1.864666666666667e-05,
|
||
|
|
"loss": 0.5037,
|
||
|
|
"mean_token_accuracy": 0.8492428727447987,
|
||
|
|
"num_tokens": 46090713.0,
|
||
|
|
"step": 1610
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5139110781252384,
|
||
|
|
"epoch": 2.592,
|
||
|
|
"grad_norm": 1.25,
|
||
|
|
"learning_rate": 1.8624444444444446e-05,
|
||
|
|
"loss": 0.4983,
|
||
|
|
"mean_token_accuracy": 0.8500846106559038,
|
||
|
|
"num_tokens": 46384084.0,
|
||
|
|
"step": 1620
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5300323162227869,
|
||
|
|
"epoch": 2.608,
|
||
|
|
"grad_norm": 1.59375,
|
||
|
|
"learning_rate": 1.8602222222222224e-05,
|
||
|
|
"loss": 0.5166,
|
||
|
|
"mean_token_accuracy": 0.8454120934009552,
|
||
|
|
"num_tokens": 46667552.0,
|
||
|
|
"step": 1630
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.498782429471612,
|
||
|
|
"epoch": 2.624,
|
||
|
|
"grad_norm": 1.421875,
|
||
|
|
"learning_rate": 1.858e-05,
|
||
|
|
"loss": 0.4798,
|
||
|
|
"mean_token_accuracy": 0.8554852791130543,
|
||
|
|
"num_tokens": 46961124.0,
|
||
|
|
"step": 1640
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5273189852014184,
|
||
|
|
"epoch": 2.64,
|
||
|
|
"grad_norm": 1.46875,
|
||
|
|
"learning_rate": 1.855777777777778e-05,
|
||
|
|
"loss": 0.5138,
|
||
|
|
"mean_token_accuracy": 0.8468208122998476,
|
||
|
|
"num_tokens": 47245326.0,
|
||
|
|
"step": 1650
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5071797709912061,
|
||
|
|
"epoch": 2.656,
|
||
|
|
"grad_norm": 1.5,
|
||
|
|
"learning_rate": 1.8535555555555557e-05,
|
||
|
|
"loss": 0.4904,
|
||
|
|
"mean_token_accuracy": 0.8513832915574312,
|
||
|
|
"num_tokens": 47527347.0,
|
||
|
|
"step": 1660
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5276256233453751,
|
||
|
|
"epoch": 2.672,
|
||
|
|
"grad_norm": 1.6484375,
|
||
|
|
"learning_rate": 1.8513333333333335e-05,
|
||
|
|
"loss": 0.5105,
|
||
|
|
"mean_token_accuracy": 0.8466553438454867,
|
||
|
|
"num_tokens": 47808167.0,
|
||
|
|
"step": 1670
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5247279232367873,
|
||
|
|
"epoch": 2.6879999999999997,
|
||
|
|
"grad_norm": 1.390625,
|
||
|
|
"learning_rate": 1.8491111111111112e-05,
|
||
|
|
"loss": 0.5049,
|
||
|
|
"mean_token_accuracy": 0.8479447506368161,
|
||
|
|
"num_tokens": 48092035.0,
|
||
|
|
"step": 1680
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5260777238756418,
|
||
|
|
"epoch": 2.7039999999999997,
|
||
|
|
"grad_norm": 1.5234375,
|
||
|
|
"learning_rate": 1.846888888888889e-05,
|
||
|
|
"loss": 0.5153,
|
||
|
|
"mean_token_accuracy": 0.8464129194617271,
|
||
|
|
"num_tokens": 48386231.0,
|
||
|
|
"step": 1690
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5193733531981707,
|
||
|
|
"epoch": 2.7199999999999998,
|
||
|
|
"grad_norm": 1.4609375,
|
||
|
|
"learning_rate": 1.8446666666666667e-05,
|
||
|
|
"loss": 0.5044,
|
||
|
|
"mean_token_accuracy": 0.8482546083629131,
|
||
|
|
"num_tokens": 48663262.0,
|
||
|
|
"step": 1700
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.7199999999999998,
|
||
|
|
"eval_math_entropy": 0.555855593919754,
|
||
|
|
"eval_math_loss": 0.6335127949714661,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8283462662696839,
|
||
|
|
"eval_math_num_tokens": 48663262.0,
|
||
|
|
"eval_math_runtime": 35.7836,
|
||
|
|
"eval_math_samples_per_second": 13.973,
|
||
|
|
"eval_math_steps_per_second": 3.493,
|
||
|
|
"step": 1700
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.7199999999999998,
|
||
|
|
"eval_physics_entropy": 0.5213737454414368,
|
||
|
|
"eval_physics_loss": 0.6999514698982239,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8152164473533631,
|
||
|
|
"eval_physics_num_tokens": 48663262.0,
|
||
|
|
"eval_physics_runtime": 41.4953,
|
||
|
|
"eval_physics_samples_per_second": 12.05,
|
||
|
|
"eval_physics_steps_per_second": 3.012,
|
||
|
|
"step": 1700
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5167077284306287,
|
||
|
|
"epoch": 2.7359999999999998,
|
||
|
|
"grad_norm": 1.5546875,
|
||
|
|
"learning_rate": 1.842444444444445e-05,
|
||
|
|
"loss": 0.5014,
|
||
|
|
"mean_token_accuracy": 0.8498601946979761,
|
||
|
|
"num_tokens": 48949456.0,
|
||
|
|
"step": 1710
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.522574995085597,
|
||
|
|
"epoch": 2.752,
|
||
|
|
"grad_norm": 1.4921875,
|
||
|
|
"learning_rate": 1.8402222222222223e-05,
|
||
|
|
"loss": 0.5015,
|
||
|
|
"mean_token_accuracy": 0.8494253471493721,
|
||
|
|
"num_tokens": 49235751.0,
|
||
|
|
"step": 1720
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5236585384234786,
|
||
|
|
"epoch": 2.768,
|
||
|
|
"grad_norm": 1.3515625,
|
||
|
|
"learning_rate": 1.8380000000000004e-05,
|
||
|
|
"loss": 0.5081,
|
||
|
|
"mean_token_accuracy": 0.8480564054101706,
|
||
|
|
"num_tokens": 49524628.0,
|
||
|
|
"step": 1730
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.536912607960403,
|
||
|
|
"epoch": 2.784,
|
||
|
|
"grad_norm": 1.484375,
|
||
|
|
"learning_rate": 1.8357777777777778e-05,
|
||
|
|
"loss": 0.5166,
|
||
|
|
"mean_token_accuracy": 0.8444275613874197,
|
||
|
|
"num_tokens": 49799604.0,
|
||
|
|
"step": 1740
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5194279940798878,
|
||
|
|
"epoch": 2.8,
|
||
|
|
"grad_norm": 1.234375,
|
||
|
|
"learning_rate": 1.833555555555556e-05,
|
||
|
|
"loss": 0.5044,
|
||
|
|
"mean_token_accuracy": 0.8491090513765812,
|
||
|
|
"num_tokens": 50079307.0,
|
||
|
|
"step": 1750
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5226334543898702,
|
||
|
|
"epoch": 2.816,
|
||
|
|
"grad_norm": 1.46875,
|
||
|
|
"learning_rate": 1.8313333333333333e-05,
|
||
|
|
"loss": 0.5013,
|
||
|
|
"mean_token_accuracy": 0.8487468950450421,
|
||
|
|
"num_tokens": 50372216.0,
|
||
|
|
"step": 1760
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5100100792944431,
|
||
|
|
"epoch": 2.832,
|
||
|
|
"grad_norm": 1.5390625,
|
||
|
|
"learning_rate": 1.8291111111111114e-05,
|
||
|
|
"loss": 0.4977,
|
||
|
|
"mean_token_accuracy": 0.8502487774938345,
|
||
|
|
"num_tokens": 50657211.0,
|
||
|
|
"step": 1770
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5339175367727875,
|
||
|
|
"epoch": 2.848,
|
||
|
|
"grad_norm": 1.3828125,
|
||
|
|
"learning_rate": 1.8268888888888888e-05,
|
||
|
|
"loss": 0.5176,
|
||
|
|
"mean_token_accuracy": 0.8462332025170326,
|
||
|
|
"num_tokens": 50946503.0,
|
||
|
|
"step": 1780
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5188839716836811,
|
||
|
|
"epoch": 2.864,
|
||
|
|
"grad_norm": 1.5234375,
|
||
|
|
"learning_rate": 1.824666666666667e-05,
|
||
|
|
"loss": 0.5046,
|
||
|
|
"mean_token_accuracy": 0.8482739225029945,
|
||
|
|
"num_tokens": 51229298.0,
|
||
|
|
"step": 1790
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.52184556145221,
|
||
|
|
"epoch": 2.88,
|
||
|
|
"grad_norm": 1.4609375,
|
||
|
|
"learning_rate": 1.8224444444444447e-05,
|
||
|
|
"loss": 0.5,
|
||
|
|
"mean_token_accuracy": 0.8485745534300804,
|
||
|
|
"num_tokens": 51509622.0,
|
||
|
|
"step": 1800
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.88,
|
||
|
|
"eval_math_entropy": 0.5322999174594879,
|
||
|
|
"eval_math_loss": 0.6332678198814392,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8286352276802063,
|
||
|
|
"eval_math_num_tokens": 51509622.0,
|
||
|
|
"eval_math_runtime": 35.803,
|
||
|
|
"eval_math_samples_per_second": 13.965,
|
||
|
|
"eval_math_steps_per_second": 3.491,
|
||
|
|
"step": 1800
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.88,
|
||
|
|
"eval_physics_entropy": 0.5120384588241577,
|
||
|
|
"eval_physics_loss": 0.7020947337150574,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8149435911178589,
|
||
|
|
"eval_physics_num_tokens": 51509622.0,
|
||
|
|
"eval_physics_runtime": 41.5069,
|
||
|
|
"eval_physics_samples_per_second": 12.046,
|
||
|
|
"eval_physics_steps_per_second": 3.012,
|
||
|
|
"step": 1800
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5138611877337098,
|
||
|
|
"epoch": 2.896,
|
||
|
|
"grad_norm": 1.4765625,
|
||
|
|
"learning_rate": 1.8202222222222225e-05,
|
||
|
|
"loss": 0.496,
|
||
|
|
"mean_token_accuracy": 0.8509018927812576,
|
||
|
|
"num_tokens": 51790856.0,
|
||
|
|
"step": 1810
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5193969309329987,
|
||
|
|
"epoch": 2.912,
|
||
|
|
"grad_norm": 1.484375,
|
||
|
|
"learning_rate": 1.8180000000000002e-05,
|
||
|
|
"loss": 0.5039,
|
||
|
|
"mean_token_accuracy": 0.8491050932556391,
|
||
|
|
"num_tokens": 52074393.0,
|
||
|
|
"step": 1820
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5282592326402664,
|
||
|
|
"epoch": 2.928,
|
||
|
|
"grad_norm": 1.4140625,
|
||
|
|
"learning_rate": 1.815777777777778e-05,
|
||
|
|
"loss": 0.514,
|
||
|
|
"mean_token_accuracy": 0.8482961695641279,
|
||
|
|
"num_tokens": 52366467.0,
|
||
|
|
"step": 1830
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.512021827697754,
|
||
|
|
"epoch": 2.944,
|
||
|
|
"grad_norm": 1.4375,
|
||
|
|
"learning_rate": 1.8135555555555557e-05,
|
||
|
|
"loss": 0.4995,
|
||
|
|
"mean_token_accuracy": 0.8494243275374174,
|
||
|
|
"num_tokens": 52657863.0,
|
||
|
|
"step": 1840
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5220472222194076,
|
||
|
|
"epoch": 2.96,
|
||
|
|
"grad_norm": 1.40625,
|
||
|
|
"learning_rate": 1.8113333333333335e-05,
|
||
|
|
"loss": 0.5045,
|
||
|
|
"mean_token_accuracy": 0.848219645395875,
|
||
|
|
"num_tokens": 52947183.0,
|
||
|
|
"step": 1850
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5195147603750229,
|
||
|
|
"epoch": 2.976,
|
||
|
|
"grad_norm": 1.3203125,
|
||
|
|
"learning_rate": 1.8091111111111113e-05,
|
||
|
|
"loss": 0.5009,
|
||
|
|
"mean_token_accuracy": 0.8488207392394542,
|
||
|
|
"num_tokens": 53226984.0,
|
||
|
|
"step": 1860
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5123168766498566,
|
||
|
|
"epoch": 2.992,
|
||
|
|
"grad_norm": 1.3828125,
|
||
|
|
"learning_rate": 1.806888888888889e-05,
|
||
|
|
"loss": 0.4941,
|
||
|
|
"mean_token_accuracy": 0.8505234856158495,
|
||
|
|
"num_tokens": 53515204.0,
|
||
|
|
"step": 1870
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4752411070279777,
|
||
|
|
"epoch": 3.008,
|
||
|
|
"grad_norm": 1.875,
|
||
|
|
"learning_rate": 1.8046666666666668e-05,
|
||
|
|
"loss": 0.4333,
|
||
|
|
"mean_token_accuracy": 0.8668791756033898,
|
||
|
|
"num_tokens": 53800592.0,
|
||
|
|
"step": 1880
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3587687503546476,
|
||
|
|
"epoch": 3.024,
|
||
|
|
"grad_norm": 1.8828125,
|
||
|
|
"learning_rate": 1.8024444444444445e-05,
|
||
|
|
"loss": 0.3423,
|
||
|
|
"mean_token_accuracy": 0.888729453831911,
|
||
|
|
"num_tokens": 54092017.0,
|
||
|
|
"step": 1890
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3438520409166813,
|
||
|
|
"epoch": 3.04,
|
||
|
|
"grad_norm": 1.8359375,
|
||
|
|
"learning_rate": 1.8002222222222223e-05,
|
||
|
|
"loss": 0.3233,
|
||
|
|
"mean_token_accuracy": 0.8938188333064317,
|
||
|
|
"num_tokens": 54380158.0,
|
||
|
|
"step": 1900
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.04,
|
||
|
|
"eval_math_entropy": 0.406534277677536,
|
||
|
|
"eval_math_loss": 0.7412194609642029,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8225762667655945,
|
||
|
|
"eval_math_num_tokens": 54380158.0,
|
||
|
|
"eval_math_runtime": 35.7958,
|
||
|
|
"eval_math_samples_per_second": 13.968,
|
||
|
|
"eval_math_steps_per_second": 3.492,
|
||
|
|
"step": 1900
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.04,
|
||
|
|
"eval_physics_entropy": 0.38361185932159425,
|
||
|
|
"eval_physics_loss": 0.8625921607017517,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8045027327537536,
|
||
|
|
"eval_physics_num_tokens": 54380158.0,
|
||
|
|
"eval_physics_runtime": 41.4955,
|
||
|
|
"eval_physics_samples_per_second": 12.05,
|
||
|
|
"eval_physics_steps_per_second": 3.012,
|
||
|
|
"step": 1900
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.34686512807384134,
|
||
|
|
"epoch": 3.056,
|
||
|
|
"grad_norm": 1.8984375,
|
||
|
|
"learning_rate": 1.798e-05,
|
||
|
|
"loss": 0.3202,
|
||
|
|
"mean_token_accuracy": 0.8946099061518907,
|
||
|
|
"num_tokens": 54658093.0,
|
||
|
|
"step": 1910
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.34148691557347777,
|
||
|
|
"epoch": 3.072,
|
||
|
|
"grad_norm": 1.953125,
|
||
|
|
"learning_rate": 1.7957777777777778e-05,
|
||
|
|
"loss": 0.3202,
|
||
|
|
"mean_token_accuracy": 0.8947911463677883,
|
||
|
|
"num_tokens": 54949938.0,
|
||
|
|
"step": 1920
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3277126123197377,
|
||
|
|
"epoch": 3.088,
|
||
|
|
"grad_norm": 2.21875,
|
||
|
|
"learning_rate": 1.7935555555555556e-05,
|
||
|
|
"loss": 0.3064,
|
||
|
|
"mean_token_accuracy": 0.8989322237670422,
|
||
|
|
"num_tokens": 55234643.0,
|
||
|
|
"step": 1930
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.34841074971482155,
|
||
|
|
"epoch": 3.104,
|
||
|
|
"grad_norm": 1.9609375,
|
||
|
|
"learning_rate": 1.7913333333333337e-05,
|
||
|
|
"loss": 0.3262,
|
||
|
|
"mean_token_accuracy": 0.8924810353666544,
|
||
|
|
"num_tokens": 55519679.0,
|
||
|
|
"step": 1940
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.33432568768039345,
|
||
|
|
"epoch": 3.12,
|
||
|
|
"grad_norm": 1.9453125,
|
||
|
|
"learning_rate": 1.789111111111111e-05,
|
||
|
|
"loss": 0.3163,
|
||
|
|
"mean_token_accuracy": 0.896980819106102,
|
||
|
|
"num_tokens": 55812397.0,
|
||
|
|
"step": 1950
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.33816060032695533,
|
||
|
|
"epoch": 3.136,
|
||
|
|
"grad_norm": 1.8671875,
|
||
|
|
"learning_rate": 1.7868888888888892e-05,
|
||
|
|
"loss": 0.3182,
|
||
|
|
"mean_token_accuracy": 0.8948573391884566,
|
||
|
|
"num_tokens": 56101550.0,
|
||
|
|
"step": 1960
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3384258272126317,
|
||
|
|
"epoch": 3.152,
|
||
|
|
"grad_norm": 2.078125,
|
||
|
|
"learning_rate": 1.7846666666666666e-05,
|
||
|
|
"loss": 0.3209,
|
||
|
|
"mean_token_accuracy": 0.8947299372404813,
|
||
|
|
"num_tokens": 56387662.0,
|
||
|
|
"step": 1970
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3522730448283255,
|
||
|
|
"epoch": 3.168,
|
||
|
|
"grad_norm": 2.03125,
|
||
|
|
"learning_rate": 1.7824444444444447e-05,
|
||
|
|
"loss": 0.333,
|
||
|
|
"mean_token_accuracy": 0.892043673619628,
|
||
|
|
"num_tokens": 56673426.0,
|
||
|
|
"step": 1980
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3521577626466751,
|
||
|
|
"epoch": 3.184,
|
||
|
|
"grad_norm": 2.25,
|
||
|
|
"learning_rate": 1.780222222222222e-05,
|
||
|
|
"loss": 0.3268,
|
||
|
|
"mean_token_accuracy": 0.89161705262959,
|
||
|
|
"num_tokens": 56957044.0,
|
||
|
|
"step": 1990
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3363432651385665,
|
||
|
|
"epoch": 3.2,
|
||
|
|
"grad_norm": 1.8671875,
|
||
|
|
"learning_rate": 1.7780000000000003e-05,
|
||
|
|
"loss": 0.3174,
|
||
|
|
"mean_token_accuracy": 0.8954976741224527,
|
||
|
|
"num_tokens": 57250097.0,
|
||
|
|
"step": 2000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.2,
|
||
|
|
"eval_math_entropy": 0.3876710822582245,
|
||
|
|
"eval_math_loss": 0.7634654641151428,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8221899409294129,
|
||
|
|
"eval_math_num_tokens": 57250097.0,
|
||
|
|
"eval_math_runtime": 35.8054,
|
||
|
|
"eval_math_samples_per_second": 13.964,
|
||
|
|
"eval_math_steps_per_second": 3.491,
|
||
|
|
"step": 2000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.2,
|
||
|
|
"eval_physics_entropy": 0.35663985908031465,
|
||
|
|
"eval_physics_loss": 0.8939647674560547,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8043588905334472,
|
||
|
|
"eval_physics_num_tokens": 57250097.0,
|
||
|
|
"eval_physics_runtime": 41.1357,
|
||
|
|
"eval_physics_samples_per_second": 12.155,
|
||
|
|
"eval_physics_steps_per_second": 3.039,
|
||
|
|
"step": 2000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3453281221911311,
|
||
|
|
"epoch": 3.216,
|
||
|
|
"grad_norm": 2.046875,
|
||
|
|
"learning_rate": 1.7757777777777777e-05,
|
||
|
|
"loss": 0.3293,
|
||
|
|
"mean_token_accuracy": 0.8925925586372614,
|
||
|
|
"num_tokens": 57532026.0,
|
||
|
|
"step": 2010
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.34009323371574285,
|
||
|
|
"epoch": 3.232,
|
||
|
|
"grad_norm": 2.21875,
|
||
|
|
"learning_rate": 1.7735555555555558e-05,
|
||
|
|
"loss": 0.3216,
|
||
|
|
"mean_token_accuracy": 0.895359569042921,
|
||
|
|
"num_tokens": 57814853.0,
|
||
|
|
"step": 2020
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.32144517404958606,
|
||
|
|
"epoch": 3.248,
|
||
|
|
"grad_norm": 2.046875,
|
||
|
|
"learning_rate": 1.7713333333333335e-05,
|
||
|
|
"loss": 0.3013,
|
||
|
|
"mean_token_accuracy": 0.8993485007435084,
|
||
|
|
"num_tokens": 58094190.0,
|
||
|
|
"step": 2030
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3362208865582943,
|
||
|
|
"epoch": 3.2640000000000002,
|
||
|
|
"grad_norm": 2.15625,
|
||
|
|
"learning_rate": 1.7691111111111113e-05,
|
||
|
|
"loss": 0.3179,
|
||
|
|
"mean_token_accuracy": 0.8959619626402855,
|
||
|
|
"num_tokens": 58374378.0,
|
||
|
|
"step": 2040
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.34206875674426557,
|
||
|
|
"epoch": 3.2800000000000002,
|
||
|
|
"grad_norm": 1.96875,
|
||
|
|
"learning_rate": 1.766888888888889e-05,
|
||
|
|
"loss": 0.3233,
|
||
|
|
"mean_token_accuracy": 0.8937642563134431,
|
||
|
|
"num_tokens": 58652789.0,
|
||
|
|
"step": 2050
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3506787528283894,
|
||
|
|
"epoch": 3.296,
|
||
|
|
"grad_norm": 2.078125,
|
||
|
|
"learning_rate": 1.7646666666666668e-05,
|
||
|
|
"loss": 0.331,
|
||
|
|
"mean_token_accuracy": 0.8917676325887441,
|
||
|
|
"num_tokens": 58930014.0,
|
||
|
|
"step": 2060
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.345881634298712,
|
||
|
|
"epoch": 3.312,
|
||
|
|
"grad_norm": 2.015625,
|
||
|
|
"learning_rate": 1.7624444444444446e-05,
|
||
|
|
"loss": 0.3262,
|
||
|
|
"mean_token_accuracy": 0.8924973297864198,
|
||
|
|
"num_tokens": 59210270.0,
|
||
|
|
"step": 2070
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.34120792699977753,
|
||
|
|
"epoch": 3.328,
|
||
|
|
"grad_norm": 2.140625,
|
||
|
|
"learning_rate": 1.7602222222222223e-05,
|
||
|
|
"loss": 0.3192,
|
||
|
|
"mean_token_accuracy": 0.8952000044286251,
|
||
|
|
"num_tokens": 59496408.0,
|
||
|
|
"step": 2080
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3278652913868427,
|
||
|
|
"epoch": 3.344,
|
||
|
|
"grad_norm": 2.15625,
|
||
|
|
"learning_rate": 1.758e-05,
|
||
|
|
"loss": 0.3112,
|
||
|
|
"mean_token_accuracy": 0.8966204669326544,
|
||
|
|
"num_tokens": 59779722.0,
|
||
|
|
"step": 2090
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3657307820394635,
|
||
|
|
"epoch": 3.36,
|
||
|
|
"grad_norm": 2.21875,
|
||
|
|
"learning_rate": 1.755777777777778e-05,
|
||
|
|
"loss": 0.3419,
|
||
|
|
"mean_token_accuracy": 0.8882275156676769,
|
||
|
|
"num_tokens": 60061562.0,
|
||
|
|
"step": 2100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.36,
|
||
|
|
"eval_math_entropy": 0.3830914467573166,
|
||
|
|
"eval_math_loss": 0.7686993479728699,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8224415044784545,
|
||
|
|
"eval_math_num_tokens": 60061562.0,
|
||
|
|
"eval_math_runtime": 35.8062,
|
||
|
|
"eval_math_samples_per_second": 13.964,
|
||
|
|
"eval_math_steps_per_second": 3.491,
|
||
|
|
"step": 2100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.36,
|
||
|
|
"eval_physics_entropy": 0.3522386626005173,
|
||
|
|
"eval_physics_loss": 0.9039635062217712,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8043717379570008,
|
||
|
|
"eval_physics_num_tokens": 60061562.0,
|
||
|
|
"eval_physics_runtime": 41.4883,
|
||
|
|
"eval_physics_samples_per_second": 12.052,
|
||
|
|
"eval_physics_steps_per_second": 3.013,
|
||
|
|
"step": 2100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.35135511131957176,
|
||
|
|
"epoch": 3.376,
|
||
|
|
"grad_norm": 2.015625,
|
||
|
|
"learning_rate": 1.7535555555555556e-05,
|
||
|
|
"loss": 0.3328,
|
||
|
|
"mean_token_accuracy": 0.8913474876433611,
|
||
|
|
"num_tokens": 60346574.0,
|
||
|
|
"step": 2110
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3290985696949065,
|
||
|
|
"epoch": 3.392,
|
||
|
|
"grad_norm": 2.015625,
|
||
|
|
"learning_rate": 1.7513333333333334e-05,
|
||
|
|
"loss": 0.3122,
|
||
|
|
"mean_token_accuracy": 0.8976335555315018,
|
||
|
|
"num_tokens": 60636820.0,
|
||
|
|
"step": 2120
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3349749938584864,
|
||
|
|
"epoch": 3.408,
|
||
|
|
"grad_norm": 2.09375,
|
||
|
|
"learning_rate": 1.749111111111111e-05,
|
||
|
|
"loss": 0.3148,
|
||
|
|
"mean_token_accuracy": 0.8964381668716669,
|
||
|
|
"num_tokens": 60922033.0,
|
||
|
|
"step": 2130
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3461784217506647,
|
||
|
|
"epoch": 3.424,
|
||
|
|
"grad_norm": 2.25,
|
||
|
|
"learning_rate": 1.746888888888889e-05,
|
||
|
|
"loss": 0.327,
|
||
|
|
"mean_token_accuracy": 0.8919767573475837,
|
||
|
|
"num_tokens": 61209156.0,
|
||
|
|
"step": 2140
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3424711484462023,
|
||
|
|
"epoch": 3.44,
|
||
|
|
"grad_norm": 1.9609375,
|
||
|
|
"learning_rate": 1.7446666666666667e-05,
|
||
|
|
"loss": 0.3233,
|
||
|
|
"mean_token_accuracy": 0.8935233734548091,
|
||
|
|
"num_tokens": 61491228.0,
|
||
|
|
"step": 2150
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3479755653068423,
|
||
|
|
"epoch": 3.456,
|
||
|
|
"grad_norm": 2.140625,
|
||
|
|
"learning_rate": 1.7424444444444444e-05,
|
||
|
|
"loss": 0.3292,
|
||
|
|
"mean_token_accuracy": 0.8924077559262514,
|
||
|
|
"num_tokens": 61773033.0,
|
||
|
|
"step": 2160
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3536012376658618,
|
||
|
|
"epoch": 3.472,
|
||
|
|
"grad_norm": 2.0625,
|
||
|
|
"learning_rate": 1.7402222222222222e-05,
|
||
|
|
"loss": 0.3402,
|
||
|
|
"mean_token_accuracy": 0.8894314222037792,
|
||
|
|
"num_tokens": 62061802.0,
|
||
|
|
"step": 2170
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.34079490676522256,
|
||
|
|
"epoch": 3.488,
|
||
|
|
"grad_norm": 2.078125,
|
||
|
|
"learning_rate": 1.7380000000000003e-05,
|
||
|
|
"loss": 0.3233,
|
||
|
|
"mean_token_accuracy": 0.8940149489790201,
|
||
|
|
"num_tokens": 62347490.0,
|
||
|
|
"step": 2180
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.34653188129886986,
|
||
|
|
"epoch": 3.504,
|
||
|
|
"grad_norm": 2.0,
|
||
|
|
"learning_rate": 1.735777777777778e-05,
|
||
|
|
"loss": 0.3243,
|
||
|
|
"mean_token_accuracy": 0.8935138266533613,
|
||
|
|
"num_tokens": 62636811.0,
|
||
|
|
"step": 2190
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3420496225357056,
|
||
|
|
"epoch": 3.52,
|
||
|
|
"grad_norm": 2.03125,
|
||
|
|
"learning_rate": 1.7335555555555558e-05,
|
||
|
|
"loss": 0.3255,
|
||
|
|
"mean_token_accuracy": 0.8933457151055336,
|
||
|
|
"num_tokens": 62928536.0,
|
||
|
|
"step": 2200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.52,
|
||
|
|
"eval_math_entropy": 0.39784613919258116,
|
||
|
|
"eval_math_loss": 0.7561159133911133,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8226356348991394,
|
||
|
|
"eval_math_num_tokens": 62928536.0,
|
||
|
|
"eval_math_runtime": 35.8487,
|
||
|
|
"eval_math_samples_per_second": 13.948,
|
||
|
|
"eval_math_steps_per_second": 3.487,
|
||
|
|
"step": 2200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.52,
|
||
|
|
"eval_physics_entropy": 0.3542559086084366,
|
||
|
|
"eval_physics_loss": 0.8993895053863525,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8046151103973389,
|
||
|
|
"eval_physics_num_tokens": 62928536.0,
|
||
|
|
"eval_physics_runtime": 41.5136,
|
||
|
|
"eval_physics_samples_per_second": 12.044,
|
||
|
|
"eval_physics_steps_per_second": 3.011,
|
||
|
|
"step": 2200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3441290996037424,
|
||
|
|
"epoch": 3.536,
|
||
|
|
"grad_norm": 2.234375,
|
||
|
|
"learning_rate": 1.7313333333333336e-05,
|
||
|
|
"loss": 0.3257,
|
||
|
|
"mean_token_accuracy": 0.8930490992963314,
|
||
|
|
"num_tokens": 63212496.0,
|
||
|
|
"step": 2210
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3405540899373591,
|
||
|
|
"epoch": 3.552,
|
||
|
|
"grad_norm": 2.0,
|
||
|
|
"learning_rate": 1.7291111111111113e-05,
|
||
|
|
"loss": 0.3211,
|
||
|
|
"mean_token_accuracy": 0.8948382496833801,
|
||
|
|
"num_tokens": 63506921.0,
|
||
|
|
"step": 2220
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.33624334512278437,
|
||
|
|
"epoch": 3.568,
|
||
|
|
"grad_norm": 1.9375,
|
||
|
|
"learning_rate": 1.726888888888889e-05,
|
||
|
|
"loss": 0.3159,
|
||
|
|
"mean_token_accuracy": 0.8952401872724295,
|
||
|
|
"num_tokens": 63803208.0,
|
||
|
|
"step": 2230
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.34232397861778735,
|
||
|
|
"epoch": 3.584,
|
||
|
|
"grad_norm": 2.25,
|
||
|
|
"learning_rate": 1.724666666666667e-05,
|
||
|
|
"loss": 0.3238,
|
||
|
|
"mean_token_accuracy": 0.8939516879618168,
|
||
|
|
"num_tokens": 64084867.0,
|
||
|
|
"step": 2240
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.33109099259600044,
|
||
|
|
"epoch": 3.6,
|
||
|
|
"grad_norm": 2.109375,
|
||
|
|
"learning_rate": 1.7224444444444446e-05,
|
||
|
|
"loss": 0.3153,
|
||
|
|
"mean_token_accuracy": 0.8967003002762794,
|
||
|
|
"num_tokens": 64377046.0,
|
||
|
|
"step": 2250
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.35013388488441705,
|
||
|
|
"epoch": 3.616,
|
||
|
|
"grad_norm": 2.109375,
|
||
|
|
"learning_rate": 1.7202222222222224e-05,
|
||
|
|
"loss": 0.331,
|
||
|
|
"mean_token_accuracy": 0.8905543919652701,
|
||
|
|
"num_tokens": 64664301.0,
|
||
|
|
"step": 2260
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3353199644945562,
|
||
|
|
"epoch": 3.632,
|
||
|
|
"grad_norm": 2.140625,
|
||
|
|
"learning_rate": 1.718e-05,
|
||
|
|
"loss": 0.316,
|
||
|
|
"mean_token_accuracy": 0.895779125764966,
|
||
|
|
"num_tokens": 64946172.0,
|
||
|
|
"step": 2270
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3319018458016217,
|
||
|
|
"epoch": 3.648,
|
||
|
|
"grad_norm": 2.0,
|
||
|
|
"learning_rate": 1.715777777777778e-05,
|
||
|
|
"loss": 0.3138,
|
||
|
|
"mean_token_accuracy": 0.8972285516560078,
|
||
|
|
"num_tokens": 65230709.0,
|
||
|
|
"step": 2280
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.36236064694821835,
|
||
|
|
"epoch": 3.664,
|
||
|
|
"grad_norm": 2.109375,
|
||
|
|
"learning_rate": 1.7135555555555557e-05,
|
||
|
|
"loss": 0.344,
|
||
|
|
"mean_token_accuracy": 0.8877242401242256,
|
||
|
|
"num_tokens": 65515738.0,
|
||
|
|
"step": 2290
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3466769499704242,
|
||
|
|
"epoch": 3.68,
|
||
|
|
"grad_norm": 2.0625,
|
||
|
|
"learning_rate": 1.7113333333333334e-05,
|
||
|
|
"loss": 0.3276,
|
||
|
|
"mean_token_accuracy": 0.8919828839600086,
|
||
|
|
"num_tokens": 65802680.0,
|
||
|
|
"step": 2300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.68,
|
||
|
|
"eval_math_entropy": 0.3900481255054474,
|
||
|
|
"eval_math_loss": 0.7627390623092651,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8218873319625855,
|
||
|
|
"eval_math_num_tokens": 65802680.0,
|
||
|
|
"eval_math_runtime": 35.8389,
|
||
|
|
"eval_math_samples_per_second": 13.951,
|
||
|
|
"eval_math_steps_per_second": 3.488,
|
||
|
|
"step": 2300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.68,
|
||
|
|
"eval_physics_entropy": 0.3501615208387375,
|
||
|
|
"eval_physics_loss": 0.9098384380340576,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8042201838493347,
|
||
|
|
"eval_physics_num_tokens": 65802680.0,
|
||
|
|
"eval_physics_runtime": 41.525,
|
||
|
|
"eval_physics_samples_per_second": 12.041,
|
||
|
|
"eval_physics_steps_per_second": 3.01,
|
||
|
|
"step": 2300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3516978560946882,
|
||
|
|
"epoch": 3.6959999999999997,
|
||
|
|
"grad_norm": 2.0625,
|
||
|
|
"learning_rate": 1.7091111111111112e-05,
|
||
|
|
"loss": 0.3334,
|
||
|
|
"mean_token_accuracy": 0.8908947184681892,
|
||
|
|
"num_tokens": 66089104.0,
|
||
|
|
"step": 2310
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3427901498042047,
|
||
|
|
"epoch": 3.7119999999999997,
|
||
|
|
"grad_norm": 2.0,
|
||
|
|
"learning_rate": 1.706888888888889e-05,
|
||
|
|
"loss": 0.3228,
|
||
|
|
"mean_token_accuracy": 0.8933374501764775,
|
||
|
|
"num_tokens": 66369895.0,
|
||
|
|
"step": 2320
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3467458762228489,
|
||
|
|
"epoch": 3.7279999999999998,
|
||
|
|
"grad_norm": 2.21875,
|
||
|
|
"learning_rate": 1.704666666666667e-05,
|
||
|
|
"loss": 0.327,
|
||
|
|
"mean_token_accuracy": 0.8924272887408733,
|
||
|
|
"num_tokens": 66653165.0,
|
||
|
|
"step": 2330
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.34478482212871314,
|
||
|
|
"epoch": 3.7439999999999998,
|
||
|
|
"grad_norm": 2.109375,
|
||
|
|
"learning_rate": 1.7024444444444445e-05,
|
||
|
|
"loss": 0.3245,
|
||
|
|
"mean_token_accuracy": 0.8931741625070572,
|
||
|
|
"num_tokens": 66941476.0,
|
||
|
|
"step": 2340
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.34707736801356076,
|
||
|
|
"epoch": 3.76,
|
||
|
|
"grad_norm": 2.140625,
|
||
|
|
"learning_rate": 1.7002222222222226e-05,
|
||
|
|
"loss": 0.3306,
|
||
|
|
"mean_token_accuracy": 0.8919444873929023,
|
||
|
|
"num_tokens": 67225380.0,
|
||
|
|
"step": 2350
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.35979456212371586,
|
||
|
|
"epoch": 3.776,
|
||
|
|
"grad_norm": 2.234375,
|
||
|
|
"learning_rate": 1.698e-05,
|
||
|
|
"loss": 0.3425,
|
||
|
|
"mean_token_accuracy": 0.888241409137845,
|
||
|
|
"num_tokens": 67516641.0,
|
||
|
|
"step": 2360
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.337408459559083,
|
||
|
|
"epoch": 3.792,
|
||
|
|
"grad_norm": 2.234375,
|
||
|
|
"learning_rate": 1.695777777777778e-05,
|
||
|
|
"loss": 0.3207,
|
||
|
|
"mean_token_accuracy": 0.8952056996524334,
|
||
|
|
"num_tokens": 67808009.0,
|
||
|
|
"step": 2370
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.35541942873969673,
|
||
|
|
"epoch": 3.808,
|
||
|
|
"grad_norm": 2.0625,
|
||
|
|
"learning_rate": 1.6935555555555555e-05,
|
||
|
|
"loss": 0.3354,
|
||
|
|
"mean_token_accuracy": 0.8904807876795531,
|
||
|
|
"num_tokens": 68092996.0,
|
||
|
|
"step": 2380
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.338986696396023,
|
||
|
|
"epoch": 3.824,
|
||
|
|
"grad_norm": 2.03125,
|
||
|
|
"learning_rate": 1.6913333333333336e-05,
|
||
|
|
"loss": 0.3217,
|
||
|
|
"mean_token_accuracy": 0.8948666993528605,
|
||
|
|
"num_tokens": 68380631.0,
|
||
|
|
"step": 2390
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.34510293062776326,
|
||
|
|
"epoch": 3.84,
|
||
|
|
"grad_norm": 2.140625,
|
||
|
|
"learning_rate": 1.689111111111111e-05,
|
||
|
|
"loss": 0.327,
|
||
|
|
"mean_token_accuracy": 0.8922771938145161,
|
||
|
|
"num_tokens": 68666505.0,
|
||
|
|
"step": 2400
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.84,
|
||
|
|
"eval_math_entropy": 0.3931693367958069,
|
||
|
|
"eval_math_loss": 0.7599309682846069,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8226110072135925,
|
||
|
|
"eval_math_num_tokens": 68666505.0,
|
||
|
|
"eval_math_runtime": 35.819,
|
||
|
|
"eval_math_samples_per_second": 13.959,
|
||
|
|
"eval_math_steps_per_second": 3.49,
|
||
|
|
"step": 2400
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.84,
|
||
|
|
"eval_physics_entropy": 0.3486952732801437,
|
||
|
|
"eval_physics_loss": 0.9106001257896423,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8040544500350952,
|
||
|
|
"eval_physics_num_tokens": 68666505.0,
|
||
|
|
"eval_physics_runtime": 41.4809,
|
||
|
|
"eval_physics_samples_per_second": 12.054,
|
||
|
|
"eval_physics_steps_per_second": 3.013,
|
||
|
|
"step": 2400
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.341458183247596,
|
||
|
|
"epoch": 3.856,
|
||
|
|
"grad_norm": 2.1875,
|
||
|
|
"learning_rate": 1.686888888888889e-05,
|
||
|
|
"loss": 0.3259,
|
||
|
|
"mean_token_accuracy": 0.8934879627078771,
|
||
|
|
"num_tokens": 68951258.0,
|
||
|
|
"step": 2410
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.35674845492467283,
|
||
|
|
"epoch": 3.872,
|
||
|
|
"grad_norm": 2.234375,
|
||
|
|
"learning_rate": 1.684666666666667e-05,
|
||
|
|
"loss": 0.3383,
|
||
|
|
"mean_token_accuracy": 0.8894801624119282,
|
||
|
|
"num_tokens": 69243705.0,
|
||
|
|
"step": 2420
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3472815524786711,
|
||
|
|
"epoch": 3.888,
|
||
|
|
"grad_norm": 2.171875,
|
||
|
|
"learning_rate": 1.6824444444444447e-05,
|
||
|
|
"loss": 0.3299,
|
||
|
|
"mean_token_accuracy": 0.8912571128457785,
|
||
|
|
"num_tokens": 69529393.0,
|
||
|
|
"step": 2430
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.345150712877512,
|
||
|
|
"epoch": 3.904,
|
||
|
|
"grad_norm": 2.28125,
|
||
|
|
"learning_rate": 1.6802222222222224e-05,
|
||
|
|
"loss": 0.3269,
|
||
|
|
"mean_token_accuracy": 0.8925980541855096,
|
||
|
|
"num_tokens": 69820950.0,
|
||
|
|
"step": 2440
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3369919521734118,
|
||
|
|
"epoch": 3.92,
|
||
|
|
"grad_norm": 2.21875,
|
||
|
|
"learning_rate": 1.6780000000000002e-05,
|
||
|
|
"loss": 0.317,
|
||
|
|
"mean_token_accuracy": 0.8948526129126548,
|
||
|
|
"num_tokens": 70097736.0,
|
||
|
|
"step": 2450
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3590950695797801,
|
||
|
|
"epoch": 3.936,
|
||
|
|
"grad_norm": 2.046875,
|
||
|
|
"learning_rate": 1.675777777777778e-05,
|
||
|
|
"loss": 0.34,
|
||
|
|
"mean_token_accuracy": 0.8890990480780602,
|
||
|
|
"num_tokens": 70395123.0,
|
||
|
|
"step": 2460
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.345857547596097,
|
||
|
|
"epoch": 3.952,
|
||
|
|
"grad_norm": 2.015625,
|
||
|
|
"learning_rate": 1.6735555555555557e-05,
|
||
|
|
"loss": 0.3301,
|
||
|
|
"mean_token_accuracy": 0.8929780598729848,
|
||
|
|
"num_tokens": 70678153.0,
|
||
|
|
"step": 2470
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3558691343292594,
|
||
|
|
"epoch": 3.968,
|
||
|
|
"grad_norm": 2.078125,
|
||
|
|
"learning_rate": 1.6713333333333335e-05,
|
||
|
|
"loss": 0.3375,
|
||
|
|
"mean_token_accuracy": 0.8891723170876503,
|
||
|
|
"num_tokens": 70963030.0,
|
||
|
|
"step": 2480
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.33871191181242466,
|
||
|
|
"epoch": 3.984,
|
||
|
|
"grad_norm": 1.953125,
|
||
|
|
"learning_rate": 1.6691111111111112e-05,
|
||
|
|
"loss": 0.3215,
|
||
|
|
"mean_token_accuracy": 0.8948656614869833,
|
||
|
|
"num_tokens": 71264525.0,
|
||
|
|
"step": 2490
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3556901432573795,
|
||
|
|
"epoch": 4.0,
|
||
|
|
"grad_norm": 2.265625,
|
||
|
|
"learning_rate": 1.666888888888889e-05,
|
||
|
|
"loss": 0.337,
|
||
|
|
"mean_token_accuracy": 0.8900394972413779,
|
||
|
|
"num_tokens": 71545696.0,
|
||
|
|
"step": 2500
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 4.0,
|
||
|
|
"eval_math_entropy": 0.3949365179538727,
|
||
|
|
"eval_math_loss": 0.7542520761489868,
|
||
|
|
"eval_math_mean_token_accuracy": 0.8234356184005738,
|
||
|
|
"eval_math_num_tokens": 71545696.0,
|
||
|
|
"eval_math_runtime": 35.7966,
|
||
|
|
"eval_math_samples_per_second": 13.968,
|
||
|
|
"eval_math_steps_per_second": 3.492,
|
||
|
|
"step": 2500
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 4.0,
|
||
|
|
"eval_physics_entropy": 0.3519064650535583,
|
||
|
|
"eval_physics_loss": 0.9068982005119324,
|
||
|
|
"eval_physics_mean_token_accuracy": 0.8042608909606933,
|
||
|
|
"eval_physics_num_tokens": 71545696.0,
|
||
|
|
"eval_physics_runtime": 41.1579,
|
||
|
|
"eval_physics_samples_per_second": 12.148,
|
||
|
|
"eval_physics_steps_per_second": 3.037,
|
||
|
|
"step": 2500
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"logging_steps": 10,
|
||
|
|
"max_steps": 10000,
|
||
|
|
"num_input_tokens_seen": 0,
|
||
|
|
"num_train_epochs": 16,
|
||
|
|
"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": 5.428120947106271e+18,
|
||
|
|
"train_batch_size": 4,
|
||
|
|
"trial_name": null,
|
||
|
|
"trial_params": null
|
||
|
|
}
|