Publish llama3b_code_shuffled_lr1e-5_bs256_seed42/support

runs/llama3b_code_shuffled_lr1e-5_bs256_seed42/metrics/train_metrics.jsonl

@@ -0,0 +1,52 @@
+{"global_step": 14, "epoch": 0, "batch_id": 895, "num_supervised_tokens": 40412, "learning_rate": 1e-05, "batch_loss": 0.4612032019945366, "token_entropy": 0.4433785201794533, "gradient_norm_pre_clip": 1.087500598402832, "gradient_norm_post_clip": 1.087500598402832, "parameter_l2_norm": 1111.3914642212828, "parameter_l2_from_init": 2.4949073994249464, "relative_parameter_l2_from_init": 0.002244829022380549, "update_norm": 0.117726137142581, "relative_update_norm": 0.00010592679621223113}
+{"global_step": 15, "epoch": 0, "batch_id": 959, "num_supervised_tokens": 39608, "learning_rate": 1e-05, "batch_loss": 0.28639072508268454, "token_entropy": 0.27585561866351893, "gradient_norm_pre_clip": 0.9909000469751281, "gradient_norm_post_clip": 0.9909000469751281, "parameter_l2_norm": 1111.391055420089, "parameter_l2_from_init": 2.572233787656467, "relative_parameter_l2_from_init": 0.0023144045587463447, "update_norm": 0.10973823748771615, "relative_update_norm": 9.87395363248058e-05}
+{"global_step": 16, "epoch": 0, "batch_id": 1023, "num_supervised_tokens": 41941, "learning_rate": 1e-05, "batch_loss": 0.321947337630045, "token_entropy": 0.3117776223226768, "gradient_norm_pre_clip": 0.9980872012713379, "gradient_norm_post_clip": 0.9980872012713379, "parameter_l2_norm": 1111.3906310872976, "parameter_l2_from_init": 2.6437197314855596, "relative_parameter_l2_from_init": 0.0023787250707768917, "update_norm": 0.1033714283681166, "relative_update_norm": 9.301088697048497e-05}
+{"global_step": 17, "epoch": 0, "batch_id": 1087, "num_supervised_tokens": 39016, "learning_rate": 1e-05, "batch_loss": 0.5190669899151953, "token_entropy": 0.5194097300496824, "gradient_norm_pre_clip": 1.0362881007815545, "gradient_norm_post_clip": 1.0362881007815545, "parameter_l2_norm": 1111.3902140189455, "parameter_l2_from_init": 2.7121272329579567, "relative_parameter_l2_from_init": 0.002440275709766207, "update_norm": 0.10373121863028563, "relative_update_norm": 9.333465179181194e-05}
+{"global_step": 18, "epoch": 0, "batch_id": 1151, "num_supervised_tokens": 40414, "learning_rate": 1e-05, "batch_loss": 0.3134142085619482, "token_entropy": 0.31879683702368533, "gradient_norm_pre_clip": 0.9251000539159745, "gradient_norm_post_clip": 0.9251000539159745, "parameter_l2_norm": 1111.3897851139204, "parameter_l2_from_init": 2.775610971006578, "relative_parameter_l2_from_init": 0.002497396121390942, "update_norm": 0.09833420034483423, "relative_update_norm": 8.847858929597298e-05}
+{"global_step": 19, "epoch": 0, "batch_id": 1215, "num_supervised_tokens": 39653, "learning_rate": 1e-05, "batch_loss": 0.32447727807441124, "token_entropy": 0.34026951549711754, "gradient_norm_pre_clip": 0.8987405152371549, "gradient_norm_post_clip": 0.8987405152371549, "parameter_l2_norm": 1111.3893308229358, "parameter_l2_from_init": 2.8345680847391312, "relative_parameter_l2_from_init": 0.0025504436373080183, "update_norm": 0.09352760461339875, "relative_update_norm": 8.415377223762406e-05}
+{"global_step": 20, "epoch": 0, "batch_id": 1279, "num_supervised_tokens": 39429, "learning_rate": 1e-05, "batch_loss": 0.3013725602910659, "token_entropy": 0.3104709460384668, "gradient_norm_pre_clip": 0.8534413996576412, "gradient_norm_post_clip": 0.8534413996576412, "parameter_l2_norm": 1111.3888696018116, "parameter_l2_from_init": 2.889408496956239, "relative_parameter_l2_from_init": 0.002599787091486979, "update_norm": 0.08866025800453341, "relative_update_norm": 7.977429001632759e-05}
+{"global_step": 21, "epoch": 0, "batch_id": 1343, "num_supervised_tokens": 41149, "learning_rate": 1e-05, "batch_loss": 0.36207159741494055, "token_entropy": 0.370782965550449, "gradient_norm_pre_clip": 0.8723788778594365, "gradient_norm_post_clip": 0.8723788778594365, "parameter_l2_norm": 1111.3883834803605, "parameter_l2_from_init": 2.9406745026363605, "relative_parameter_l2_from_init": 0.0026459144216791895, "update_norm": 0.08538157837907408, "relative_update_norm": 7.682424942367851e-05}
+{"global_step": 22, "epoch": 0, "batch_id": 1407, "num_supervised_tokens": 39265, "learning_rate": 1e-05, "batch_loss": 0.3019186931141513, "token_entropy": 0.30840078528999415, "gradient_norm_pre_clip": 0.8916829990896697, "gradient_norm_post_clip": 0.8916829990896697, "parameter_l2_norm": 1111.3878829925018, "parameter_l2_from_init": 2.9885175863940425, "relative_parameter_l2_from_init": 0.0026889619283578673, "update_norm": 0.08230019266339776, "relative_update_norm": 7.405172750471044e-05}
+{"global_step": 23, "epoch": 0, "batch_id": 1471, "num_supervised_tokens": 40470, "learning_rate": 1e-05, "batch_loss": 0.30771712068673607, "token_entropy": 0.3016034212765176, "gradient_norm_pre_clip": 0.8723659238150858, "gradient_norm_post_clip": 0.8723659238150858, "parameter_l2_norm": 1111.387379140691, "parameter_l2_from_init": 3.033382159179985, "relative_parameter_l2_from_init": 0.0027293294767044723, "update_norm": 0.0804849195452285, "relative_update_norm": 7.241842138558231e-05}
+{"global_step": 24, "epoch": 0, "batch_id": 1535, "num_supervised_tokens": 38912, "learning_rate": 1e-05, "batch_loss": 0.35337233472608587, "token_entropy": 0.3371439698446346, "gradient_norm_pre_clip": 0.9313698676142831, "gradient_norm_post_clip": 0.9313698676142831, "parameter_l2_norm": 1111.3868502872272, "parameter_l2_from_init": 3.075547031139828, "relative_parameter_l2_from_init": 0.002767267930180634, "update_norm": 0.07991548370285796, "relative_update_norm": 7.190609073897589e-05}
+{"global_step": 25, "epoch": 0, "batch_id": 1599, "num_supervised_tokens": 41429, "learning_rate": 1e-05, "batch_loss": 0.5806565873734276, "token_entropy": 0.5597535758824291, "gradient_norm_pre_clip": 0.9250791226471478, "gradient_norm_post_clip": 0.9250791226471478, "parameter_l2_norm": 1111.386250168774, "parameter_l2_from_init": 3.1189148388772514, "relative_parameter_l2_from_init": 0.0028062887425235796, "update_norm": 0.08617807625017539, "relative_update_norm": 7.754106750653832e-05}
+{"global_step": 26, "epoch": 0, "batch_id": 1663, "num_supervised_tokens": 39262, "learning_rate": 1e-05, "batch_loss": 0.5155096809260157, "token_entropy": 0.5092184722505533, "gradient_norm_pre_clip": 0.8358542771410669, "gradient_norm_post_clip": 0.8358542771410669, "parameter_l2_norm": 1111.3856703370757, "parameter_l2_from_init": 3.1623988976405326, "relative_parameter_l2_from_init": 0.002845414153408655, "update_norm": 0.08792299428079091, "relative_update_norm": 7.911114622714585e-05}
+{"global_step": 27, "epoch": 0, "batch_id": 1727, "num_supervised_tokens": 38545, "learning_rate": 1e-05, "batch_loss": 0.32572473063934504, "token_entropy": 0.3182208392635241, "gradient_norm_pre_clip": 0.8588616663921642, "gradient_norm_post_clip": 0.8588616663921642, "parameter_l2_norm": 1111.3851255342402, "parameter_l2_from_init": 3.2040350530814323, "relative_parameter_l2_from_init": 0.0028828768865488195, "update_norm": 0.08566161229038136, "relative_update_norm": 7.707644300998183e-05}
+{"global_step": 28, "epoch": 0, "batch_id": 1791, "num_supervised_tokens": 43013, "learning_rate": 1e-05, "batch_loss": 0.4262012253165892, "token_entropy": 0.4179435634686204, "gradient_norm_pre_clip": 0.834003972826843, "gradient_norm_post_clip": 0.834003972826843, "parameter_l2_norm": 1111.384542727795, "parameter_l2_from_init": 3.244742777056313, "relative_parameter_l2_from_init": 0.0029195042500473334, "update_norm": 0.0853216693004558, "relative_update_norm": 7.677061000960236e-05}
+{"global_step": 29, "epoch": 0, "batch_id": 1855, "num_supervised_tokens": 38324, "learning_rate": 1e-05, "batch_loss": 0.3160643122628348, "token_entropy": 0.3218754889409884, "gradient_norm_pre_clip": 0.8569055056712461, "gradient_norm_post_clip": 0.8569055056712461, "parameter_l2_norm": 1111.3839774759756, "parameter_l2_from_init": 3.283325691878964, "relative_parameter_l2_from_init": 0.0029542197857749864, "update_norm": 0.08255200698927383, "relative_update_norm": 7.427856498053421e-05}
+{"global_step": 30, "epoch": 0, "batch_id": 1919, "num_supervised_tokens": 38828, "learning_rate": 1e-05, "batch_loss": 0.8852764786756163, "token_entropy": 0.8943308849861316, "gradient_norm_pre_clip": 0.8139658418581828, "gradient_norm_post_clip": 0.8139658418581828, "parameter_l2_norm": 1111.383350827056, "parameter_l2_from_init": 3.326264902334066, "relative_parameter_l2_from_init": 0.002992854961513347, "update_norm": 0.09451026113430223, "relative_update_norm": 8.503839927417548e-05}
+{"global_step": 31, "epoch": 0, "batch_id": 1983, "num_supervised_tokens": 40369, "learning_rate": 1e-05, "batch_loss": 0.3613994264167132, "token_entropy": 0.3625531777388488, "gradient_norm_pre_clip": 0.8542292790190319, "gradient_norm_post_clip": 0.8542292790190319, "parameter_l2_norm": 1111.382751684944, "parameter_l2_from_init": 3.3672295948391495, "relative_parameter_l2_from_init": 0.0030297135361640544, "update_norm": 0.09078379330572157, "relative_update_norm": 8.168544380240397e-05}
+{"global_step": 32, "epoch": 0, "batch_id": 2047, "num_supervised_tokens": 37272, "learning_rate": 1e-05, "batch_loss": 0.2946084138688322, "token_entropy": 0.29734649484724057, "gradient_norm_pre_clip": 0.8519896088520588, "gradient_norm_post_clip": 0.8519896088520588, "parameter_l2_norm": 1111.382157979415, "parameter_l2_from_init": 3.406031993536535, "relative_parameter_l2_from_init": 0.0030646265556829146, "update_norm": 0.08759297117786509, "relative_update_norm": 7.881444789172824e-05}
+{"global_step": 33, "epoch": 0, "batch_id": 2111, "num_supervised_tokens": 40831, "learning_rate": 1e-05, "batch_loss": 0.3546911130001533, "token_entropy": 0.35302886961489227, "gradient_norm_pre_clip": 0.8150291300310106, "gradient_norm_post_clip": 0.8150291300310106, "parameter_l2_norm": 1111.3815773595638, "parameter_l2_from_init": 3.4429346442677806, "relative_parameter_l2_from_init": 0.0030978302494886916, "update_norm": 0.084589456921709, "relative_update_norm": 7.611198407902156e-05}
+{"global_step": 34, "epoch": 0, "batch_id": 2175, "num_supervised_tokens": 39754, "learning_rate": 1e-05, "batch_loss": 0.329923191575837, "token_entropy": 0.32514480484883157, "gradient_norm_pre_clip": 0.8059542892390243, "gradient_norm_post_clip": 0.8059542892390243, "parameter_l2_norm": 1111.3810083220221, "parameter_l2_from_init": 3.4781055037162747, "relative_parameter_l2_from_init": 0.0031294757390368188, "update_norm": 0.08191302660666067, "relative_update_norm": 7.37038207359095e-05}
+{"global_step": 35, "epoch": 0, "batch_id": 2239, "num_supervised_tokens": 39094, "learning_rate": 1e-05, "batch_loss": 0.3233221953750669, "token_entropy": 0.3136613100039878, "gradient_norm_pre_clip": 0.8585930528788227, "gradient_norm_post_clip": 0.8585930528788227, "parameter_l2_norm": 1111.3804500739243, "parameter_l2_from_init": 3.5114077930053202, "relative_parameter_l2_from_init": 0.003159439955554431, "update_norm": 0.07984392828902125, "relative_update_norm": 7.184212056610351e-05}
+{"global_step": 36, "epoch": 0, "batch_id": 2303, "num_supervised_tokens": 40998, "learning_rate": 1e-05, "batch_loss": 0.3198851030650224, "token_entropy": 0.3074530119476996, "gradient_norm_pre_clip": 0.8457476115192365, "gradient_norm_post_clip": 0.8457476115192365, "parameter_l2_norm": 1111.3798955599943, "parameter_l2_from_init": 3.542969008790096, "relative_parameter_l2_from_init": 0.003187837616001312, "update_norm": 0.07792331763122023, "relative_update_norm": 7.011402486451923e-05}
+{"global_step": 37, "epoch": 0, "batch_id": 2367, "num_supervised_tokens": 40817, "learning_rate": 1e-05, "batch_loss": 0.8428550172598265, "token_entropy": 0.8373980356150028, "gradient_norm_pre_clip": 0.7946845919712369, "gradient_norm_post_clip": 0.7946845919712369, "parameter_l2_norm": 1111.3793620228935, "parameter_l2_from_init": 3.5772183229226395, "relative_parameter_l2_from_init": 0.003218653931820922, "update_norm": 0.0833283970068241, "relative_update_norm": 7.497745581234538e-05}
+{"global_step": 38, "epoch": 0, "batch_id": 2431, "num_supervised_tokens": 39943, "learning_rate": 1e-05, "batch_loss": 0.4020401740610507, "token_entropy": 0.4035425257307352, "gradient_norm_pre_clip": 0.8208162721289158, "gradient_norm_post_clip": 0.8208162721289158, "parameter_l2_norm": 1111.3788452579615, "parameter_l2_from_init": 3.6109889425244583, "relative_parameter_l2_from_init": 0.0032490395353120213, "update_norm": 0.08287951743896496, "relative_update_norm": 7.457359638668292e-05}
+{"global_step": 39, "epoch": 0, "batch_id": 2495, "num_supervised_tokens": 35981, "learning_rate": 1e-05, "batch_loss": 0.32615276219188216, "token_entropy": 0.3314569177524459, "gradient_norm_pre_clip": 0.9084558252548246, "gradient_norm_post_clip": 0.9084558252548246, "parameter_l2_norm": 1111.3783308251066, "parameter_l2_from_init": 3.6431293301834256, "relative_parameter_l2_from_init": 0.003277958314030639, "update_norm": 0.08103313846816762, "relative_update_norm": 7.291228938034739e-05}
+{"global_step": 40, "epoch": 0, "batch_id": 2559, "num_supervised_tokens": 38408, "learning_rate": 1e-05, "batch_loss": 0.32482856501866086, "token_entropy": 0.3292709243617885, "gradient_norm_pre_clip": 0.803978039857223, "gradient_norm_post_clip": 0.803978039857223, "parameter_l2_norm": 1111.3778275593386, "parameter_l2_from_init": 3.6740527676734165, "relative_parameter_l2_from_init": 0.0033057821242310904, "update_norm": 0.07943392194146198, "relative_update_norm": 7.14733729355608e-05}
+{"global_step": 41, "epoch": 0, "batch_id": 2623, "num_supervised_tokens": 40832, "learning_rate": 1e-05, "batch_loss": 0.31328207808372466, "token_entropy": 0.3147029693363976, "gradient_norm_pre_clip": 0.7622509020376526, "gradient_norm_post_clip": 0.7622509020376526, "parameter_l2_norm": 1111.3773326443845, "parameter_l2_from_init": 3.7034537503711076, "relative_parameter_l2_from_init": 0.0033322360837093053, "update_norm": 0.07718247560091389, "relative_update_norm": 6.944758844169312e-05}
+{"global_step": 42, "epoch": 0, "batch_id": 2687, "num_supervised_tokens": 43825, "learning_rate": 1e-05, "batch_loss": 0.6791087345636986, "token_entropy": 0.6755356403139748, "gradient_norm_pre_clip": 0.7819675845906413, "gradient_norm_post_clip": 0.7819675845906413, "parameter_l2_norm": 1111.376877007955, "parameter_l2_from_init": 3.7353790203008392, "relative_parameter_l2_from_init": 0.003360961307139286, "update_norm": 0.08022182423384303, "relative_update_norm": 7.218237655782073e-05}
+{"global_step": 43, "epoch": 0, "batch_id": 2751, "num_supervised_tokens": 40744, "learning_rate": 1e-05, "batch_loss": 0.8228945560515989, "token_entropy": 0.8112946806547569, "gradient_norm_pre_clip": 0.843029586852585, "gradient_norm_post_clip": 0.843029586852585, "parameter_l2_norm": 1111.3764757595577, "parameter_l2_from_init": 3.770124054159218, "relative_parameter_l2_from_init": 0.003392223653952984, "update_norm": 0.08610157516327212, "relative_update_norm": 7.74729149313935e-05}
+{"global_step": 44, "epoch": 0, "batch_id": 2815, "num_supervised_tokens": 38458, "learning_rate": 1e-05, "batch_loss": 0.2788850388586191, "token_entropy": 0.2809187384314827, "gradient_norm_pre_clip": 0.7624924627356249, "gradient_norm_post_clip": 0.7624924627356249, "parameter_l2_norm": 1111.3760749555122, "parameter_l2_from_init": 3.803039102572164, "relative_parameter_l2_from_init": 0.003421839444890745, "update_norm": 0.08276508422521224, "relative_update_norm": 7.447081693613504e-05}
+{"global_step": 45, "epoch": 0, "batch_id": 2879, "num_supervised_tokens": 38610, "learning_rate": 1e-05, "batch_loss": 0.33248843019666585, "token_entropy": 0.3266126643805515, "gradient_norm_pre_clip": 0.8601164421104649, "gradient_norm_post_clip": 0.8601164421104649, "parameter_l2_norm": 1111.3756226222622, "parameter_l2_from_init": 3.834422100027337, "relative_parameter_l2_from_init": 0.0034500767508175724, "update_norm": 0.08087030697210641, "relative_update_norm": 7.276595358578678e-05}
+{"global_step": 46, "epoch": 0, "batch_id": 2943, "num_supervised_tokens": 39547, "learning_rate": 1e-05, "batch_loss": 0.7547508802270675, "token_entropy": 0.7432767411931642, "gradient_norm_pre_clip": 0.8241506275731846, "gradient_norm_post_clip": 0.8241506275731846, "parameter_l2_norm": 1111.3752613301153, "parameter_l2_from_init": 3.8698417441001847, "relative_parameter_l2_from_init": 0.003481946087930222, "update_norm": 0.087142140138561, "relative_update_norm": 7.840928547776708e-05}
+{"global_step": 47, "epoch": 0, "batch_id": 3007, "num_supervised_tokens": 41809, "learning_rate": 1e-05, "batch_loss": 0.30026848570880804, "token_entropy": 0.2855471509722568, "gradient_norm_pre_clip": 0.8204597970983664, "gradient_norm_post_clip": 0.8204597970983664, "parameter_l2_norm": 1111.374884301955, "parameter_l2_from_init": 3.9033218359574966, "relative_parameter_l2_from_init": 0.0035120702848806112, "update_norm": 0.08385199614452032, "relative_update_norm": 7.544888527617486e-05}
+{"global_step": 48, "epoch": 0, "batch_id": 3071, "num_supervised_tokens": 39441, "learning_rate": 1e-05, "batch_loss": 0.3346621644175761, "token_entropy": 0.3238676108693996, "gradient_norm_pre_clip": 0.9174978078306488, "gradient_norm_post_clip": 0.9174978078306488, "parameter_l2_norm": 1111.374500075895, "parameter_l2_from_init": 3.935139906006575, "relative_parameter_l2_from_init": 0.003540699053667289, "update_norm": 0.08179564974196582, "relative_update_norm": 7.359863820555542e-05}
+{"global_step": 49, "epoch": 0, "batch_id": 3135, "num_supervised_tokens": 38556, "learning_rate": 1e-05, "batch_loss": 0.29028012816085896, "token_entropy": 0.2904777863495874, "gradient_norm_pre_clip": 0.8158777145010899, "gradient_norm_post_clip": 0.8158777145010899, "parameter_l2_norm": 1111.374133191979, "parameter_l2_from_init": 3.965279745528411, "relative_parameter_l2_from_init": 0.003567817810260862, "update_norm": 0.07952575718404478, "relative_update_norm": 7.155624268097618e-05}
+{"global_step": 50, "epoch": 0, "batch_id": 3199, "num_supervised_tokens": 40743, "learning_rate": 1e-05, "batch_loss": 0.32469227994270694, "token_entropy": 0.32338379681866725, "gradient_norm_pre_clip": 0.8254355019539719, "gradient_norm_post_clip": 0.8254355019539719, "parameter_l2_norm": 1111.373745615621, "parameter_l2_from_init": 3.9938642999971625, "relative_parameter_l2_from_init": 0.0035935371766301553, "update_norm": 0.07757271548551037, "relative_update_norm": 6.979894548664245e-05}
+{"global_step": 51, "epoch": 0, "batch_id": 3263, "num_supervised_tokens": 40553, "learning_rate": 1e-05, "batch_loss": 0.3623727856130562, "token_entropy": 0.3627632084814208, "gradient_norm_pre_clip": 0.8109962853067555, "gradient_norm_post_clip": 0.8109962853067555, "parameter_l2_norm": 1111.373369494821, "parameter_l2_from_init": 4.021706188325537, "relative_parameter_l2_from_init": 0.0036185883183967077, "update_norm": 0.07643370906476078, "relative_update_norm": 6.877410523117359e-05}
+{"global_step": 52, "epoch": 0, "batch_id": 3327, "num_supervised_tokens": 41338, "learning_rate": 1e-05, "batch_loss": 0.34058005133124397, "token_entropy": 0.34146710190534696, "gradient_norm_pre_clip": 0.8057442159640869, "gradient_norm_post_clip": 0.8057442159640869, "parameter_l2_norm": 1111.372997199324, "parameter_l2_from_init": 4.048722697039907, "relative_parameter_l2_from_init": 0.0036428968129161407, "update_norm": 0.0757059338420721, "relative_update_norm": 6.811928491411268e-05}
+{"global_step": 53, "epoch": 0, "batch_id": 3391, "num_supervised_tokens": 40788, "learning_rate": 1e-05, "batch_loss": 0.31408596454168863, "token_entropy": 0.31541321774937536, "gradient_norm_pre_clip": 0.7779131896896184, "gradient_norm_post_clip": 0.7779131896896184, "parameter_l2_norm": 1111.3726109217407, "parameter_l2_from_init": 4.074596039883554, "relative_parameter_l2_from_init": 0.0036661767274070783, "update_norm": 0.07500924773577247, "relative_update_norm": 6.749243862826702e-05}
+{"global_step": 54, "epoch": 0, "batch_id": 3455, "num_supervised_tokens": 43295, "learning_rate": 1e-05, "batch_loss": 0.8633354564870767, "token_entropy": 0.8582622241500288, "gradient_norm_pre_clip": 0.7978883312908075, "gradient_norm_post_clip": 0.7978883312908075, "parameter_l2_norm": 1111.3722687042932, "parameter_l2_from_init": 4.105170242477762, "relative_parameter_l2_from_init": 0.0036936863084582384, "update_norm": 0.08148758574816783, "relative_update_norm": 7.332159353154558e-05}
+{"global_step": 55, "epoch": 0, "batch_id": 3519, "num_supervised_tokens": 41589, "learning_rate": 1e-05, "batch_loss": 0.896627443543704, "token_entropy": 0.8886718982060979, "gradient_norm_pre_clip": 0.7921144658033267, "gradient_norm_post_clip": 0.7921144658033267, "parameter_l2_norm": 1111.3720015623846, "parameter_l2_from_init": 4.141565951738828, "relative_parameter_l2_from_init": 0.003726433874343128, "update_norm": 0.08950271040713419, "relative_update_norm": 8.053352998033946e-05}
+{"global_step": 56, "epoch": 0, "batch_id": 3583, "num_supervised_tokens": 36972, "learning_rate": 1e-05, "batch_loss": 0.3557583080050695, "token_entropy": 0.3465007816598643, "gradient_norm_pre_clip": 0.833158272483286, "gradient_norm_post_clip": 0.833158272483286, "parameter_l2_norm": 1111.3717177336898, "parameter_l2_from_init": 4.1763900679134585, "relative_parameter_l2_from_init": 0.003757767376614828, "update_norm": 0.08730331717072877, "relative_update_norm": 7.855456079875576e-05}
+{"global_step": 57, "epoch": 0, "batch_id": 3647, "num_supervised_tokens": 40103, "learning_rate": 1e-05, "batch_loss": 0.32647902025127923, "token_entropy": 0.3185660972668459, "gradient_norm_pre_clip": 0.8562844622244082, "gradient_norm_post_clip": 0.8562844622244082, "parameter_l2_norm": 1111.3714299318926, "parameter_l2_from_init": 4.209263314621674, "relative_parameter_l2_from_init": 0.00378734555586408, "update_norm": 0.08457713271137743, "relative_update_norm": 7.610158983172755e-05}
+{"global_step": 58, "epoch": 0, "batch_id": 3711, "num_supervised_tokens": 40479, "learning_rate": 1e-05, "batch_loss": 0.7181024808862639, "token_entropy": 0.7071846531044074, "gradient_norm_pre_clip": 0.8121140035022043, "gradient_norm_post_clip": 0.8121140035022043, "parameter_l2_norm": 1111.3711488266806, "parameter_l2_from_init": 4.244442094468648, "relative_parameter_l2_from_init": 0.0038189981719053113, "update_norm": 0.0884230593203882, "relative_update_norm": 7.956213314853458e-05}
+{"global_step": 59, "epoch": 0, "batch_id": 3775, "num_supervised_tokens": 40628, "learning_rate": 1e-05, "batch_loss": 0.33156284170024086, "token_entropy": 0.3310607892504308, "gradient_norm_pre_clip": 0.770338544436683, "gradient_norm_post_clip": 0.770338544436683, "parameter_l2_norm": 1111.370851585567, "parameter_l2_from_init": 4.277895709146856, "relative_parameter_l2_from_init": 0.0038490985456307997, "update_norm": 0.08523893376446091, "relative_update_norm": 7.669711117837265e-05}
+{"global_step": 60, "epoch": 0, "batch_id": 3839, "num_supervised_tokens": 41323, "learning_rate": 1e-05, "batch_loss": 0.3467256864139547, "token_entropy": 0.3577566686888313, "gradient_norm_pre_clip": 0.7904698878852525, "gradient_norm_post_clip": 0.7904698878852525, "parameter_l2_norm": 1111.370564739476, "parameter_l2_from_init": 4.309957600998914, "relative_parameter_l2_from_init": 0.0038779466966116795, "update_norm": 0.0824427111023627, "relative_update_norm": 7.418111808790676e-05}
+{"global_step": 61, "epoch": 0, "batch_id": 3903, "num_supervised_tokens": 39408, "learning_rate": 1e-05, "batch_loss": 0.3039487356133702, "token_entropy": 0.30650235122652786, "gradient_norm_pre_clip": 0.8226817607351489, "gradient_norm_post_clip": 0.8226817607351489, "parameter_l2_norm": 1111.3702618398504, "parameter_l2_from_init": 4.340223841801995, "relative_parameter_l2_from_init": 0.003905179184586494, "update_norm": 0.08012306574310141, "relative_update_norm": 7.209394429041077e-05}
+{"global_step": 62, "epoch": 0, "batch_id": 3967, "num_supervised_tokens": 39380, "learning_rate": 1e-05, "batch_loss": 0.344430457377724, "token_entropy": 0.3376035106188153, "gradient_norm_pre_clip": 0.7639938841782681, "gradient_norm_post_clip": 0.7639938841782681, "parameter_l2_norm": 1111.3699483167777, "parameter_l2_from_init": 4.368779342883799, "relative_parameter_l2_from_init": 0.003930872409750613, "update_norm": 0.07840275606331235, "relative_update_norm": 7.054604650958668e-05}
+{"global_step": 63, "epoch": 0, "batch_id": 4031, "num_supervised_tokens": 39405, "learning_rate": 1e-05, "batch_loss": 0.28901337974308566, "token_entropy": 0.28835931336754184, "gradient_norm_pre_clip": 0.7279837717994925, "gradient_norm_post_clip": 0.7279837717994925, "parameter_l2_norm": 1111.3696275940968, "parameter_l2_from_init": 4.395711823456157, "relative_parameter_l2_from_init": 0.003955105298733773, "update_norm": 0.07644632676211903, "relative_update_norm": 6.878569007469706e-05}
+{"global_step": 64, "epoch": 0, "batch_id": 4095, "num_supervised_tokens": 40217, "learning_rate": 1e-05, "batch_loss": 0.2697283982564398, "token_entropy": 0.263231818980769, "gradient_norm_pre_clip": 0.7524662875224158, "gradient_norm_post_clip": 0.7524662875224158, "parameter_l2_norm": 1111.369301942354, "parameter_l2_from_init": 4.421121173142959, "relative_parameter_l2_from_init": 0.0039779677286700095, "update_norm": 0.07468420973627342, "relative_update_norm": 6.720017334089298e-05}
+{"global_step": 65, "epoch": 0, "batch_id": 4105, "num_supervised_tokens": 5109, "learning_rate": 1e-05, "batch_loss": 0.2012191668844055, "token_entropy": 0.19790141646421966, "gradient_norm_pre_clip": 1.8771221805800455, "gradient_norm_post_clip": 1.8771221805800455, "parameter_l2_norm": 1111.3689795942305, "parameter_l2_from_init": 4.4453503427087755, "relative_parameter_l2_from_init": 0.003999768274470652, "update_norm": 0.09515196959799986, "relative_update_norm": 8.561690252749415e-05}