grpo_test / trainer_state.json
minosu's picture
Upload folder using huggingface_hub
8f2a06c verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.8028904054596547,
"eval_steps": 500,
"global_step": 100,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"completion_length": 155.3333339691162,
"epoch": 0.008028904054596548,
"grad_norm": 0.8266991671440796,
"kl": 0.0,
"learning_rate": 3.0000000000000004e-08,
"loss": 0.0,
"reward": 2.3694444835186004,
"reward_std": 0.706648226082325,
"rewards/accuracy_reward_log": 1.4888889163732528,
"rewards/format_number_reward": 0.4138888940215111,
"rewards/format_reasoning_reward": 0.46666666865348816,
"step": 1
},
{
"completion_length": 177.53333435058593,
"epoch": 0.016057808109193095,
"grad_norm": 0.853670554176349,
"kl": 0.0,
"learning_rate": 6.000000000000001e-08,
"loss": 0.0,
"reward": 2.3472222685813904,
"reward_std": 0.6901098385453224,
"rewards/accuracy_reward_log": 1.500000025331974,
"rewards/format_number_reward": 0.4000000059604645,
"rewards/format_reasoning_reward": 0.4472222253680229,
"step": 2
},
{
"completion_length": 160.18333435058594,
"epoch": 0.02408671216378964,
"grad_norm": 0.6327729129725634,
"kl": 0.00010757744312286376,
"learning_rate": 9e-08,
"loss": 0.0,
"reward": 2.4138889491558073,
"reward_std": 0.5638244189321995,
"rewards/accuracy_reward_log": 1.5222222447395324,
"rewards/format_number_reward": 0.4194444492459297,
"rewards/format_reasoning_reward": 0.4722222238779068,
"step": 3
},
{
"completion_length": 167.59444580078124,
"epoch": 0.03211561621838619,
"grad_norm": 0.6798573023319466,
"kl": 0.00010120868682861328,
"learning_rate": 1.2000000000000002e-07,
"loss": 0.0,
"reward": 2.20833335518837,
"reward_std": 0.5129089742898941,
"rewards/accuracy_reward_log": 1.333333359658718,
"rewards/format_number_reward": 0.4027777835726738,
"rewards/format_reasoning_reward": 0.4722222238779068,
"step": 4
},
{
"completion_length": 147.56666870117186,
"epoch": 0.04014452027298274,
"grad_norm": 0.5872669908882053,
"kl": 9.968876838684082e-05,
"learning_rate": 1.5000000000000002e-07,
"loss": 0.0,
"reward": 2.4361111402511595,
"reward_std": 0.48325310088694096,
"rewards/accuracy_reward_log": 1.5333333551883697,
"rewards/format_number_reward": 0.42222222685813904,
"rewards/format_reasoning_reward": 0.48055555671453476,
"step": 5
},
{
"completion_length": 168.4333324432373,
"epoch": 0.04817342432757928,
"grad_norm": 0.842231210713151,
"kl": 0.000108298659324646,
"learning_rate": 1.8e-07,
"loss": 0.0,
"reward": 1.9972222298383713,
"reward_std": 0.6861602704972029,
"rewards/accuracy_reward_log": 1.188888917118311,
"rewards/format_number_reward": 0.35555556118488313,
"rewards/format_reasoning_reward": 0.45277778059244156,
"step": 6
},
{
"completion_length": 161.47222175598145,
"epoch": 0.056202328382175835,
"grad_norm": 0.6301911132102126,
"kl": 0.00010362863540649414,
"learning_rate": 2.1000000000000003e-07,
"loss": 0.0,
"reward": 2.344444477558136,
"reward_std": 0.5787256445735693,
"rewards/accuracy_reward_log": 1.455555585026741,
"rewards/format_number_reward": 0.41111111640930176,
"rewards/format_reasoning_reward": 0.47777777910232544,
"step": 7
},
{
"completion_length": 178.705558013916,
"epoch": 0.06423123243677238,
"grad_norm": 0.7904000046666478,
"kl": 0.00015439987182617188,
"learning_rate": 2.4000000000000003e-07,
"loss": 0.0,
"reward": 2.033333358168602,
"reward_std": 0.6374398373067379,
"rewards/accuracy_reward_log": 1.2000000223517417,
"rewards/format_number_reward": 0.37777778320014477,
"rewards/format_reasoning_reward": 0.4555555582046509,
"step": 8
},
{
"completion_length": 154.6055564880371,
"epoch": 0.07226013649136893,
"grad_norm": 0.8113384704385274,
"kl": 0.00010915249586105347,
"learning_rate": 2.7e-07,
"loss": 0.0,
"reward": 2.4722222566604612,
"reward_std": 0.38092764765024184,
"rewards/accuracy_reward_log": 1.5777777969837188,
"rewards/format_number_reward": 0.4194444492459297,
"rewards/format_reasoning_reward": 0.4750000014901161,
"step": 9
},
{
"completion_length": 155.86111221313476,
"epoch": 0.08028904054596547,
"grad_norm": 0.6734680070959974,
"kl": 0.00010448098182678223,
"learning_rate": 3.0000000000000004e-07,
"loss": 0.0,
"reward": 2.3611111462116243,
"reward_std": 0.5814844127744436,
"rewards/accuracy_reward_log": 1.4777778029441833,
"rewards/format_number_reward": 0.413888893276453,
"rewards/format_reasoning_reward": 0.4694444462656975,
"step": 10
},
{
"completion_length": 167.8611110687256,
"epoch": 0.08831794460056203,
"grad_norm": 0.6192549717392329,
"kl": 0.0002144932746887207,
"learning_rate": 3.3e-07,
"loss": 0.0,
"reward": 2.3972222745418548,
"reward_std": 0.5403781462460756,
"rewards/accuracy_reward_log": 1.511111134290695,
"rewards/format_number_reward": 0.4194444492459297,
"rewards/format_reasoning_reward": 0.46666666865348816,
"step": 11
},
{
"completion_length": 159.29444694519043,
"epoch": 0.09634684865515857,
"grad_norm": 0.6219588398194654,
"kl": 0.00014747381210327147,
"learning_rate": 3.6e-07,
"loss": 0.0,
"reward": 2.22777781188488,
"reward_std": 0.6599382009357214,
"rewards/accuracy_reward_log": 1.3777778007090091,
"rewards/format_number_reward": 0.3805555604398251,
"rewards/format_reasoning_reward": 0.4694444462656975,
"step": 12
},
{
"completion_length": 163.36666717529297,
"epoch": 0.10437575270975512,
"grad_norm": 0.5303599869712791,
"kl": 0.0001802980899810791,
"learning_rate": 3.9e-07,
"loss": 0.0,
"reward": 2.388888931274414,
"reward_std": 0.38289184793829917,
"rewards/accuracy_reward_log": 1.4888889119029045,
"rewards/format_number_reward": 0.4111111145466566,
"rewards/format_reasoning_reward": 0.4888888895511627,
"step": 13
},
{
"completion_length": 153.45000038146972,
"epoch": 0.11240465676435167,
"grad_norm": 0.5530377333718409,
"kl": 0.0004338264465332031,
"learning_rate": 4.2000000000000006e-07,
"loss": 0.0,
"reward": 2.569444465637207,
"reward_std": 0.42339018881320956,
"rewards/accuracy_reward_log": 1.6333333551883698,
"rewards/format_number_reward": 0.45277778059244156,
"rewards/format_reasoning_reward": 0.4833333343267441,
"step": 14
},
{
"completion_length": 165.07777938842773,
"epoch": 0.12043356081894821,
"grad_norm": 0.49102029519105816,
"kl": 0.0006180524826049805,
"learning_rate": 4.5e-07,
"loss": 0.0,
"reward": 2.597222250699997,
"reward_std": 0.3478126596659422,
"rewards/accuracy_reward_log": 1.6444444566965104,
"rewards/format_number_reward": 0.46388889104127884,
"rewards/format_reasoning_reward": 0.4888888895511627,
"step": 15
},
{
"completion_length": 164.92778129577636,
"epoch": 0.12846246487354476,
"grad_norm": 0.6338598904220616,
"kl": 0.0007488012313842773,
"learning_rate": 4.800000000000001e-07,
"loss": 0.0,
"reward": 2.319444465637207,
"reward_std": 0.5170031324028969,
"rewards/accuracy_reward_log": 1.4222222447395325,
"rewards/format_number_reward": 0.4277777820825577,
"rewards/format_reasoning_reward": 0.4694444462656975,
"step": 16
},
{
"completion_length": 156.144446182251,
"epoch": 0.13649136892814132,
"grad_norm": 0.5574851133438189,
"kl": 0.0012537002563476562,
"learning_rate": 5.100000000000001e-07,
"loss": 0.0001,
"reward": 2.5555556058883666,
"reward_std": 0.44343185126781465,
"rewards/accuracy_reward_log": 1.6222222417593002,
"rewards/format_number_reward": 0.45000000223517417,
"rewards/format_reasoning_reward": 0.4833333343267441,
"step": 17
},
{
"completion_length": 153.38888931274414,
"epoch": 0.14452027298273787,
"grad_norm": 0.6029065735373917,
"kl": 0.001799297332763672,
"learning_rate": 5.4e-07,
"loss": 0.0001,
"reward": 2.4750000178813933,
"reward_std": 0.5368539825081825,
"rewards/accuracy_reward_log": 1.5444444686174392,
"rewards/format_number_reward": 0.45000000298023224,
"rewards/format_reasoning_reward": 0.48055555671453476,
"step": 18
},
{
"completion_length": 170.45000228881835,
"epoch": 0.1525491770373344,
"grad_norm": 0.6238637844164427,
"kl": 0.0022980213165283204,
"learning_rate": 5.7e-07,
"loss": 0.0001,
"reward": 2.3472222328186034,
"reward_std": 0.5638264730572701,
"rewards/accuracy_reward_log": 1.4444444686174394,
"rewards/format_number_reward": 0.4333333373069763,
"rewards/format_reasoning_reward": 0.4694444462656975,
"step": 19
},
{
"completion_length": 146.12777786254884,
"epoch": 0.16057808109193095,
"grad_norm": 0.4196477311949672,
"kl": 0.00455629825592041,
"learning_rate": 6.000000000000001e-07,
"loss": 0.0002,
"reward": 2.5250000238418577,
"reward_std": 0.27503596656024454,
"rewards/accuracy_reward_log": 1.5888888999819755,
"rewards/format_number_reward": 0.44722222462296485,
"rewards/format_reasoning_reward": 0.4888888895511627,
"step": 20
},
{
"completion_length": 179.50000228881837,
"epoch": 0.1686069851465275,
"grad_norm": 0.5891348987729047,
"kl": 0.0013088226318359376,
"learning_rate": 6.3e-07,
"loss": 0.0001,
"reward": 2.4500000417232513,
"reward_std": 0.5685264855623245,
"rewards/accuracy_reward_log": 1.5666666895151138,
"rewards/format_number_reward": 0.43611111491918564,
"rewards/format_reasoning_reward": 0.4472222253680229,
"step": 21
},
{
"completion_length": 167.82222518920898,
"epoch": 0.17663588920112405,
"grad_norm": 0.5080236888533528,
"kl": 0.0015691757202148438,
"learning_rate": 6.6e-07,
"loss": 0.0001,
"reward": 2.4805555999279023,
"reward_std": 0.38282057382166385,
"rewards/accuracy_reward_log": 1.5333333551883697,
"rewards/format_number_reward": 0.4694444462656975,
"rewards/format_reasoning_reward": 0.47777777910232544,
"step": 22
},
{
"completion_length": 159.62222480773926,
"epoch": 0.1846647932557206,
"grad_norm": 0.5410426478942184,
"kl": 0.0024078369140625,
"learning_rate": 6.900000000000001e-07,
"loss": 0.0001,
"reward": 2.6277777940034865,
"reward_std": 0.332701800763607,
"rewards/accuracy_reward_log": 1.6777777917683125,
"rewards/format_number_reward": 0.46388889029622077,
"rewards/format_reasoning_reward": 0.4861111119389534,
"step": 23
},
{
"completion_length": 157.90000114440917,
"epoch": 0.19269369731031713,
"grad_norm": 0.5882381452711394,
"kl": 0.002088165283203125,
"learning_rate": 7.2e-07,
"loss": 0.0001,
"reward": 2.6722222566604614,
"reward_std": 0.38149141892790794,
"rewards/accuracy_reward_log": 1.7111111283302307,
"rewards/format_number_reward": 0.48055555671453476,
"rewards/format_reasoning_reward": 0.48055555671453476,
"step": 24
},
{
"completion_length": 148.97222328186035,
"epoch": 0.20072260136491368,
"grad_norm": 0.46880008245690374,
"kl": 0.002790069580078125,
"learning_rate": 7.5e-07,
"loss": 0.0001,
"reward": 2.6666667222976685,
"reward_std": 0.42418686002492906,
"rewards/accuracy_reward_log": 1.7111111283302307,
"rewards/format_number_reward": 0.46388889104127884,
"rewards/format_reasoning_reward": 0.49166666716337204,
"step": 25
},
{
"completion_length": 166.62222595214843,
"epoch": 0.20875150541951024,
"grad_norm": 0.6233073148075161,
"kl": 0.003025054931640625,
"learning_rate": 7.8e-07,
"loss": 0.0001,
"reward": 2.544444477558136,
"reward_std": 0.5540193915367126,
"rewards/accuracy_reward_log": 1.6111111342906952,
"rewards/format_number_reward": 0.45277778059244156,
"rewards/format_reasoning_reward": 0.48055555671453476,
"step": 26
},
{
"completion_length": 153.51666755676268,
"epoch": 0.2167804094741068,
"grad_norm": 0.4559712721597118,
"kl": 0.003208160400390625,
"learning_rate": 8.100000000000001e-07,
"loss": 0.0001,
"reward": 2.5555555880069734,
"reward_std": 0.34781265556812285,
"rewards/accuracy_reward_log": 1.5888889104127883,
"rewards/format_number_reward": 0.4750000014901161,
"rewards/format_reasoning_reward": 0.49166666716337204,
"step": 27
},
{
"completion_length": 172.8500011444092,
"epoch": 0.22480931352870334,
"grad_norm": 0.5449553128247491,
"kl": 0.0030437469482421874,
"learning_rate": 8.400000000000001e-07,
"loss": 0.0001,
"reward": 2.4750000476837157,
"reward_std": 0.3911139152944088,
"rewards/accuracy_reward_log": 1.544444465637207,
"rewards/format_number_reward": 0.45000000223517417,
"rewards/format_reasoning_reward": 0.48055555671453476,
"step": 28
},
{
"completion_length": 164.09999961853026,
"epoch": 0.23283821758329987,
"grad_norm": 0.5619230788736052,
"kl": 0.00468292236328125,
"learning_rate": 8.699999999999999e-07,
"loss": 0.0002,
"reward": 2.502777820825577,
"reward_std": 0.46127643398940565,
"rewards/accuracy_reward_log": 1.5666666895151138,
"rewards/format_number_reward": 0.46388889104127884,
"rewards/format_reasoning_reward": 0.4722222238779068,
"step": 29
},
{
"completion_length": 151.90000228881837,
"epoch": 0.24086712163789642,
"grad_norm": 0.7524148953577687,
"kl": 0.0029834747314453126,
"learning_rate": 9e-07,
"loss": 0.0001,
"reward": 2.7583333969116213,
"reward_std": 0.31546305269002917,
"rewards/accuracy_reward_log": 1.7777777910232544,
"rewards/format_number_reward": 0.4861111119389534,
"rewards/format_reasoning_reward": 0.49444444477558136,
"step": 30
},
{
"completion_length": 150.7055564880371,
"epoch": 0.24889602569249297,
"grad_norm": 1.70651807869166,
"kl": 0.013134002685546875,
"learning_rate": 9.3e-07,
"loss": 0.0005,
"reward": 2.8583333492279053,
"reward_std": 0.1264950528740883,
"rewards/accuracy_reward_log": 1.8666666746139526,
"rewards/format_number_reward": 0.49444444477558136,
"rewards/format_reasoning_reward": 0.4972222223877907,
"step": 31
},
{
"completion_length": 169.57222366333008,
"epoch": 0.2569249297470895,
"grad_norm": 0.41158376714433204,
"kl": 0.004754638671875,
"learning_rate": 9.600000000000001e-07,
"loss": 0.0002,
"reward": 2.6888889491558077,
"reward_std": 0.29488888159394266,
"rewards/accuracy_reward_log": 1.722222238779068,
"rewards/format_number_reward": 0.4833333343267441,
"rewards/format_reasoning_reward": 0.4833333343267441,
"step": 32
},
{
"completion_length": 160.33333473205568,
"epoch": 0.2649538338016861,
"grad_norm": 0.5042425185498995,
"kl": 0.005771636962890625,
"learning_rate": 9.9e-07,
"loss": 0.0002,
"reward": 2.6055556178092956,
"reward_std": 0.33884221240878104,
"rewards/accuracy_reward_log": 1.644444465637207,
"rewards/format_number_reward": 0.4750000014901161,
"rewards/format_reasoning_reward": 0.4861111119389534,
"step": 33
},
{
"completion_length": 165.0555561065674,
"epoch": 0.27298273785628263,
"grad_norm": 0.5229992920065437,
"kl": 0.006603240966796875,
"learning_rate": 1.0200000000000002e-06,
"loss": 0.0003,
"reward": 2.663888943195343,
"reward_std": 0.4261931136250496,
"rewards/accuracy_reward_log": 1.6888889074325562,
"rewards/format_number_reward": 0.4833333343267441,
"rewards/format_reasoning_reward": 0.49166666716337204,
"step": 34
},
{
"completion_length": 164.63888969421387,
"epoch": 0.2810116419108792,
"grad_norm": 0.4700519781333619,
"kl": 0.006055450439453125,
"learning_rate": 1.05e-06,
"loss": 0.0002,
"reward": 2.6194444715976717,
"reward_std": 0.34433056265115736,
"rewards/accuracy_reward_log": 1.6444444596767425,
"rewards/format_number_reward": 0.4861111119389534,
"rewards/format_reasoning_reward": 0.4888888895511627,
"step": 35
},
{
"completion_length": 195.3277786254883,
"epoch": 0.28904054596547574,
"grad_norm": 0.5312518629436677,
"kl": 0.005257415771484375,
"learning_rate": 1.08e-06,
"loss": 0.0002,
"reward": 2.4138889372348786,
"reward_std": 0.4737018562853336,
"rewards/accuracy_reward_log": 1.5000000193715095,
"rewards/format_number_reward": 0.45277778059244156,
"rewards/format_reasoning_reward": 0.4611111134290695,
"step": 36
},
{
"completion_length": 162.9777805328369,
"epoch": 0.29706945002007223,
"grad_norm": 0.3947867705421231,
"kl": 0.00660400390625,
"learning_rate": 1.11e-06,
"loss": 0.0003,
"reward": 2.8111111521720886,
"reward_std": 0.21169509664177893,
"rewards/accuracy_reward_log": 1.8222222328186035,
"rewards/format_number_reward": 0.49166666716337204,
"rewards/format_reasoning_reward": 0.4972222223877907,
"step": 37
},
{
"completion_length": 171.81111373901368,
"epoch": 0.3050983540746688,
"grad_norm": 0.49680622935913366,
"kl": 0.0083709716796875,
"learning_rate": 1.14e-06,
"loss": 0.0003,
"reward": 2.600000059604645,
"reward_std": 0.3381901502609253,
"rewards/accuracy_reward_log": 1.6333333551883698,
"rewards/format_number_reward": 0.47777777910232544,
"rewards/format_reasoning_reward": 0.4888888895511627,
"step": 38
},
{
"completion_length": 159.78333473205566,
"epoch": 0.31312725812926534,
"grad_norm": 0.4841125000891351,
"kl": 0.01015167236328125,
"learning_rate": 1.17e-06,
"loss": 0.0004,
"reward": 2.7166666865348814,
"reward_std": 0.25784134939312936,
"rewards/accuracy_reward_log": 1.7555555701255798,
"rewards/format_number_reward": 0.4750000014901161,
"rewards/format_reasoning_reward": 0.4861111119389534,
"step": 39
},
{
"completion_length": 160.47777938842773,
"epoch": 0.3211561621838619,
"grad_norm": 0.4051878172293322,
"kl": 0.00718231201171875,
"learning_rate": 1.2000000000000002e-06,
"loss": 0.0003,
"reward": 2.6527778089046476,
"reward_std": 0.27083261907100675,
"rewards/accuracy_reward_log": 1.6777777969837189,
"rewards/format_number_reward": 0.4861111119389534,
"rewards/format_reasoning_reward": 0.4888888895511627,
"step": 40
},
{
"completion_length": 164.8000015258789,
"epoch": 0.32918506623845845,
"grad_norm": 0.474618622817464,
"kl": 0.006967926025390625,
"learning_rate": 1.2299999999999999e-06,
"loss": 0.0003,
"reward": 2.7750000238418577,
"reward_std": 0.24537386298179625,
"rewards/accuracy_reward_log": 1.800000011920929,
"rewards/format_number_reward": 0.4861111119389534,
"rewards/format_reasoning_reward": 0.4888888895511627,
"step": 41
},
{
"completion_length": 166.5111141204834,
"epoch": 0.337213970293055,
"grad_norm": 0.3629585253598538,
"kl": 0.007363128662109375,
"learning_rate": 1.26e-06,
"loss": 0.0003,
"reward": 2.63055557012558,
"reward_std": 0.22272009253501893,
"rewards/accuracy_reward_log": 1.6555555701255797,
"rewards/format_number_reward": 0.4861111119389534,
"rewards/format_reasoning_reward": 0.4888888895511627,
"step": 42
},
{
"completion_length": 166.7388900756836,
"epoch": 0.34524287434765155,
"grad_norm": 0.4008326525358567,
"kl": 0.009012603759765625,
"learning_rate": 1.29e-06,
"loss": 0.0004,
"reward": 2.708333361148834,
"reward_std": 0.29148011431097987,
"rewards/accuracy_reward_log": 1.7444444596767426,
"rewards/format_number_reward": 0.47777777910232544,
"rewards/format_reasoning_reward": 0.4861111119389534,
"step": 43
},
{
"completion_length": 185.93888893127442,
"epoch": 0.3532717784022481,
"grad_norm": 0.5017342406963871,
"kl": 0.00979156494140625,
"learning_rate": 1.32e-06,
"loss": 0.0004,
"reward": 2.5722222566604613,
"reward_std": 0.3904368232935667,
"rewards/accuracy_reward_log": 1.6555555760860443,
"rewards/format_number_reward": 0.4472222253680229,
"rewards/format_reasoning_reward": 0.4694444462656975,
"step": 44
},
{
"completion_length": 170.97777862548827,
"epoch": 0.36130068245684466,
"grad_norm": 0.36508724491351696,
"kl": 0.00971527099609375,
"learning_rate": 1.35e-06,
"loss": 0.0004,
"reward": 2.680555593967438,
"reward_std": 0.1924500897526741,
"rewards/accuracy_reward_log": 1.7111111253499984,
"rewards/format_number_reward": 0.48055555671453476,
"rewards/format_reasoning_reward": 0.4888888895511627,
"step": 45
},
{
"completion_length": 200.09444770812988,
"epoch": 0.3693295865114412,
"grad_norm": 0.5859797241606328,
"kl": 0.01236724853515625,
"learning_rate": 1.3800000000000001e-06,
"loss": 0.0005,
"reward": 2.4972222447395325,
"reward_std": 0.4179751716554165,
"rewards/accuracy_reward_log": 1.588888904452324,
"rewards/format_number_reward": 0.4527777798473835,
"rewards/format_reasoning_reward": 0.4555555582046509,
"step": 46
},
{
"completion_length": 190.55000267028808,
"epoch": 0.37735849056603776,
"grad_norm": 0.6735251651031329,
"kl": 0.013592529296875,
"learning_rate": 1.41e-06,
"loss": 0.0005,
"reward": 2.525000035762787,
"reward_std": 0.39592516496777536,
"rewards/accuracy_reward_log": 1.6000000178813933,
"rewards/format_number_reward": 0.4611111134290695,
"rewards/format_reasoning_reward": 0.46388889104127884,
"step": 47
},
{
"completion_length": 177.14444732666016,
"epoch": 0.38538739462063426,
"grad_norm": 0.4013696331659362,
"kl": 0.01161346435546875,
"learning_rate": 1.44e-06,
"loss": 0.0005,
"reward": 2.725000035762787,
"reward_std": 0.30932264029979706,
"rewards/accuracy_reward_log": 1.7666666805744171,
"rewards/format_number_reward": 0.4750000014901161,
"rewards/format_reasoning_reward": 0.4833333343267441,
"step": 48
},
{
"completion_length": 183.58333511352538,
"epoch": 0.3934162986752308,
"grad_norm": 0.413245950296645,
"kl": 0.0140655517578125,
"learning_rate": 1.4700000000000001e-06,
"loss": 0.0006,
"reward": 2.547222238779068,
"reward_std": 0.28799803368747234,
"rewards/accuracy_reward_log": 1.6000000208616256,
"rewards/format_number_reward": 0.4694444462656975,
"rewards/format_reasoning_reward": 0.47777777910232544,
"step": 49
},
{
"completion_length": 179.96666717529297,
"epoch": 0.40144520272982737,
"grad_norm": 0.41408601639844395,
"kl": 0.01335296630859375,
"learning_rate": 1.5e-06,
"loss": 0.0005,
"reward": 2.658333349227905,
"reward_std": 0.3410413548350334,
"rewards/accuracy_reward_log": 1.7222222357988357,
"rewards/format_number_reward": 0.46388889104127884,
"rewards/format_reasoning_reward": 0.4722222238779068,
"step": 50
},
{
"completion_length": 185.75556030273438,
"epoch": 0.4094741067844239,
"grad_norm": 0.4053783571755214,
"kl": 0.0148193359375,
"learning_rate": 1.53e-06,
"loss": 0.0006,
"reward": 2.6555556058883667,
"reward_std": 0.35194680131971834,
"rewards/accuracy_reward_log": 1.722222238779068,
"rewards/format_number_reward": 0.46388889104127884,
"rewards/format_reasoning_reward": 0.4694444462656975,
"step": 51
},
{
"completion_length": 209.23333358764648,
"epoch": 0.41750301083902047,
"grad_norm": 0.55024118806201,
"kl": 0.018719482421875,
"learning_rate": 1.56e-06,
"loss": 0.0007,
"reward": 2.4083333551883697,
"reward_std": 0.5498852420598268,
"rewards/accuracy_reward_log": 1.5333333522081376,
"rewards/format_number_reward": 0.430555559694767,
"rewards/format_reasoning_reward": 0.4444444477558136,
"step": 52
},
{
"completion_length": 201.15000343322754,
"epoch": 0.425531914893617,
"grad_norm": 0.4665326355575322,
"kl": 0.01983795166015625,
"learning_rate": 1.59e-06,
"loss": 0.0008,
"reward": 2.5222222208976746,
"reward_std": 0.5347743809223175,
"rewards/accuracy_reward_log": 1.6333333492279052,
"rewards/format_number_reward": 0.4416666693985462,
"rewards/format_reasoning_reward": 0.44722222462296485,
"step": 53
},
{
"completion_length": 192.3111145019531,
"epoch": 0.4335608189482136,
"grad_norm": 0.4264344321145415,
"kl": 0.02133026123046875,
"learning_rate": 1.6200000000000002e-06,
"loss": 0.0009,
"reward": 2.566666692495346,
"reward_std": 0.47150271385908127,
"rewards/accuracy_reward_log": 1.6555555760860443,
"rewards/format_number_reward": 0.45277778059244156,
"rewards/format_reasoning_reward": 0.4583333358168602,
"step": 54
},
{
"completion_length": 207.86111373901366,
"epoch": 0.44158972300281013,
"grad_norm": 0.41848310682566847,
"kl": 0.0194488525390625,
"learning_rate": 1.65e-06,
"loss": 0.0008,
"reward": 2.525000047683716,
"reward_std": 0.5073827020823956,
"rewards/accuracy_reward_log": 1.6333333492279052,
"rewards/format_number_reward": 0.4444444477558136,
"rewards/format_reasoning_reward": 0.4472222253680229,
"step": 55
},
{
"completion_length": 218.4888916015625,
"epoch": 0.4496186270574067,
"grad_norm": 0.493209444659726,
"kl": 0.0301055908203125,
"learning_rate": 1.6800000000000002e-06,
"loss": 0.0012,
"reward": 2.4250000417232513,
"reward_std": 0.6275906786322594,
"rewards/accuracy_reward_log": 1.5777777969837188,
"rewards/format_number_reward": 0.41388889253139494,
"rewards/format_reasoning_reward": 0.4333333373069763,
"step": 56
},
{
"completion_length": 207.85555877685547,
"epoch": 0.45764753111200324,
"grad_norm": 0.4781117283691274,
"kl": 0.028985595703125,
"learning_rate": 1.71e-06,
"loss": 0.0012,
"reward": 2.4833333492279053,
"reward_std": 0.4261951830238104,
"rewards/accuracy_reward_log": 1.588888907432556,
"rewards/format_number_reward": 0.4444444477558136,
"rewards/format_reasoning_reward": 0.45000000298023224,
"step": 57
},
{
"completion_length": 207.41666984558105,
"epoch": 0.46567643516659973,
"grad_norm": 9.7179275371274,
"kl": 0.066973876953125,
"learning_rate": 1.7399999999999999e-06,
"loss": 0.0027,
"reward": 2.4972222566604616,
"reward_std": 0.6165656700730324,
"rewards/accuracy_reward_log": 1.6222222447395325,
"rewards/format_number_reward": 0.4333333373069763,
"rewards/format_reasoning_reward": 0.4416666701436043,
"step": 58
},
{
"completion_length": 199.16111221313477,
"epoch": 0.4737053392211963,
"grad_norm": 0.4594537727679852,
"kl": 0.03179931640625,
"learning_rate": 1.77e-06,
"loss": 0.0013,
"reward": 2.5444444954395293,
"reward_std": 0.4083905890583992,
"rewards/accuracy_reward_log": 1.6333333522081375,
"rewards/format_number_reward": 0.45000000298023224,
"rewards/format_reasoning_reward": 0.4611111134290695,
"step": 59
},
{
"completion_length": 185.04444580078126,
"epoch": 0.48173424327579284,
"grad_norm": 0.35591417052929925,
"kl": 0.02755126953125,
"learning_rate": 1.8e-06,
"loss": 0.0011,
"reward": 2.708333361148834,
"reward_std": 0.30322221145033834,
"rewards/accuracy_reward_log": 1.7555555701255798,
"rewards/format_number_reward": 0.4750000014901161,
"rewards/format_reasoning_reward": 0.47777777910232544,
"step": 60
},
{
"completion_length": 217.3444480895996,
"epoch": 0.4897631473303894,
"grad_norm": 0.47620034091897007,
"kl": 0.043768310546875,
"learning_rate": 1.83e-06,
"loss": 0.0017,
"reward": 2.45833335518837,
"reward_std": 0.48677313327789307,
"rewards/accuracy_reward_log": 1.5777777925133705,
"rewards/format_number_reward": 0.4305555589497089,
"rewards/format_reasoning_reward": 0.45000000298023224,
"step": 61
},
{
"completion_length": 190.76111221313477,
"epoch": 0.49779205138498595,
"grad_norm": 0.381306777730851,
"kl": 0.03514404296875,
"learning_rate": 1.86e-06,
"loss": 0.0014,
"reward": 2.694444453716278,
"reward_std": 0.2467763565480709,
"rewards/accuracy_reward_log": 1.7555555671453476,
"rewards/format_number_reward": 0.4694444462656975,
"rewards/format_reasoning_reward": 0.4694444462656975,
"step": 62
},
{
"completion_length": 187.97777671813964,
"epoch": 0.5058209554395825,
"grad_norm": 0.594069782615144,
"kl": 0.039031982421875,
"learning_rate": 1.8900000000000001e-06,
"loss": 0.0016,
"reward": 2.6111111283302306,
"reward_std": 0.39384557902812956,
"rewards/accuracy_reward_log": 1.6888889014720916,
"rewards/format_number_reward": 0.4555555582046509,
"rewards/format_reasoning_reward": 0.46666666865348816,
"step": 63
},
{
"completion_length": 193.00555725097655,
"epoch": 0.513849859494179,
"grad_norm": 0.5261946833334382,
"kl": 0.04935302734375,
"learning_rate": 1.9200000000000003e-06,
"loss": 0.002,
"reward": 2.586111146211624,
"reward_std": 0.4021389245986938,
"rewards/accuracy_reward_log": 1.6777777940034866,
"rewards/format_number_reward": 0.45277778059244156,
"rewards/format_reasoning_reward": 0.4555555582046509,
"step": 64
},
{
"completion_length": 193.02222366333007,
"epoch": 0.5218787635487756,
"grad_norm": 0.4720147986556917,
"kl": 0.0471923828125,
"learning_rate": 1.95e-06,
"loss": 0.0019,
"reward": 2.6138889133930205,
"reward_std": 0.444037689268589,
"rewards/accuracy_reward_log": 1.7000000149011611,
"rewards/format_number_reward": 0.4555555582046509,
"rewards/format_reasoning_reward": 0.4583333358168602,
"step": 65
},
{
"completion_length": 210.9277801513672,
"epoch": 0.5299076676033722,
"grad_norm": 0.5124703016443702,
"kl": 0.06097412109375,
"learning_rate": 1.98e-06,
"loss": 0.0024,
"reward": 2.4833333492279053,
"reward_std": 0.5216697975993156,
"rewards/accuracy_reward_log": 1.6111111342906952,
"rewards/format_number_reward": 0.4277777820825577,
"rewards/format_reasoning_reward": 0.4444444477558136,
"step": 66
},
{
"completion_length": 195.95000152587892,
"epoch": 0.5379365716579687,
"grad_norm": 0.3914341348179692,
"kl": 0.05845947265625,
"learning_rate": 2.0100000000000002e-06,
"loss": 0.0023,
"reward": 2.5666666865348815,
"reward_std": 0.4269164353609085,
"rewards/accuracy_reward_log": 1.666666680574417,
"rewards/format_number_reward": 0.4416666701436043,
"rewards/format_reasoning_reward": 0.4583333358168602,
"step": 67
},
{
"completion_length": 189.8333354949951,
"epoch": 0.5459654757125653,
"grad_norm": 1.8409712600042354,
"kl": 0.104864501953125,
"learning_rate": 2.0400000000000004e-06,
"loss": 0.0042,
"reward": 2.6055555939674377,
"reward_std": 0.3877051591873169,
"rewards/accuracy_reward_log": 1.7000000119209289,
"rewards/format_number_reward": 0.45000000298023224,
"rewards/format_reasoning_reward": 0.4555555582046509,
"step": 68
},
{
"completion_length": 216.18333587646484,
"epoch": 0.5539943797671618,
"grad_norm": 1.234592815069376,
"kl": 0.11241455078125,
"learning_rate": 2.07e-06,
"loss": 0.0045,
"reward": 2.3138889193534853,
"reward_std": 0.5511893726885319,
"rewards/accuracy_reward_log": 1.46666669100523,
"rewards/format_number_reward": 0.4138888940215111,
"rewards/format_reasoning_reward": 0.4333333373069763,
"step": 69
},
{
"completion_length": 216.01666946411132,
"epoch": 0.5620232838217584,
"grad_norm": 0.47189027920553445,
"kl": 0.086163330078125,
"learning_rate": 2.1e-06,
"loss": 0.0034,
"reward": 2.4277777910232543,
"reward_std": 0.5567510481923819,
"rewards/accuracy_reward_log": 1.588888907432556,
"rewards/format_number_reward": 0.413888893276453,
"rewards/format_reasoning_reward": 0.4250000037252903,
"step": 70
},
{
"completion_length": 180.42222366333007,
"epoch": 0.5700521878763549,
"grad_norm": 0.5661866241275311,
"kl": 0.055902099609375,
"learning_rate": 2.13e-06,
"loss": 0.0022,
"reward": 2.508333349227905,
"reward_std": 0.416572679579258,
"rewards/accuracy_reward_log": 1.6000000208616256,
"rewards/format_number_reward": 0.4472222253680229,
"rewards/format_reasoning_reward": 0.4611111134290695,
"step": 71
},
{
"completion_length": 212.58333511352538,
"epoch": 0.5780810919309515,
"grad_norm": 0.5852592346007427,
"kl": 0.108624267578125,
"learning_rate": 2.16e-06,
"loss": 0.0043,
"reward": 2.2777777791023253,
"reward_std": 0.53003438860178,
"rewards/accuracy_reward_log": 1.4888889133930205,
"rewards/format_number_reward": 0.38611111640930174,
"rewards/format_reasoning_reward": 0.40277778208255766,
"step": 72
},
{
"completion_length": 203.32777824401856,
"epoch": 0.586109995985548,
"grad_norm": 1.0262474142127427,
"kl": 0.104498291015625,
"learning_rate": 2.19e-06,
"loss": 0.0042,
"reward": 2.4166666686534883,
"reward_std": 0.40687683820724485,
"rewards/accuracy_reward_log": 1.5777777969837188,
"rewards/format_number_reward": 0.4166666708886623,
"rewards/format_reasoning_reward": 0.42222222611308097,
"step": 73
},
{
"completion_length": 205.4111114501953,
"epoch": 0.5941389000401445,
"grad_norm": 0.6749287604370076,
"kl": 0.106732177734375,
"learning_rate": 2.22e-06,
"loss": 0.0043,
"reward": 2.2611111223697664,
"reward_std": 0.5402376987040043,
"rewards/accuracy_reward_log": 1.4333333566784858,
"rewards/format_number_reward": 0.4027777828276157,
"rewards/format_reasoning_reward": 0.42500000447034836,
"step": 74
},
{
"completion_length": 204.6277805328369,
"epoch": 0.602167804094741,
"grad_norm": 1.3836757790747922,
"kl": 0.118878173828125,
"learning_rate": 2.25e-06,
"loss": 0.0048,
"reward": 2.211111146211624,
"reward_std": 0.7402307014912367,
"rewards/accuracy_reward_log": 1.4000000268220902,
"rewards/format_number_reward": 0.38055556192994117,
"rewards/format_reasoning_reward": 0.430555559694767,
"step": 75
},
{
"completion_length": 192.80555725097656,
"epoch": 0.6101967081493376,
"grad_norm": 0.5732652733178961,
"kl": 0.081439208984375,
"learning_rate": 2.28e-06,
"loss": 0.0033,
"reward": 2.444444465637207,
"reward_std": 0.5093889623880387,
"rewards/accuracy_reward_log": 1.533333358168602,
"rewards/format_number_reward": 0.45277778059244156,
"rewards/format_reasoning_reward": 0.4583333358168602,
"step": 76
},
{
"completion_length": 159.32222518920898,
"epoch": 0.6182256122039341,
"grad_norm": 0.4884865339246385,
"kl": 0.057391357421875,
"learning_rate": 2.31e-06,
"loss": 0.0023,
"reward": 2.5805555820465087,
"reward_std": 0.16979632191359997,
"rewards/accuracy_reward_log": 1.611111131310463,
"rewards/format_number_reward": 0.47777777910232544,
"rewards/format_reasoning_reward": 0.49166666716337204,
"step": 77
},
{
"completion_length": 190.20000076293945,
"epoch": 0.6262545162585307,
"grad_norm": 0.5209974936705293,
"kl": 0.067266845703125,
"learning_rate": 2.34e-06,
"loss": 0.0027,
"reward": 2.4777778029441833,
"reward_std": 0.4935527116060257,
"rewards/accuracy_reward_log": 1.5555555790662765,
"rewards/format_number_reward": 0.4583333358168602,
"rewards/format_reasoning_reward": 0.46388889104127884,
"step": 78
},
{
"completion_length": 169.23889122009277,
"epoch": 0.6342834203131272,
"grad_norm": 0.4459209602371647,
"kl": 0.05482177734375,
"learning_rate": 2.37e-06,
"loss": 0.0022,
"reward": 2.6750000178813935,
"reward_std": 0.23374302312731743,
"rewards/accuracy_reward_log": 1.7000000149011611,
"rewards/format_number_reward": 0.4833333343267441,
"rewards/format_reasoning_reward": 0.49166666716337204,
"step": 79
},
{
"completion_length": 182.30000267028808,
"epoch": 0.6423123243677238,
"grad_norm": 0.45736704423541613,
"kl": 0.07120361328125,
"learning_rate": 2.4000000000000003e-06,
"loss": 0.0028,
"reward": 2.5750000059604643,
"reward_std": 0.31974178850650786,
"rewards/accuracy_reward_log": 1.6444444626569747,
"rewards/format_number_reward": 0.46388889104127884,
"rewards/format_reasoning_reward": 0.46666666865348816,
"step": 80
},
{
"completion_length": 163.07778053283693,
"epoch": 0.6503412284223203,
"grad_norm": 0.4682785728868757,
"kl": 0.06102294921875,
"learning_rate": 2.43e-06,
"loss": 0.0024,
"reward": 2.686111146211624,
"reward_std": 0.3670576632022858,
"rewards/accuracy_reward_log": 1.7444444596767426,
"rewards/format_number_reward": 0.4694444462656975,
"rewards/format_reasoning_reward": 0.4722222238779068,
"step": 81
},
{
"completion_length": 165.8722225189209,
"epoch": 0.6583701324769169,
"grad_norm": 0.509179920003658,
"kl": 0.061614990234375,
"learning_rate": 2.4599999999999997e-06,
"loss": 0.0025,
"reward": 2.6250000596046448,
"reward_std": 0.402138914167881,
"rewards/accuracy_reward_log": 1.6777777969837189,
"rewards/format_number_reward": 0.4694444462656975,
"rewards/format_reasoning_reward": 0.47777777910232544,
"step": 82
},
{
"completion_length": 159.80555725097656,
"epoch": 0.6663990365315134,
"grad_norm": 0.5994985713489502,
"kl": 0.054705810546875,
"learning_rate": 2.49e-06,
"loss": 0.0022,
"reward": 2.6194444835186004,
"reward_std": 0.4117614269256592,
"rewards/accuracy_reward_log": 1.6555555760860443,
"rewards/format_number_reward": 0.48055555671453476,
"rewards/format_reasoning_reward": 0.4833333343267441,
"step": 83
},
{
"completion_length": 175.34444770812988,
"epoch": 0.67442794058611,
"grad_norm": 0.518979425053283,
"kl": 0.08248291015625,
"learning_rate": 2.52e-06,
"loss": 0.0033,
"reward": 2.5138889074325563,
"reward_std": 0.4138893112540245,
"rewards/accuracy_reward_log": 1.6000000208616256,
"rewards/format_number_reward": 0.45277778059244156,
"rewards/format_reasoning_reward": 0.4611111134290695,
"step": 84
},
{
"completion_length": 165.01666755676268,
"epoch": 0.6824568446407066,
"grad_norm": 0.49612998471580416,
"kl": 0.065625,
"learning_rate": 2.55e-06,
"loss": 0.0026,
"reward": 2.469444477558136,
"reward_std": 0.5141268767416477,
"rewards/accuracy_reward_log": 1.588888907432556,
"rewards/format_number_reward": 0.43333333656191825,
"rewards/format_reasoning_reward": 0.44722222462296485,
"step": 85
},
{
"completion_length": 165.6388900756836,
"epoch": 0.6904857486953031,
"grad_norm": 0.552553452651074,
"kl": 0.06807861328125,
"learning_rate": 2.58e-06,
"loss": 0.0027,
"reward": 2.4944444805383683,
"reward_std": 0.37747681848704817,
"rewards/accuracy_reward_log": 1.577777798473835,
"rewards/format_number_reward": 0.4527777798473835,
"rewards/format_reasoning_reward": 0.46388889104127884,
"step": 86
},
{
"completion_length": 155.51666641235352,
"epoch": 0.6985146527498997,
"grad_norm": 0.5440640607374735,
"kl": 0.060748291015625,
"learning_rate": 2.61e-06,
"loss": 0.0024,
"reward": 2.4611111283302307,
"reward_std": 0.5409881249070168,
"rewards/accuracy_reward_log": 1.577777799963951,
"rewards/format_number_reward": 0.43888889253139496,
"rewards/format_reasoning_reward": 0.4444444477558136,
"step": 87
},
{
"completion_length": 150.6777774810791,
"epoch": 0.7065435568044962,
"grad_norm": 0.4538101109952547,
"kl": 0.05191650390625,
"learning_rate": 2.64e-06,
"loss": 0.0021,
"reward": 2.6416666924953462,
"reward_std": 0.2997001264244318,
"rewards/accuracy_reward_log": 1.7000000149011611,
"rewards/format_number_reward": 0.4694444462656975,
"rewards/format_reasoning_reward": 0.4722222238779068,
"step": 88
},
{
"completion_length": 165.80000228881835,
"epoch": 0.7145724608590928,
"grad_norm": 0.5131296250407812,
"kl": 0.06552734375,
"learning_rate": 2.6700000000000003e-06,
"loss": 0.0026,
"reward": 2.600000035762787,
"reward_std": 0.48052144795656204,
"rewards/accuracy_reward_log": 1.6777777969837189,
"rewards/format_number_reward": 0.4583333358168602,
"rewards/format_reasoning_reward": 0.46388889104127884,
"step": 89
},
{
"completion_length": 151.10555686950684,
"epoch": 0.7226013649136893,
"grad_norm": 0.5165638439339917,
"kl": 0.0578857421875,
"learning_rate": 2.7e-06,
"loss": 0.0023,
"reward": 2.6555556058883667,
"reward_std": 0.362246410548687,
"rewards/accuracy_reward_log": 1.7111111283302307,
"rewards/format_number_reward": 0.4722222238779068,
"rewards/format_reasoning_reward": 0.4722222238779068,
"step": 90
},
{
"completion_length": 159.4166687011719,
"epoch": 0.7306302689682859,
"grad_norm": 0.49295723950300413,
"kl": 0.058404541015625,
"learning_rate": 2.73e-06,
"loss": 0.0023,
"reward": 2.5833333671092986,
"reward_std": 0.44544019252061845,
"rewards/accuracy_reward_log": 1.6444444626569747,
"rewards/format_number_reward": 0.4694444462656975,
"rewards/format_reasoning_reward": 0.4694444462656975,
"step": 91
},
{
"completion_length": 159.53333625793456,
"epoch": 0.7386591730228824,
"grad_norm": 1.0273725644592662,
"kl": 0.06968994140625,
"learning_rate": 2.7600000000000003e-06,
"loss": 0.0028,
"reward": 2.5111111402511597,
"reward_std": 0.44476102106273174,
"rewards/accuracy_reward_log": 1.577777799963951,
"rewards/format_number_reward": 0.4611111134290695,
"rewards/format_reasoning_reward": 0.4722222238779068,
"step": 92
},
{
"completion_length": 154.76666793823242,
"epoch": 0.746688077077479,
"grad_norm": 0.3153647070205455,
"kl": 0.053741455078125,
"learning_rate": 2.7900000000000004e-06,
"loss": 0.0021,
"reward": 2.688888907432556,
"reward_std": 0.2062800731509924,
"rewards/accuracy_reward_log": 1.733333346247673,
"rewards/format_number_reward": 0.4750000014901161,
"rewards/format_reasoning_reward": 0.48055555671453476,
"step": 93
},
{
"completion_length": 166.42777900695802,
"epoch": 0.7547169811320755,
"grad_norm": 0.5748531305032764,
"kl": 0.066802978515625,
"learning_rate": 2.82e-06,
"loss": 0.0027,
"reward": 2.600000041723251,
"reward_std": 0.3458063915371895,
"rewards/accuracy_reward_log": 1.655555573105812,
"rewards/format_number_reward": 0.4722222238779068,
"rewards/format_reasoning_reward": 0.4722222238779068,
"step": 94
},
{
"completion_length": 158.03888969421388,
"epoch": 0.762745885186672,
"grad_norm": 0.5545909018402391,
"kl": 0.066021728515625,
"learning_rate": 2.85e-06,
"loss": 0.0026,
"reward": 2.7111111342906953,
"reward_std": 0.3430013954639435,
"rewards/accuracy_reward_log": 1.7666666775941848,
"rewards/format_number_reward": 0.4722222238779068,
"rewards/format_reasoning_reward": 0.4722222238779068,
"step": 95
},
{
"completion_length": 165.80000228881835,
"epoch": 0.7707747892412685,
"grad_norm": 0.42499585100246984,
"kl": 0.05565185546875,
"learning_rate": 2.88e-06,
"loss": 0.0022,
"reward": 2.602777808904648,
"reward_std": 0.3154630549252033,
"rewards/accuracy_reward_log": 1.6444444626569747,
"rewards/format_number_reward": 0.47777777910232544,
"rewards/format_reasoning_reward": 0.48055555671453476,
"step": 96
},
{
"completion_length": 174.88333435058593,
"epoch": 0.7788036932958651,
"grad_norm": 0.5630736371263277,
"kl": 0.07933349609375,
"learning_rate": 2.91e-06,
"loss": 0.0032,
"reward": 2.394444453716278,
"reward_std": 0.4866618663072586,
"rewards/accuracy_reward_log": 1.4777778044342995,
"rewards/format_number_reward": 0.4555555582046509,
"rewards/format_reasoning_reward": 0.4611111134290695,
"step": 97
},
{
"completion_length": 194.8888916015625,
"epoch": 0.7868325973504616,
"grad_norm": 0.5354591296941339,
"kl": 0.08121337890625,
"learning_rate": 2.9400000000000002e-06,
"loss": 0.0032,
"reward": 2.4166667103767394,
"reward_std": 0.467490179464221,
"rewards/accuracy_reward_log": 1.5444444686174392,
"rewards/format_number_reward": 0.4333333373069763,
"rewards/format_reasoning_reward": 0.43888889253139496,
"step": 98
},
{
"completion_length": 198.76111335754393,
"epoch": 0.7948615014050582,
"grad_norm": 0.7579425034647131,
"kl": 0.092218017578125,
"learning_rate": 2.97e-06,
"loss": 0.0037,
"reward": 2.472222238779068,
"reward_std": 0.44544018656015394,
"rewards/accuracy_reward_log": 1.577777799963951,
"rewards/format_number_reward": 0.4472222253680229,
"rewards/format_reasoning_reward": 0.4472222253680229,
"step": 99
},
{
"completion_length": 181.00000228881837,
"epoch": 0.8028904054596547,
"grad_norm": 3.752908048104343,
"kl": 0.29696044921875,
"learning_rate": 3e-06,
"loss": 0.0119,
"reward": 2.644444489479065,
"reward_std": 0.3877051673829556,
"rewards/accuracy_reward_log": 1.722222238779068,
"rewards/format_number_reward": 0.4611111134290695,
"rewards/format_reasoning_reward": 0.4611111134290695,
"step": 100
}
],
"logging_steps": 1.0,
"max_steps": 1000,
"num_input_tokens_seen": 0,
"num_train_epochs": 9,
"save_steps": 50,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}