802 lines
24 KiB
JSON
802 lines
24 KiB
JSON
|
{
|
||
|
|
"best_global_step": 335,
|
||
|
|
"best_metric": 0.2155456691980362,
|
||
|
|
"best_model_checkpoint": "/content/DualMind100_agentic/checkpoint-335",
|
||
|
|
"epoch": 5.0,
|
||
|
|
"eval_steps": 50,
|
||
|
|
"global_step": 335,
|
||
|
|
"is_hyper_param_search": false,
|
||
|
|
"is_local_process_zero": true,
|
||
|
|
"is_world_process_zero": true,
|
||
|
|
"log_history": [
|
||
|
|
{
|
||
|
|
"entropy": 0.16182193332642783,
|
||
|
|
"epoch": 0.07092198581560284,
|
||
|
|
"grad_norm": 0.5555469989776611,
|
||
|
|
"learning_rate": 5e-05,
|
||
|
|
"loss": 0.43506731986999514,
|
||
|
|
"mean_token_accuracy": 0.8946891106665135,
|
||
|
|
"num_tokens": 123430.0,
|
||
|
|
"step": 5
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.26333686951547863,
|
||
|
|
"epoch": 0.14184397163120568,
|
||
|
|
"grad_norm": 0.2982931435108185,
|
||
|
|
"learning_rate": 9.99862592554908e-05,
|
||
|
|
"loss": 0.3709733247756958,
|
||
|
|
"mean_token_accuracy": 0.898155527561903,
|
||
|
|
"num_tokens": 240929.0,
|
||
|
|
"step": 10
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.34238293119706215,
|
||
|
|
"epoch": 0.2127659574468085,
|
||
|
|
"grad_norm": 0.27323758602142334,
|
||
|
|
"learning_rate": 9.950612572673255e-05,
|
||
|
|
"loss": 0.3364511489868164,
|
||
|
|
"mean_token_accuracy": 0.906800027191639,
|
||
|
|
"num_tokens": 373968.0,
|
||
|
|
"step": 15
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2675066887401044,
|
||
|
|
"epoch": 0.28368794326241137,
|
||
|
|
"grad_norm": 0.21929873526096344,
|
||
|
|
"learning_rate": 9.834648865751254e-05,
|
||
|
|
"loss": 0.2727418184280396,
|
||
|
|
"mean_token_accuracy": 0.9184959597885609,
|
||
|
|
"num_tokens": 494846.0,
|
||
|
|
"step": 20
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2389053585124202,
|
||
|
|
"epoch": 0.3546099290780142,
|
||
|
|
"grad_norm": 0.19836418330669403,
|
||
|
|
"learning_rate": 9.652326481535435e-05,
|
||
|
|
"loss": 0.3025625705718994,
|
||
|
|
"mean_token_accuracy": 0.9214970305562019,
|
||
|
|
"num_tokens": 630306.0,
|
||
|
|
"step": 25
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.25811560889706014,
|
||
|
|
"epoch": 0.425531914893617,
|
||
|
|
"grad_norm": 0.19982172548770905,
|
||
|
|
"learning_rate": 9.406147912470143e-05,
|
||
|
|
"loss": 0.330981969833374,
|
||
|
|
"mean_token_accuracy": 0.9110702231526375,
|
||
|
|
"num_tokens": 758739.0,
|
||
|
|
"step": 30
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.24841026657959447,
|
||
|
|
"epoch": 0.49645390070921985,
|
||
|
|
"grad_norm": 0.16917800903320312,
|
||
|
|
"learning_rate": 9.099492118367123e-05,
|
||
|
|
"loss": 0.26140174865722654,
|
||
|
|
"mean_token_accuracy": 0.9228238843381404,
|
||
|
|
"num_tokens": 886920.0,
|
||
|
|
"step": 35
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2569453418778721,
|
||
|
|
"epoch": 0.5673758865248227,
|
||
|
|
"grad_norm": 0.2460022270679474,
|
||
|
|
"learning_rate": 8.736568148000386e-05,
|
||
|
|
"loss": 0.28326830863952634,
|
||
|
|
"mean_token_accuracy": 0.9195088751614093,
|
||
|
|
"num_tokens": 1010775.0,
|
||
|
|
"step": 40
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.29926245249807837,
|
||
|
|
"epoch": 0.6382978723404256,
|
||
|
|
"grad_norm": 0.17467764019966125,
|
||
|
|
"learning_rate": 8.322357367194109e-05,
|
||
|
|
"loss": 0.3357896566390991,
|
||
|
|
"mean_token_accuracy": 0.9085973531007767,
|
||
|
|
"num_tokens": 1134373.0,
|
||
|
|
"step": 45
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2604390908149071,
|
||
|
|
"epoch": 0.7092198581560284,
|
||
|
|
"grad_norm": 0.16242940723896027,
|
||
|
|
"learning_rate": 7.86254508636036e-05,
|
||
|
|
"loss": 0.2815252304077148,
|
||
|
|
"mean_token_accuracy": 0.9152372993528843,
|
||
|
|
"num_tokens": 1256155.0,
|
||
|
|
"step": 50
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.7092198581560284,
|
||
|
|
"eval_entropy": 0.22815538255963475,
|
||
|
|
"eval_loss": 0.23423044383525848,
|
||
|
|
"eval_mean_token_accuracy": 0.9300008341670036,
|
||
|
|
"eval_num_tokens": 1256155.0,
|
||
|
|
"eval_runtime": 5.3417,
|
||
|
|
"eval_samples_per_second": 8.986,
|
||
|
|
"eval_steps_per_second": 8.986,
|
||
|
|
"step": 50
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2784213587117847,
|
||
|
|
"epoch": 0.7801418439716312,
|
||
|
|
"grad_norm": 0.1878730058670044,
|
||
|
|
"learning_rate": 7.363442525942826e-05,
|
||
|
|
"loss": 0.2845616340637207,
|
||
|
|
"mean_token_accuracy": 0.9107947684824467,
|
||
|
|
"num_tokens": 1378868.0,
|
||
|
|
"step": 55
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.29253384250914677,
|
||
|
|
"epoch": 0.851063829787234,
|
||
|
|
"grad_norm": 0.17584209144115448,
|
||
|
|
"learning_rate": 6.831900190841232e-05,
|
||
|
|
"loss": 0.31951193809509276,
|
||
|
|
"mean_token_accuracy": 0.9116400554776192,
|
||
|
|
"num_tokens": 1509648.0,
|
||
|
|
"step": 60
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.27282086764462293,
|
||
|
|
"epoch": 0.9219858156028369,
|
||
|
|
"grad_norm": 0.1782023161649704,
|
||
|
|
"learning_rate": 6.275213842808383e-05,
|
||
|
|
"loss": 0.28685288429260253,
|
||
|
|
"mean_token_accuracy": 0.9180420659482479,
|
||
|
|
"num_tokens": 1640179.0,
|
||
|
|
"step": 65
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.26427618057350627,
|
||
|
|
"epoch": 0.9929078014184397,
|
||
|
|
"grad_norm": 0.2283456027507782,
|
||
|
|
"learning_rate": 5.70102436140943e-05,
|
||
|
|
"loss": 0.29142701625823975,
|
||
|
|
"mean_token_accuracy": 0.9174253843724728,
|
||
|
|
"num_tokens": 1764592.0,
|
||
|
|
"step": 70
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2524041742629682,
|
||
|
|
"epoch": 1.0567375886524824,
|
||
|
|
"grad_norm": 0.16881126165390015,
|
||
|
|
"learning_rate": 5.117212868016303e-05,
|
||
|
|
"loss": 0.276317834854126,
|
||
|
|
"mean_token_accuracy": 0.9248060741358333,
|
||
|
|
"num_tokens": 1874267.0,
|
||
|
|
"step": 75
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.29945559421321377,
|
||
|
|
"epoch": 1.127659574468085,
|
||
|
|
"grad_norm": 0.2186647355556488,
|
||
|
|
"learning_rate": 4.5317925523282464e-05,
|
||
|
|
"loss": 0.29511370658874514,
|
||
|
|
"mean_token_accuracy": 0.907706456631422,
|
||
|
|
"num_tokens": 1991985.0,
|
||
|
|
"step": 80
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.24618784408958164,
|
||
|
|
"epoch": 1.198581560283688,
|
||
|
|
"grad_norm": 0.15828007459640503,
|
||
|
|
"learning_rate": 3.952798686169279e-05,
|
||
|
|
"loss": 0.2731397867202759,
|
||
|
|
"mean_token_accuracy": 0.9240930497646331,
|
||
|
|
"num_tokens": 2118168.0,
|
||
|
|
"step": 85
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2373207072203513,
|
||
|
|
"epoch": 1.2695035460992907,
|
||
|
|
"grad_norm": 0.20958004891872406,
|
||
|
|
"learning_rate": 3.388178334194232e-05,
|
||
|
|
"loss": 0.27325007915496824,
|
||
|
|
"mean_token_accuracy": 0.9249208129942417,
|
||
|
|
"num_tokens": 2240951.0,
|
||
|
|
"step": 90
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.22738294367445633,
|
||
|
|
"epoch": 1.3404255319148937,
|
||
|
|
"grad_norm": 0.19565178453922272,
|
||
|
|
"learning_rate": 2.8456812752951485e-05,
|
||
|
|
"loss": 0.25463979244232177,
|
||
|
|
"mean_token_accuracy": 0.9295149140059948,
|
||
|
|
"num_tokens": 2364856.0,
|
||
|
|
"step": 95
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.247153190523386,
|
||
|
|
"epoch": 1.4113475177304964,
|
||
|
|
"grad_norm": 0.1792992502450943,
|
||
|
|
"learning_rate": 2.3327536318821495e-05,
|
||
|
|
"loss": 0.2785398244857788,
|
||
|
|
"mean_token_accuracy": 0.9210605584084988,
|
||
|
|
"num_tokens": 2496977.0,
|
||
|
|
"step": 100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.4113475177304964,
|
||
|
|
"eval_entropy": 0.2170733372719648,
|
||
|
|
"eval_loss": 0.22858719527721405,
|
||
|
|
"eval_mean_token_accuracy": 0.9310164228081703,
|
||
|
|
"eval_num_tokens": 2496977.0,
|
||
|
|
"eval_runtime": 5.44,
|
||
|
|
"eval_samples_per_second": 8.824,
|
||
|
|
"eval_steps_per_second": 8.824,
|
||
|
|
"step": 100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2818379683652893,
|
||
|
|
"epoch": 1.4822695035460993,
|
||
|
|
"grad_norm": 0.17297224700450897,
|
||
|
|
"learning_rate": 1.8564356670455767e-05,
|
||
|
|
"loss": 0.28323366641998293,
|
||
|
|
"mean_token_accuracy": 0.9150983460247517,
|
||
|
|
"num_tokens": 2625956.0,
|
||
|
|
"step": 105
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.25849783700541595,
|
||
|
|
"epoch": 1.5531914893617023,
|
||
|
|
"grad_norm": 0.17231720685958862,
|
||
|
|
"learning_rate": 1.4232651523993634e-05,
|
||
|
|
"loss": 0.26291708946228026,
|
||
|
|
"mean_token_accuracy": 0.9214990958571434,
|
||
|
|
"num_tokens": 2762411.0,
|
||
|
|
"step": 110
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.23584751789458097,
|
||
|
|
"epoch": 1.624113475177305,
|
||
|
|
"grad_norm": 0.19976992905139923,
|
||
|
|
"learning_rate": 1.0391876329443533e-05,
|
||
|
|
"loss": 0.24894940853118896,
|
||
|
|
"mean_token_accuracy": 0.9259149916470051,
|
||
|
|
"num_tokens": 2887514.0,
|
||
|
|
"step": 115
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2430717541486956,
|
||
|
|
"epoch": 1.6950354609929077,
|
||
|
|
"grad_norm": 0.21702121198177338,
|
||
|
|
"learning_rate": 7.094748206242796e-06,
|
||
|
|
"loss": 0.2538208246231079,
|
||
|
|
"mean_token_accuracy": 0.9232472732663155,
|
||
|
|
"num_tokens": 3013243.0,
|
||
|
|
"step": 120
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.23513522319262847,
|
||
|
|
"epoch": 1.7659574468085106,
|
||
|
|
"grad_norm": 0.16074544191360474,
|
||
|
|
"learning_rate": 4.386522366755169e-06,
|
||
|
|
"loss": 0.22128725051879883,
|
||
|
|
"mean_token_accuracy": 0.9245993517339229,
|
||
|
|
"num_tokens": 3134144.0,
|
||
|
|
"step": 125
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.24327249150956048,
|
||
|
|
"epoch": 1.8368794326241136,
|
||
|
|
"grad_norm": 0.16139043867588043,
|
||
|
|
"learning_rate": 2.3043709592610485e-06,
|
||
|
|
"loss": 0.26496360301971433,
|
||
|
|
"mean_token_accuracy": 0.9262310199439525,
|
||
|
|
"num_tokens": 3260737.0,
|
||
|
|
"step": 130
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.23524304252350703,
|
||
|
|
"epoch": 1.9078014184397163,
|
||
|
|
"grad_norm": 0.16922682523727417,
|
||
|
|
"learning_rate": 8.768728562211947e-07,
|
||
|
|
"loss": 0.2534662961959839,
|
||
|
|
"mean_token_accuracy": 0.924710838496685,
|
||
|
|
"num_tokens": 3383649.0,
|
||
|
|
"step": 135
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2577046745980624,
|
||
|
|
"epoch": 1.978723404255319,
|
||
|
|
"grad_norm": 0.18219797313213348,
|
||
|
|
"learning_rate": 1.236213907994943e-07,
|
||
|
|
"loss": 0.2727572679519653,
|
||
|
|
"mean_token_accuracy": 0.9200983986258506,
|
||
|
|
"num_tokens": 3511363.0,
|
||
|
|
"step": 140
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.0,
|
||
|
|
"eval_entropy": 0.21362574538215995,
|
||
|
|
"eval_loss": 0.22699420154094696,
|
||
|
|
"eval_mean_token_accuracy": 0.9318861390153567,
|
||
|
|
"eval_num_tokens": 3550882.0,
|
||
|
|
"eval_runtime": 5.4445,
|
||
|
|
"eval_samples_per_second": 8.816,
|
||
|
|
"eval_steps_per_second": 8.816,
|
||
|
|
"step": 142
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.6595380927125613,
|
||
|
|
"epoch": 2.1641791044776117,
|
||
|
|
"grad_norm": 0.4891488254070282,
|
||
|
|
"learning_rate": 6.554477648417657e-05,
|
||
|
|
"loss": 0.3567054271697998,
|
||
|
|
"mean_token_accuracy": 0.8980764100948969,
|
||
|
|
"num_tokens": 74137.0,
|
||
|
|
"step": 145
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3680407987907529,
|
||
|
|
"epoch": 2.2388059701492535,
|
||
|
|
"grad_norm": 0.2562958300113678,
|
||
|
|
"learning_rate": 6.317935830345338e-05,
|
||
|
|
"loss": 0.2808666467666626,
|
||
|
|
"mean_token_accuracy": 0.9229060940444469,
|
||
|
|
"num_tokens": 201191.0,
|
||
|
|
"step": 150
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.2388059701492535,
|
||
|
|
"eval_entropy": 0.29109591808584,
|
||
|
|
"eval_loss": 0.25212642550468445,
|
||
|
|
"eval_mean_token_accuracy": 0.9247896313667298,
|
||
|
|
"eval_num_tokens": 201191.0,
|
||
|
|
"eval_runtime": 5.3589,
|
||
|
|
"eval_samples_per_second": 8.397,
|
||
|
|
"eval_steps_per_second": 8.397,
|
||
|
|
"step": 150
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.28654855480417607,
|
||
|
|
"epoch": 2.3134328358208958,
|
||
|
|
"grad_norm": 0.22995419800281525,
|
||
|
|
"learning_rate": 6.078178934582885e-05,
|
||
|
|
"loss": 0.2800537347793579,
|
||
|
|
"mean_token_accuracy": 0.9219874046742916,
|
||
|
|
"num_tokens": 319457.0,
|
||
|
|
"step": 155
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2231586970272474,
|
||
|
|
"epoch": 2.388059701492537,
|
||
|
|
"grad_norm": 0.17113032937049866,
|
||
|
|
"learning_rate": 5.835791843207916e-05,
|
||
|
|
"loss": 0.2297281265258789,
|
||
|
|
"mean_token_accuracy": 0.9307773187756538,
|
||
|
|
"num_tokens": 441969.0,
|
||
|
|
"step": 160
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.20086921484908088,
|
||
|
|
"epoch": 2.4626865671641793,
|
||
|
|
"grad_norm": 0.13817362487316132,
|
||
|
|
"learning_rate": 5.5913658546068295e-05,
|
||
|
|
"loss": 0.20193655490875245,
|
||
|
|
"mean_token_accuracy": 0.9379530139267445,
|
||
|
|
"num_tokens": 564392.0,
|
||
|
|
"step": 165
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2754073503077962,
|
||
|
|
"epoch": 2.5373134328358207,
|
||
|
|
"grad_norm": 0.19452188909053802,
|
||
|
|
"learning_rate": 5.34549724101439e-05,
|
||
|
|
"loss": 0.27802190780639646,
|
||
|
|
"mean_token_accuracy": 0.9231111846864224,
|
||
|
|
"num_tokens": 696793.0,
|
||
|
|
"step": 170
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.24629125613719224,
|
||
|
|
"epoch": 2.611940298507463,
|
||
|
|
"grad_norm": 0.1801047921180725,
|
||
|
|
"learning_rate": 5.0987857939197324e-05,
|
||
|
|
"loss": 0.24364988803863524,
|
||
|
|
"mean_token_accuracy": 0.9304406151175499,
|
||
|
|
"num_tokens": 832284.0,
|
||
|
|
"step": 175
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.24637198023265228,
|
||
|
|
"epoch": 2.6865671641791042,
|
||
|
|
"grad_norm": 0.15811710059642792,
|
||
|
|
"learning_rate": 4.851833360887201e-05,
|
||
|
|
"loss": 0.2485351085662842,
|
||
|
|
"mean_token_accuracy": 0.9251487389206886,
|
||
|
|
"num_tokens": 953807.0,
|
||
|
|
"step": 180
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.19791743826935998,
|
||
|
|
"epoch": 2.7611940298507465,
|
||
|
|
"grad_norm": 0.17821364104747772,
|
||
|
|
"learning_rate": 4.6052423773614404e-05,
|
||
|
|
"loss": 0.21749985218048096,
|
||
|
|
"mean_token_accuracy": 0.9387934006750583,
|
||
|
|
"num_tokens": 1074631.0,
|
||
|
|
"step": 185
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.23313339608721434,
|
||
|
|
"epoch": 2.835820895522388,
|
||
|
|
"grad_norm": 0.22559864819049835,
|
||
|
|
"learning_rate": 4.3596143970383664e-05,
|
||
|
|
"loss": 0.23709943294525146,
|
||
|
|
"mean_token_accuracy": 0.9306622236967087,
|
||
|
|
"num_tokens": 1203325.0,
|
||
|
|
"step": 190
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.26784485091629906,
|
||
|
|
"epoch": 2.91044776119403,
|
||
|
|
"grad_norm": 0.20925050973892212,
|
||
|
|
"learning_rate": 4.115548624387137e-05,
|
||
|
|
"loss": 0.2629614114761353,
|
||
|
|
"mean_token_accuracy": 0.9191216990351677,
|
||
|
|
"num_tokens": 1332870.0,
|
||
|
|
"step": 195
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2317873791558668,
|
||
|
|
"epoch": 2.9850746268656714,
|
||
|
|
"grad_norm": 0.15967224538326263,
|
||
|
|
"learning_rate": 3.873640452903026e-05,
|
||
|
|
"loss": 0.23617920875549317,
|
||
|
|
"mean_token_accuracy": 0.931964573264122,
|
||
|
|
"num_tokens": 1450793.0,
|
||
|
|
"step": 200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.9850746268656714,
|
||
|
|
"eval_entropy": 0.2280733192132579,
|
||
|
|
"eval_loss": 0.21912212669849396,
|
||
|
|
"eval_mean_token_accuracy": 0.9303096612294515,
|
||
|
|
"eval_num_tokens": 1450793.0,
|
||
|
|
"eval_runtime": 5.2231,
|
||
|
|
"eval_samples_per_second": 8.615,
|
||
|
|
"eval_steps_per_second": 8.615,
|
||
|
|
"step": 200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.24150694359559566,
|
||
|
|
"epoch": 3.0597014925373136,
|
||
|
|
"grad_norm": 0.17629101872444153,
|
||
|
|
"learning_rate": 3.6344800126570844e-05,
|
||
|
|
"loss": 0.24918053150177003,
|
||
|
|
"mean_token_accuracy": 0.929844381660223,
|
||
|
|
"num_tokens": 1577936.0,
|
||
|
|
"step": 205
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.20565980894316455,
|
||
|
|
"epoch": 3.1343283582089554,
|
||
|
|
"grad_norm": 0.15850670635700226,
|
||
|
|
"learning_rate": 3.3986507306858125e-05,
|
||
|
|
"loss": 0.2212688684463501,
|
||
|
|
"mean_token_accuracy": 0.9367655150592327,
|
||
|
|
"num_tokens": 1709033.0,
|
||
|
|
"step": 210
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.21145587725332007,
|
||
|
|
"epoch": 3.208955223880597,
|
||
|
|
"grad_norm": 0.2352381944656372,
|
||
|
|
"learning_rate": 3.16672790773276e-05,
|
||
|
|
"loss": 0.22223784923553466,
|
||
|
|
"mean_token_accuracy": 0.9365130670368671,
|
||
|
|
"num_tokens": 1831296.0,
|
||
|
|
"step": 215
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.22000169460952748,
|
||
|
|
"epoch": 3.283582089552239,
|
||
|
|
"grad_norm": 0.1637473851442337,
|
||
|
|
"learning_rate": 2.9392773148140408e-05,
|
||
|
|
"loss": 0.22411210536956788,
|
||
|
|
"mean_token_accuracy": 0.9312412738800049,
|
||
|
|
"num_tokens": 1950727.0,
|
||
|
|
"step": 220
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.21870606549200602,
|
||
|
|
"epoch": 3.3582089552238807,
|
||
|
|
"grad_norm": 0.18144957721233368,
|
||
|
|
"learning_rate": 2.716853813031435e-05,
|
||
|
|
"loss": 0.21267392635345458,
|
||
|
|
"mean_token_accuracy": 0.9330131947994232,
|
||
|
|
"num_tokens": 2084554.0,
|
||
|
|
"step": 225
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.18131871867808513,
|
||
|
|
"epoch": 3.4328358208955225,
|
||
|
|
"grad_norm": 0.14299768209457397,
|
||
|
|
"learning_rate": 2.500000000000001e-05,
|
||
|
|
"loss": 0.19915605783462526,
|
||
|
|
"mean_token_accuracy": 0.9421401508152485,
|
||
|
|
"num_tokens": 2207986.0,
|
||
|
|
"step": 230
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.18844851908506827,
|
||
|
|
"epoch": 3.5074626865671643,
|
||
|
|
"grad_norm": 0.21421347558498383,
|
||
|
|
"learning_rate": 2.289244886192207e-05,
|
||
|
|
"loss": 0.18854159116744995,
|
||
|
|
"mean_token_accuracy": 0.9412525840103626,
|
||
|
|
"num_tokens": 2330341.0,
|
||
|
|
"step": 235
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2162812261347426,
|
||
|
|
"epoch": 3.582089552238806,
|
||
|
|
"grad_norm": 0.19194981455802917,
|
||
|
|
"learning_rate": 2.0851026044276406e-05,
|
||
|
|
"loss": 0.2259037971496582,
|
||
|
|
"mean_token_accuracy": 0.931151507049799,
|
||
|
|
"num_tokens": 2454431.0,
|
||
|
|
"step": 240
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.20769512332044543,
|
||
|
|
"epoch": 3.656716417910448,
|
||
|
|
"grad_norm": 0.1685398668050766,
|
||
|
|
"learning_rate": 1.888071155656421e-05,
|
||
|
|
"loss": 0.2129582643508911,
|
||
|
|
"mean_token_accuracy": 0.934239100664854,
|
||
|
|
"num_tokens": 2578412.0,
|
||
|
|
"step": 245
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2097844194737263,
|
||
|
|
"epoch": 3.7313432835820897,
|
||
|
|
"grad_norm": 0.1834094077348709,
|
||
|
|
"learning_rate": 1.6986311940960147e-05,
|
||
|
|
"loss": 0.21508474349975587,
|
||
|
|
"mean_token_accuracy": 0.9365896977484226,
|
||
|
|
"num_tokens": 2704058.0,
|
||
|
|
"step": 250
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.7313432835820897,
|
||
|
|
"eval_entropy": 0.21005761296384864,
|
||
|
|
"eval_loss": 0.21609170734882355,
|
||
|
|
"eval_mean_token_accuracy": 0.9321770442856683,
|
||
|
|
"eval_num_tokens": 2704058.0,
|
||
|
|
"eval_runtime": 5.3012,
|
||
|
|
"eval_samples_per_second": 8.489,
|
||
|
|
"eval_steps_per_second": 8.489,
|
||
|
|
"step": 250
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.22172395456582308,
|
||
|
|
"epoch": 3.8059701492537314,
|
||
|
|
"grad_norm": 0.16993354260921478,
|
||
|
|
"learning_rate": 1.5172448546850165e-05,
|
||
|
|
"loss": 0.2224128484725952,
|
||
|
|
"mean_token_accuracy": 0.9313701257109642,
|
||
|
|
"num_tokens": 2822117.0,
|
||
|
|
"step": 255
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.20137031929916702,
|
||
|
|
"epoch": 3.8805970149253732,
|
||
|
|
"grad_norm": 0.18573690950870514,
|
||
|
|
"learning_rate": 1.3443546257143624e-05,
|
||
|
|
"loss": 0.20644056797027588,
|
||
|
|
"mean_token_accuracy": 0.935058569163084,
|
||
|
|
"num_tokens": 2954055.0,
|
||
|
|
"step": 260
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.21610731023247354,
|
||
|
|
"epoch": 3.955223880597015,
|
||
|
|
"grad_norm": 0.18307460844516754,
|
||
|
|
"learning_rate": 1.1803822693861378e-05,
|
||
|
|
"loss": 0.21918556690216065,
|
||
|
|
"mean_token_accuracy": 0.9331639289855957,
|
||
|
|
"num_tokens": 3076367.0,
|
||
|
|
"step": 265
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.18582435508724301,
|
||
|
|
"epoch": 4.029850746268656,
|
||
|
|
"grad_norm": 0.17089995741844177,
|
||
|
|
"learning_rate": 1.0257277929332332e-05,
|
||
|
|
"loss": 0.19541841745376587,
|
||
|
|
"mean_token_accuracy": 0.9410657376050949,
|
||
|
|
"num_tokens": 3200217.0,
|
||
|
|
"step": 270
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.1961809309606906,
|
||
|
|
"epoch": 4.104477611940299,
|
||
|
|
"grad_norm": 0.19942045211791992,
|
||
|
|
"learning_rate": 8.80768472809842e-06,
|
||
|
|
"loss": 0.18296244144439697,
|
||
|
|
"mean_token_accuracy": 0.9419995754957199,
|
||
|
|
"num_tokens": 3325910.0,
|
||
|
|
"step": 275
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.17922255574376322,
|
||
|
|
"epoch": 4.17910447761194,
|
||
|
|
"grad_norm": 0.17642585933208466,
|
||
|
|
"learning_rate": 7.458579343331995e-06,
|
||
|
|
"loss": 0.1955660581588745,
|
||
|
|
"mean_token_accuracy": 0.9431271247565747,
|
||
|
|
"num_tokens": 3449422.0,
|
||
|
|
"step": 280
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.20833609464461916,
|
||
|
|
"epoch": 4.253731343283582,
|
||
|
|
"grad_norm": 0.18559329211711884,
|
||
|
|
"learning_rate": 6.213252890218163e-06,
|
||
|
|
"loss": 0.20198218822479247,
|
||
|
|
"mean_token_accuracy": 0.934450775384903,
|
||
|
|
"num_tokens": 3569595.0,
|
||
|
|
"step": 285
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.20185105923446828,
|
||
|
|
"epoch": 4.3283582089552235,
|
||
|
|
"grad_norm": 0.14244937896728516,
|
||
|
|
"learning_rate": 5.074743317346009e-06,
|
||
|
|
"loss": 0.20316836833953858,
|
||
|
|
"mean_token_accuracy": 0.9381563618779183,
|
||
|
|
"num_tokens": 3688537.0,
|
||
|
|
"step": 290
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.21535796255921014,
|
||
|
|
"epoch": 4.402985074626866,
|
||
|
|
"grad_norm": 0.17846742272377014,
|
||
|
|
"learning_rate": 4.045827995694834e-06,
|
||
|
|
"loss": 0.2133563280105591,
|
||
|
|
"mean_token_accuracy": 0.9316843308508396,
|
||
|
|
"num_tokens": 3819875.0,
|
||
|
|
"step": 295
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.17974913779762575,
|
||
|
|
"epoch": 4.477611940298507,
|
||
|
|
"grad_norm": 0.20783886313438416,
|
||
|
|
"learning_rate": 3.1290169432939553e-06,
|
||
|
|
"loss": 0.2002777338027954,
|
||
|
|
"mean_token_accuracy": 0.9437027871608734,
|
||
|
|
"num_tokens": 3951096.0,
|
||
|
|
"step": 300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 4.477611940298507,
|
||
|
|
"eval_entropy": 0.20027935401433045,
|
||
|
|
"eval_loss": 0.2157754898071289,
|
||
|
|
"eval_mean_token_accuracy": 0.9326928456624349,
|
||
|
|
"eval_num_tokens": 3951096.0,
|
||
|
|
"eval_runtime": 5.1907,
|
||
|
|
"eval_samples_per_second": 8.669,
|
||
|
|
"eval_steps_per_second": 8.669,
|
||
|
|
"step": 300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.1885046276351204,
|
||
|
|
"epoch": 4.552238805970149,
|
||
|
|
"grad_norm": 0.20049376785755157,
|
||
|
|
"learning_rate": 2.3265467020847866e-06,
|
||
|
|
"loss": 0.1992756724357605,
|
||
|
|
"mean_token_accuracy": 0.9404157467186451,
|
||
|
|
"num_tokens": 4073686.0,
|
||
|
|
"step": 305
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.20385300312773325,
|
||
|
|
"epoch": 4.6268656716417915,
|
||
|
|
"grad_norm": 0.22206291556358337,
|
||
|
|
"learning_rate": 1.6403748819221466e-06,
|
||
|
|
"loss": 0.21113824844360352,
|
||
|
|
"mean_token_accuracy": 0.937010295689106,
|
||
|
|
"num_tokens": 4198560.0,
|
||
|
|
"step": 310
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.20165168155217544,
|
||
|
|
"epoch": 4.701492537313433,
|
||
|
|
"grad_norm": 0.22487279772758484,
|
||
|
|
"learning_rate": 1.0721753850247984e-06,
|
||
|
|
"loss": 0.2212970495223999,
|
||
|
|
"mean_token_accuracy": 0.9365487344563007,
|
||
|
|
"num_tokens": 4320893.0,
|
||
|
|
"step": 315
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2170650392654352,
|
||
|
|
"epoch": 4.776119402985074,
|
||
|
|
"grad_norm": 0.1543811559677124,
|
||
|
|
"learning_rate": 6.233343225249933e-07,
|
||
|
|
"loss": 0.22227189540863038,
|
||
|
|
"mean_token_accuracy": 0.9313660271465778,
|
||
|
|
"num_tokens": 4453160.0,
|
||
|
|
"step": 320
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.18664323054545093,
|
||
|
|
"epoch": 4.850746268656716,
|
||
|
|
"grad_norm": 0.20742292702198029,
|
||
|
|
"learning_rate": 2.9494663307847447e-07,
|
||
|
|
"loss": 0.2171410322189331,
|
||
|
|
"mean_token_accuracy": 0.939918453246355,
|
||
|
|
"num_tokens": 4580581.0,
|
||
|
|
"step": 325
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.21867067418352235,
|
||
|
|
"epoch": 4.925373134328359,
|
||
|
|
"grad_norm": 0.17955397069454193,
|
||
|
|
"learning_rate": 8.781341178393244e-08,
|
||
|
|
"loss": 0.2199192762374878,
|
||
|
|
"mean_token_accuracy": 0.9308036096394062,
|
||
|
|
"num_tokens": 4705647.0,
|
||
|
|
"step": 330
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.19878723044821528,
|
||
|
|
"epoch": 5.0,
|
||
|
|
"grad_norm": 0.20007997751235962,
|
||
|
|
"learning_rate": 2.4399559277132888e-09,
|
||
|
|
"loss": 0.19873000383377076,
|
||
|
|
"mean_token_accuracy": 0.9358412712812424,
|
||
|
|
"num_tokens": 4828163.0,
|
||
|
|
"step": 335
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 5.0,
|
||
|
|
"eval_entropy": 0.20022574655918612,
|
||
|
|
"eval_loss": 0.2155456691980362,
|
||
|
|
"eval_mean_token_accuracy": 0.9328206459681193,
|
||
|
|
"eval_num_tokens": 4828163.0,
|
||
|
|
"eval_runtime": 5.2104,
|
||
|
|
"eval_samples_per_second": 8.637,
|
||
|
|
"eval_steps_per_second": 8.637,
|
||
|
|
"step": 335
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 5.0,
|
||
|
|
"step": 335,
|
||
|
|
"total_flos": 7.27718510456832e+16,
|
||
|
|
"train_loss": 0.12883905314687472,
|
||
|
|
"train_runtime": 1744.2725,
|
||
|
|
"train_samples_per_second": 3.073,
|
||
|
|
"train_steps_per_second": 0.192
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"logging_steps": 5,
|
||
|
|
"max_steps": 335,
|
||
|
|
"num_input_tokens_seen": 0,
|
||
|
|
"num_train_epochs": 5,
|
||
|
|
"save_steps": 50,
|
||
|
|
"stateful_callbacks": {
|
||
|
|
"TrainerControl": {
|
||
|
|
"args": {
|
||
|
|
"should_epoch_stop": false,
|
||
|
|
"should_evaluate": false,
|
||
|
|
"should_log": false,
|
||
|
|
"should_save": true,
|
||
|
|
"should_training_stop": true
|
||
|
|
},
|
||
|
|
"attributes": {}
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"total_flos": 7.27718510456832e+16,
|
||
|
|
"train_batch_size": 1,
|
||
|
|
"trial_name": null,
|
||
|
|
"trial_params": null
|
||
|
|
}
|