{"event": "start", "run": "exp32_onechunk_full_s0", "condition": "A", "params": 82470528, "langs": ["en", "de", "vi", "tr", "fi", "ru", "el", "hi", "ta", "th", "zh", "my"], "premium": {"en": 1.0, "de": 1.184, "vi": 1.39, "tr": 1.12, "fi": 1.104, "ru": 1.971, "el": 2.174, "hi": 2.552, "ta": 3.192, "th": 2.762, "zh": 0.915, "my": 3.524}, "schedule": {"tau1": 0.4, "tau2": 0.7, "p1_objective": "global", "p2_objective": "global", "p3_objective": "global", "freeze_in_p3": false, "freeze_scope": "router"}, "corpus": "/local/scratch/sas245/beetle-hnet/scratch/data/tier1_corpus", "premium_kind": "flores_per_sent", "doc_filtered": {"my": 37511}, "weights": {"en": 0.2946, "de": 0.0909, "vi": 0.0576, "tr": 0.0612, "fi": 0.0424, "ru": 0.1476, "el": 0.0522, "hi": 0.042, "ta": 0.0313, "th": 0.0542, "zh": 0.1038, "my": 0.0222}, "t": 1.9}
{"event": "probe_schedule", "n": 30, "steps": [0, 1, 2, 4, 8, 16, 32, 64, 128, 256, 4000, 8000, 12000, 16000, 20000, 24000, 28000, 32000, 36000, 40000, 44000, 48000, 52000, 56000, 60000, 64000, 68000, 72000, 76000, 76300], "t": 1.9}
{"event": "probe", "step": 0, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 8739.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.502155, "cps": 68.273438, "mean_p": 0.50363, "boundary_entropy": 0.687415, "bpb": 8.146538}, "de": {"chunks": 12222.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.595237, "cps": 95.484375, "mean_p": 0.507356, "boundary_entropy": 0.688163, "bpb": 8.115859}, "vi": {"chunks": 6386.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.346406, "cps": 49.890625, "mean_p": 0.491233, "boundary_entropy": 0.688507, "bpb": 8.208555}, "tr": {"chunks": 11403.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.629896, "cps": 89.085938, "mean_p": 0.50659, "boundary_entropy": 0.687851, "bpb": 8.138757}, "fi": {"chunks": 7137.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.379406, "cps": 55.757812, "mean_p": 0.497164, "boundary_entropy": 0.68792, "bpb": 8.132766}, "ru": {"chunks": 10473.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.552665, "cps": 81.820312, "mean_p": 0.496692, "boundary_entropy": 0.690192, "bpb": 8.035314}, "el": {"chunks": 36555.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.781172, "cps": 285.585938, "mean_p": 0.530217, "boundary_entropy": 0.688772, "bpb": 8.179564}, "hi": {"chunks": 42584.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.530996, "cps": 332.6875, "mean_p": 0.521073, "boundary_entropy": 0.69018, "bpb": 7.981719}, "ta": {"chunks": 49758.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.661567, "cps": 388.734375, "mean_p": 0.521091, "boundary_entropy": 0.690409, "bpb": 7.914469}, "th": {"chunks": 45716.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.842151, "cps": 357.15625, "mean_p": 0.525713, "boundary_entropy": 0.689893, "bpb": 8.034608}, "zh": {"chunks": 6218.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.052828, "cps": 48.578125, "mean_p": 0.496822, "boundary_entropy": 0.686552, "bpb": 8.096223}, "my": {"chunks": 53266.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.811167, "cps": 416.140625, "mean_p": 0.521325, "boundary_entropy": 0.690575, "bpb": 8.110618}}, "bpb": {"en": 8.14654, "de": 8.11586, "vi": 8.20855, "tr": 8.13876, "fi": 8.13277, "ru": 8.03531, "el": 8.17956, "hi": 7.98172, "ta": 7.91447, "th": 8.03461, "zh": 8.09622, "my": 8.11062}, "gini_cpc": 0.391191, "gini_cps": 0.409188, "bpic": 1.321301, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1853654183968407, "vi": 0.6898396505624507, "tr": 1.2543852937189204, "fi": 0.755555191784985, "ru": 1.1005867247811874, "el": 3.54705827773647, "hi": 5.040269530092341, "ta": 5.300292378412309, "th": 5.659910185771798, "zh": 2.096619669942964, "my": 5.598208760550946}, "t": 4.6}
{"event": "train", "step": 0, "phase": 1, "objective": "global", "lm_loss": 5.62259, "bpb": 8.11168, "chunk_loss": 0.017982, "F": 0.00098, "G": 0.50248, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.69172, "router_noise": 0.0, "grad_norm": 4.926, "lr": 1e-07, "router_trainable": true, "sec_per_step": 0.001, "alpha": 0.03, "t": 5.4}
{"event": "probe", "step": 1, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 8748.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.502672, "cps": 68.34375, "mean_p": 0.503656, "boundary_entropy": 0.687415, "bpb": 8.142322}, "de": {"chunks": 12229.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.595578, "cps": 95.539062, "mean_p": 0.507374, "boundary_entropy": 0.688162, "bpb": 8.11158}, "vi": {"chunks": 6393.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.346786, "cps": 49.945312, "mean_p": 0.491256, "boundary_entropy": 0.688508, "bpb": 8.204543}, "tr": {"chunks": 11412.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.630393, "cps": 89.15625, "mean_p": 0.506612, "boundary_entropy": 0.687851, "bpb": 8.135617}, "fi": {"chunks": 7146.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.379884, "cps": 55.828125, "mean_p": 0.497187, "boundary_entropy": 0.68792, "bpb": 8.129767}, "ru": {"chunks": 10493.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.55372, "cps": 81.976562, "mean_p": 0.496718, "boundary_entropy": 0.690193, "bpb": 8.030103}, "el": {"chunks": 36556.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.781221, "cps": 285.59375, "mean_p": 0.530212, "boundary_entropy": 0.688773, "bpb": 8.177577}, "hi": {"chunks": 42587.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.531174, "cps": 332.710938, "mean_p": 0.5211, "boundary_entropy": 0.690178, "bpb": 7.979981}, "ta": {"chunks": 49764.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.661888, "cps": 388.78125, "mean_p": 0.521106, "boundary_entropy": 0.690407, "bpb": 7.911907}, "th": {"chunks": 45718.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.842275, "cps": 357.171875, "mean_p": 0.525741, "boundary_entropy": 0.689891, "bpb": 8.031333}, "zh": {"chunks": 6220.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.053166, "cps": 48.59375, "mean_p": 0.496831, "boundary_entropy": 0.686552, "bpb": 8.096042}, "my": {"chunks": 53266.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.811167, "cps": 416.140625, "mean_p": 0.521313, "boundary_entropy": 0.690576, "bpb": 8.106692}}, "bpb": {"en": 8.14232, "de": 8.11158, "vi": 8.20454, "tr": 8.13562, "fi": 8.12977, "ru": 8.0301, "el": 8.17758, "hi": 7.97998, "ta": 7.91191, "th": 8.03133, "zh": 8.09604, "my": 8.10669}, "gini_cpc": 0.391021, "gini_cps": 0.409034, "bpic": 1.320959, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.184824111636145, "vi": 0.6898853272704342, "tr": 1.2540837989474496, "fi": 0.7557296722235678, "ru": 1.1015540317998054, "el": 3.5435059744605857, "hi": 5.0354387842247625, "ta": 5.29547787364197, "th": 5.654334589946329, "zh": 2.095136331098504, "my": 5.592449286517458}, "t": 7.7}
{"event": "probe", "step": 2, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 8764.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.503591, "cps": 68.46875, "mean_p": 0.503715, "boundary_entropy": 0.687415, "bpb": 8.131242}, "de": {"chunks": 12239.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.596065, "cps": 95.617188, "mean_p": 0.50741, "boundary_entropy": 0.688162, "bpb": 8.101659}, "vi": {"chunks": 6420.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.348251, "cps": 50.15625, "mean_p": 0.491304, "boundary_entropy": 0.68851, "bpb": 8.198075}, "tr": {"chunks": 11425.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.631111, "cps": 89.257812, "mean_p": 0.506644, "boundary_entropy": 0.68785, "bpb": 8.128362}, "fi": {"chunks": 7166.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.380947, "cps": 55.984375, "mean_p": 0.497235, "boundary_entropy": 0.687921, "bpb": 8.120704}, "ru": {"chunks": 10540.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.556201, "cps": 82.34375, "mean_p": 0.496763, "boundary_entropy": 0.690194, "bpb": 8.02066}, "el": {"chunks": 36555.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.781172, "cps": 285.585938, "mean_p": 0.530194, "boundary_entropy": 0.688775, "bpb": 8.173161}, "hi": {"chunks": 42591.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.531412, "cps": 332.742188, "mean_p": 0.521137, "boundary_entropy": 0.690175, "bpb": 7.973888}, "ta": {"chunks": 49772.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.662316, "cps": 388.84375, "mean_p": 0.521135, "boundary_entropy": 0.690405, "bpb": 7.90428}, "th": {"chunks": 45719.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.842338, "cps": 357.179688, "mean_p": 0.525787, "boundary_entropy": 0.689886, "bpb": 8.023717}, "zh": {"chunks": 6223.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.053674, "cps": 48.617188, "mean_p": 0.496846, "boundary_entropy": 0.686552, "bpb": 8.095042}, "my": {"chunks": 53266.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.811167, "cps": 416.140625, "mean_p": 0.521313, "boundary_entropy": 0.690576, "bpb": 8.101196}}, "bpb": {"en": 8.13124, "de": 8.10166, "vi": 8.19808, "tr": 8.12836, "fi": 8.1207, "ru": 8.02066, "el": 8.17316, "hi": 7.97389, "ta": 7.90428, "th": 8.02372, "zh": 8.09504, "my": 8.1012}, "gini_cpc": 0.39062, "gini_cps": 0.40867, "bpic": 1.320287, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1836281324962679, "vi": 0.6915341590848574, "tr": 1.253220264043625, "fi": 0.7564612250312089, "ru": 1.1044680264309858, "el": 3.5369400147351677, "hi": 5.026717925486065, "ta": 5.286659922811462, "th": 5.644135204535127, "zh": 2.0923200157093724, "my": 5.582239429307932}, "t": 10.7}
{"event": "probe", "step": 4, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 8831.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.507441, "cps": 68.992188, "mean_p": 0.503866, "boundary_entropy": 0.687414, "bpb": 8.097246}, "de": {"chunks": 12275.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.597818, "cps": 95.898438, "mean_p": 0.507482, "boundary_entropy": 0.688161, "bpb": 8.069407}, "vi": {"chunks": 6472.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.351071, "cps": 50.5625, "mean_p": 0.491443, "boundary_entropy": 0.688516, "bpb": 8.175298}, "tr": {"chunks": 11454.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.632713, "cps": 89.484375, "mean_p": 0.506747, "boundary_entropy": 0.687848, "bpb": 8.105044}, "fi": {"chunks": 7227.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.38419, "cps": 56.460938, "mean_p": 0.49736, "boundary_entropy": 0.687924, "bpb": 8.093508}, "ru": {"chunks": 10654.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.562216, "cps": 83.234375, "mean_p": 0.496858, "boundary_entropy": 0.690196, "bpb": 7.986125}, "el": {"chunks": 36550.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.780929, "cps": 285.546875, "mean_p": 0.530065, "boundary_entropy": 0.68879, "bpb": 8.155456}, "hi": {"chunks": 42600.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.531947, "cps": 332.8125, "mean_p": 0.521257, "boundary_entropy": 0.690165, "bpb": 7.953401}, "ta": {"chunks": 49799.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.66376, "cps": 389.054688, "mean_p": 0.521224, "boundary_entropy": 0.690398, "bpb": 7.879973}, "th": {"chunks": 45723.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.842586, "cps": 357.210938, "mean_p": 0.525996, "boundary_entropy": 0.689866, "bpb": 7.996255}, "zh": {"chunks": 6237.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.056045, "cps": 48.726562, "mean_p": 0.496895, "boundary_entropy": 0.686554, "bpb": 8.091757}, "my": {"chunks": 53267.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.81122, "cps": 416.148438, "mean_p": 0.521269, "boundary_entropy": 0.69058, "bpb": 8.081738}}, "bpb": {"en": 8.09725, "de": 8.06941, "vi": 8.1753, "tr": 8.10504, "fi": 8.09351, "ru": 7.98612, "el": 8.15546, "hi": 7.9534, "ta": 7.87997, "th": 7.99625, "zh": 8.09176, "my": 8.08174}, "gini_cpc": 0.389536, "gini_cps": 0.407693, "bpic": 1.318432, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.178103182173339, "vi": 0.6918462672041007, "tr": 1.2468691012317925, "fi": 0.7571124829619341, "ru": 1.1079437541474337, "el": 3.5096254571314778, "hi": 4.989634818449057, "ta": 5.249396626662458, "th": 5.601803724530328, "zh": 2.081117191643528, "my": 5.539991486751898}, "t": 14.0}
{"event": "probe", "step": 8, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 9051.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.520083, "cps": 70.710938, "mean_p": 0.504457, "boundary_entropy": 0.687412, "bpb": 7.968007}, "de": {"chunks": 12410.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.604393, "cps": 96.953125, "mean_p": 0.507789, "boundary_entropy": 0.688157, "bpb": 7.94816}, "vi": {"chunks": 6654.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.360944, "cps": 51.984375, "mean_p": 0.491935, "boundary_entropy": 0.688539, "bpb": 8.08812}, "tr": {"chunks": 11619.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.641827, "cps": 90.773438, "mean_p": 0.507155, "boundary_entropy": 0.68784, "bpb": 8.016203}, "fi": {"chunks": 7482.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.397746, "cps": 58.453125, "mean_p": 0.497866, "boundary_entropy": 0.687933, "bpb": 7.989154}, "ru": {"chunks": 11211.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.591609, "cps": 87.585938, "mean_p": 0.497354, "boundary_entropy": 0.690205, "bpb": 7.854394}, "el": {"chunks": 36510.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.77898, "cps": 285.234375, "mean_p": 0.529611, "boundary_entropy": 0.688841, "bpb": 8.090206}, "hi": {"chunks": 42641.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.534383, "cps": 333.132812, "mean_p": 0.5217, "boundary_entropy": 0.690128, "bpb": 7.8733}, "ta": {"chunks": 49851.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.666542, "cps": 389.460938, "mean_p": 0.521434, "boundary_entropy": 0.690382, "bpb": 7.792733}, "th": {"chunks": 45732.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.843146, "cps": 357.28125, "mean_p": 0.526506, "boundary_entropy": 0.689815, "bpb": 7.914746}, "zh": {"chunks": 6310.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.068405, "cps": 49.296875, "mean_p": 0.497112, "boundary_entropy": 0.686562, "bpb": 8.082346}, "my": {"chunks": 53262.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.810956, "cps": 416.109375, "mean_p": 0.521177, "boundary_entropy": 0.690587, "bpb": 8.038737}}, "bpb": {"en": 7.96801, "de": 7.94816, "vi": 8.08812, "tr": 8.0162, "fi": 7.98915, "ru": 7.85439, "el": 8.09021, "hi": 7.8733, "ta": 7.79273, "th": 7.91475, "zh": 8.08235, "my": 8.03874}, "gini_cpc": 0.385043, "gini_cps": 0.403563, "bpic": 1.311027, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1621091740062695, "vi": 0.6940123676708039, "tr": 1.2340869379066504, "fi": 0.7647744594463703, "ru": 1.137529566406021, "el": 3.42057047776385, "hi": 4.873038733949914, "ta": 5.127149256646144, "th": 5.466718169106798, "zh": 2.054298135152636, "my": 5.404825159015525}, "t": 18.2}
{"event": "probe", "step": 16, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 9985.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.573752, "cps": 78.007812, "mean_p": 0.506866, "boundary_entropy": 0.687389, "bpb": 7.527685}, "de": {"chunks": 13039.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.635027, "cps": 101.867188, "mean_p": 0.509193, "boundary_entropy": 0.688133, "bpb": 7.539079}, "vi": {"chunks": 7302.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.396094, "cps": 57.046875, "mean_p": 0.493531, "boundary_entropy": 0.688608, "bpb": 7.79385}, "tr": {"chunks": 12159.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.671657, "cps": 94.992188, "mean_p": 0.508635, "boundary_entropy": 0.687805, "bpb": 7.711454}, "fi": {"chunks": 8331.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.442879, "cps": 65.085938, "mean_p": 0.499666, "boundary_entropy": 0.687955, "bpb": 7.624599}, "ru": {"chunks": 13551.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.715092, "cps": 105.867188, "mean_p": 0.499328, "boundary_entropy": 0.690236, "bpb": 7.409384}, "el": {"chunks": 36384.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.77284, "cps": 284.25, "mean_p": 0.528395, "boundary_entropy": 0.688978, "bpb": 7.888284}, "hi": {"chunks": 42765.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.541753, "cps": 334.101562, "mean_p": 0.523287, "boundary_entropy": 0.689989, "bpb": 7.552659}, "ta": {"chunks": 50022.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.675689, "cps": 390.796875, "mean_p": 0.522055, "boundary_entropy": 0.690336, "bpb": 7.483386}, "th": {"chunks": 45780.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.84613, "cps": 357.65625, "mean_p": 0.528401, "boundary_entropy": 0.689618, "bpb": 7.573519}, "zh": {"chunks": 6556.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.110058, "cps": 51.21875, "mean_p": 0.497962, "boundary_entropy": 0.68659, "bpb": 8.050941}, "my": {"chunks": 53252.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.810429, "cps": 416.03125, "mean_p": 0.520784, "boundary_entropy": 0.690617, "bpb": 7.882852}}, "bpb": {"en": 7.52769, "de": 7.53908, "vi": 7.79385, "tr": 7.71145, "fi": 7.6246, "ru": 7.40938, "el": 7.88828, "hi": 7.55266, "ta": 7.48339, "th": 7.57352, "zh": 8.05094, "my": 7.88285}, "gini_cpc": 0.36939, "gini_cps": 0.389291, "bpic": 1.283008, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1067968874853213, "vi": 0.6903585972938252, "tr": 1.1706399957888554, "fi": 0.7719004640606116, "ru": 1.2463447308165414, "el": 3.0899087119203874, "hi": 4.4300584308855555, "ta": 4.663496266062641, "th": 4.960560766546151, "zh": 1.934735289524195, "my": 4.898336294815879}, "t": 23.4}
{"event": "probe", "step": 32, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 12909.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.741769, "cps": 100.851562, "mean_p": 0.514383, "boundary_entropy": 0.687206, "bpb": 6.530342}, "de": {"chunks": 15754.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.767253, "cps": 123.078125, "mean_p": 0.514561, "boundary_entropy": 0.688031, "bpb": 6.617548}, "vi": {"chunks": 8710.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.472471, "cps": 68.046875, "mean_p": 0.496792, "boundary_entropy": 0.688751, "bpb": 7.078966}, "tr": {"chunks": 13168.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.727393, "cps": 102.875, "mean_p": 0.51158, "boundary_entropy": 0.687736, "bpb": 6.960988}, "fi": {"chunks": 10884.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.578598, "cps": 85.03125, "mean_p": 0.505236, "boundary_entropy": 0.687971, "bpb": 6.741016}, "ru": {"chunks": 23782.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.254987, "cps": 185.796875, "mean_p": 0.507295, "boundary_entropy": 0.690209, "bpb": 6.567182}, "el": {"chunks": 35418.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.725771, "cps": 276.703125, "mean_p": 0.522646, "boundary_entropy": 0.689545, "bpb": 7.462208}, "hi": {"chunks": 42920.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.550966, "cps": 335.3125, "mean_p": 0.527305, "boundary_entropy": 0.689586, "bpb": 6.838323}, "ta": {"chunks": 50143.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.682161, "cps": 391.742188, "mean_p": 0.521874, "boundary_entropy": 0.690373, "bpb": 6.851098}, "th": {"chunks": 45876.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.852098, "cps": 358.40625, "mean_p": 0.532959, "boundary_entropy": 0.689088, "bpb": 6.704495}, "zh": {"chunks": 7345.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.243651, "cps": 57.382812, "mean_p": 0.500682, "boundary_entropy": 0.686665, "bpb": 8.030038}, "my": {"chunks": 53084.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.801562, "cps": 414.71875, "mean_p": 0.518571, "boundary_entropy": 0.690778, "bpb": 7.274269}}, "bpb": {"en": 6.53034, "de": 6.61755, "vi": 7.07897, "tr": 6.96099, "fi": 6.74102, "ru": 6.56718, "el": 7.46221, "hi": 6.83832, "ta": 6.8511, "th": 6.7045, "zh": 8.03004, "my": 7.27427}, "gini_cpc": 0.325199, "gini_cps": 0.350477, "bpic": 1.199342, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0343557984047693, "vi": 0.6369517460301852, "tr": 0.9806200887120358, "fi": 0.7800243658147074, "ru": 1.6918843759370186, "el": 2.326562414700798, "hi": 3.439031547498393, "ta": 3.615899603435085, "th": 3.8449969373862176, "zh": 1.676601602344075, "my": 3.776867801310592}, "t": 32.1}
{"event": "probe", "step": 64, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17216.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.989255, "cps": 134.5, "mean_p": 0.562916, "boundary_entropy": 0.680388, "bpb": 5.841323}, "de": {"chunks": 20634.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.004919, "cps": 161.203125, "mean_p": 0.564854, "boundary_entropy": 0.680646, "bpb": 5.992346}, "vi": {"chunks": 22178.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.203038, "cps": 173.265625, "mean_p": 0.528907, "boundary_entropy": 0.687367, "bpb": 6.72299}, "tr": {"chunks": 19322.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.067337, "cps": 150.953125, "mean_p": 0.555522, "boundary_entropy": 0.682321, "bpb": 6.392593}, "fi": {"chunks": 19132.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.017064, "cps": 149.46875, "mean_p": 0.555383, "boundary_entropy": 0.682396, "bpb": 6.045062}, "ru": {"chunks": 33793.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.783272, "cps": 264.007812, "mean_p": 0.525514, "boundary_entropy": 0.689268, "bpb": 5.509834}, "el": {"chunks": 6651.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.324075, "cps": 51.960938, "mean_p": 0.491735, "boundary_entropy": 0.690279, "bpb": 6.390755}, "hi": {"chunks": 43142.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.56416, "cps": 337.046875, "mean_p": 0.547004, "boundary_entropy": 0.686729, "bpb": 5.744061}, "ta": {"chunks": 50855.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.720246, "cps": 397.304688, "mean_p": 0.536716, "boundary_entropy": 0.688783, "bpb": 5.676793}, "th": {"chunks": 46048.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.862791, "cps": 359.75, "mean_p": 0.542548, "boundary_entropy": 0.68771, "bpb": 5.557942}, "zh": {"chunks": 12311.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.08449, "cps": 96.179688, "mean_p": 0.519296, "boundary_entropy": 0.686187, "bpb": 7.597142}, "my": {"chunks": 53370.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.816656, "cps": 416.953125, "mean_p": 0.523306, "boundary_entropy": 0.690393, "bpb": 5.912742}}, "bpb": {"en": 5.84132, "de": 5.99235, "vi": 6.72299, "tr": 6.39259, "fi": 6.04506, "ru": 5.50983, "el": 6.39076, "hi": 5.74406, "ta": 5.67679, "th": 5.55794, "zh": 7.59714, "my": 5.91274}, "gini_cpc": 0.276239, "gini_cps": 0.29683, "bpic": 1.113532, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0158343290255218, "vi": 1.2161050821080397, "tr": 1.0789303070492742, "fi": 1.0281118266244202, "ru": 1.8026416458641894, "el": 0.3275955312491, "hi": 2.5920123579710226, "ta": 2.7497933374080694, "th": 2.8938870290614567, "zh": 2.1071320597193703, "my": 2.8472505987331402}, "t": 46.6}
{"event": "probe", "step": 128, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17383.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.998851, "cps": 135.804688, "mean_p": 0.781587, "boundary_entropy": 0.517479, "bpb": 5.400876}, "de": {"chunks": 20823.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.014124, "cps": 162.679688, "mean_p": 0.785948, "boundary_entropy": 0.512924, "bpb": 5.547965}, "vi": {"chunks": 24065.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.305397, "cps": 188.007812, "mean_p": 0.633597, "boundary_entropy": 0.648501, "bpb": 6.102857}, "tr": {"chunks": 19667.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.086395, "cps": 153.648438, "mean_p": 0.758669, "boundary_entropy": 0.543778, "bpb": 5.937732}, "fi": {"chunks": 19497.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.036468, "cps": 152.320312, "mean_p": 0.768262, "boundary_entropy": 0.533315, "bpb": 5.606037}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.666712, "boundary_entropy": 0.63455, "bpb": 4.760145}, "el": {"chunks": 37059.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.80573, "cps": 289.523438, "mean_p": 0.524371, "boundary_entropy": 0.689575, "bpb": 5.191834}, "hi": {"chunks": 43165.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.565527, "cps": 337.226562, "mean_p": 0.598487, "boundary_entropy": 0.671966, "bpb": 4.483432}, "ta": {"chunks": 50884.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.721797, "cps": 397.53125, "mean_p": 0.593797, "boundary_entropy": 0.673994, "bpb": 4.517573}, "th": {"chunks": 46083.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.864967, "cps": 360.023438, "mean_p": 0.573865, "boundary_entropy": 0.680422, "bpb": 4.428636}, "zh": {"chunks": 16108.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.727396, "cps": 125.84375, "mean_p": 0.610654, "boundary_entropy": 0.662696, "bpb": 6.804355}, "my": {"chunks": 53696.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.833861, "cps": 419.5, "mean_p": 0.575809, "boundary_entropy": 0.680108, "bpb": 4.382354}}, "bpb": {"en": 5.40088, "de": 5.54796, "vi": 6.10286, "tr": 5.93773, "fi": 5.60604, "ru": 4.76015, "el": 5.19183, "hi": 4.48343, "ta": 4.51757, "th": 4.42864, "zh": 6.80435, "my": 4.38235}, "gini_cpc": 0.222072, "gini_cps": 0.234508, "bpic": 1.002065, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.015290405196144, "vi": 1.3068992661245542, "tr": 1.0876444707949326, "fi": 1.0376605316785297, "ru": 1.8258450124974255, "el": 1.8078077381350268, "hi": 2.5684792549363573, "ta": 2.7249288341816005, "th": 2.8682636473370495, "zh": 2.7305338722538353, "my": 2.837121590664039}, "t": 74.7}
{"event": "probe", "step": 256, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.872537, "boundary_entropy": 0.37491, "bpb": 4.24804}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.866353, "boundary_entropy": 0.383951, "bpb": 4.46678}, "vi": {"chunks": 24270.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316517, "cps": 189.609375, "mean_p": 0.732635, "boundary_entropy": 0.569202, "bpb": 4.318637}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.8232, "boundary_entropy": 0.448918, "bpb": 4.560246}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.858508, "boundary_entropy": 0.395631, "bpb": 4.472297}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.809793, "boundary_entropy": 0.483519, "bpb": 3.134515}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.746828, "boundary_entropy": 0.563695, "bpb": 3.364958}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.734049, "boundary_entropy": 0.576231, "bpb": 2.603986}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.733203, "boundary_entropy": 0.578116, "bpb": 2.616693}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.744419, "boundary_entropy": 0.565301, "bpb": 2.600731}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.836682, "boundary_entropy": 0.439207, "bpb": 5.883889}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.759935, "boundary_entropy": 0.549106, "bpb": 2.660104}}, "bpb": {"en": 4.24804, "de": 4.46678, "vi": 4.31864, "tr": 4.56025, "fi": 4.4723, "ru": 3.13451, "el": 3.36496, "hi": 2.60399, "ta": 2.61669, "th": 2.60073, "zh": 5.88389, "my": 2.6601}, "gini_cpc": 0.22152, "gini_cps": 0.233857, "bpic": 1.00001, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3153837378744861, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 127.3}
{"event": "train", "step": 500, "phase": 1, "objective": "global", "lm_loss": 2.27458, "bpb": 3.28153, "chunk_loss": 0.001225, "F": 0.00098, "G": 0.97069, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.12861, "router_noise": 0.0, "grad_norm": 1.8494, "lr": 7.51e-05, "router_trainable": true, "sec_per_step": 0.425, "alpha": 0.03, "t": 218.1}
{"event": "train", "step": 1000, "phase": 1, "objective": "global", "lm_loss": 1.75053, "bpb": 2.52548, "chunk_loss": 0.000274, "F": 0.00098, "G": 0.99726, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.0186, "router_noise": 0.0, "grad_norm": 1.9074, "lr": 0.0001501, "router_trainable": true, "sec_per_step": 0.335, "alpha": 0.03, "t": 385.9}
{"event": "train", "step": 1500, "phase": 1, "objective": "global", "lm_loss": 1.61068, "bpb": 2.32371, "chunk_loss": 0.000195, "F": 0.00098, "G": 0.99947, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00449, "router_noise": 0.0, "grad_norm": 0.8842, "lr": 0.0002251, "router_trainable": true, "sec_per_step": 0.289, "alpha": 0.03, "t": 530.6}
{"event": "train", "step": 2000, "phase": 1, "objective": "global", "lm_loss": 1.5071, "bpb": 2.17429, "chunk_loss": 0.000181, "F": 0.00098, "G": 0.99984, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00152, "router_noise": 0.0, "grad_norm": 0.9456, "lr": 0.0003, "router_trainable": true, "sec_per_step": 0.239, "alpha": 0.03, "t": 650.0}
{"event": "train", "step": 2500, "phase": 1, "objective": "global", "lm_loss": 1.37264, "bpb": 1.9803, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00105, "router_noise": 0.0, "grad_norm": 0.6328, "lr": 0.0003, "router_trainable": true, "sec_per_step": 0.221, "alpha": 0.03, "t": 760.3}
{"event": "train", "step": 3000, "phase": 1, "objective": "global", "lm_loss": 1.41488, "bpb": 2.04123, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.558, "lr": 0.0002999, "router_trainable": true, "sec_per_step": 0.217, "alpha": 0.03, "t": 868.7}
{"event": "train", "step": 3500, "phase": 1, "objective": "global", "lm_loss": 1.39633, "bpb": 2.01447, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.498, "lr": 0.0002997, "router_trainable": true, "sec_per_step": 0.216, "alpha": 0.03, "t": 976.5}
{"event": "probe", "step": 4000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 2.014759}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 2.193994}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001017, "bpb": 1.86607}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.9999, "boundary_entropy": 0.001017, "bpb": 2.153767}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.9999, "boundary_entropy": 0.001017, "bpb": 2.33251}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.9999, "boundary_entropy": 0.001019, "bpb": 1.338835}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.9999, "boundary_entropy": 0.00102, "bpb": 1.366243}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001019, "bpb": 1.121035}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.954891}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001019, "bpb": 1.085235}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 3.008486}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.9999, "boundary_entropy": 0.00102, "bpb": 1.087695}}, "bpb": {"en": 2.01476, "de": 2.19399, "vi": 1.86607, "tr": 2.15377, "fi": 2.33251, "ru": 1.33883, "el": 1.36624, "hi": 1.12103, "ta": 0.95489, "th": 1.08524, "zh": 3.00849, "my": 1.0877}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 1087.1}
{"event": "eval", "step": 4000, "phase": 1, "bpb": {"en": 2.02184, "de": 2.185, "vi": 1.87254, "tr": 2.14585, "fi": 2.34821, "ru": 1.3373, "el": 1.36327, "hi": 1.13119, "ta": 0.95528, "th": 1.08086, "zh": 3.01432, "my": 1.0982}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.71282, "t": 1090.6}
{"event": "train", "step": 4000, "phase": 1, "objective": "global", "lm_loss": 1.38822, "bpb": 2.00277, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4777, "lr": 0.0002995, "router_trainable": true, "sec_per_step": 0.229, "alpha": 0.03, "t": 1090.9}
{"event": "train", "step": 4500, "phase": 1, "objective": "global", "lm_loss": 1.31885, "bpb": 1.9027, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4714, "lr": 0.0002992, "router_trainable": true, "sec_per_step": 0.218, "alpha": 0.03, "t": 1199.8}
{"event": "train", "step": 5000, "phase": 1, "objective": "global", "lm_loss": 1.27138, "bpb": 1.83422, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4424, "lr": 0.0002989, "router_trainable": true, "sec_per_step": 0.217, "alpha": 0.03, "t": 1308.1}
{"event": "train", "step": 5500, "phase": 1, "objective": "global", "lm_loss": 1.23345, "bpb": 1.77949, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4412, "lr": 0.0002985, "router_trainable": true, "sec_per_step": 0.216, "alpha": 0.03, "t": 1416.4}
{"event": "train", "step": 6000, "phase": 1, "objective": "global", "lm_loss": 1.19735, "bpb": 1.72741, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4381, "lr": 0.0002981, "router_trainable": true, "sec_per_step": 0.218, "alpha": 0.03, "t": 1525.6}
{"event": "train", "step": 6500, "phase": 1, "objective": "global", "lm_loss": 1.12921, "bpb": 1.62911, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4227, "lr": 0.0002976, "router_trainable": true, "sec_per_step": 0.22, "alpha": 0.03, "t": 1635.8}
{"event": "train", "step": 7000, "phase": 1, "objective": "global", "lm_loss": 1.15677, "bpb": 1.66886, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4027, "lr": 0.000297, "router_trainable": true, "sec_per_step": 0.22, "alpha": 0.03, "t": 1745.8}
{"event": "train", "step": 7500, "phase": 1, "objective": "global", "lm_loss": 1.14075, "bpb": 1.64575, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3633, "lr": 0.0002964, "router_trainable": true, "sec_per_step": 0.222, "alpha": 0.03, "t": 1856.6}
{"event": "probe", "step": 8000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.78099}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.919919}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.571943}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.815175}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 2.038556}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 1.181113}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 1.202795}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.990248}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.840922}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.954653}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.640377}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.9394}}, "bpb": {"en": 1.78099, "de": 1.91992, "vi": 1.57194, "tr": 1.81517, "fi": 2.03856, "ru": 1.18111, "el": 1.2028, "hi": 0.99025, "ta": 0.84092, "th": 0.95465, "zh": 2.64038, "my": 0.9394}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 1965.7}
{"event": "eval", "step": 8000, "phase": 1, "bpb": {"en": 1.7817, "de": 1.92374, "vi": 1.58936, "tr": 1.80739, "fi": 2.04917, "ru": 1.18006, "el": 1.20113, "hi": 0.9946, "ta": 0.84369, "th": 0.95467, "zh": 2.64443, "my": 0.9461}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.493, "t": 1969.7}
{"event": "train", "step": 8000, "phase": 1, "objective": "global", "lm_loss": 1.15993, "bpb": 1.67343, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3682, "lr": 0.0002957, "router_trainable": true, "sec_per_step": 0.226, "alpha": 0.03, "t": 1969.9}
{"event": "train", "step": 8500, "phase": 1, "objective": "global", "lm_loss": 1.13879, "bpb": 1.64293, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3814, "lr": 0.0002949, "router_trainable": true, "sec_per_step": 0.216, "alpha": 0.03, "t": 2077.8}
{"event": "train", "step": 9000, "phase": 1, "objective": "global", "lm_loss": 1.12127, "bpb": 1.61765, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3591, "lr": 0.0002941, "router_trainable": true, "sec_per_step": 0.218, "alpha": 0.03, "t": 2187.0}
{"event": "train", "step": 9500, "phase": 1, "objective": "global", "lm_loss": 1.17138, "bpb": 1.68994, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3574, "lr": 0.0002933, "router_trainable": true, "sec_per_step": 0.221, "alpha": 0.03, "t": 2297.5}
{"event": "train", "step": 10000, "phase": 1, "objective": "global", "lm_loss": 1.1062, "bpb": 1.59591, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3396, "lr": 0.0002924, "router_trainable": true, "sec_per_step": 0.221, "alpha": 0.03, "t": 2408.1}
{"event": "train", "step": 10500, "phase": 1, "objective": "global", "lm_loss": 1.10329, "bpb": 1.59171, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3388, "lr": 0.0002914, "router_trainable": true, "sec_per_step": 0.227, "alpha": 0.03, "t": 2521.6}
{"event": "train", "step": 11000, "phase": 1, "objective": "global", "lm_loss": 1.14079, "bpb": 1.64581, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3411, "lr": 0.0002903, "router_trainable": true, "sec_per_step": 0.223, "alpha": 0.03, "t": 2633.2}
{"event": "train", "step": 11500, "phase": 1, "objective": "global", "lm_loss": 1.04231, "bpb": 1.50373, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3233, "lr": 0.0002893, "router_trainable": true, "sec_per_step": 0.221, "alpha": 0.03, "t": 2743.9}
{"event": "probe", "step": 12000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.728047}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.852458}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.492436}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.744193}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.944011}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 1.123151}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 1.13048}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.926171}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.803184}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.899788}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.539111}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.885468}}, "bpb": {"en": 1.72805, "de": 1.85246, "vi": 1.49244, "tr": 1.74419, "fi": 1.94401, "ru": 1.12315, "el": 1.13048, "hi": 0.92617, "ta": 0.80318, "th": 0.89979, "zh": 2.53911, "my": 0.88547}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 2855.2}
{"event": "eval", "step": 12000, "phase": 1, "bpb": {"en": 1.7282, "de": 1.85761, "vi": 1.5102, "tr": 1.73765, "fi": 1.95507, "ru": 1.12178, "el": 1.12913, "hi": 0.93178, "ta": 0.807, "th": 0.90008, "zh": 2.54317, "my": 0.89157}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.4261, "t": 2858.9}
{"event": "train", "step": 12000, "phase": 1, "objective": "global", "lm_loss": 1.01812, "bpb": 1.46884, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3247, "lr": 0.0002881, "router_trainable": true, "sec_per_step": 0.23, "alpha": 0.03, "t": 2859.1}
{"event": "train", "step": 12500, "phase": 1, "objective": "global", "lm_loss": 1.03938, "bpb": 1.4995, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3239, "lr": 0.0002869, "router_trainable": true, "sec_per_step": 0.22, "alpha": 0.03, "t": 2968.9}
{"event": "train", "step": 13000, "phase": 1, "objective": "global", "lm_loss": 1.12915, "bpb": 1.62901, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3478, "lr": 0.0002857, "router_trainable": true, "sec_per_step": 0.219, "alpha": 0.03, "t": 3078.2}
{"event": "train", "step": 13500, "phase": 1, "objective": "global", "lm_loss": 1.02237, "bpb": 1.47496, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3489, "lr": 0.0002844, "router_trainable": true, "sec_per_step": 0.22, "alpha": 0.03, "t": 3188.0}
{"event": "train", "step": 14000, "phase": 1, "objective": "global", "lm_loss": 1.0833, "bpb": 1.56287, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3204, "lr": 0.000283, "router_trainable": true, "sec_per_step": 0.217, "alpha": 0.03, "t": 3296.5}
{"event": "train", "step": 14500, "phase": 1, "objective": "global", "lm_loss": 1.03992, "bpb": 1.50028, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3306, "lr": 0.0002816, "router_trainable": true, "sec_per_step": 0.22, "alpha": 0.03, "t": 3406.2}
{"event": "train", "step": 15000, "phase": 1, "objective": "global", "lm_loss": 1.06684, "bpb": 1.53913, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3424, "lr": 0.0002801, "router_trainable": true, "sec_per_step": 0.221, "alpha": 0.03, "t": 3516.5}
{"event": "train", "step": 15500, "phase": 1, "objective": "global", "lm_loss": 1.00978, "bpb": 1.45681, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3413, "lr": 0.0002786, "router_trainable": true, "sec_per_step": 0.22, "alpha": 0.03, "t": 3626.5}
{"event": "probe", "step": 16000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.697555}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.809778}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.455973}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.706427}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.892486}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 1.094718}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 1.102459}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.899695}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.783922}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.879555}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.489251}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.860952}}, "bpb": {"en": 1.69755, "de": 1.80978, "vi": 1.45597, "tr": 1.70643, "fi": 1.89249, "ru": 1.09472, "el": 1.10246, "hi": 0.8997, "ta": 0.78392, "th": 0.87956, "zh": 2.48925, "my": 0.86095}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 3738.1}
{"event": "eval", "step": 16000, "phase": 1, "bpb": {"en": 1.69783, "de": 1.8163, "vi": 1.47327, "tr": 1.69785, "fi": 1.90543, "ru": 1.09382, "el": 1.10054, "hi": 0.90501, "ta": 0.78755, "th": 0.8798, "zh": 2.49518, "my": 0.86667}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.39327, "t": 3741.8}
{"event": "train", "step": 16000, "phase": 1, "objective": "global", "lm_loss": 0.98611, "bpb": 1.42265, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3195, "lr": 0.000277, "router_trainable": true, "sec_per_step": 0.231, "alpha": 0.03, "t": 3742.0}
{"event": "train", "step": 16500, "phase": 1, "objective": "global", "lm_loss": 1.03459, "bpb": 1.4926, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3302, "lr": 0.0002754, "router_trainable": true, "sec_per_step": 0.22, "alpha": 0.03, "t": 3852.0}
{"event": "train", "step": 17000, "phase": 1, "objective": "global", "lm_loss": 1.0824, "bpb": 1.56157, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3445, "lr": 0.0002737, "router_trainable": true, "sec_per_step": 0.22, "alpha": 0.03, "t": 3962.0}
{"event": "train", "step": 17500, "phase": 1, "objective": "global", "lm_loss": 0.96408, "bpb": 1.39088, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3167, "lr": 0.000272, "router_trainable": true, "sec_per_step": 0.22, "alpha": 0.03, "t": 4071.7}
{"event": "train", "step": 18000, "phase": 1, "objective": "global", "lm_loss": 1.03335, "bpb": 1.49082, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3284, "lr": 0.0002703, "router_trainable": true, "sec_per_step": 0.21, "alpha": 0.03, "t": 4176.9}
{"event": "train", "step": 18500, "phase": 1, "objective": "global", "lm_loss": 1.05379, "bpb": 1.5203, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3252, "lr": 0.0002685, "router_trainable": true, "sec_per_step": 0.226, "alpha": 0.03, "t": 4289.8}
{"event": "train", "step": 19000, "phase": 1, "objective": "global", "lm_loss": 1.0783, "bpb": 1.55566, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3394, "lr": 0.0002666, "router_trainable": true, "sec_per_step": 0.221, "alpha": 0.03, "t": 4400.3}
{"event": "train", "step": 19500, "phase": 1, "objective": "global", "lm_loss": 1.06552, "bpb": 1.53722, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3295, "lr": 0.0002647, "router_trainable": true, "sec_per_step": 0.223, "alpha": 0.03, "t": 4511.8}
{"event": "probe", "step": 20000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.673958}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.786895}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.433922}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.681648}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.857025}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 1.078181}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 1.084436}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.888832}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.766917}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.865857}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.455075}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.84045}}, "bpb": {"en": 1.67396, "de": 1.7869, "vi": 1.43392, "tr": 1.68165, "fi": 1.85702, "ru": 1.07818, "el": 1.08444, "hi": 0.88883, "ta": 0.76692, "th": 0.86586, "zh": 2.45507, "my": 0.84045}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 4622.8}
{"event": "eval", "step": 20000, "phase": 1, "bpb": {"en": 1.67418, "de": 1.79186, "vi": 1.45186, "tr": 1.6731, "fi": 1.87007, "ru": 1.07716, "el": 1.08274, "hi": 0.89359, "ta": 0.77204, "th": 0.86673, "zh": 2.45998, "my": 0.84423}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.37146, "t": 4627.1}
{"event": "train", "step": 20000, "phase": 1, "objective": "global", "lm_loss": 1.05789, "bpb": 1.52621, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3151, "lr": 0.0002628, "router_trainable": true, "sec_per_step": 0.231, "alpha": 0.03, "t": 4627.3}
{"event": "train", "step": 20500, "phase": 1, "objective": "global", "lm_loss": 1.01247, "bpb": 1.46068, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3313, "lr": 0.0002608, "router_trainable": true, "sec_per_step": 0.219, "alpha": 0.03, "t": 4736.9}
{"event": "train", "step": 21000, "phase": 1, "objective": "global", "lm_loss": 1.05071, "bpb": 1.51586, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3298, "lr": 0.0002587, "router_trainable": true, "sec_per_step": 0.216, "alpha": 0.03, "t": 4844.7}
{"event": "train", "step": 21500, "phase": 1, "objective": "global", "lm_loss": 1.0263, "bpb": 1.48063, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3244, "lr": 0.0002567, "router_trainable": true, "sec_per_step": 0.217, "alpha": 0.03, "t": 4953.4}
{"event": "train", "step": 22000, "phase": 1, "objective": "global", "lm_loss": 1.05559, "bpb": 1.5229, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3256, "lr": 0.0002545, "router_trainable": true, "sec_per_step": 0.216, "alpha": 0.03, "t": 5061.5}
{"event": "train", "step": 22500, "phase": 1, "objective": "global", "lm_loss": 1.02565, "bpb": 1.47971, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3217, "lr": 0.0002524, "router_trainable": true, "sec_per_step": 0.217, "alpha": 0.03, "t": 5170.1}
{"event": "train", "step": 23000, "phase": 1, "objective": "global", "lm_loss": 1.07119, "bpb": 1.5454, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3396, "lr": 0.0002502, "router_trainable": true, "sec_per_step": 0.219, "alpha": 0.03, "t": 5279.7}
{"event": "train", "step": 23500, "phase": 1, "objective": "global", "lm_loss": 0.99262, "bpb": 1.43205, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3575, "lr": 0.000248, "router_trainable": true, "sec_per_step": 0.217, "alpha": 0.03, "t": 5388.2}
{"event": "probe", "step": 24000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.658475}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.767446}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.418585}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.659959}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.829733}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 1.063983}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 1.06756}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.880793}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.757004}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.857555}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.421162}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.817688}}, "bpb": {"en": 1.65847, "de": 1.76745, "vi": 1.41859, "tr": 1.65996, "fi": 1.82973, "ru": 1.06398, "el": 1.06756, "hi": 0.88079, "ta": 0.757, "th": 0.85756, "zh": 2.42116, "my": 0.81769}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 5499.6}
{"event": "eval", "step": 24000, "phase": 1, "bpb": {"en": 1.65889, "de": 1.77245, "vi": 1.43579, "tr": 1.65306, "fi": 1.84403, "ru": 1.06272, "el": 1.06708, "hi": 0.88545, "ta": 0.76251, "th": 0.85786, "zh": 2.42551, "my": 0.82215}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.35396, "t": 5503.3}
{"event": "train", "step": 24000, "phase": 1, "objective": "global", "lm_loss": 1.06015, "bpb": 1.52948, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3344, "lr": 0.0002457, "router_trainable": true, "sec_per_step": 0.231, "alpha": 0.03, "t": 5503.5}
{"event": "train", "step": 24500, "phase": 1, "objective": "global", "lm_loss": 1.03738, "bpb": 1.49663, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3354, "lr": 0.0002434, "router_trainable": true, "sec_per_step": 0.217, "alpha": 0.03, "t": 5612.2}
{"event": "train", "step": 25000, "phase": 1, "objective": "global", "lm_loss": 1.03992, "bpb": 1.50029, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3237, "lr": 0.000241, "router_trainable": true, "sec_per_step": 0.221, "alpha": 0.03, "t": 5722.5}
{"event": "train", "step": 25500, "phase": 1, "objective": "global", "lm_loss": 1.04082, "bpb": 1.50159, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.331, "lr": 0.0002387, "router_trainable": true, "sec_per_step": 0.22, "alpha": 0.03, "t": 5832.4}
{"event": "train", "step": 26000, "phase": 1, "objective": "global", "lm_loss": 1.00795, "bpb": 1.45417, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3446, "lr": 0.0002363, "router_trainable": true, "sec_per_step": 0.22, "alpha": 0.03, "t": 5942.4}
{"event": "train", "step": 26500, "phase": 1, "objective": "global", "lm_loss": 0.99234, "bpb": 1.43164, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.33, "lr": 0.0002338, "router_trainable": true, "sec_per_step": 0.22, "alpha": 0.03, "t": 6052.5}
{"event": "train", "step": 27000, "phase": 1, "objective": "global", "lm_loss": 0.94556, "bpb": 1.36415, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3066, "lr": 0.0002313, "router_trainable": true, "sec_per_step": 0.216, "alpha": 0.03, "t": 6160.5}
{"event": "train", "step": 27500, "phase": 1, "objective": "global", "lm_loss": 1.01842, "bpb": 1.46927, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3394, "lr": 0.0002288, "router_trainable": true, "sec_per_step": 0.218, "alpha": 0.03, "t": 6269.5}
{"event": "probe", "step": 28000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.645219}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.753983}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.400641}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.641791}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.82195}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 1.049934}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 1.051378}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.865731}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.748096}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.841856}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.400818}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.813081}}, "bpb": {"en": 1.64522, "de": 1.75398, "vi": 1.40064, "tr": 1.64179, "fi": 1.82195, "ru": 1.04993, "el": 1.05138, "hi": 0.86573, "ta": 0.7481, "th": 0.84186, "zh": 2.40082, "my": 0.81308}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 6381.1}
{"event": "eval", "step": 28000, "phase": 1, "bpb": {"en": 1.64461, "de": 1.76022, "vi": 1.41812, "tr": 1.63458, "fi": 1.836, "ru": 1.04908, "el": 1.04827, "hi": 0.87148, "ta": 0.75354, "th": 0.84225, "zh": 2.40496, "my": 0.81661}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.33998, "t": 6384.8}
{"event": "train", "step": 28000, "phase": 1, "objective": "global", "lm_loss": 1.06689, "bpb": 1.5392, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3341, "lr": 0.0002263, "router_trainable": true, "sec_per_step": 0.231, "alpha": 0.03, "t": 6385.0}
{"event": "train", "step": 28500, "phase": 1, "objective": "global", "lm_loss": 1.07451, "bpb": 1.55019, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3334, "lr": 0.0002238, "router_trainable": true, "sec_per_step": 0.222, "alpha": 0.03, "t": 6496.1}
{"event": "train", "step": 29000, "phase": 1, "objective": "global", "lm_loss": 0.98582, "bpb": 1.42223, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3297, "lr": 0.0002212, "router_trainable": true, "sec_per_step": 0.223, "alpha": 0.03, "t": 6607.8}
{"event": "train", "step": 29500, "phase": 1, "objective": "global", "lm_loss": 1.06788, "bpb": 1.54062, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3472, "lr": 0.0002186, "router_trainable": true, "sec_per_step": 0.227, "alpha": 0.03, "t": 6721.4}
{"event": "train", "step": 30000, "phase": 1, "objective": "global", "lm_loss": 1.0208, "bpb": 1.47271, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3253, "lr": 0.0002159, "router_trainable": true, "sec_per_step": 0.348, "alpha": 0.03, "t": 6895.2}
{"event": "train", "step": 30500, "phase": 1, "objective": "global", "lm_loss": 1.09221, "bpb": 1.57573, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3669, "lr": 0.0002133, "router_trainable": true, "sec_per_step": 0.344, "alpha": 0.03, "t": 7067.0}
{"event": "phase_transition", "step": 30520, "phase": 2, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 7073.9}
{"event": "train", "step": 31000, "phase": 2, "objective": "global", "lm_loss": 1.0378, "bpb": 1.49723, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3544, "lr": 0.0002106, "router_trainable": true, "sec_per_step": 0.363, "alpha": 0.03, "t": 7248.5}
{"event": "train", "step": 31500, "phase": 2, "objective": "global", "lm_loss": 0.94686, "bpb": 1.36604, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3205, "lr": 0.0002079, "router_trainable": true, "sec_per_step": 0.338, "alpha": 0.03, "t": 7417.6}
{"event": "probe", "step": 32000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.634387}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.736177}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.385617}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.628038}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.805244}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 1.04638}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 1.041229}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.854669}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.741013}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.83557}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.385737}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.799425}}, "bpb": {"en": 1.63439, "de": 1.73618, "vi": 1.38562, "tr": 1.62804, "fi": 1.80524, "ru": 1.04638, "el": 1.04123, "hi": 0.85467, "ta": 0.74101, "th": 0.83557, "zh": 2.38574, "my": 0.79943}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 7595.7}
{"event": "eval", "step": 32000, "phase": 2, "bpb": {"en": 1.63359, "de": 1.7428, "vi": 1.40303, "tr": 1.62093, "fi": 1.81785, "ru": 1.04402, "el": 1.03865, "hi": 0.86052, "ta": 0.74553, "th": 0.83533, "zh": 2.3884, "my": 0.80427}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.32791, "t": 7601.3}
{"event": "train", "step": 32000, "phase": 2, "objective": "global", "lm_loss": 1.02002, "bpb": 1.47158, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3349, "lr": 0.0002052, "router_trainable": true, "sec_per_step": 0.368, "alpha": 0.03, "t": 7601.6}
{"event": "train", "step": 32500, "phase": 2, "objective": "global", "lm_loss": 0.94903, "bpb": 1.36917, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3384, "lr": 0.0002025, "router_trainable": true, "sec_per_step": 0.365, "alpha": 0.03, "t": 7784.0}
{"event": "train", "step": 33000, "phase": 2, "objective": "global", "lm_loss": 1.08391, "bpb": 1.56375, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3556, "lr": 0.0001997, "router_trainable": true, "sec_per_step": 0.34, "alpha": 0.03, "t": 7953.9}
{"event": "train", "step": 33500, "phase": 2, "objective": "global", "lm_loss": 1.02295, "bpb": 1.47581, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3408, "lr": 0.0001969, "router_trainable": true, "sec_per_step": 0.358, "alpha": 0.03, "t": 8133.1}
{"event": "train", "step": 34000, "phase": 2, "objective": "global", "lm_loss": 0.96506, "bpb": 1.39229, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3241, "lr": 0.0001942, "router_trainable": true, "sec_per_step": 0.357, "alpha": 0.03, "t": 8311.8}
{"event": "train", "step": 34500, "phase": 2, "objective": "global", "lm_loss": 1.01124, "bpb": 1.45891, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.35, "lr": 0.0001914, "router_trainable": true, "sec_per_step": 0.342, "alpha": 0.03, "t": 8482.7}
{"event": "train", "step": 35000, "phase": 2, "objective": "global", "lm_loss": 1.01553, "bpb": 1.4651, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3466, "lr": 0.0001886, "router_trainable": true, "sec_per_step": 0.359, "alpha": 0.03, "t": 8662.2}
{"event": "train", "step": 35500, "phase": 2, "objective": "global", "lm_loss": 1.05443, "bpb": 1.52122, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3541, "lr": 0.0001858, "router_trainable": true, "sec_per_step": 0.345, "alpha": 0.03, "t": 8835.0}
{"event": "probe", "step": 36000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.623672}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.726485}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.37718}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.615728}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.78513}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 1.036104}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 1.028934}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.849225}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.734363}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.826237}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.367016}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.794905}}, "bpb": {"en": 1.62367, "de": 1.72649, "vi": 1.37718, "tr": 1.61573, "fi": 1.78513, "ru": 1.0361, "el": 1.02893, "hi": 0.84922, "ta": 0.73436, "th": 0.82624, "zh": 2.36702, "my": 0.79491}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 9012.2}
{"event": "eval", "step": 36000, "phase": 2, "bpb": {"en": 1.62297, "de": 1.73276, "vi": 1.39357, "tr": 1.60894, "fi": 1.799, "ru": 1.03455, "el": 1.02661, "hi": 0.85458, "ta": 0.73899, "th": 0.82682, "zh": 2.3714, "my": 0.79898}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.31743, "t": 9018.2}
{"event": "train", "step": 36000, "phase": 2, "objective": "global", "lm_loss": 1.00003, "bpb": 1.44273, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3446, "lr": 0.0001829, "router_trainable": true, "sec_per_step": 0.367, "alpha": 0.03, "t": 9018.5}
{"event": "train", "step": 36500, "phase": 2, "objective": "global", "lm_loss": 0.98195, "bpb": 1.41665, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3515, "lr": 0.0001801, "router_trainable": true, "sec_per_step": 0.37, "alpha": 0.03, "t": 9203.4}
{"event": "train", "step": 37000, "phase": 2, "objective": "global", "lm_loss": 1.00795, "bpb": 1.45416, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3476, "lr": 0.0001773, "router_trainable": true, "sec_per_step": 0.341, "alpha": 0.03, "t": 9373.7}
{"event": "train", "step": 37500, "phase": 2, "objective": "global", "lm_loss": 0.99912, "bpb": 1.44142, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3725, "lr": 0.0001744, "router_trainable": true, "sec_per_step": 0.34, "alpha": 0.03, "t": 9543.8}
{"event": "train", "step": 38000, "phase": 2, "objective": "global", "lm_loss": 0.98829, "bpb": 1.4258, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3435, "lr": 0.0001716, "router_trainable": true, "sec_per_step": 0.365, "alpha": 0.03, "t": 9726.3}
{"event": "train", "step": 38500, "phase": 2, "objective": "global", "lm_loss": 0.98902, "bpb": 1.42686, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.346, "lr": 0.0001687, "router_trainable": true, "sec_per_step": 0.344, "alpha": 0.03, "t": 9898.1}
{"event": "train", "step": 39000, "phase": 2, "objective": "global", "lm_loss": 1.00676, "bpb": 1.45244, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3574, "lr": 0.0001659, "router_trainable": true, "sec_per_step": 0.343, "alpha": 0.03, "t": 10069.4}
{"event": "train", "step": 39500, "phase": 2, "objective": "global", "lm_loss": 0.94484, "bpb": 1.36312, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3704, "lr": 0.000163, "router_trainable": true, "sec_per_step": 0.342, "alpha": 0.03, "t": 10240.3}
{"event": "probe", "step": 40000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.614222}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.713837}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.366857}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.603351}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.764334}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 1.025368}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 1.020651}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.839806}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.727854}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.819226}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.360929}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.784484}}, "bpb": {"en": 1.61422, "de": 1.71384, "vi": 1.36686, "tr": 1.60335, "fi": 1.76433, "ru": 1.02537, "el": 1.02065, "hi": 0.83981, "ta": 0.72785, "th": 0.81923, "zh": 2.36093, "my": 0.78448}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 10425.6}
{"event": "eval", "step": 40000, "phase": 2, "bpb": {"en": 1.61354, "de": 1.71957, "vi": 1.38399, "tr": 1.59676, "fi": 1.77977, "ru": 1.02339, "el": 1.01884, "hi": 0.84567, "ta": 0.73323, "th": 0.8192, "zh": 2.36333, "my": 0.78978}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.30725, "t": 10431.5}
{"event": "train", "step": 40000, "phase": 2, "objective": "global", "lm_loss": 1.00954, "bpb": 1.45646, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3546, "lr": 0.0001601, "router_trainable": true, "sec_per_step": 0.383, "alpha": 0.03, "t": 10431.9}
{"event": "train", "step": 40500, "phase": 2, "objective": "global", "lm_loss": 0.99067, "bpb": 1.42924, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3716, "lr": 0.0001573, "router_trainable": true, "sec_per_step": 0.362, "alpha": 0.03, "t": 10612.7}
{"event": "train", "step": 41000, "phase": 2, "objective": "global", "lm_loss": 0.98765, "bpb": 1.42488, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.348, "lr": 0.0001545, "router_trainable": true, "sec_per_step": 0.358, "alpha": 0.03, "t": 10791.7}
{"event": "train", "step": 41500, "phase": 2, "objective": "global", "lm_loss": 1.00571, "bpb": 1.45093, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3851, "lr": 0.0001516, "router_trainable": true, "sec_per_step": 0.344, "alpha": 0.03, "t": 10963.8}
{"event": "train", "step": 42000, "phase": 2, "objective": "global", "lm_loss": 0.93298, "bpb": 1.34601, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3476, "lr": 0.0001488, "router_trainable": true, "sec_per_step": 0.347, "alpha": 0.03, "t": 11137.5}
{"event": "train", "step": 42500, "phase": 2, "objective": "global", "lm_loss": 1.05547, "bpb": 1.52272, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3632, "lr": 0.0001459, "router_trainable": true, "sec_per_step": 0.366, "alpha": 0.03, "t": 11320.6}
{"event": "train", "step": 43000, "phase": 2, "objective": "global", "lm_loss": 1.01976, "bpb": 1.47121, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3571, "lr": 0.0001431, "router_trainable": true, "sec_per_step": 0.348, "alpha": 0.03, "t": 11494.8}
{"event": "train", "step": 43500, "phase": 2, "objective": "global", "lm_loss": 1.0476, "bpb": 1.51137, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3748, "lr": 0.0001403, "router_trainable": true, "sec_per_step": 0.344, "alpha": 0.03, "t": 11666.9}
{"event": "probe", "step": 44000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.607977}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.705693}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.355888}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.590783}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.749973}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 1.019044}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 1.011626}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.833797}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.721255}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.813965}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.348259}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.77657}}, "bpb": {"en": 1.60798, "de": 1.70569, "vi": 1.35589, "tr": 1.59078, "fi": 1.74997, "ru": 1.01904, "el": 1.01163, "hi": 0.8338, "ta": 0.72126, "th": 0.81397, "zh": 2.34826, "my": 0.77657}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 11778.4}
{"event": "eval", "step": 44000, "phase": 2, "bpb": {"en": 1.6068, "de": 1.71158, "vi": 1.37328, "tr": 1.5847, "fi": 1.76378, "ru": 1.01682, "el": 1.00973, "hi": 0.83997, "ta": 0.72654, "th": 0.81415, "zh": 2.35295, "my": 0.78181}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.29851, "t": 11782.1}
{"event": "train", "step": 44000, "phase": 2, "objective": "global", "lm_loss": 0.94894, "bpb": 1.36904, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3755, "lr": 0.0001375, "router_trainable": true, "sec_per_step": 0.231, "alpha": 0.03, "t": 11782.3}
{"event": "train", "step": 44500, "phase": 2, "objective": "global", "lm_loss": 0.93577, "bpb": 1.35003, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3541, "lr": 0.0001347, "router_trainable": true, "sec_per_step": 0.222, "alpha": 0.03, "t": 11893.5}
{"event": "train", "step": 45000, "phase": 2, "objective": "global", "lm_loss": 0.98429, "bpb": 1.42003, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3733, "lr": 0.0001319, "router_trainable": true, "sec_per_step": 0.217, "alpha": 0.03, "t": 12002.1}
{"event": "train", "step": 45500, "phase": 2, "objective": "global", "lm_loss": 0.96751, "bpb": 1.39582, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3774, "lr": 0.0001292, "router_trainable": true, "sec_per_step": 0.218, "alpha": 0.03, "t": 12111.0}
{"event": "train", "step": 46000, "phase": 2, "objective": "global", "lm_loss": 1.01279, "bpb": 1.46115, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3722, "lr": 0.0001264, "router_trainable": true, "sec_per_step": 0.214, "alpha": 0.03, "t": 12218.3}
{"event": "train", "step": 46500, "phase": 2, "objective": "global", "lm_loss": 0.98445, "bpb": 1.42026, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3759, "lr": 0.0001237, "router_trainable": true, "sec_per_step": 0.217, "alpha": 0.03, "t": 12327.0}
{"event": "train", "step": 47000, "phase": 2, "objective": "global", "lm_loss": 0.99172, "bpb": 1.43075, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3984, "lr": 0.000121, "router_trainable": true, "sec_per_step": 0.215, "alpha": 0.03, "t": 12434.4}
{"event": "train", "step": 47500, "phase": 2, "objective": "global", "lm_loss": 0.99913, "bpb": 1.44144, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3998, "lr": 0.0001183, "router_trainable": true, "sec_per_step": 0.22, "alpha": 0.03, "t": 12544.6}
{"event": "probe", "step": 48000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.595592}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.693373}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.344724}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.581239}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.74025}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 1.012197}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 1.002397}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.825867}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.717642}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.806193}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.341567}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.769286}}, "bpb": {"en": 1.59559, "de": 1.69337, "vi": 1.34472, "tr": 1.58124, "fi": 1.74025, "ru": 1.0122, "el": 1.0024, "hi": 0.82587, "ta": 0.71764, "th": 0.80619, "zh": 2.34157, "my": 0.76929}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 12657.0}
{"event": "eval", "step": 48000, "phase": 2, "bpb": {"en": 1.59425, "de": 1.69919, "vi": 1.36126, "tr": 1.57509, "fi": 1.75525, "ru": 1.00942, "el": 1.00134, "hi": 0.83203, "ta": 0.7232, "th": 0.80607, "zh": 2.3453, "my": 0.77415}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.28971, "t": 12660.7}
{"event": "train", "step": 48000, "phase": 2, "objective": "global", "lm_loss": 0.98127, "bpb": 1.41568, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3676, "lr": 0.0001157, "router_trainable": true, "sec_per_step": 0.233, "alpha": 0.03, "t": 12660.9}
{"event": "train", "step": 48500, "phase": 2, "objective": "global", "lm_loss": 1.00379, "bpb": 1.44816, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3975, "lr": 0.000113, "router_trainable": true, "sec_per_step": 0.22, "alpha": 0.03, "t": 12771.0}
{"event": "train", "step": 49000, "phase": 2, "objective": "global", "lm_loss": 0.90378, "bpb": 1.30388, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3704, "lr": 0.0001104, "router_trainable": true, "sec_per_step": 0.226, "alpha": 0.03, "t": 12884.0}
{"event": "train", "step": 49500, "phase": 2, "objective": "global", "lm_loss": 1.05462, "bpb": 1.5215, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3924, "lr": 0.0001078, "router_trainable": true, "sec_per_step": 0.232, "alpha": 0.03, "t": 13000.1}
{"event": "train", "step": 50000, "phase": 2, "objective": "global", "lm_loss": 0.9803, "bpb": 1.41428, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.366, "lr": 0.0001052, "router_trainable": true, "sec_per_step": 0.239, "alpha": 0.03, "t": 13119.8}
{"event": "train", "step": 50500, "phase": 2, "objective": "global", "lm_loss": 1.00006, "bpb": 1.44278, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3916, "lr": 0.0001027, "router_trainable": true, "sec_per_step": 0.221, "alpha": 0.03, "t": 13230.3}
{"event": "train", "step": 51000, "phase": 2, "objective": "global", "lm_loss": 0.92594, "bpb": 1.33585, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3818, "lr": 0.0001002, "router_trainable": true, "sec_per_step": 0.237, "alpha": 0.03, "t": 13348.7}
{"event": "train", "step": 51500, "phase": 2, "objective": "global", "lm_loss": 0.94425, "bpb": 1.36226, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3895, "lr": 9.77e-05, "router_trainable": true, "sec_per_step": 0.235, "alpha": 0.03, "t": 13466.1}
{"event": "probe", "step": 52000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.588648}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.684238}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.335844}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.575302}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.722837}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 1.004394}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 0.993574}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.818386}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.708865}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.798838}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.320253}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.763436}}, "bpb": {"en": 1.58865, "de": 1.68424, "vi": 1.33584, "tr": 1.5753, "fi": 1.72284, "ru": 1.00439, "el": 0.99357, "hi": 0.81839, "ta": 0.70886, "th": 0.79884, "zh": 2.32025, "my": 0.76344}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 13585.1}
{"event": "eval", "step": 52000, "phase": 2, "bpb": {"en": 1.58716, "de": 1.68973, "vi": 1.35312, "tr": 1.56827, "fi": 1.73891, "ru": 1.00142, "el": 0.99217, "hi": 0.82421, "ta": 0.71442, "th": 0.79907, "zh": 2.32335, "my": 0.76793}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.27998, "t": 13588.9}
{"event": "train", "step": 52000, "phase": 2, "objective": "global", "lm_loss": 0.94748, "bpb": 1.36693, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3825, "lr": 9.52e-05, "router_trainable": true, "sec_per_step": 0.246, "alpha": 0.03, "t": 13589.1}
{"event": "train", "step": 52500, "phase": 2, "objective": "global", "lm_loss": 0.96063, "bpb": 1.38589, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.388, "lr": 9.28e-05, "router_trainable": true, "sec_per_step": 0.228, "alpha": 0.03, "t": 13703.0}
{"event": "train", "step": 53000, "phase": 2, "objective": "global", "lm_loss": 0.96601, "bpb": 1.39366, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.403, "lr": 9.04e-05, "router_trainable": true, "sec_per_step": 0.229, "alpha": 0.03, "t": 13817.4}
{"event": "phase_transition", "step": 53410, "phase": 3, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 13910.2}
{"event": "train", "step": 53500, "phase": 3, "objective": "global", "lm_loss": 1.00332, "bpb": 1.44749, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4145, "lr": 8.8e-05, "router_trainable": true, "sec_per_step": 0.227, "alpha": 0.03, "t": 13930.8}
{"event": "train", "step": 54000, "phase": 3, "objective": "global", "lm_loss": 0.94463, "bpb": 1.36281, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4215, "lr": 8.57e-05, "router_trainable": true, "sec_per_step": 0.226, "alpha": 0.03, "t": 14044.0}
{"event": "train", "step": 54500, "phase": 3, "objective": "global", "lm_loss": 0.99495, "bpb": 1.4354, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3881, "lr": 8.34e-05, "router_trainable": true, "sec_per_step": 0.228, "alpha": 0.03, "t": 14157.8}
{"event": "train", "step": 55000, "phase": 3, "objective": "global", "lm_loss": 0.94657, "bpb": 1.36561, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3987, "lr": 8.11e-05, "router_trainable": true, "sec_per_step": 0.226, "alpha": 0.03, "t": 14271.0}
{"event": "train", "step": 55500, "phase": 3, "objective": "global", "lm_loss": 1.00862, "bpb": 1.45513, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4126, "lr": 7.89e-05, "router_trainable": true, "sec_per_step": 0.226, "alpha": 0.03, "t": 14383.8}
{"event": "probe", "step": 56000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.581403}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.676919}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.326198}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.566054}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.714405}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 0.997053}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 0.98676}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.813531}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.705025}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.794335}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.315291}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.754878}}, "bpb": {"en": 1.5814, "de": 1.67692, "vi": 1.3262, "tr": 1.56605, "fi": 1.7144, "ru": 0.99705, "el": 0.98676, "hi": 0.81353, "ta": 0.70503, "th": 0.79434, "zh": 2.31529, "my": 0.75488}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 14497.1}
{"event": "eval", "step": 56000, "phase": 3, "bpb": {"en": 1.58015, "de": 1.68279, "vi": 1.34402, "tr": 1.5595, "fi": 1.73048, "ru": 0.99516, "el": 0.98514, "hi": 0.81995, "ta": 0.70998, "th": 0.79453, "zh": 2.31915, "my": 0.75985}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.27339, "t": 14500.6}
{"event": "train", "step": 56000, "phase": 3, "objective": "global", "lm_loss": 0.9959, "bpb": 1.43678, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4203, "lr": 7.68e-05, "router_trainable": true, "sec_per_step": 0.234, "alpha": 0.03, "t": 14500.8}
{"event": "train", "step": 56500, "phase": 3, "objective": "global", "lm_loss": 0.98112, "bpb": 1.41546, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4113, "lr": 7.46e-05, "router_trainable": true, "sec_per_step": 0.224, "alpha": 0.03, "t": 14612.9}
{"event": "train", "step": 57000, "phase": 3, "objective": "global", "lm_loss": 0.99729, "bpb": 1.43879, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4075, "lr": 7.25e-05, "router_trainable": true, "sec_per_step": 0.227, "alpha": 0.03, "t": 14726.5}
{"event": "train", "step": 57500, "phase": 3, "objective": "global", "lm_loss": 0.95831, "bpb": 1.38255, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.3929, "lr": 7.05e-05, "router_trainable": true, "sec_per_step": 0.228, "alpha": 0.03, "t": 14840.6}
{"event": "train", "step": 58000, "phase": 3, "objective": "global", "lm_loss": 1.0098, "bpb": 1.45683, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4099, "lr": 6.84e-05, "router_trainable": true, "sec_per_step": 0.228, "alpha": 0.03, "t": 14954.7}
{"event": "train", "step": 58500, "phase": 3, "objective": "global", "lm_loss": 0.99918, "bpb": 1.44151, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4234, "lr": 6.65e-05, "router_trainable": true, "sec_per_step": 0.227, "alpha": 0.03, "t": 15068.2}
{"event": "train", "step": 59000, "phase": 3, "objective": "global", "lm_loss": 1.01593, "bpb": 1.46567, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4487, "lr": 6.45e-05, "router_trainable": true, "sec_per_step": 0.223, "alpha": 0.03, "t": 15179.8}
{"event": "train", "step": 59500, "phase": 3, "objective": "global", "lm_loss": 0.97423, "bpb": 1.40552, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4346, "lr": 6.27e-05, "router_trainable": true, "sec_per_step": 0.226, "alpha": 0.03, "t": 15292.8}
{"event": "probe", "step": 60000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.57532}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.669001}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.320996}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.556827}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.701354}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 0.991443}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 0.980928}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.808399}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.699557}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.789243}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.30054}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.749641}}, "bpb": {"en": 1.57532, "de": 1.669, "vi": 1.321, "tr": 1.55683, "fi": 1.70135, "ru": 0.99144, "el": 0.98093, "hi": 0.8084, "ta": 0.69956, "th": 0.78924, "zh": 2.30054, "my": 0.74964}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 15408.8}
{"event": "eval", "step": 60000, "phase": 3, "bpb": {"en": 1.57442, "de": 1.67442, "vi": 1.3384, "tr": 1.55064, "fi": 1.71789, "ru": 0.98866, "el": 0.97959, "hi": 0.81487, "ta": 0.70506, "th": 0.78897, "zh": 2.30431, "my": 0.75501}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.26602, "t": 15413.0}
{"event": "train", "step": 60000, "phase": 3, "objective": "global", "lm_loss": 0.9808, "bpb": 1.41499, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4237, "lr": 6.08e-05, "router_trainable": true, "sec_per_step": 0.241, "alpha": 0.03, "t": 15413.2}
{"event": "train", "step": 60500, "phase": 3, "objective": "global", "lm_loss": 0.91468, "bpb": 1.31961, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4118, "lr": 5.9e-05, "router_trainable": true, "sec_per_step": 0.228, "alpha": 0.03, "t": 15527.3}
{"event": "train", "step": 61000, "phase": 3, "objective": "global", "lm_loss": 0.92009, "bpb": 1.32742, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4294, "lr": 5.73e-05, "router_trainable": true, "sec_per_step": 0.23, "alpha": 0.03, "t": 15642.3}
{"event": "train", "step": 61500, "phase": 3, "objective": "global", "lm_loss": 0.98574, "bpb": 1.42213, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.418, "lr": 5.56e-05, "router_trainable": true, "sec_per_step": 0.233, "alpha": 0.03, "t": 15758.9}
{"event": "train", "step": 62000, "phase": 3, "objective": "global", "lm_loss": 0.97408, "bpb": 1.4053, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4888, "lr": 5.39e-05, "router_trainable": true, "sec_per_step": 0.229, "alpha": 0.03, "t": 15873.2}
{"event": "train", "step": 62500, "phase": 3, "objective": "global", "lm_loss": 0.99733, "bpb": 1.43884, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4522, "lr": 5.23e-05, "router_trainable": true, "sec_per_step": 0.225, "alpha": 0.03, "t": 15985.6}
{"event": "train", "step": 63000, "phase": 3, "objective": "global", "lm_loss": 0.98193, "bpb": 1.41662, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4837, "lr": 5.08e-05, "router_trainable": true, "sec_per_step": 0.221, "alpha": 0.03, "t": 16096.0}
{"event": "train", "step": 63500, "phase": 3, "objective": "global", "lm_loss": 0.89086, "bpb": 1.28523, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4268, "lr": 4.93e-05, "router_trainable": true, "sec_per_step": 0.224, "alpha": 0.03, "t": 16207.9}
{"event": "probe", "step": 64000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.569704}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.662878}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.315205}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.549721}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.692091}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 0.987222}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 0.974941}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.801619}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.694984}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.785162}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.294897}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.744948}}, "bpb": {"en": 1.5697, "de": 1.66288, "vi": 1.3152, "tr": 1.54972, "fi": 1.69209, "ru": 0.98722, "el": 0.97494, "hi": 0.80162, "ta": 0.69498, "th": 0.78516, "zh": 2.2949, "my": 0.74495}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 16323.4}
{"event": "eval", "step": 64000, "phase": 3, "bpb": {"en": 1.56846, "de": 1.66837, "vi": 1.33194, "tr": 1.54317, "fi": 1.70833, "ru": 0.98446, "el": 0.97311, "hi": 0.80808, "ta": 0.70068, "th": 0.78541, "zh": 2.29767, "my": 0.74988}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.25996, "t": 16327.1}
{"event": "train", "step": 64000, "phase": 3, "objective": "global", "lm_loss": 0.9701, "bpb": 1.39955, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4309, "lr": 4.78e-05, "router_trainable": true, "sec_per_step": 0.239, "alpha": 0.03, "t": 16327.3}
{"event": "train", "step": 64500, "phase": 3, "objective": "global", "lm_loss": 0.91231, "bpb": 1.31619, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4176, "lr": 4.65e-05, "router_trainable": true, "sec_per_step": 0.224, "alpha": 0.03, "t": 16439.3}
{"event": "train", "step": 65000, "phase": 3, "objective": "global", "lm_loss": 0.96232, "bpb": 1.38833, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4219, "lr": 4.51e-05, "router_trainable": true, "sec_per_step": 0.234, "alpha": 0.03, "t": 16556.4}
{"event": "train", "step": 65500, "phase": 3, "objective": "global", "lm_loss": 0.92076, "bpb": 1.32838, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4276, "lr": 4.38e-05, "router_trainable": true, "sec_per_step": 0.228, "alpha": 0.03, "t": 16670.2}
{"event": "train", "step": 66000, "phase": 3, "objective": "global", "lm_loss": 0.91404, "bpb": 1.31868, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4426, "lr": 4.26e-05, "router_trainable": true, "sec_per_step": 0.225, "alpha": 0.03, "t": 16782.9}
{"event": "train", "step": 66500, "phase": 3, "objective": "global", "lm_loss": 1.01428, "bpb": 1.46329, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4361, "lr": 4.14e-05, "router_trainable": true, "sec_per_step": 0.225, "alpha": 0.03, "t": 16895.2}
{"event": "train", "step": 67000, "phase": 3, "objective": "global", "lm_loss": 0.92961, "bpb": 1.34115, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4775, "lr": 4.03e-05, "router_trainable": true, "sec_per_step": 0.224, "alpha": 0.03, "t": 17007.1}
{"event": "train", "step": 67500, "phase": 3, "objective": "global", "lm_loss": 0.92638, "bpb": 1.33649, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4469, "lr": 3.92e-05, "router_trainable": true, "sec_per_step": 0.224, "alpha": 0.03, "t": 17118.9}
{"event": "probe", "step": 68000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.565724}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.657781}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.308694}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.54371}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.687261}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 0.982975}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 0.969583}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.798889}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.692383}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.78076}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.294817}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.740946}}, "bpb": {"en": 1.56572, "de": 1.65778, "vi": 1.30869, "tr": 1.54371, "fi": 1.68726, "ru": 0.98298, "el": 0.96958, "hi": 0.79889, "ta": 0.69238, "th": 0.78076, "zh": 2.29482, "my": 0.74095}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 17227.2}
{"event": "eval", "step": 68000, "phase": 3, "bpb": {"en": 1.56411, "de": 1.66302, "vi": 1.3255, "tr": 1.53771, "fi": 1.7033, "ru": 0.97998, "el": 0.96736, "hi": 0.80516, "ta": 0.69733, "th": 0.78072, "zh": 2.29799, "my": 0.74609}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.25569, "t": 17230.9}
{"event": "train", "step": 68000, "phase": 3, "objective": "global", "lm_loss": 0.97618, "bpb": 1.40832, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4393, "lr": 3.82e-05, "router_trainable": true, "sec_per_step": 0.224, "alpha": 0.03, "t": 17231.1}
{"event": "train", "step": 68500, "phase": 3, "objective": "global", "lm_loss": 0.97326, "bpb": 1.40412, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4289, "lr": 3.73e-05, "router_trainable": true, "sec_per_step": 0.227, "alpha": 0.03, "t": 17344.5}
{"event": "train", "step": 69000, "phase": 3, "objective": "global", "lm_loss": 0.9927, "bpb": 1.43216, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4497, "lr": 3.64e-05, "router_trainable": true, "sec_per_step": 0.226, "alpha": 0.03, "t": 17457.7}
{"event": "train", "step": 69500, "phase": 3, "objective": "global", "lm_loss": 0.94864, "bpb": 1.36859, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4635, "lr": 3.55e-05, "router_trainable": true, "sec_per_step": 0.225, "alpha": 0.03, "t": 17570.4}
{"event": "train", "step": 70000, "phase": 3, "objective": "global", "lm_loss": 1.0058, "bpb": 1.45106, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4919, "lr": 3.48e-05, "router_trainable": true, "sec_per_step": 0.223, "alpha": 0.03, "t": 17681.9}
{"event": "train", "step": 70500, "phase": 3, "objective": "global", "lm_loss": 0.97695, "bpb": 1.40944, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4543, "lr": 3.4e-05, "router_trainable": true, "sec_per_step": 0.226, "alpha": 0.03, "t": 17795.1}
{"event": "train", "step": 71000, "phase": 3, "objective": "global", "lm_loss": 1.01534, "bpb": 1.46483, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4423, "lr": 3.34e-05, "router_trainable": true, "sec_per_step": 0.222, "alpha": 0.03, "t": 17905.9}
{"event": "train", "step": 71500, "phase": 3, "objective": "global", "lm_loss": 0.94483, "bpb": 1.36311, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4723, "lr": 3.28e-05, "router_trainable": true, "sec_per_step": 0.224, "alpha": 0.03, "t": 18017.7}
{"event": "probe", "step": 72000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.562088}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.653274}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.305625}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.541575}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.682232}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 0.980594}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 0.965655}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.795736}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.690118}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.777393}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.283211}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.7378}}, "bpb": {"en": 1.56209, "de": 1.65327, "vi": 1.30562, "tr": 1.54157, "fi": 1.68223, "ru": 0.98059, "el": 0.96565, "hi": 0.79574, "ta": 0.69012, "th": 0.77739, "zh": 2.28321, "my": 0.7378}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 18129.9}
{"event": "eval", "step": 72000, "phase": 3, "bpb": {"en": 1.56031, "de": 1.6588, "vi": 1.32247, "tr": 1.5352, "fi": 1.69812, "ru": 0.97723, "el": 0.96405, "hi": 0.80194, "ta": 0.69518, "th": 0.77798, "zh": 2.2857, "my": 0.74271}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.25164, "t": 18133.6}
{"event": "train", "step": 72000, "phase": 3, "objective": "global", "lm_loss": 0.93607, "bpb": 1.35047, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4563, "lr": 3.22e-05, "router_trainable": true, "sec_per_step": 0.232, "alpha": 0.03, "t": 18133.8}
{"event": "train", "step": 72500, "phase": 3, "objective": "global", "lm_loss": 0.91151, "bpb": 1.31503, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4762, "lr": 3.17e-05, "router_trainable": true, "sec_per_step": 0.222, "alpha": 0.03, "t": 18244.6}
{"event": "train", "step": 73000, "phase": 3, "objective": "global", "lm_loss": 1.02775, "bpb": 1.48273, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4807, "lr": 3.13e-05, "router_trainable": true, "sec_per_step": 0.223, "alpha": 0.03, "t": 18355.9}
{"event": "train", "step": 73500, "phase": 3, "objective": "global", "lm_loss": 1.0245, "bpb": 1.47804, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4898, "lr": 3.09e-05, "router_trainable": true, "sec_per_step": 0.222, "alpha": 0.03, "t": 18467.0}
{"event": "train", "step": 74000, "phase": 3, "objective": "global", "lm_loss": 0.98972, "bpb": 1.42786, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4826, "lr": 3.06e-05, "router_trainable": true, "sec_per_step": 0.222, "alpha": 0.03, "t": 18577.9}
{"event": "train", "step": 74500, "phase": 3, "objective": "global", "lm_loss": 0.94551, "bpb": 1.36408, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4551, "lr": 3.04e-05, "router_trainable": true, "sec_per_step": 0.224, "alpha": 0.03, "t": 18689.6}
{"event": "train", "step": 75000, "phase": 3, "objective": "global", "lm_loss": 0.94038, "bpb": 1.35669, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4943, "lr": 3.02e-05, "router_trainable": true, "sec_per_step": 0.224, "alpha": 0.03, "t": 18801.4}
{"event": "train", "step": 75500, "phase": 3, "objective": "global", "lm_loss": 0.96105, "bpb": 1.3865, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4915, "lr": 3.01e-05, "router_trainable": true, "sec_per_step": 0.223, "alpha": 0.03, "t": 18913.1}
{"event": "probe", "step": 76000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.560575}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.652698}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.302994}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.537401}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.67814}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 0.978262}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 0.964077}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.794831}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.687652}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.776512}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.276007}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.736259}}, "bpb": {"en": 1.56057, "de": 1.6527, "vi": 1.30299, "tr": 1.5374, "fi": 1.67814, "ru": 0.97826, "el": 0.96408, "hi": 0.79483, "ta": 0.68765, "th": 0.77651, "zh": 2.27601, "my": 0.73626}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 19028.0}
{"event": "eval", "step": 76000, "phase": 3, "bpb": {"en": 1.55863, "de": 1.65811, "vi": 1.32044, "tr": 1.53158, "fi": 1.69419, "ru": 0.97578, "el": 0.96292, "hi": 0.80097, "ta": 0.69342, "th": 0.77641, "zh": 2.27956, "my": 0.74096}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.24941, "t": 19031.8}
{"event": "train", "step": 76000, "phase": 3, "objective": "global", "lm_loss": 0.96237, "bpb": 1.38841, "chunk_loss": 0.000179, "F": 0.00098, "G": 0.9999, "chunks": 1.0, "bytes_per_chunk": 1024.0, "boundary_entropy": 0.00102, "router_noise": 0.0, "grad_norm": 0.4979, "lr": 3e-05, "router_trainable": true, "sec_per_step": 0.238, "alpha": 0.03, "t": 19032.0}
{"event": "probe", "step": 76300, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.559452}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.651009}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.302872}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.539152}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.680107}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 0.97902}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 0.965519}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.794732}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.687757}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.776659}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.279954}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.735461}}, "bpb": {"en": 1.55945, "de": 1.65101, "vi": 1.30287, "tr": 1.53915, "fi": 1.68011, "ru": 0.97902, "el": 0.96552, "hi": 0.79473, "ta": 0.68776, "th": 0.77666, "zh": 2.27995, "my": 0.73546}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 19101.6}
{"event": "eval", "step": 76300, "phase": 3, "bpb": {"en": 1.55834, "de": 1.65648, "vi": 1.32016, "tr": 1.53289, "fi": 1.69581, "ru": 0.97586, "el": 0.96398, "hi": 0.80115, "ta": 0.69313, "th": 0.77658, "zh": 2.28275, "my": 0.74076}, "bytes_per_chunk": {"en": 1024.0, "de": 1024.0, "vi": 1024.0, "tr": 1024.0, "fi": 1024.0, "ru": 1024.0, "el": 1024.0, "hi": 1024.0, "ta": 1024.0, "th": 1024.0, "zh": 1024.0, "my": 1024.0}, "bpb_mean": 1.24982, "t": 19105.7}
{"event": "probe", "step": 76300, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 17418.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 1.000862, "cps": 136.078125, "mean_p": 0.999901, "boundary_entropy": 0.001014, "bpb": 1.559452}, "de": {"chunks": 20862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 1.016023, "cps": 162.984375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.651009}, "vi": {"chunks": 24274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 1.316734, "cps": 189.640625, "mean_p": 0.9999, "boundary_entropy": 0.001016, "bpb": 1.302872}, "tr": {"chunks": 19742.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 1.090537, "cps": 154.234375, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.539152}, "fi": {"chunks": 19572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 1.040455, "cps": 152.90625, "mean_p": 0.999901, "boundary_entropy": 0.001015, "bpb": 1.680107}, "ru": {"chunks": 34560.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.823747, "cps": 270.0, "mean_p": 0.999901, "boundary_entropy": 0.001017, "bpb": 0.97902}, "el": {"chunks": 37264.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.815719, "cps": 291.125, "mean_p": 0.999901, "boundary_entropy": 0.001018, "bpb": 0.965519}, "hi": {"chunks": 43173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 2.566003, "cps": 337.289062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.794732}, "ta": {"chunks": 50915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.723455, "cps": 397.773438, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.687757}, "th": {"chunks": 46085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.865092, "cps": 360.039062, "mean_p": 0.9999, "boundary_entropy": 0.001018, "bpb": 0.776659}, "zh": {"chunks": 16173.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.738402, "cps": 126.351562, "mean_p": 0.999901, "boundary_entropy": 0.001013, "bpb": 2.279954}, "my": {"chunks": 53743.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.836342, "cps": 419.867188, "mean_p": 0.999901, "boundary_entropy": 0.001019, "bpb": 0.735461}}, "bpb": {"en": 1.55945, "de": 1.65101, "vi": 1.30287, "tr": 1.53915, "fi": 1.68011, "ru": 0.97902, "el": 0.96552, "hi": 0.79473, "ta": 0.68776, "th": 0.77666, "zh": 2.27995, "my": 0.73546}, "gini_cpc": 0.221516, "gini_cps": 0.233853, "bpic": 1.0, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.01514801064883, "vi": 1.3156005295906583, "tr": 1.0895983329898389, "fi": 1.0395590357977051, "ru": 1.8221761311426539, "el": 1.8141552920914839, "hi": 2.563793186225859, "ta": 2.721110089193083, "th": 2.8626243461390897, "zh": 2.7360433739796948, "my": 2.8338989711748765}, "t": 19107.8}
{"event": "reps_dumped", "dirs": ["/local/scratch/sas245/beetle-hnet/scratch/reps/exp32_onechunk_full_s0/th0.30", "/local/scratch/sas245/beetle-hnet/scratch/reps/exp32_onechunk_full_s0/th0.40", "/local/scratch/sas245/beetle-hnet/scratch/reps/exp32_onechunk_full_s0/th0.50", "/local/scratch/sas245/beetle-hnet/scratch/reps/exp32_onechunk_full_s0/th0.60", "/local/scratch/sas245/beetle-hnet/scratch/reps/exp32_onechunk_full_s0/th0.70"], "t": 19125.1}
{"event": "final", "run": "exp32_onechunk_full_s0", "condition": "A", "step": 76300, "params": 82470528, "per_lang": {"en": {"cpc": 1.000861920358559, "cps": 136.078125, "bpb": 1.55465540263535, "bytes_per_chunk": 1024.0}, "de": {"cpc": 1.0160229873861588, "cps": 162.984375, "bpb": 1.6536014824767307, "bytes_per_chunk": 1024.0}, "vi": {"cpc": 1.3167344724708434, "cps": 189.640625, "bpb": 1.3188714878687111, "bytes_per_chunk": 1024.0}, "tr": {"cpc": 1.0905374799756946, "cps": 154.234375, "bpb": 1.5317365538758294, "bytes_per_chunk": 1024.0}, "fi": {"cpc": 1.040455052894583, "cps": 152.90625, "bpb": 1.7016902501639997, "bytes_per_chunk": 1024.0}, "ru": {"cpc": 1.8237467018469657, "cps": 270.0, "bpb": 0.9783672510003824, "bytes_per_chunk": 1024.0}, "el": {"cpc": 1.8157189494713248, "cps": 291.125, "bpb": 0.9624589417868981, "bytes_per_chunk": 1024.0}, "hi": {"cpc": 2.566002971768202, "cps": 337.2890625, "bpb": 0.8020311979218256, "bytes_per_chunk": 1024.0}, "ta": {"cpc": 2.7234554693768387, "cps": 397.7734375, "bpb": 0.6915733115473524, "bytes_per_chunk": 1024.0}, "th": {"cpc": 2.8650917003419334, "cps": 360.0390625, "bpb": 0.7796792523496693, "bytes_per_chunk": 1024.0}, "zh": {"cpc": 2.7384016254656283, "cps": 126.3515625, "bpb": 2.283702253142574, "bytes_per_chunk": 1024.0}, "my": {"cpc": 2.8363415663922313, "cps": 419.8671875, "bpb": 0.7417384499635484, "bytes_per_chunk": 1024.0}}, "gini_cpc": 0.22151585453568945, "gini_cps": 0.2338532044751218, "bpic": 1.0, "seed": 0, "mean_boundary_entropy": 0.0010202066041529179, "dechunk": "ema", "router_noise": 0.0, "transitions": [{"step": 0, "phase": 1, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0}, {"step": 30520, "phase": 2, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0}, {"step": 53410, "phase": 3, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0}], "t": 19132.2}