Blame for log.jsonl

{"event": "start", "run": "specfull_A_s2", "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": 3.0}
{"event": "probe_schedule", "n": 44, "steps": [0, 1, 2, 4, 8, 16, 32, 64, 128, 256, 500, 1000, 1500, 2000, 2500, 3000, 3500, 4000, 4500, 5000, 5500, 6000, 6500, 7000, 7500, 8000, 12000, 16000, 20000, 24000, 28000, 32000, 36000, 40000, 44000, 48000, 52000, 56000, 60000, 64000, 68000, 72000, 76000, 76300], "t": 3.0}
{"event": "probe", "step": 0, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 10185.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.585244, "cps": 79.570312, "mean_p": 0.50663, "boundary_entropy": 0.687329, "bpb": 8.017475}, "de": {"chunks": 10966.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.534067, "cps": 85.671875, "mean_p": 0.503879, "boundary_entropy": 0.688251, "bpb": 8.111776}, "vi": {"chunks": 5741.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.311418, "cps": 44.851562, "mean_p": 0.489692, "boundary_entropy": 0.688475, "bpb": 7.989146}, "tr": {"chunks": 6370.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.351875, "cps": 49.765625, "mean_p": 0.495802, "boundary_entropy": 0.687878, "bpb": 8.107657}, "fi": {"chunks": 2115.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.112434, "cps": 16.523438, "mean_p": 0.479404, "boundary_entropy": 0.686744, "bpb": 8.107682}, "ru": {"chunks": 23602.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.245488, "cps": 184.390625, "mean_p": 0.507156, "boundary_entropy": 0.690197, "bpb": 8.281387}, "el": {"chunks": 1455.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.070896, "cps": 11.367188, "mean_p": 0.480429, "boundary_entropy": 0.689544, "bpb": 7.890102}, "hi": {"chunks": 7663.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.455453, "cps": 59.867188, "mean_p": 0.494285, "boundary_entropy": 0.690846, "bpb": 7.810752}, "ta": {"chunks": 357.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.019096, "cps": 2.789062, "mean_p": 0.475379, "boundary_entropy": 0.689908, "bpb": 7.609497}, "th": {"chunks": 27234.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 1.69313, "cps": 212.765625, "mean_p": 0.50318, "boundary_entropy": 0.691076, "bpb": 7.876604}, "zh": {"chunks": 14702.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.489333, "cps": 114.859375, "mean_p": 0.531852, "boundary_entropy": 0.685234, "bpb": 8.024454}, "my": {"chunks": 789.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.04164, "cps": 6.164062, "mean_p": 0.475663, "boundary_entropy": 0.690063, "bpb": 8.245358}}, "bpb": {"en": 8.01748, "de": 8.11178, "vi": 7.98915, "tr": 8.10766, "fi": 8.10768, "ru": 8.28139, "el": 7.8901, "hi": 7.81075, "ta": 7.6095, "th": 7.8766, "zh": 8.02445, "my": 8.24536}, "gini_cpc": 0.562508, "gini_cps": 0.489131, "bpic": 3.45192, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9125547315729478, "vi": 0.5321174384544991, "tr": 0.6012456783662331, "fi": 0.19211513271032077, "ru": 2.1281521716321925, "el": 0.1211393488838307, "hi": 0.7782279770749311, "ta": 0.032629155488401715, "th": 2.8930334475858417, "zh": 4.253496332079356, "my": 0.07115029646287195}, "t": 8.5}
{"event": "train", "step": 0, "phase": 1, "objective": "global", "lm_loss": 5.5908, "bpb": 8.06582, "chunk_loss": 0.059712, "F": 0.56091, "G": 0.50224, "chunks": 574.38, "bytes_per_chunk": 1.783, "boundary_entropy": 0.6917, "router_noise": 0.0, "grad_norm": 9.2823, "lr": 1e-07, "router_trainable": true, "sec_per_step": 0.001, "alpha": 0.03, "t": 9.4}
{"event": "probe", "step": 1, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 10165.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.584095, "cps": 79.414062, "mean_p": 0.506586, "boundary_entropy": 0.687329, "bpb": 8.008794}, "de": {"chunks": 10960.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.533775, "cps": 85.625, "mean_p": 0.503847, "boundary_entropy": 0.688251, "bpb": 8.103936}, "vi": {"chunks": 5730.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.310822, "cps": 44.765625, "mean_p": 0.489665, "boundary_entropy": 0.688474, "bpb": 7.984204}, "tr": {"chunks": 6353.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.350936, "cps": 49.632812, "mean_p": 0.495761, "boundary_entropy": 0.687876, "bpb": 8.101684}, "fi": {"chunks": 2112.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.112275, "cps": 16.5, "mean_p": 0.479375, "boundary_entropy": 0.686741, "bpb": 8.102087}, "ru": {"chunks": 23603.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.245541, "cps": 184.398438, "mean_p": 0.507155, "boundary_entropy": 0.690197, "bpb": 8.268137}, "el": {"chunks": 1452.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.07075, "cps": 11.34375, "mean_p": 0.480416, "boundary_entropy": 0.689543, "bpb": 7.887171}, "hi": {"chunks": 7655.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.454978, "cps": 59.804688, "mean_p": 0.494273, "boundary_entropy": 0.690846, "bpb": 7.805385}, "ta": {"chunks": 357.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.019096, "cps": 2.789062, "mean_p": 0.475359, "boundary_entropy": 0.689906, "bpb": 7.606587}, "th": {"chunks": 27168.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 1.689027, "cps": 212.25, "mean_p": 0.503154, "boundary_entropy": 0.691076, "bpb": 7.867793}, "zh": {"chunks": 14703.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.489502, "cps": 114.867188, "mean_p": 0.531828, "boundary_entropy": 0.685236, "bpb": 8.02313}, "my": {"chunks": 790.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.041693, "cps": 6.171875, "mean_p": 0.475663, "boundary_entropy": 0.690063, "bpb": 8.245417}}, "bpb": {"en": 8.00879, "de": 8.10394, "vi": 7.9842, "tr": 8.10168, "fi": 8.10209, "ru": 8.26814, "el": 7.88717, "hi": 7.80538, "ta": 7.60659, "th": 7.86779, "zh": 8.02313, "my": 8.24542}, "gini_cpc": 0.56265, "gini_cps": 0.489159, "bpic": 3.455992, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9138499327405701, "vi": 0.532142832843117, "tr": 0.600820913307491, "fi": 0.19222008618990213, "ru": 2.1324297326305928, "el": 0.12112743158438319, "hi": 0.7789451173029852, "ta": 0.03269335451543251, "th": 2.8917007027808643, "zh": 4.2621551211645174, "my": 0.07138064245210518}, "t": 14.4}
{"event": "probe", "step": 2, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 10126.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.581854, "cps": 79.109375, "mean_p": 0.506493, "boundary_entropy": 0.687331, "bpb": 7.991795}, "de": {"chunks": 10925.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.53207, "cps": 85.351562, "mean_p": 0.503777, "boundary_entropy": 0.688251, "bpb": 8.088732}, "vi": {"chunks": 5706.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.30952, "cps": 44.578125, "mean_p": 0.489605, "boundary_entropy": 0.68847, "bpb": 7.974587}, "tr": {"chunks": 6334.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.349887, "cps": 49.484375, "mean_p": 0.495677, "boundary_entropy": 0.687873, "bpb": 8.090093}, "fi": {"chunks": 2109.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.112115, "cps": 16.476562, "mean_p": 0.479308, "boundary_entropy": 0.686733, "bpb": 8.092405}, "ru": {"chunks": 23599.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.24533, "cps": 184.367188, "mean_p": 0.507146, "boundary_entropy": 0.690197, "bpb": 8.241596}, "el": {"chunks": 1449.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.070604, "cps": 11.320312, "mean_p": 0.480409, "boundary_entropy": 0.689542, "bpb": 7.880888}, "hi": {"chunks": 7641.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.454146, "cps": 59.695312, "mean_p": 0.494264, "boundary_entropy": 0.690845, "bpb": 7.795115}, "ta": {"chunks": 357.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.019096, "cps": 2.789062, "mean_p": 0.475348, "boundary_entropy": 0.689905, "bpb": 7.598956}, "th": {"chunks": 27112.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 1.685546, "cps": 211.8125, "mean_p": 0.50313, "boundary_entropy": 0.691076, "bpb": 7.852349}, "zh": {"chunks": 14695.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.488148, "cps": 114.804688, "mean_p": 0.531777, "boundary_entropy": 0.685243, "bpb": 8.019945}, "my": {"chunks": 787.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.041535, "cps": 6.148438, "mean_p": 0.475637, "boundary_entropy": 0.690061, "bpb": 8.244649}}, "bpb": {"en": 7.9918, "de": 8.08873, "vi": 7.97459, "tr": 8.09009, "fi": 8.0924, "ru": 8.2416, "el": 7.88089, "hi": 7.79512, "ta": 7.59896, "th": 7.85235, "zh": 8.01994, "my": 8.24465}, "gini_cpc": 0.562953, "gini_cps": 0.489431, "bpic": 3.462477, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9144400434715445, "vi": 0.5319549108410593, "tr": 0.6013311543185373, "fi": 0.19268632486530646, "ru": 2.1402799502000387, "el": 0.12134272325979603, "hi": 0.7805151260536263, "ta": 0.032819272037267574, "th": 2.896854534942411, "zh": 4.2762426912270115, "my": 0.07138345322652126}, "t": 20.1}
{"event": "probe", "step": 4, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 9994.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.574269, "cps": 78.078125, "mean_p": 0.506127, "boundary_entropy": 0.687334, "bpb": 7.92982}, "de": {"chunks": 10795.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.525739, "cps": 84.335938, "mean_p": 0.503497, "boundary_entropy": 0.68825, "bpb": 8.032906}, "vi": {"chunks": 5596.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.303553, "cps": 43.71875, "mean_p": 0.489353, "boundary_entropy": 0.688457, "bpb": 7.940291}, "tr": {"chunks": 6190.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.341932, "cps": 48.359375, "mean_p": 0.49533, "boundary_entropy": 0.687861, "bpb": 8.048401}, "fi": {"chunks": 2074.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.110255, "cps": 16.203125, "mean_p": 0.479034, "boundary_entropy": 0.686703, "bpb": 8.056182}, "ru": {"chunks": 23587.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.244697, "cps": 184.273438, "mean_p": 0.50713, "boundary_entropy": 0.690199, "bpb": 8.161028}, "el": {"chunks": 1435.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.069922, "cps": 11.210938, "mean_p": 0.480358, "boundary_entropy": 0.689538, "bpb": 7.861634}, "hi": {"chunks": 7625.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.453195, "cps": 59.570312, "mean_p": 0.494241, "boundary_entropy": 0.690844, "bpb": 7.756058}, "ta": {"chunks": 355.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.018989, "cps": 2.773438, "mean_p": 0.475328, "boundary_entropy": 0.689904, "bpb": 7.572663}, "th": {"chunks": 26844.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 1.668884, "cps": 209.71875, "mean_p": 0.503021, "boundary_entropy": 0.691077, "bpb": 7.796914}, "zh": {"chunks": 14686.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.486624, "cps": 114.734375, "mean_p": 0.531608, "boundary_entropy": 0.685263, "bpb": 8.010271}, "my": {"chunks": 676.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.035677, "cps": 5.28125, "mean_p": 0.47554, "boundary_entropy": 0.690051, "bpb": 8.239559}}, "bpb": {"en": 7.92982, "de": 8.03291, "vi": 7.94029, "tr": 8.0484, "fi": 8.05618, "ru": 8.16103, "el": 7.86163, "hi": 7.75606, "ta": 7.57266, "th": 7.79691, "zh": 8.01027, "my": 8.23956}, "gini_cpc": 0.565547, "gini_cps": 0.49198, "bpic": 3.49346, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9154929718167919, "vi": 0.5285904821981879, "tr": 0.5954219938470291, "fi": 0.191991341740348, "ru": 2.167445908178152, "el": 0.12175753047209117, "hi": 0.7891681517082612, "ta": 0.03306645661641435, "th": 2.9061025797946907, "zh": 4.330069392803985, "my": 0.06212524221309226}, "t": 26.2}
{"event": "probe", "step": 8, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 9505.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.54617, "cps": 74.257812, "mean_p": 0.504804, "boundary_entropy": 0.687345, "bpb": 7.710362}, "de": {"chunks": 10305.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.501875, "cps": 80.507812, "mean_p": 0.502489, "boundary_entropy": 0.688246, "bpb": 7.83432}, "vi": {"chunks": 5307.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.287876, "cps": 41.460938, "mean_p": 0.488558, "boundary_entropy": 0.688414, "bpb": 7.818913}, "tr": {"chunks": 5764.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.3184, "cps": 45.03125, "mean_p": 0.494145, "boundary_entropy": 0.687814, "bpb": 7.896799}, "fi": {"chunks": 1999.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.106268, "cps": 15.617188, "mean_p": 0.478137, "boundary_entropy": 0.686601, "bpb": 7.927454}, "ru": {"chunks": 23473.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.238681, "cps": 183.382812, "mean_p": 0.507025, "boundary_entropy": 0.690206, "bpb": 7.863763}, "el": {"chunks": 1385.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.067485, "cps": 10.820312, "mean_p": 0.480058, "boundary_entropy": 0.689511, "bpb": 7.785512}, "hi": {"chunks": 7572.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.450045, "cps": 59.15625, "mean_p": 0.494141, "boundary_entropy": 0.69084, "bpb": 7.59775}, "ta": {"chunks": 351.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.018775, "cps": 2.742188, "mean_p": 0.475337, "boundary_entropy": 0.689907, "bpb": 7.471064}, "th": {"chunks": 25794.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 1.603606, "cps": 201.515625, "mean_p": 0.502639, "boundary_entropy": 0.691078, "bpb": 7.594774}, "zh": {"chunks": 14624.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.476126, "cps": 114.25, "mean_p": 0.530955, "boundary_entropy": 0.685341, "bpb": 7.976611}, "my": {"chunks": 663.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.034991, "cps": 5.179688, "mean_p": 0.475317, "boundary_entropy": 0.690028, "bpb": 8.209359}}, "bpb": {"en": 7.71036, "de": 7.83432, "vi": 7.81891, "tr": 7.8968, "fi": 7.92745, "ru": 7.86376, "el": 7.78551, "hi": 7.59775, "ta": 7.47106, "th": 7.59477, "zh": 7.97661, "my": 8.20936}, "gini_cpc": 0.569912, "gini_cps": 0.494782, "bpic": 3.595408, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9188985959733339, "vi": 0.5270817080984741, "tr": 0.5829689441762655, "fi": 0.1945686697523926, "ru": 2.2679390738661365, "el": 0.1235608613812455, "hi": 0.8240006065471133, "ta": 0.03437586585740054, "th": 2.9360917940380182, "zh": 4.533616025073156, "my": 0.06406519484595122}, "t": 33.2}
{"event": "probe", "step": 16, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 7718.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.443487, "cps": 60.296875, "mean_p": 0.500198, "boundary_entropy": 0.687333, "bpb": 7.094041}, "de": {"chunks": 8671.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.422296, "cps": 67.742188, "mean_p": 0.499086, "boundary_entropy": 0.688209, "bpb": 7.236975}, "vi": {"chunks": 4296.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.233035, "cps": 33.5625, "mean_p": 0.485888, "boundary_entropy": 0.688256, "bpb": 7.450285}, "tr": {"chunks": 4455.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.246092, "cps": 34.804688, "mean_p": 0.490191, "boundary_entropy": 0.687621, "bpb": 7.411554}, "fi": {"chunks": 1742.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.092605, "cps": 13.609375, "mean_p": 0.47527, "boundary_entropy": 0.686257, "bpb": 7.519829}, "ru": {"chunks": 21906.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.155989, "cps": 171.140625, "mean_p": 0.505754, "boundary_entropy": 0.690247, "bpb": 6.961035}, "el": {"chunks": 1242.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.060517, "cps": 9.703125, "mean_p": 0.479028, "boundary_entropy": 0.689418, "bpb": 7.49428}, "hi": {"chunks": 7017.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.417058, "cps": 54.820312, "mean_p": 0.49347, "boundary_entropy": 0.690815, "bpb": 7.01133}, "ta": {"chunks": 334.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.017866, "cps": 2.609375, "mean_p": 0.475376, "boundary_entropy": 0.68992, "bpb": 7.123181}, "th": {"chunks": 20377.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 1.266832, "cps": 159.195312, "mean_p": 0.500676, "boundary_entropy": 0.691079, "bpb": 6.821602}, "zh": {"chunks": 14342.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.428378, "cps": 112.046875, "mean_p": 0.528308, "boundary_entropy": 0.685635, "bpb": 7.85335}, "my": {"chunks": 712.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.037577, "cps": 5.5625, "mean_p": 0.474656, "boundary_entropy": 0.689959, "bpb": 8.062811}}, "bpb": {"en": 7.09404, "de": 7.23697, "vi": 7.45028, "tr": 7.41155, "fi": 7.51983, "ru": 6.96103, "el": 7.49428, "hi": 7.01133, "ta": 7.12318, "th": 6.8216, "zh": 7.85335, "my": 8.06281}, "gini_cpc": 0.58319, "gini_cps": 0.496927, "bpic": 4.135036, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9522174260669408, "vi": 0.5254609863430431, "tr": 0.55490227188503, "fi": 0.2088120769924926, "ru": 2.6065929421063463, "el": 0.13645834402643484, "hi": 0.9404067756931078, "ta": 0.040284719430256195, "th": 2.8565282680125654, "zh": 5.475649255631129, "my": 0.0847297575180594}, "t": 41.9}
{"event": "probe", "step": 32, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 5312.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.305235, "cps": 41.5, "mean_p": 0.494957, "boundary_entropy": 0.687303, "bpb": 6.284583}, "de": {"chunks": 7600.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.370136, "cps": 59.375, "mean_p": 0.497477, "boundary_entropy": 0.688246, "bpb": 6.375336}, "vi": {"chunks": 2733.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.148251, "cps": 21.351562, "mean_p": 0.481671, "boundary_entropy": 0.688017, "bpb": 6.958707}, "tr": {"chunks": 3604.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.199083, "cps": 28.15625, "mean_p": 0.487633, "boundary_entropy": 0.687528, "bpb": 6.664219}, "fi": {"chunks": 1555.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.082664, "cps": 12.148438, "mean_p": 0.475629, "boundary_entropy": 0.686362, "bpb": 6.665813}, "ru": {"chunks": 11721.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.618522, "cps": 91.570312, "mean_p": 0.4977, "boundary_entropy": 0.690281, "bpb": 5.627702}, "el": {"chunks": 1380.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.067242, "cps": 10.78125, "mean_p": 0.479186, "boundary_entropy": 0.689427, "bpb": 6.80534}, "hi": {"chunks": 7750.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.460624, "cps": 60.546875, "mean_p": 0.493603, "boundary_entropy": 0.690808, "bpb": 5.972686}, "ta": {"chunks": 434.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.023215, "cps": 3.390625, "mean_p": 0.480497, "boundary_entropy": 0.690422, "bpb": 6.39714}, "th": {"chunks": 4866.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.302518, "cps": 38.015625, "mean_p": 0.492143, "boundary_entropy": 0.69092, "bpb": 5.49088}, "zh": {"chunks": 12212.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.067728, "cps": 95.40625, "mean_p": 0.517166, "boundary_entropy": 0.686551, "bpb": 7.475833}, "my": {"chunks": 659.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.034779, "cps": 5.148438, "mean_p": 0.473333, "boundary_entropy": 0.68981, "bpb": 7.371093}}, "bpb": {"en": 6.28458, "de": 6.37534, "vi": 6.95871, "tr": 6.66422, "fi": 6.66581, "ru": 5.6277, "el": 6.80534, "hi": 5.97269, "ta": 6.39714, "th": 5.49088, "zh": 7.47583, "my": 7.37109}, "gini_cpc": 0.57886, "gini_cps": 0.435889, "bpic": 6.414953, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.2126270141687874, "vi": 0.4856937820443695, "tr": 0.6522292700604106, "fi": 0.27082238835198325, "ru": 2.0263828698699813, "el": 0.22029482496570113, "hi": 1.5090814595678403, "ta": 0.07605544537067768, "th": 0.9910991260659674, "zh": 6.774221717550867, "my": 0.113943116115507}, "t": 53.4}
{"event": "probe", "step": 64, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4508.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.259036, "cps": 35.21875, "mean_p": 0.495509, "boundary_entropy": 0.687561, "bpb": 5.795383}, "de": {"chunks": 6724.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.327473, "cps": 52.53125, "mean_p": 0.49747, "boundary_entropy": 0.688439, "bpb": 5.915382}, "vi": {"chunks": 2617.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.141958, "cps": 20.445312, "mean_p": 0.484832, "boundary_entropy": 0.688431, "bpb": 6.445748}, "tr": {"chunks": 3976.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.219632, "cps": 31.0625, "mean_p": 0.491598, "boundary_entropy": 0.687944, "bpb": 6.241339}, "fi": {"chunks": 2655.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.141141, "cps": 20.742188, "mean_p": 0.488975, "boundary_entropy": 0.687795, "bpb": 5.9663}, "ru": {"chunks": 7384.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.389657, "cps": 57.6875, "mean_p": 0.494709, "boundary_entropy": 0.690293, "bpb": 5.096043}, "el": {"chunks": 14510.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.707012, "cps": 113.359375, "mean_p": 0.49824, "boundary_entropy": 0.690464, "bpb": 5.331456}, "hi": {"chunks": 11172.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.664012, "cps": 87.28125, "mean_p": 0.49611, "boundary_entropy": 0.690883, "bpb": 5.069969}, "ta": {"chunks": 11419.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.610805, "cps": 89.210938, "mean_p": 0.495293, "boundary_entropy": 0.691223, "bpb": 4.851253}, "th": {"chunks": 6879.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.427666, "cps": 53.742188, "mean_p": 0.494655, "boundary_entropy": 0.69102, "bpb": 4.741185}, "zh": {"chunks": 3678.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.622757, "cps": 28.734375, "mean_p": 0.491356, "boundary_entropy": 0.686742, "bpb": 6.978806}, "my": {"chunks": 9574.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.505278, "cps": 74.796875, "mean_p": 0.495244, "boundary_entropy": 0.691313, "bpb": 4.716072}}, "bpb": {"en": 5.79538, "de": 5.91538, "vi": 6.44575, "tr": 6.24134, "fi": 5.9663, "ru": 5.09604, "el": 5.33146, "hi": 5.06997, "ta": 4.85125, "th": 4.74118, "zh": 6.97881, "my": 4.71607}, "gini_cpc": 0.267043, "gini_cps": 0.292738, "bpic": 4.509977, "worst_lang": "el", "cpc_parity_vs_en": {"en": 1.0, "de": 1.26419919785455, "vi": 0.5480255334878849, "tr": 0.8478832134819974, "fi": 0.5448699473872567, "ru": 1.504259235413794, "el": 2.729397441466434, "hi": 2.5633981523354223, "ta": 2.3579946547731967, "th": 1.6509900542233626, "zh": 2.404133029409883, "my": 1.9506091626286224}, "t": 70.2}
{"event": "probe", "step": 128, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2688.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.154456, "cps": 21.0, "mean_p": 0.487062, "boundary_entropy": 0.686997, "bpb": 5.346095}, "de": {"chunks": 3841.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.187065, "cps": 30.007812, "mean_p": 0.487767, "boundary_entropy": 0.687801, "bpb": 5.49151}, "vi": {"chunks": 3708.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.201139, "cps": 28.96875, "mean_p": 0.487387, "boundary_entropy": 0.68854, "bpb": 5.861594}, "tr": {"chunks": 2771.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.153069, "cps": 21.648438, "mean_p": 0.484472, "boundary_entropy": 0.687359, "bpb": 5.766382}, "fi": {"chunks": 2233.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.118707, "cps": 17.445312, "mean_p": 0.483372, "boundary_entropy": 0.687268, "bpb": 5.45939}, "ru": {"chunks": 4377.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.230976, "cps": 34.195312, "mean_p": 0.49162, "boundary_entropy": 0.690204, "bpb": 4.674302}, "el": {"chunks": 9838.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.479365, "cps": 76.859375, "mean_p": 0.495204, "boundary_entropy": 0.690459, "bpb": 4.997911}, "hi": {"chunks": 9445.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.561367, "cps": 73.789062, "mean_p": 0.492603, "boundary_entropy": 0.690712, "bpb": 4.362636}, "ta": {"chunks": 7667.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.41011, "cps": 59.898438, "mean_p": 0.494001, "boundary_entropy": 0.691183, "bpb": 4.299411}, "th": {"chunks": 7098.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.441281, "cps": 55.453125, "mean_p": 0.487657, "boundary_entropy": 0.690459, "bpb": 4.133718}, "zh": {"chunks": 5790.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.980359, "cps": 45.234375, "mean_p": 0.499601, "boundary_entropy": 0.68725, "bpb": 6.654368}, "my": {"chunks": 15228.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.803673, "cps": 118.96875, "mean_p": 0.496972, "boundary_entropy": 0.691328, "bpb": 4.171322}}, "bpb": {"en": 5.3461, "de": 5.49151, "vi": 5.86159, "tr": 5.76638, "fi": 5.45939, "ru": 4.6743, "el": 4.99791, "hi": 4.36264, "ta": 4.29941, "th": 4.13372, "zh": 6.65437, "my": 4.17132}, "gini_cpc": 0.362575, "gini_cps": 0.317998, "bpic": 5.138731, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.211118828314699, "vi": 1.3022412239916308, "tr": 0.9910163740619156, "fi": 0.7685492364839721, "ru": 1.4954165685073502, "el": 3.103564881039391, "hi": 3.634475496179155, "ta": 2.6551853890459634, "th": 2.856997008082064, "zh": 6.347167756155967, "my": 5.203245866506831}, "t": 95.6}
{"event": "probe", "step": 256, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3094.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.177785, "cps": 24.171875, "mean_p": 0.399395, "boundary_entropy": 0.643928, "bpb": 4.243834}, "de": {"chunks": 4842.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.235816, "cps": 37.828125, "mean_p": 0.411667, "boundary_entropy": 0.649758, "bpb": 4.443774}, "vi": {"chunks": 5108.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.277082, "cps": 39.90625, "mean_p": 0.43067, "boundary_entropy": 0.665619, "bpb": 4.235251}, "tr": {"chunks": 3683.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.203447, "cps": 28.773438, "mean_p": 0.426669, "boundary_entropy": 0.661716, "bpb": 4.538577}, "fi": {"chunks": 3804.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.202222, "cps": 29.71875, "mean_p": 0.426498, "boundary_entropy": 0.66244, "bpb": 4.430324}, "ru": {"chunks": 5605.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.295778, "cps": 43.789062, "mean_p": 0.472936, "boundary_entropy": 0.684995, "bpb": 3.128142}, "el": {"chunks": 7619.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.371242, "cps": 59.523438, "mean_p": 0.468623, "boundary_entropy": 0.68511, "bpb": 3.338928}, "hi": {"chunks": 8036.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.477623, "cps": 62.78125, "mean_p": 0.484702, "boundary_entropy": 0.688837, "bpb": 2.698886}, "ta": {"chunks": 9376.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.501524, "cps": 73.25, "mean_p": 0.491885, "boundary_entropy": 0.690805, "bpb": 2.635399}, "th": {"chunks": 8003.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.497544, "cps": 62.523438, "mean_p": 0.477056, "boundary_entropy": 0.688403, "bpb": 2.575749}, "zh": {"chunks": 4066.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.688452, "cps": 31.765625, "mean_p": 0.493599, "boundary_entropy": 0.686933, "bpb": 5.968163}, "my": {"chunks": 11685.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.616688, "cps": 91.289062, "mean_p": 0.481398, "boundary_entropy": 0.69, "bpb": 2.633376}}, "bpb": {"en": 4.24383, "de": 4.44377, "vi": 4.23525, "tr": 4.53858, "fi": 4.43032, "ru": 3.12814, "el": 3.33893, "hi": 2.69889, "ta": 2.6354, "th": 2.57575, "zh": 5.96816, "my": 2.63338}, "gini_cpc": 0.246043, "gini_cps": 0.226767, "bpic": 5.122476, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.3264051174531006, "vi": 1.558517048930106, "tr": 1.1443397350526816, "fi": 1.1374503159159821, "ru": 1.66368160010097, "el": 2.088146378158158, "hi": 2.6865112651529923, "ta": 2.8209535819182943, "th": 2.7985660575413545, "zh": 3.872377985553022, "my": 3.4687192580544663}, "t": 133.5}
{"event": "probe", "step": 500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3610.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.207435, "cps": 28.203125, "mean_p": 0.328005, "boundary_entropy": 0.541571, "bpb": 3.372283}, "de": {"chunks": 5168.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.251692, "cps": 40.375, "mean_p": 0.346538, "boundary_entropy": 0.549308, "bpb": 3.62611}, "vi": {"chunks": 5300.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.287497, "cps": 41.40625, "mean_p": 0.38123, "boundary_entropy": 0.593958, "bpb": 2.972794}, "tr": {"chunks": 5655.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.312379, "cps": 44.179688, "mean_p": 0.356401, "boundary_entropy": 0.561769, "bpb": 3.599329}, "fi": {"chunks": 4374.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.232524, "cps": 34.171875, "mean_p": 0.353635, "boundary_entropy": 0.586276, "bpb": 3.713427}, "ru": {"chunks": 6249.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.329763, "cps": 48.820312, "mean_p": 0.313531, "boundary_entropy": 0.559628, "bpb": 2.589485}, "el": {"chunks": 8874.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.432393, "cps": 69.328125, "mean_p": 0.369887, "boundary_entropy": 0.60205, "bpb": 2.660681}, "hi": {"chunks": 8950.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.531947, "cps": 69.921875, "mean_p": 0.420862, "boundary_entropy": 0.639441, "bpb": 2.050839}, "ta": {"chunks": 12336.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.659856, "cps": 96.375, "mean_p": 0.477461, "boundary_entropy": 0.681155, "bpb": 1.852142}, "th": {"chunks": 12474.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.775505, "cps": 97.453125, "mean_p": 0.344248, "boundary_entropy": 0.599756, "bpb": 2.00725}, "zh": {"chunks": 8690.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.471385, "cps": 67.890625, "mean_p": 0.525222, "boundary_entropy": 0.607205, "bpb": 4.731925}, "my": {"chunks": 18150.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.957885, "cps": 141.796875, "mean_p": 0.384294, "boundary_entropy": 0.606195, "bpb": 1.999621}}, "bpb": {"en": 3.37228, "de": 3.62611, "vi": 2.97279, "tr": 3.59933, "fi": 3.71343, "ru": 2.58949, "el": 2.66068, "hi": 2.05084, "ta": 1.85214, "th": 2.00725, "zh": 4.73192, "my": 1.99962}, "gini_cpc": 0.345225, "gini_cps": 0.261581, "bpic": 3.844345, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.2133525749307277, "vi": 1.3859566478613563, "tr": 1.5059098574592324, "fi": 1.120943733782217, "ru": 1.5897111804647015, "el": 2.084469272566702, "hi": 2.5643947594802285, "ta": 3.1810156773812603, "th": 3.7385362216135367, "zh": 7.093217095530814, "my": 4.617747390876445}, "t": 196.5}
{"event": "train", "step": 500, "phase": 1, "objective": "global", "lm_loss": 2.17966, "bpb": 3.14458, "chunk_loss": 0.035818, "F": 0.2654, "G": 0.38269, "chunks": 271.77, "bytes_per_chunk": 3.768, "boundary_entropy": 0.58448, "router_noise": 0.0, "grad_norm": 2.3009, "lr": 7.51e-05, "router_trainable": true, "sec_per_step": 0.376, "alpha": 0.03, "t": 197.2}
{"event": "probe", "step": 1000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3075.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.176694, "cps": 24.023438, "mean_p": 0.225119, "boundary_entropy": 0.271395, "bpb": 2.663089}, "de": {"chunks": 4170.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.203088, "cps": 32.578125, "mean_p": 0.231127, "boundary_entropy": 0.27515, "bpb": 2.981327}, "vi": {"chunks": 4656.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.252563, "cps": 36.375, "mean_p": 0.239061, "boundary_entropy": 0.327108, "bpb": 2.356721}, "tr": {"chunks": 3307.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.182677, "cps": 25.835938, "mean_p": 0.221637, "boundary_entropy": 0.309526, "bpb": 3.009348}, "fi": {"chunks": 3669.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.195045, "cps": 28.664062, "mean_p": 0.237354, "boundary_entropy": 0.312604, "bpb": 3.235394}, "ru": {"chunks": 6803.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.358997, "cps": 53.148438, "mean_p": 0.297116, "boundary_entropy": 0.438905, "bpb": 1.901474}, "el": {"chunks": 6554.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.319349, "cps": 51.203125, "mean_p": 0.277416, "boundary_entropy": 0.397838, "bpb": 2.033423}, "hi": {"chunks": 7021.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.417296, "cps": 54.851562, "mean_p": 0.240272, "boundary_entropy": 0.3826, "bpb": 1.565059}, "ta": {"chunks": 7788.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.416582, "cps": 60.84375, "mean_p": 0.243754, "boundary_entropy": 0.408598, "bpb": 1.344913}, "th": {"chunks": 6358.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.395275, "cps": 49.671875, "mean_p": 0.221681, "boundary_entropy": 0.391351, "bpb": 1.71883}, "zh": {"chunks": 2363.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.400102, "cps": 18.460938, "mean_p": 0.305585, "boundary_entropy": 0.536487, "bpb": 3.751827}, "my": {"chunks": 6456.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.340722, "cps": 50.4375, "mean_p": 0.272321, "boundary_entropy": 0.486923, "bpb": 1.624014}}, "bpb": {"en": 2.66309, "de": 2.98133, "vi": 2.35672, "tr": 3.00935, "fi": 3.23539, "ru": 1.90147, "el": 2.03342, "hi": 1.56506, "ta": 1.34491, "th": 1.71883, "zh": 3.75183, "my": 1.62401}, "gini_cpc": 0.170584, "gini_cps": 0.191051, "bpic": 6.168129, "worst_lang": "hi", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1493773853630027, "vi": 1.4293837146996369, "tr": 1.0338621681085063, "fi": 1.1038621149275658, "ru": 2.031749945298925, "el": 1.8073596904365477, "hi": 2.361690051824738, "ta": 2.357650760931252, "th": 2.237064254683579, "zh": 2.2643798369578683, "my": 1.9283201779452877}, "t": 310.8}
{"event": "train", "step": 1000, "phase": 1, "objective": "global", "lm_loss": 1.78026, "bpb": 2.56837, "chunk_loss": 0.030355, "F": 0.18291, "G": 0.25608, "chunks": 187.3, "bytes_per_chunk": 5.467, "boundary_entropy": 0.35824, "router_noise": 0.0, "grad_norm": 1.5743, "lr": 0.0001501, "router_trainable": true, "sec_per_step": 0.228, "alpha": 0.03, "t": 311.3}
{"event": "probe", "step": 1500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3005.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.172671, "cps": 23.476562, "mean_p": 0.201026, "boundary_entropy": 0.177733, "bpb": 2.365838}, "de": {"chunks": 3899.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.189889, "cps": 30.460938, "mean_p": 0.21169, "boundary_entropy": 0.182006, "bpb": 2.657337}, "vi": {"chunks": 4422.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.23987, "cps": 34.546875, "mean_p": 0.211965, "boundary_entropy": 0.214906, "bpb": 2.124649}, "tr": {"chunks": 3607.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.199249, "cps": 28.179688, "mean_p": 0.209372, "boundary_entropy": 0.188088, "bpb": 2.645969}, "fi": {"chunks": 3572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.189889, "cps": 27.90625, "mean_p": 0.211879, "boundary_entropy": 0.202543, "bpb": 2.871671}, "ru": {"chunks": 6069.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.320264, "cps": 47.414062, "mean_p": 0.225388, "boundary_entropy": 0.271262, "bpb": 1.658772}, "el": {"chunks": 7235.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.352531, "cps": 56.523438, "mean_p": 0.252938, "boundary_entropy": 0.311415, "bpb": 1.753464}, "hi": {"chunks": 7101.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.422051, "cps": 55.476562, "mean_p": 0.221187, "boundary_entropy": 0.340412, "bpb": 1.392611}, "ta": {"chunks": 8399.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.449265, "cps": 65.617188, "mean_p": 0.223111, "boundary_entropy": 0.367001, "bpb": 1.141358}, "th": {"chunks": 8223.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.511222, "cps": 64.242188, "mean_p": 0.204521, "boundary_entropy": 0.343195, "bpb": 1.408204}, "zh": {"chunks": 2648.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.448358, "cps": 20.6875, "mean_p": 0.299416, "boundary_entropy": 0.473215, "bpb": 3.521554}, "my": {"chunks": 9499.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.501319, "cps": 74.210938, "mean_p": 0.27714, "boundary_entropy": 0.473664, "bpb": 1.297227}}, "bpb": {"en": 2.36584, "de": 2.65734, "vi": 2.12465, "tr": 2.64597, "fi": 2.87167, "ru": 1.65877, "el": 1.75346, "hi": 1.39261, "ta": 1.14136, "th": 1.4082, "zh": 3.52155, "my": 1.29723}, "gini_cpc": 0.211992, "gini_cps": 0.228846, "bpic": 5.670607, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.099715818041539, "vi": 1.3891695018744186, "tr": 1.1539187619595503, "fi": 1.0997126243353734, "ru": 1.8547593412913392, "el": 2.041631389092048, "hi": 2.444241331312597, "ta": 2.601847006348962, "th": 2.96066226789151, "zh": 2.5965947937543983, "my": 2.9033149837878387}, "t": 411.1}
{"event": "train", "step": 1500, "phase": 1, "objective": "global", "lm_loss": 1.65277, "bpb": 2.38445, "chunk_loss": 0.030121, "F": 0.17738, "G": 0.21909, "chunks": 181.64, "bytes_per_chunk": 5.638, "boundary_entropy": 0.2407, "router_noise": 0.0, "grad_norm": 1.0027, "lr": 0.0002251, "router_trainable": true, "sec_per_step": 0.201, "alpha": 0.03, "t": 411.6}
{"event": "probe", "step": 2000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3061.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.175889, "cps": 23.914062, "mean_p": 0.194749, "boundary_entropy": 0.146815, "bpb": 2.235832}, "de": {"chunks": 3926.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.191204, "cps": 30.671875, "mean_p": 0.204637, "boundary_entropy": 0.144092, "bpb": 2.487173}, "vi": {"chunks": 4335.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.235151, "cps": 33.867188, "mean_p": 0.203449, "boundary_entropy": 0.186335, "bpb": 2.04474}, "tr": {"chunks": 3598.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.198752, "cps": 28.109375, "mean_p": 0.202852, "boundary_entropy": 0.154752, "bpb": 2.460428}, "fi": {"chunks": 3399.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180692, "cps": 26.554688, "mean_p": 0.19632, "boundary_entropy": 0.158043, "bpb": 2.691931}, "ru": {"chunks": 5882.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.310396, "cps": 45.953125, "mean_p": 0.199492, "boundary_entropy": 0.216861, "bpb": 1.539626}, "el": {"chunks": 6448.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.314184, "cps": 50.375, "mean_p": 0.210689, "boundary_entropy": 0.252172, "bpb": 1.60687}, "hi": {"chunks": 5268.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.313105, "cps": 41.15625, "mean_p": 0.195335, "boundary_entropy": 0.331736, "bpb": 1.327621}, "ta": {"chunks": 9571.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.511955, "cps": 74.773438, "mean_p": 0.240112, "boundary_entropy": 0.370796, "bpb": 1.084718}, "th": {"chunks": 8389.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.521542, "cps": 65.539062, "mean_p": 0.197973, "boundary_entropy": 0.31776, "bpb": 1.244545}, "zh": {"chunks": 2802.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.474433, "cps": 21.890625, "mean_p": 0.254443, "boundary_entropy": 0.37511, "bpb": 3.405926}, "my": {"chunks": 7356.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.38822, "cps": 57.46875, "mean_p": 0.265765, "boundary_entropy": 0.468542, "bpb": 1.255572}}, "bpb": {"en": 2.23583, "de": 2.48717, "vi": 2.04474, "tr": 2.46043, "fi": 2.69193, "ru": 1.53963, "el": 1.60687, "hi": 1.32762, "ta": 1.08472, "th": 1.24454, "zh": 3.40593, "my": 1.25557}, "gini_cpc": 0.217386, "gini_cps": 0.222744, "bpic": 5.993301, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0870729237655743, "vi": 1.3369240948024823, "tr": 1.1299816688356112, "fi": 1.0273065845536413, "ru": 1.7647232051194748, "el": 1.7862612385562173, "hi": 1.7801290355151056, "ta": 2.91066777258593, "th": 2.965172200080083, "zh": 2.6973386717509302, "my": 2.2071870282098076}, "t": 506.6}
{"event": "train", "step": 2000, "phase": 1, "objective": "global", "lm_loss": 1.53056, "bpb": 2.20813, "chunk_loss": 0.0301, "F": 0.17538, "G": 0.207, "chunks": 179.59, "bytes_per_chunk": 5.702, "boundary_entropy": 0.21162, "router_noise": 0.0, "grad_norm": 0.7798, "lr": 0.0003, "router_trainable": true, "sec_per_step": 0.191, "alpha": 0.03, "t": 507.2}
{"event": "probe", "step": 2500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3111.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.178762, "cps": 24.304688, "mean_p": 0.197753, "boundary_entropy": 0.140291, "bpb": 2.130459}, "de": {"chunks": 3845.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.18726, "cps": 30.039062, "mean_p": 0.201961, "boundary_entropy": 0.134637, "bpb": 2.35572}, "vi": {"chunks": 4405.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.238948, "cps": 34.414062, "mean_p": 0.202021, "boundary_entropy": 0.162619, "bpb": 1.942534}, "tr": {"chunks": 3493.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.192951, "cps": 27.289062, "mean_p": 0.193699, "boundary_entropy": 0.132475, "bpb": 2.311433}, "fi": {"chunks": 3448.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.183297, "cps": 26.9375, "mean_p": 0.197106, "boundary_entropy": 0.135773, "bpb": 2.50747}, "ru": {"chunks": 6041.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.318786, "cps": 47.195312, "mean_p": 0.201247, "boundary_entropy": 0.199596, "bpb": 1.426062}, "el": {"chunks": 6834.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.332992, "cps": 53.390625, "mean_p": 0.205961, "boundary_entropy": 0.224276, "bpb": 1.491936}, "hi": {"chunks": 6828.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.405825, "cps": 53.34375, "mean_p": 0.197527, "boundary_entropy": 0.28956, "bpb": 1.22654}, "ta": {"chunks": 9880.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.528484, "cps": 77.1875, "mean_p": 0.225639, "boundary_entropy": 0.317724, "bpb": 1.029403}, "th": {"chunks": 10020.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.622941, "cps": 78.28125, "mean_p": 0.211764, "boundary_entropy": 0.29376, "bpb": 1.16606}, "zh": {"chunks": 2753.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.466136, "cps": 21.507812, "mean_p": 0.245462, "boundary_entropy": 0.349941, "bpb": 3.252117}, "my": {"chunks": 11613.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.612888, "cps": 90.726562, "mean_p": 0.379471, "boundary_entropy": 0.573188, "bpb": 1.162191}}, "bpb": {"en": 2.13046, "de": 2.35572, "vi": 1.94253, "tr": 2.31143, "fi": 2.50747, "ru": 1.42606, "el": 1.49194, "hi": 1.22654, "ta": 1.0294, "th": 1.16606, "zh": 3.25212, "my": 1.16219}, "gini_cpc": 0.254173, "gini_cps": 0.26746, "bpic": 5.310304, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0475338027488361, "vi": 1.3366782453087145, "tr": 1.0793744740731688, "fi": 1.0253673462053918, "ru": 1.7832972116135697, "el": 1.8627657254595604, "hi": 2.270191789379869, "ta": 2.9563482003010826, "th": 3.4847430882645205, "zh": 2.6075754156814197, "my": 3.4285079359456905}, "t": 601.7}
{"event": "train", "step": 2500, "phase": 1, "objective": "global", "lm_loss": 1.30453, "bpb": 1.88205, "chunk_loss": 0.030245, "F": 0.1852, "G": 0.21302, "chunks": 189.64, "bytes_per_chunk": 5.4, "boundary_entropy": 0.21734, "router_noise": 0.0, "grad_norm": 0.5122, "lr": 0.0003, "router_trainable": true, "sec_per_step": 0.19, "alpha": 0.03, "t": 602.2}
{"event": "probe", "step": 3000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3101.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.178188, "cps": 24.226562, "mean_p": 0.191873, "boundary_entropy": 0.125778, "bpb": 2.044355}, "de": {"chunks": 3907.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.190279, "cps": 30.523438, "mean_p": 0.201624, "boundary_entropy": 0.122776, "bpb": 2.252975}, "vi": {"chunks": 4341.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.235476, "cps": 33.914062, "mean_p": 0.1984, "boundary_entropy": 0.149229, "bpb": 1.875975}, "tr": {"chunks": 3517.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.194277, "cps": 27.476562, "mean_p": 0.195293, "boundary_entropy": 0.124798, "bpb": 2.20168}, "fi": {"chunks": 3451.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.183456, "cps": 26.960938, "mean_p": 0.193541, "boundary_entropy": 0.123658, "bpb": 2.423436}, "ru": {"chunks": 5848.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.308602, "cps": 45.6875, "mean_p": 0.190396, "boundary_entropy": 0.178617, "bpb": 1.359797}, "el": {"chunks": 6745.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.328656, "cps": 52.695312, "mean_p": 0.198669, "boundary_entropy": 0.192596, "bpb": 1.406779}, "hi": {"chunks": 6841.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.406597, "cps": 53.445312, "mean_p": 0.195975, "boundary_entropy": 0.272163, "bpb": 1.159234}, "ta": {"chunks": 7728.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.413373, "cps": 60.375, "mean_p": 0.196797, "boundary_entropy": 0.287619, "bpb": 1.000326}, "th": {"chunks": 7102.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.441529, "cps": 55.484375, "mean_p": 0.171602, "boundary_entropy": 0.266443, "bpb": 1.11625}, "zh": {"chunks": 2801.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.474263, "cps": 21.882812, "mean_p": 0.245519, "boundary_entropy": 0.324438, "bpb": 3.120334}, "my": {"chunks": 8296.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.43783, "cps": 64.8125, "mean_p": 0.368156, "boundary_entropy": 0.578428, "bpb": 1.117835}}, "bpb": {"en": 2.04435, "de": 2.25298, "vi": 1.87598, "tr": 2.20168, "fi": 2.42344, "ru": 1.3598, "el": 1.40678, "hi": 1.15923, "ta": 1.00033, "th": 1.11625, "zh": 3.12033, "my": 1.11783}, "gini_cpc": 0.196466, "gini_cps": 0.201909, "bpic": 6.026901, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0678576371809936, "vi": 1.3215055696094253, "tr": 1.0902953775756246, "fi": 1.0295689340790828, "ru": 1.7318907935902879, "el": 1.8444354587391376, "hi": 2.281848840401864, "ta": 2.319871864919641, "th": 2.4778896215756405, "zh": 2.66159529500823, "my": 2.457127662550988}, "t": 697.3}
{"event": "train", "step": 3000, "phase": 1, "objective": "global", "lm_loss": 1.3345, "bpb": 1.92527, "chunk_loss": 0.030116, "F": 0.17923, "G": 0.20299, "chunks": 183.53, "bytes_per_chunk": 5.579, "boundary_entropy": 0.16717, "router_noise": 0.0, "grad_norm": 0.431, "lr": 0.0002999, "router_trainable": true, "sec_per_step": 0.191, "alpha": 0.03, "t": 697.8}
{"event": "probe", "step": 3500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3051.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.175315, "cps": 23.835938, "mean_p": 0.18573, "boundary_entropy": 0.105668, "bpb": 1.960246}, "de": {"chunks": 3764.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.183315, "cps": 29.40625, "mean_p": 0.192165, "boundary_entropy": 0.10544, "bpb": 2.146216}, "vi": {"chunks": 4184.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.22696, "cps": 32.6875, "mean_p": 0.187453, "boundary_entropy": 0.133235, "bpb": 1.790104}, "tr": {"chunks": 3386.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.187041, "cps": 26.453125, "mean_p": 0.18437, "boundary_entropy": 0.106908, "bpb": 2.080607}, "fi": {"chunks": 3452.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.18351, "cps": 26.96875, "mean_p": 0.192347, "boundary_entropy": 0.109989, "bpb": 2.311153}, "ru": {"chunks": 5842.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.308285, "cps": 45.640625, "mean_p": 0.185817, "boundary_entropy": 0.16442, "bpb": 1.313726}, "el": {"chunks": 6659.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.324465, "cps": 52.023438, "mean_p": 0.197606, "boundary_entropy": 0.176242, "bpb": 1.345421}, "hi": {"chunks": 7817.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.464606, "cps": 61.070312, "mean_p": 0.19397, "boundary_entropy": 0.239998, "bpb": 1.11563}, "ta": {"chunks": 8237.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.440599, "cps": 64.351562, "mean_p": 0.189971, "boundary_entropy": 0.242022, "bpb": 0.960237}, "th": {"chunks": 8996.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.559279, "cps": 70.28125, "mean_p": 0.187026, "boundary_entropy": 0.253993, "bpb": 1.063406}, "zh": {"chunks": 2920.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.494412, "cps": 22.8125, "mean_p": 0.24614, "boundary_entropy": 0.307161, "bpb": 3.030223}, "my": {"chunks": 7848.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.414186, "cps": 61.3125, "mean_p": 0.308315, "boundary_entropy": 0.518768, "bpb": 1.081327}}, "bpb": {"en": 1.96025, "de": 2.14622, "vi": 1.7901, "tr": 2.08061, "fi": 2.31115, "ru": 1.31373, "el": 1.34542, "hi": 1.11563, "ta": 0.96024, "th": 1.06341, "zh": 3.03022, "my": 1.08133}, "gini_cpc": 0.227934, "gini_cps": 0.222847, "bpic": 5.801152, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0456326126046658, "vi": 1.2945846297776422, "tr": 1.0668867337350878, "fi": 1.0467448098188539, "ru": 1.7584671144630983, "el": 1.850759904792527, "hi": 2.6501286159069957, "ta": 2.513191076650144, "th": 3.1901440509779504, "zh": 2.820144239429862, "my": 2.3625310818056313}, "t": 793.1}
{"event": "train", "step": 3500, "phase": 1, "objective": "global", "lm_loss": 1.22268, "bpb": 1.76396, "chunk_loss": 0.0301, "F": 0.17946, "G": 0.19726, "chunks": 183.77, "bytes_per_chunk": 5.572, "boundary_entropy": 0.15995, "router_noise": 0.0, "grad_norm": 0.4041, "lr": 0.0002997, "router_trainable": true, "sec_per_step": 0.192, "alpha": 0.03, "t": 793.6}
{"event": "probe", "step": 4000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3037.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.17451, "cps": 23.726562, "mean_p": 0.182799, "boundary_entropy": 0.095482, "bpb": 1.887711}, "de": {"chunks": 3791.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.18463, "cps": 29.617188, "mean_p": 0.192096, "boundary_entropy": 0.096162, "bpb": 2.058179}, "vi": {"chunks": 4148.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.225007, "cps": 32.40625, "mean_p": 0.18404, "boundary_entropy": 0.131204, "bpb": 1.727783}, "tr": {"chunks": 3309.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.182787, "cps": 25.851562, "mean_p": 0.179068, "boundary_entropy": 0.097836, "bpb": 1.977639}, "fi": {"chunks": 3446.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.183191, "cps": 26.921875, "mean_p": 0.190576, "boundary_entropy": 0.10317, "bpb": 2.203996}, "ru": {"chunks": 5808.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.306491, "cps": 45.375, "mean_p": 0.186135, "boundary_entropy": 0.149239, "bpb": 1.281202}, "el": {"chunks": 6669.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.324952, "cps": 52.101562, "mean_p": 0.194049, "boundary_entropy": 0.164246, "bpb": 1.308467}, "hi": {"chunks": 6949.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.413016, "cps": 54.289062, "mean_p": 0.181524, "boundary_entropy": 0.240832, "bpb": 1.09193}, "ta": {"chunks": 7783.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.416315, "cps": 60.804688, "mean_p": 0.178483, "boundary_entropy": 0.228588, "bpb": 0.971898}, "th": {"chunks": 6970.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.433323, "cps": 54.453125, "mean_p": 0.165542, "boundary_entropy": 0.255015, "bpb": 1.067579}, "zh": {"chunks": 2681.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.453945, "cps": 20.945312, "mean_p": 0.217833, "boundary_entropy": 0.258644, "bpb": 2.951983}, "my": {"chunks": 8680.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.458096, "cps": 67.8125, "mean_p": 0.351941, "boundary_entropy": 0.565997, "bpb": 1.054425}}, "bpb": {"en": 1.88771, "de": 2.05818, "vi": 1.72778, "tr": 1.97764, "fi": 2.204, "ru": 1.2812, "el": 1.30847, "hi": 1.09193, "ta": 0.9719, "th": 1.06758, "zh": 2.95198, "my": 1.05442}, "gini_cpc": 0.200814, "gini_cps": 0.213551, "bpic": 6.06567, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0579879115865765, "vi": 1.2893622003427754, "tr": 1.047431291302389, "fi": 1.0497423489261768, "ru": 1.7562919941651904, "el": 1.8620837089047968, "hi": 2.3667182901715886, "ta": 2.3856179245312097, "th": 2.483081882356746, "zh": 2.6012536321144086, "my": 2.6250385002532624}, "t": 887.8}
{"event": "eval", "step": 4000, "phase": 1, "bpb": {"en": 1.89061, "de": 2.06054, "vi": 1.74027, "tr": 1.97089, "fi": 2.21354, "ru": 1.27885, "el": 1.30631, "hi": 1.09903, "ta": 0.97108, "th": 1.06375, "zh": 2.96336, "my": 1.06309}, "bytes_per_chunk": {"en": 5.584, "de": 5.502, "vi": 5.797, "tr": 5.775, "fi": 5.67, "ru": 5.761, "el": 5.829, "hi": 6.015, "ta": 6.414, "th": 6.729, "zh": 6.009, "my": 6.213}, "bpb_mean": 1.63511, "t": 894.6}
{"event": "train", "step": 4000, "phase": 1, "objective": "global", "lm_loss": 1.18168, "bpb": 1.7048, "chunk_loss": 0.030038, "F": 0.1703, "G": 0.18723, "chunks": 174.39, "bytes_per_chunk": 5.872, "boundary_entropy": 0.15072, "router_noise": 0.0, "grad_norm": 0.3819, "lr": 0.0002995, "router_trainable": true, "sec_per_step": 0.202, "alpha": 0.03, "t": 894.8}
{"event": "probe", "step": 4500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3028.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.173993, "cps": 23.65625, "mean_p": 0.181837, "boundary_entropy": 0.083677, "bpb": 1.847302}, "de": {"chunks": 3810.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.185555, "cps": 29.765625, "mean_p": 0.190214, "boundary_entropy": 0.088591, "bpb": 2.008621}, "vi": {"chunks": 4059.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.220179, "cps": 31.710938, "mean_p": 0.177941, "boundary_entropy": 0.122767, "bpb": 1.66476}, "tr": {"chunks": 3340.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.1845, "cps": 26.09375, "mean_p": 0.178318, "boundary_entropy": 0.090973, "bpb": 1.916318}, "fi": {"chunks": 3369.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179097, "cps": 26.320312, "mean_p": 0.182563, "boundary_entropy": 0.085446, "bpb": 2.140943}, "ru": {"chunks": 5735.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.302639, "cps": 44.804688, "mean_p": 0.178024, "boundary_entropy": 0.135637, "bpb": 1.241011}, "el": {"chunks": 6633.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.323198, "cps": 51.820312, "mean_p": 0.189865, "boundary_entropy": 0.15129, "bpb": 1.262344}, "hi": {"chunks": 7250.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.430906, "cps": 56.640625, "mean_p": 0.185792, "boundary_entropy": 0.226112, "bpb": 1.050115}, "ta": {"chunks": 8634.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.461835, "cps": 67.453125, "mean_p": 0.190698, "boundary_entropy": 0.206804, "bpb": 0.897899}, "th": {"chunks": 7022.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.436556, "cps": 54.859375, "mean_p": 0.168557, "boundary_entropy": 0.25039, "bpb": 1.036477}, "zh": {"chunks": 2663.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.450897, "cps": 20.804688, "mean_p": 0.2069, "boundary_entropy": 0.226683, "bpb": 2.849555}, "my": {"chunks": 7846.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.414081, "cps": 61.296875, "mean_p": 0.358654, "boundary_entropy": 0.574135, "bpb": 1.01803}}, "bpb": {"en": 1.8473, "de": 2.00862, "vi": 1.66476, "tr": 1.91632, "fi": 2.14094, "ru": 1.24101, "el": 1.26234, "hi": 1.05012, "ta": 0.8979, "th": 1.03648, "zh": 2.84956, "my": 1.01803}, "gini_cpc": 0.203845, "gini_cps": 0.216682, "bpic": 6.054379, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0664507840939876, "vi": 1.2654475774249128, "tr": 1.060386438354255, "fi": 1.0293365092391273, "ru": 1.7393719305828101, "el": 1.8575366935353839, "hi": 2.476573280388489, "ta": 2.6543294412177967, "th": 2.509042450751847, "zh": 2.591468732288012, "my": 2.3798696857616854}, "t": 987.1}
{"event": "train", "step": 4500, "phase": 1, "objective": "global", "lm_loss": 1.29936, "bpb": 1.87459, "chunk_loss": 0.030085, "F": 0.17506, "G": 0.18789, "chunks": 179.27, "bytes_per_chunk": 5.712, "boundary_entropy": 0.11654, "router_noise": 0.0, "grad_norm": 0.3671, "lr": 0.0002992, "router_trainable": true, "sec_per_step": 0.186, "alpha": 0.03, "t": 987.6}
{"event": "probe", "step": 5000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3052.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.175372, "cps": 23.84375, "mean_p": 0.182944, "boundary_entropy": 0.077811, "bpb": 1.816396}, "de": {"chunks": 3855.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.187747, "cps": 30.117188, "mean_p": 0.192706, "boundary_entropy": 0.080805, "bpb": 1.972503}, "vi": {"chunks": 4112.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.223054, "cps": 32.125, "mean_p": 0.179598, "boundary_entropy": 0.105835, "bpb": 1.638468}, "tr": {"chunks": 3352.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.185163, "cps": 26.1875, "mean_p": 0.181572, "boundary_entropy": 0.087202, "bpb": 1.884092}, "fi": {"chunks": 3413.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.181436, "cps": 26.664062, "mean_p": 0.18575, "boundary_entropy": 0.083274, "bpb": 2.104404}, "ru": {"chunks": 5748.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.303325, "cps": 44.90625, "mean_p": 0.172693, "boundary_entropy": 0.136731, "bpb": 1.211311}, "el": {"chunks": 6635.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.323296, "cps": 51.835938, "mean_p": 0.187152, "boundary_entropy": 0.124965, "bpb": 1.237481}, "hi": {"chunks": 7699.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.457593, "cps": 60.148438, "mean_p": 0.192727, "boundary_entropy": 0.214032, "bpb": 1.029981}, "ta": {"chunks": 7903.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.422733, "cps": 61.742188, "mean_p": 0.171572, "boundary_entropy": 0.19607, "bpb": 0.881103}, "th": {"chunks": 7898.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.491016, "cps": 61.703125, "mean_p": 0.169016, "boundary_entropy": 0.23633, "bpb": 0.99075}, "zh": {"chunks": 2892.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.489672, "cps": 22.59375, "mean_p": 0.226789, "boundary_entropy": 0.238271, "bpb": 2.751209}, "my": {"chunks": 7019.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.370435, "cps": 54.835938, "mean_p": 0.268997, "boundary_entropy": 0.482355, "bpb": 1.011372}}, "bpb": {"en": 1.8164, "de": 1.9725, "vi": 1.63847, "tr": 1.88409, "fi": 2.1044, "ru": 1.21131, "el": 1.23748, "hi": 1.02998, "ta": 0.8811, "th": 0.99075, "zh": 2.75121, "my": 1.01137}, "gini_cpc": 0.214088, "gini_cps": 0.206007, "bpic": 6.036381, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0705613646644574, "vi": 1.2718900063664584, "tr": 1.0558276945763028, "fi": 1.0345798026906678, "ru": 1.7296058123571376, "el": 1.8434852748720565, "hi": 2.609268898284125, "ta": 2.4104942694288565, "th": 2.799855738538425, "zh": 2.792186589464742, "my": 2.1122798555251303}, "t": 1078.5}
{"event": "train", "step": 5000, "phase": 1, "objective": "global", "lm_loss": 1.18569, "bpb": 1.71059, "chunk_loss": 0.030033, "F": 0.17473, "G": 0.1936, "chunks": 178.92, "bytes_per_chunk": 5.723, "boundary_entropy": 0.1425, "router_noise": 0.0, "grad_norm": 0.3501, "lr": 0.0002989, "router_trainable": true, "sec_per_step": 0.183, "alpha": 0.03, "t": 1078.8}
{"event": "probe", "step": 5500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3071.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.176464, "cps": 23.992188, "mean_p": 0.18365, "boundary_entropy": 0.074346, "bpb": 1.792789}, "de": {"chunks": 3813.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.185701, "cps": 29.789062, "mean_p": 0.189679, "boundary_entropy": 0.078296, "bpb": 1.942984}, "vi": {"chunks": 4062.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.220342, "cps": 31.734375, "mean_p": 0.180574, "boundary_entropy": 0.120037, "bpb": 1.596194}, "tr": {"chunks": 3339.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.184445, "cps": 26.085938, "mean_p": 0.179754, "boundary_entropy": 0.08021, "bpb": 1.845244}, "fi": {"chunks": 3407.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.181117, "cps": 26.617188, "mean_p": 0.18588, "boundary_entropy": 0.083039, "bpb": 2.059626}, "ru": {"chunks": 5753.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.303588, "cps": 44.945312, "mean_p": 0.178651, "boundary_entropy": 0.132817, "bpb": 1.185719}, "el": {"chunks": 6619.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.322516, "cps": 51.710938, "mean_p": 0.189427, "boundary_entropy": 0.134447, "bpb": 1.21812}, "hi": {"chunks": 6784.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.40321, "cps": 53.0, "mean_p": 0.178402, "boundary_entropy": 0.21164, "bpb": 1.006047}, "ta": {"chunks": 8852.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.473496, "cps": 69.15625, "mean_p": 0.186869, "boundary_entropy": 0.189526, "bpb": 0.856972}, "th": {"chunks": 7869.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.489214, "cps": 61.476562, "mean_p": 0.180093, "boundary_entropy": 0.259931, "bpb": 0.978506}, "zh": {"chunks": 2655.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.449543, "cps": 20.742188, "mean_p": 0.202347, "boundary_entropy": 0.203996, "bpb": 2.69207}, "my": {"chunks": 9539.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.50343, "cps": 74.523438, "mean_p": 0.391816, "boundary_entropy": 0.597856, "bpb": 0.988125}}, "bpb": {"en": 1.79279, "de": 1.94298, "vi": 1.59619, "tr": 1.84524, "fi": 2.05963, "ru": 1.18572, "el": 1.21812, "hi": 1.00605, "ta": 0.85697, "th": 0.97851, "zh": 2.69207, "my": 0.98813}, "gini_cpc": 0.220041, "gini_cps": 0.236274, "bpic": 5.83582, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0523463567625801, "vi": 1.2486510332226803, "tr": 1.0452259208599268, "fi": 1.026371428441353, "ru": 1.7204001171912924, "el": 1.8276618208513602, "hi": 2.2849414186201455, "ta": 2.6832444485589293, "th": 2.7723163343710815, "zh": 2.547507002961738, "my": 2.8528817871489385}, "t": 1170.2}
{"event": "train", "step": 5500, "phase": 1, "objective": "global", "lm_loss": 1.11786, "bpb": 1.61273, "chunk_loss": 0.030079, "F": 0.17627, "G": 0.18988, "chunks": 180.5, "bytes_per_chunk": 5.673, "boundary_entropy": 0.1301, "router_noise": 0.0, "grad_norm": 0.3246, "lr": 0.0002985, "router_trainable": true, "sec_per_step": 0.184, "alpha": 0.03, "t": 1170.7}
{"event": "probe", "step": 6000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3050.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.175257, "cps": 23.828125, "mean_p": 0.182401, "boundary_entropy": 0.071047, "bpb": 1.774218}, "de": {"chunks": 3776.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.183899, "cps": 29.5, "mean_p": 0.188822, "boundary_entropy": 0.07454, "bpb": 1.921642}, "vi": {"chunks": 4204.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.228044, "cps": 32.84375, "mean_p": 0.18745, "boundary_entropy": 0.111919, "bpb": 1.562777}, "tr": {"chunks": 3454.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.190797, "cps": 26.984375, "mean_p": 0.183858, "boundary_entropy": 0.08239, "bpb": 1.821537}, "fi": {"chunks": 3441.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.182925, "cps": 26.882812, "mean_p": 0.187817, "boundary_entropy": 0.080784, "bpb": 2.048021}, "ru": {"chunks": 5852.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.308813, "cps": 45.71875, "mean_p": 0.182283, "boundary_entropy": 0.129909, "bpb": 1.167245}, "el": {"chunks": 6585.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.32086, "cps": 51.445312, "mean_p": 0.187838, "boundary_entropy": 0.125631, "bpb": 1.187493}, "hi": {"chunks": 7572.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.450045, "cps": 59.15625, "mean_p": 0.192688, "boundary_entropy": 0.201939, "bpb": 0.979151}, "ta": {"chunks": 9389.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.50222, "cps": 73.351562, "mean_p": 0.190725, "boundary_entropy": 0.167881, "bpb": 0.835269}, "th": {"chunks": 7898.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.491016, "cps": 61.703125, "mean_p": 0.173083, "boundary_entropy": 0.235241, "bpb": 0.94955}, "zh": {"chunks": 2749.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.465459, "cps": 21.476562, "mean_p": 0.207389, "boundary_entropy": 0.197264, "bpb": 2.663779}, "my": {"chunks": 8376.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.442052, "cps": 65.4375, "mean_p": 0.311771, "boundary_entropy": 0.522424, "bpb": 0.963411}}, "bpb": {"en": 1.77422, "de": 1.92164, "vi": 1.56278, "tr": 1.82154, "fi": 2.04802, "ru": 1.16725, "el": 1.18749, "hi": 0.97915, "ta": 0.83527, "th": 0.94955, "zh": 2.66378, "my": 0.96341}, "gini_cpc": 0.217577, "gini_cps": 0.229652, "bpic": 5.784539, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0493101149449147, "vi": 1.3011993757419733, "tr": 1.0886695167814773, "fi": 1.0437513976388075, "ru": 1.7620546909468402, "el": 1.8307928809180902, "hi": 2.567910086960758, "ta": 2.8656170362023685, "th": 2.801691709514516, "zh": 2.6558624460815063, "my": 2.5223048412047464}, "t": 1262.4}
{"event": "train", "step": 6000, "phase": 1, "objective": "global", "lm_loss": 1.13361, "bpb": 1.63546, "chunk_loss": 0.030062, "F": 0.17545, "G": 0.19489, "chunks": 179.66, "bytes_per_chunk": 5.7, "boundary_entropy": 0.137, "router_noise": 0.0, "grad_norm": 0.3262, "lr": 0.0002981, "router_trainable": true, "sec_per_step": 0.184, "alpha": 0.03, "t": 1262.9}
{"event": "probe", "step": 6500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3029.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.17405, "cps": 23.664062, "mean_p": 0.17896, "boundary_entropy": 0.063362, "bpb": 1.76114}, "de": {"chunks": 3833.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.186675, "cps": 29.945312, "mean_p": 0.1907, "boundary_entropy": 0.072339, "bpb": 1.902525}, "vi": {"chunks": 4044.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.219365, "cps": 31.59375, "mean_p": 0.176225, "boundary_entropy": 0.10292, "bpb": 1.545565}, "tr": {"chunks": 3303.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.182456, "cps": 25.804688, "mean_p": 0.17555, "boundary_entropy": 0.072612, "bpb": 1.806815}, "fi": {"chunks": 3425.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.182074, "cps": 26.757812, "mean_p": 0.18522, "boundary_entropy": 0.075676, "bpb": 2.012374}, "ru": {"chunks": 5752.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.303536, "cps": 44.9375, "mean_p": 0.174125, "boundary_entropy": 0.109507, "bpb": 1.153936}, "el": {"chunks": 6558.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.319544, "cps": 51.234375, "mean_p": 0.184707, "boundary_entropy": 0.122218, "bpb": 1.177244}, "hi": {"chunks": 7111.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.422645, "cps": 55.554688, "mean_p": 0.181595, "boundary_entropy": 0.207232, "bpb": 0.968332}, "ta": {"chunks": 9266.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.495641, "cps": 72.390625, "mean_p": 0.184116, "boundary_entropy": 0.159555, "bpb": 0.829753}, "th": {"chunks": 8244.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.512527, "cps": 64.40625, "mean_p": 0.175404, "boundary_entropy": 0.236844, "bpb": 0.939426}, "zh": {"chunks": 2782.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.471046, "cps": 21.734375, "mean_p": 0.205994, "boundary_entropy": 0.189904, "bpb": 2.634997}, "my": {"chunks": 8465.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.446749, "cps": 66.132812, "mean_p": 0.367958, "boundary_entropy": 0.576559, "bpb": 0.952102}}, "bpb": {"en": 1.76114, "de": 1.90252, "vi": 1.54556, "tr": 1.80681, "fi": 2.01237, "ru": 1.15394, "el": 1.17724, "hi": 0.96833, "ta": 0.82975, "th": 0.93943, "zh": 2.635, "my": 0.9521}, "gini_cpc": 0.223248, "gini_cps": 0.231908, "bpic": 5.831475, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0725344703950064, "vi": 1.2603548932061943, "tr": 1.0482934425322747, "fi": 1.0461008121323756, "ru": 1.743951929936733, "el": 1.835927022073665, "hi": 2.4282894476720087, "ta": 2.8476832007541257, "th": 2.944704803108855, "zh": 2.706378470138028, "my": 2.5667787386068093}, "t": 1354.2}
{"event": "train", "step": 6500, "phase": 1, "objective": "global", "lm_loss": 1.14498, "bpb": 1.65186, "chunk_loss": 0.030077, "F": 0.17421, "G": 0.18238, "chunks": 178.39, "bytes_per_chunk": 5.74, "boundary_entropy": 0.10895, "router_noise": 0.0, "grad_norm": 0.3228, "lr": 0.0002976, "router_trainable": true, "sec_per_step": 0.184, "alpha": 0.03, "t": 1354.8}
{"event": "probe", "step": 7000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3016.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.173303, "cps": 23.5625, "mean_p": 0.177281, "boundary_entropy": 0.060119, "bpb": 1.748972}, "de": {"chunks": 3780.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.184094, "cps": 29.53125, "mean_p": 0.187047, "boundary_entropy": 0.065385, "bpb": 1.886273}, "vi": {"chunks": 4039.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.219094, "cps": 31.554688, "mean_p": 0.176356, "boundary_entropy": 0.09991, "bpb": 1.526125}, "tr": {"chunks": 3286.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.181517, "cps": 25.671875, "mean_p": 0.173231, "boundary_entropy": 0.067395, "bpb": 1.787531}, "fi": {"chunks": 3420.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.181809, "cps": 26.71875, "mean_p": 0.18341, "boundary_entropy": 0.068275, "bpb": 1.99292}, "ru": {"chunks": 5752.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.303536, "cps": 44.9375, "mean_p": 0.171088, "boundary_entropy": 0.101476, "bpb": 1.13847}, "el": {"chunks": 6434.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.313502, "cps": 50.265625, "mean_p": 0.179747, "boundary_entropy": 0.110249, "bpb": 1.16323}, "hi": {"chunks": 6897.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.409926, "cps": 53.882812, "mean_p": 0.175776, "boundary_entropy": 0.193274, "bpb": 0.953053}, "ta": {"chunks": 8349.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.44659, "cps": 65.226562, "mean_p": 0.173369, "boundary_entropy": 0.168503, "bpb": 0.826178}, "th": {"chunks": 7028.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.436929, "cps": 54.90625, "mean_p": 0.159684, "boundary_entropy": 0.229477, "bpb": 0.939411}, "zh": {"chunks": 2702.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.457501, "cps": 21.109375, "mean_p": 0.197054, "boundary_entropy": 0.166575, "bpb": 2.616301}, "my": {"chunks": 8408.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.443741, "cps": 65.6875, "mean_p": 0.371972, "boundary_entropy": 0.580984, "bpb": 0.948237}}, "bpb": {"en": 1.74897, "de": 1.88627, "vi": 1.52613, "tr": 1.78753, "fi": 1.99292, "ru": 1.13847, "el": 1.16323, "hi": 0.95305, "ta": 0.82618, "th": 0.93941, "zh": 2.6163, "my": 0.94824}, "gini_cpc": 0.204183, "gini_cps": 0.216572, "bpic": 6.081048, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.062263295896455, "vi": 1.264222438289524, "tr": 1.0473932980904583, "fi": 1.0490761303244462, "ru": 1.751468964117495, "el": 1.8089767888030395, "hi": 2.3653637460044696, "ta": 2.576924974620514, "th": 2.5211777781992053, "zh": 2.6398830349211027, "my": 2.560484257067647}, "t": 1451.4}
{"event": "train", "step": 7000, "phase": 1, "objective": "global", "lm_loss": 1.24538, "bpb": 1.7967, "chunk_loss": 0.030079, "F": 0.17516, "G": 0.18671, "chunks": 179.36, "bytes_per_chunk": 5.709, "boundary_entropy": 0.1019, "router_noise": 0.0, "grad_norm": 0.3383, "lr": 0.000297, "router_trainable": true, "sec_per_step": 0.194, "alpha": 0.03, "t": 1451.9}
{"event": "probe", "step": 7500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3018.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.173418, "cps": 23.578125, "mean_p": 0.178502, "boundary_entropy": 0.060115, "bpb": 1.734469}, "de": {"chunks": 3751.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.182682, "cps": 29.304688, "mean_p": 0.184877, "boundary_entropy": 0.063879, "bpb": 1.867995}, "vi": {"chunks": 4077.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.221155, "cps": 31.851562, "mean_p": 0.17786, "boundary_entropy": 0.092802, "bpb": 1.509726}, "tr": {"chunks": 3274.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.180854, "cps": 25.578125, "mean_p": 0.174055, "boundary_entropy": 0.069248, "bpb": 1.764294}, "fi": {"chunks": 3342.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.177662, "cps": 26.109375, "mean_p": 0.178537, "boundary_entropy": 0.063563, "bpb": 1.976963}, "ru": {"chunks": 5722.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.301953, "cps": 44.703125, "mean_p": 0.172862, "boundary_entropy": 0.098645, "bpb": 1.130949}, "el": {"chunks": 6551.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.319203, "cps": 51.179688, "mean_p": 0.183008, "boundary_entropy": 0.105386, "bpb": 1.148178}, "hi": {"chunks": 6713.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.39899, "cps": 52.445312, "mean_p": 0.173116, "boundary_entropy": 0.197202, "bpb": 0.944449}, "ta": {"chunks": 9161.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.490024, "cps": 71.570312, "mean_p": 0.182855, "boundary_entropy": 0.160034, "bpb": 0.809847}, "th": {"chunks": 8079.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.502269, "cps": 63.117188, "mean_p": 0.17201, "boundary_entropy": 0.229213, "bpb": 0.927445}, "zh": {"chunks": 2788.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.472062, "cps": 21.78125, "mean_p": 0.200996, "boundary_entropy": 0.166859, "bpb": 2.586575}, "my": {"chunks": 13933.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.735328, "cps": 108.851562, "mean_p": 0.42637, "boundary_entropy": 0.608487, "bpb": 0.956996}}, "bpb": {"en": 1.73447, "de": 1.86799, "vi": 1.50973, "tr": 1.76429, "fi": 1.97696, "ru": 1.13095, "el": 1.14818, "hi": 0.94445, "ta": 0.80985, "th": 0.92745, "zh": 2.58657, "my": 0.957}, "gini_cpc": 0.265993, "gini_cps": 0.282181, "bpic": 5.450738, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0534151057227472, "vi": 1.2752709131702866, "tr": 1.0428768053448079, "fi": 1.024470475536024, "ru": 1.7411794142794945, "el": 1.8406517964764162, "hi": 2.300734257161341, "ta": 2.8256755801112026, "th": 2.8962858843210553, "zh": 2.7221008523865273, "my": 4.240198087388693}, "t": 1545.8}
{"event": "train", "step": 7500, "phase": 1, "objective": "global", "lm_loss": 1.08932, "bpb": 1.57155, "chunk_loss": 0.030062, "F": 0.17441, "G": 0.18253, "chunks": 178.59, "bytes_per_chunk": 5.734, "boundary_entropy": 0.10884, "router_noise": 0.0, "grad_norm": 0.2914, "lr": 0.0002964, "router_trainable": true, "sec_per_step": 0.188, "alpha": 0.03, "t": 1546.1}
{"event": "probe", "step": 8000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3075.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.176694, "cps": 24.023438, "mean_p": 0.181629, "boundary_entropy": 0.056866, "bpb": 1.726727}, "de": {"chunks": 3855.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.187747, "cps": 30.117188, "mean_p": 0.190423, "boundary_entropy": 0.061326, "bpb": 1.854592}, "vi": {"chunks": 4065.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.220504, "cps": 31.757812, "mean_p": 0.17859, "boundary_entropy": 0.093799, "bpb": 1.497324}, "tr": {"chunks": 3290.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.181738, "cps": 25.703125, "mean_p": 0.17402, "boundary_entropy": 0.066756, "bpb": 1.756411}, "fi": {"chunks": 3464.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.184148, "cps": 27.0625, "mean_p": 0.185423, "boundary_entropy": 0.066905, "bpb": 1.966449}, "ru": {"chunks": 5767.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.304327, "cps": 45.054688, "mean_p": 0.17321, "boundary_entropy": 0.107832, "bpb": 1.122348}, "el": {"chunks": 6729.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.327876, "cps": 52.570312, "mean_p": 0.188476, "boundary_entropy": 0.098708, "bpb": 1.132285}, "hi": {"chunks": 7697.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.457474, "cps": 60.132812, "mean_p": 0.188038, "boundary_entropy": 0.186675, "bpb": 0.930002}, "ta": {"chunks": 9512.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.508799, "cps": 74.3125, "mean_p": 0.188648, "boundary_entropy": 0.151017, "bpb": 0.800394}, "th": {"chunks": 7974.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.495741, "cps": 62.296875, "mean_p": 0.169948, "boundary_entropy": 0.223599, "bpb": 0.909667}, "zh": {"chunks": 3016.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.510667, "cps": 23.5625, "mean_p": 0.211922, "boundary_entropy": 0.165623, "bpb": 2.581744}, "my": {"chunks": 12316.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.649989, "cps": 96.21875, "mean_p": 0.404182, "boundary_entropy": 0.583988, "bpb": 0.929409}}, "bpb": {"en": 1.72673, "de": 1.85459, "vi": 1.49732, "tr": 1.75641, "fi": 1.96645, "ru": 1.12235, "el": 1.13229, "hi": 0.93, "ta": 0.80039, "th": 0.90967, "zh": 2.58174, "my": 0.92941}, "gini_cpc": 0.251912, "gini_cps": 0.264403, "bpic": 5.4237, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0625539138068045, "vi": 1.2479477663775824, "tr": 1.0285474850550307, "fi": 1.042185436388413, "ru": 1.7223433682991183, "el": 1.8556184554390491, "hi": 2.589079665132461, "ta": 2.8795549612195774, "th": 2.80565435150155, "zh": 2.8901267830152055, "my": 3.6786231895251187}, "t": 1635.2}
{"event": "eval", "step": 8000, "phase": 1, "bpb": {"en": 1.72587, "de": 1.86016, "vi": 1.51482, "tr": 1.7502, "fi": 1.97707, "ru": 1.12096, "el": 1.13079, "hi": 0.93475, "ta": 0.80401, "th": 0.90601, "zh": 2.58467, "my": 0.93458}, "bytes_per_chunk": {"en": 5.519, "de": 5.47, "vi": 5.88, "tr": 5.822, "fi": 5.64, "ru": 5.786, "el": 5.725, "hi": 5.529, "ta": 5.359, "th": 5.731, "zh": 5.644, "my": 4.535}, "bpb_mean": 1.43699, "t": 1640.5}
{"event": "train", "step": 8000, "phase": 1, "objective": "global", "lm_loss": 1.09038, "bpb": 1.57308, "chunk_loss": 0.030206, "F": 0.18269, "G": 0.19539, "chunks": 187.08, "bytes_per_chunk": 5.474, "boundary_entropy": 0.11603, "router_noise": 0.0, "grad_norm": 0.3302, "lr": 0.0002957, "router_trainable": true, "sec_per_step": 0.189, "alpha": 0.03, "t": 1640.7}
{"event": "train", "step": 8500, "phase": 1, "objective": "global", "lm_loss": 1.09602, "bpb": 1.58123, "chunk_loss": 0.030077, "F": 0.17644, "G": 0.18988, "chunks": 180.67, "bytes_per_chunk": 5.668, "boundary_entropy": 0.11775, "router_noise": 0.0, "grad_norm": 0.2893, "lr": 0.0002949, "router_trainable": true, "sec_per_step": 0.168, "alpha": 0.03, "t": 1724.9}
{"event": "train", "step": 9000, "phase": 1, "objective": "global", "lm_loss": 1.1049, "bpb": 1.59404, "chunk_loss": 0.030057, "F": 0.1729, "G": 0.18496, "chunks": 177.05, "bytes_per_chunk": 5.784, "boundary_entropy": 0.09534, "router_noise": 0.0, "grad_norm": 0.286, "lr": 0.0002941, "router_trainable": true, "sec_per_step": 0.171, "alpha": 0.03, "t": 1810.4}
{"event": "train", "step": 9500, "phase": 1, "objective": "global", "lm_loss": 1.04087, "bpb": 1.50165, "chunk_loss": 0.03006, "F": 0.17357, "G": 0.18442, "chunks": 177.73, "bytes_per_chunk": 5.761, "boundary_entropy": 0.10397, "router_noise": 0.0, "grad_norm": 0.2855, "lr": 0.0002933, "router_trainable": true, "sec_per_step": 0.172, "alpha": 0.03, "t": 1896.3}
{"event": "train", "step": 10000, "phase": 1, "objective": "global", "lm_loss": 1.06855, "bpb": 1.54159, "chunk_loss": 0.030073, "F": 0.17564, "G": 0.18234, "chunks": 179.86, "bytes_per_chunk": 5.693, "boundary_entropy": 0.09119, "router_noise": 0.0, "grad_norm": 0.2739, "lr": 0.0002924, "router_trainable": true, "sec_per_step": 0.172, "alpha": 0.03, "t": 1982.5}
{"event": "train", "step": 10500, "phase": 1, "objective": "global", "lm_loss": 1.0299, "bpb": 1.48584, "chunk_loss": 0.030066, "F": 0.17331, "G": 0.18409, "chunks": 177.47, "bytes_per_chunk": 5.77, "boundary_entropy": 0.10889, "router_noise": 0.0, "grad_norm": 0.287, "lr": 0.0002914, "router_trainable": true, "sec_per_step": 0.174, "alpha": 0.03, "t": 2069.3}
{"event": "train", "step": 11000, "phase": 1, "objective": "global", "lm_loss": 0.99745, "bpb": 1.43901, "chunk_loss": 0.030058, "F": 0.17343, "G": 0.1863, "chunks": 177.59, "bytes_per_chunk": 5.766, "boundary_entropy": 0.11206, "router_noise": 0.0, "grad_norm": 0.2732, "lr": 0.0002903, "router_trainable": true, "sec_per_step": 0.17, "alpha": 0.03, "t": 2154.5}
{"event": "train", "step": 11500, "phase": 1, "objective": "global", "lm_loss": 0.99426, "bpb": 1.43441, "chunk_loss": 0.030057, "F": 0.17429, "G": 0.1779, "chunks": 178.47, "bytes_per_chunk": 5.738, "boundary_entropy": 0.07553, "router_noise": 0.0, "grad_norm": 0.2695, "lr": 0.0002893, "router_trainable": true, "sec_per_step": 0.168, "alpha": 0.03, "t": 2238.4}
{"event": "probe", "step": 12000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3010.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.172959, "cps": 23.515625, "mean_p": 0.175348, "boundary_entropy": 0.045916, "bpb": 1.670114}, "de": {"chunks": 3735.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.181902, "cps": 29.179688, "mean_p": 0.18268, "boundary_entropy": 0.048855, "bpb": 1.792412}, "vi": {"chunks": 4059.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.220179, "cps": 31.710938, "mean_p": 0.17359, "boundary_entropy": 0.069127, "bpb": 1.42433}, "tr": {"chunks": 3263.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.180246, "cps": 25.492188, "mean_p": 0.169755, "boundary_entropy": 0.050314, "bpb": 1.688488}, "fi": {"chunks": 3407.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.181117, "cps": 26.617188, "mean_p": 0.179467, "boundary_entropy": 0.04968, "bpb": 1.877308}, "ru": {"chunks": 5762.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.304063, "cps": 45.015625, "mean_p": 0.170606, "boundary_entropy": 0.076205, "bpb": 1.069605}, "el": {"chunks": 6728.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.327827, "cps": 52.5625, "mean_p": 0.185543, "boundary_entropy": 0.08244, "bpb": 1.080231}, "hi": {"chunks": 6858.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.407608, "cps": 53.578125, "mean_p": 0.166511, "boundary_entropy": 0.165982, "bpb": 0.880064}, "ta": {"chunks": 8594.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.459695, "cps": 67.140625, "mean_p": 0.163078, "boundary_entropy": 0.135942, "bpb": 0.75723}, "th": {"chunks": 7411.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.46074, "cps": 57.898438, "mean_p": 0.158399, "boundary_entropy": 0.210757, "bpb": 0.869494}, "zh": {"chunks": 2791.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.47257, "cps": 21.804688, "mean_p": 0.186137, "boundary_entropy": 0.099858, "bpb": 2.480056}, "my": {"chunks": 11324.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.597636, "cps": 88.46875, "mean_p": 0.404057, "boundary_entropy": 0.589502, "bpb": 0.857288}}, "bpb": {"en": 1.67011, "de": 1.79241, "vi": 1.42433, "tr": 1.68849, "fi": 1.87731, "ru": 1.06961, "el": 1.08023, "hi": 0.88006, "ta": 0.75723, "th": 0.86949, "zh": 2.48006, "my": 0.85729}, "gini_cpc": 0.236031, "gini_cps": 0.250331, "bpic": 5.733038, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.05170956468584, "vi": 1.2730150380208094, "tr": 1.0421353961449462, "fi": 1.047171646758603, "ru": 1.7580113079532604, "el": 1.895408230907937, "hi": 2.356676832549254, "ta": 2.657831868275402, "th": 2.663872120390327, "zh": 2.7322725467540754, "my": 3.4553664268776867}, "t": 2325.1}
{"event": "eval", "step": 12000, "phase": 1, "bpb": {"en": 1.66971, "de": 1.80069, "vi": 1.44304, "tr": 1.6803, "fi": 1.89135, "ru": 1.0679, "el": 1.07925, "hi": 0.88515, "ta": 0.76155, "th": 0.86689, "zh": 2.48172, "my": 0.86366}, "bytes_per_chunk": {"en": 5.628, "de": 5.593, "vi": 5.909, "tr": 5.89, "fi": 5.733, "ru": 5.782, "el": 5.755, "hi": 6.103, "ta": 5.868, "th": 6.24, "zh": 5.771, "my": 4.917}, "bpb_mean": 1.37427, "t": 2330.3}
{"event": "train", "step": 12000, "phase": 1, "objective": "global", "lm_loss": 1.07798, "bpb": 1.5552, "chunk_loss": 0.030056, "F": 0.17442, "G": 0.17871, "chunks": 178.61, "bytes_per_chunk": 5.733, "boundary_entropy": 0.07187, "router_noise": 0.0, "grad_norm": 0.2869, "lr": 0.0002881, "router_trainable": true, "sec_per_step": 0.184, "alpha": 0.03, "t": 2330.5}
{"event": "train", "step": 12500, "phase": 1, "objective": "global", "lm_loss": 1.07831, "bpb": 1.55567, "chunk_loss": 0.030045, "F": 0.17155, "G": 0.17707, "chunks": 175.67, "bytes_per_chunk": 5.829, "boundary_entropy": 0.07978, "router_noise": 0.0, "grad_norm": 0.2904, "lr": 0.0002869, "router_trainable": true, "sec_per_step": 0.169, "alpha": 0.03, "t": 2414.9}
{"event": "train", "step": 13000, "phase": 1, "objective": "global", "lm_loss": 1.0506, "bpb": 1.51569, "chunk_loss": 0.030058, "F": 0.17506, "G": 0.18078, "chunks": 179.27, "bytes_per_chunk": 5.712, "boundary_entropy": 0.07844, "router_noise": 0.0, "grad_norm": 0.2809, "lr": 0.0002857, "router_trainable": true, "sec_per_step": 0.168, "alpha": 0.03, "t": 2499.0}
{"event": "train", "step": 13500, "phase": 1, "objective": "global", "lm_loss": 1.01543, "bpb": 1.46496, "chunk_loss": 0.030057, "F": 0.17331, "G": 0.17624, "chunks": 177.47, "bytes_per_chunk": 5.77, "boundary_entropy": 0.07171, "router_noise": 0.0, "grad_norm": 0.2731, "lr": 0.0002844, "router_trainable": true, "sec_per_step": 0.169, "alpha": 0.03, "t": 2583.3}
{"event": "train", "step": 14000, "phase": 1, "objective": "global", "lm_loss": 1.05023, "bpb": 1.51517, "chunk_loss": 0.030116, "F": 0.17496, "G": 0.1856, "chunks": 179.16, "bytes_per_chunk": 5.716, "boundary_entropy": 0.08855, "router_noise": 0.0, "grad_norm": 0.2925, "lr": 0.000283, "router_trainable": true, "sec_per_step": 0.168, "alpha": 0.03, "t": 2667.1}
{"event": "train", "step": 14500, "phase": 1, "objective": "global", "lm_loss": 1.09671, "bpb": 1.58222, "chunk_loss": 0.030018, "F": 0.17079, "G": 0.18061, "chunks": 174.89, "bytes_per_chunk": 5.855, "boundary_entropy": 0.08796, "router_noise": 0.0, "grad_norm": 0.2739, "lr": 0.0002816, "router_trainable": true, "sec_per_step": 0.169, "alpha": 0.03, "t": 2751.4}
{"event": "train", "step": 15000, "phase": 1, "objective": "global", "lm_loss": 0.97569, "bpb": 1.40762, "chunk_loss": 0.030035, "F": 0.17052, "G": 0.17335, "chunks": 174.61, "bytes_per_chunk": 5.865, "boundary_entropy": 0.06184, "router_noise": 0.0, "grad_norm": 0.2633, "lr": 0.0002801, "router_trainable": true, "sec_per_step": 0.169, "alpha": 0.03, "t": 2836.1}
{"event": "train", "step": 15500, "phase": 1, "objective": "global", "lm_loss": 1.02226, "bpb": 1.47481, "chunk_loss": 0.030043, "F": 0.17307, "G": 0.17851, "chunks": 177.22, "bytes_per_chunk": 5.778, "boundary_entropy": 0.08142, "router_noise": 0.0, "grad_norm": 0.2906, "lr": 0.0002786, "router_trainable": true, "sec_per_step": 0.168, "alpha": 0.03, "t": 2920.3}
{"event": "probe", "step": 16000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3008.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.172844, "cps": 23.5, "mean_p": 0.174155, "boundary_entropy": 0.038935, "bpb": 1.635504}, "de": {"chunks": 3748.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.182535, "cps": 29.28125, "mean_p": 0.181963, "boundary_entropy": 0.042553, "bpb": 1.753795}, "vi": {"chunks": 4050.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.219691, "cps": 31.640625, "mean_p": 0.173164, "boundary_entropy": 0.064902, "bpb": 1.38777}, "tr": {"chunks": 3226.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.178203, "cps": 25.203125, "mean_p": 0.167421, "boundary_entropy": 0.043292, "bpb": 1.6477}, "fi": {"chunks": 3355.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.178353, "cps": 26.210938, "mean_p": 0.175297, "boundary_entropy": 0.043142, "bpb": 1.824413}, "ru": {"chunks": 5714.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.30153, "cps": 44.640625, "mean_p": 0.167823, "boundary_entropy": 0.059277, "bpb": 1.042304}, "el": {"chunks": 6652.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.324124, "cps": 51.96875, "mean_p": 0.179423, "boundary_entropy": 0.065558, "bpb": 1.050011}, "hi": {"chunks": 7042.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.418544, "cps": 55.015625, "mean_p": 0.164221, "boundary_entropy": 0.139418, "bpb": 0.856916}, "ta": {"chunks": 8786.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.469965, "cps": 68.640625, "mean_p": 0.165153, "boundary_entropy": 0.124207, "bpb": 0.73258}, "th": {"chunks": 8653.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.537955, "cps": 67.601562, "mean_p": 0.168192, "boundary_entropy": 0.186565, "bpb": 0.839905}, "zh": {"chunks": 2851.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.482729, "cps": 22.273438, "mean_p": 0.187058, "boundary_entropy": 0.083073, "bpb": 2.44536}, "my": {"chunks": 9723.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.513141, "cps": 75.960938, "mean_p": 0.386561, "boundary_entropy": 0.584761, "bpb": 0.827276}}, "bpb": {"en": 1.6355, "de": 1.7538, "vi": 1.38777, "tr": 1.6477, "fi": 1.82441, "ru": 1.0423, "el": 1.05001, "hi": 0.85692, "ta": 0.73258, "th": 0.8399, "zh": 2.44536, "my": 0.82728}, "gini_cpc": 0.234897, "gini_cps": 0.242478, "bpic": 5.744537, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0560718423171416, "vi": 1.2710369310804492, "tr": 1.0310034057479598, "fi": 1.0318746168493538, "ru": 1.7445254519171391, "el": 1.8752435643040863, "hi": 2.42151540419209, "ta": 2.7190175934514293, "th": 3.1123750611776533, "zh": 2.792865768043317, "my": 2.968815426121434}, "t": 3007.3}
{"event": "eval", "step": 16000, "phase": 1, "bpb": {"en": 1.63324, "de": 1.76126, "vi": 1.40616, "tr": 1.6405, "fi": 1.83923, "ru": 1.04191, "el": 1.04753, "hi": 0.86248, "ta": 0.73718, "th": 0.83648, "zh": 2.4476, "my": 0.83508}, "bytes_per_chunk": {"en": 5.623, "de": 5.619, "vi": 5.922, "tr": 5.936, "fi": 5.773, "ru": 5.827, "el": 5.83, "hi": 5.981, "ta": 5.678, "th": 5.427, "zh": 5.693, "my": 5.745}, "bpb_mean": 1.34072, "t": 3012.7}
{"event": "train", "step": 16000, "phase": 1, "objective": "global", "lm_loss": 0.98346, "bpb": 1.41884, "chunk_loss": 0.030071, "F": 0.17773, "G": 0.18136, "chunks": 182.0, "bytes_per_chunk": 5.626, "boundary_entropy": 0.08375, "router_noise": 0.0, "grad_norm": 0.2798, "lr": 0.000277, "router_trainable": true, "sec_per_step": 0.185, "alpha": 0.03, "t": 3012.9}
{"event": "train", "step": 16500, "phase": 1, "objective": "global", "lm_loss": 1.01248, "bpb": 1.46071, "chunk_loss": 0.030092, "F": 0.17438, "G": 0.17679, "chunks": 178.56, "bytes_per_chunk": 5.735, "boundary_entropy": 0.06657, "router_noise": 0.0, "grad_norm": 0.2537, "lr": 0.0002754, "router_trainable": true, "sec_per_step": 0.167, "alpha": 0.03, "t": 3096.6}
{"event": "train", "step": 17000, "phase": 1, "objective": "global", "lm_loss": 0.97968, "bpb": 1.41338, "chunk_loss": 0.030055, "F": 0.17384, "G": 0.18269, "chunks": 178.02, "bytes_per_chunk": 5.752, "boundary_entropy": 0.0796, "router_noise": 0.0, "grad_norm": 0.2775, "lr": 0.0002737, "router_trainable": true, "sec_per_step": 0.167, "alpha": 0.03, "t": 3180.4}
{"event": "train", "step": 17500, "phase": 1, "objective": "global", "lm_loss": 0.9788, "bpb": 1.41211, "chunk_loss": 0.030054, "F": 0.17502, "G": 0.17741, "chunks": 179.22, "bytes_per_chunk": 5.714, "boundary_entropy": 0.06976, "router_noise": 0.0, "grad_norm": 0.265, "lr": 0.000272, "router_trainable": true, "sec_per_step": 0.167, "alpha": 0.03, "t": 3263.8}
{"event": "train", "step": 18000, "phase": 1, "objective": "global", "lm_loss": 1.04324, "bpb": 1.50508, "chunk_loss": 0.030043, "F": 0.17316, "G": 0.1747, "chunks": 177.31, "bytes_per_chunk": 5.775, "boundary_entropy": 0.06083, "router_noise": 0.0, "grad_norm": 0.2708, "lr": 0.0002703, "router_trainable": true, "sec_per_step": 0.167, "alpha": 0.03, "t": 3347.6}
{"event": "train", "step": 18500, "phase": 1, "objective": "global", "lm_loss": 1.09238, "bpb": 1.57597, "chunk_loss": 0.030099, "F": 0.17352, "G": 0.17545, "chunks": 177.69, "bytes_per_chunk": 5.763, "boundary_entropy": 0.06062, "router_noise": 0.0, "grad_norm": 0.4478, "lr": 0.0002685, "router_trainable": true, "sec_per_step": 0.167, "alpha": 0.03, "t": 3431.2}
{"event": "train", "step": 19000, "phase": 1, "objective": "global", "lm_loss": 0.99999, "bpb": 1.44267, "chunk_loss": 0.030101, "F": 0.17249, "G": 0.17687, "chunks": 176.62, "bytes_per_chunk": 5.798, "boundary_entropy": 0.06849, "router_noise": 0.0, "grad_norm": 0.2634, "lr": 0.0002666, "router_trainable": true, "sec_per_step": 0.169, "alpha": 0.03, "t": 3515.5}
{"event": "train", "step": 19500, "phase": 1, "objective": "global", "lm_loss": 0.97252, "bpb": 1.40305, "chunk_loss": 0.030024, "F": 0.17361, "G": 0.18142, "chunks": 177.78, "bytes_per_chunk": 5.76, "boundary_entropy": 0.08024, "router_noise": 0.0, "grad_norm": 0.263, "lr": 0.0002647, "router_trainable": true, "sec_per_step": 0.167, "alpha": 0.03, "t": 3598.8}
{"event": "probe", "step": 20000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3008.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.172844, "cps": 23.5, "mean_p": 0.174275, "boundary_entropy": 0.033412, "bpb": 1.611346}, "de": {"chunks": 3723.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.181318, "cps": 29.085938, "mean_p": 0.180791, "boundary_entropy": 0.037081, "bpb": 1.722618}, "vi": {"chunks": 4057.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.220071, "cps": 31.695312, "mean_p": 0.173789, "boundary_entropy": 0.059993, "bpb": 1.360847}, "tr": {"chunks": 3259.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.180025, "cps": 25.460938, "mean_p": 0.168299, "boundary_entropy": 0.041187, "bpb": 1.61987}, "fi": {"chunks": 3374.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179363, "cps": 26.359375, "mean_p": 0.175704, "boundary_entropy": 0.038319, "bpb": 1.792101}, "ru": {"chunks": 5714.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.30153, "cps": 44.640625, "mean_p": 0.167009, "boundary_entropy": 0.055168, "bpb": 1.022414}, "el": {"chunks": 6705.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.326707, "cps": 52.382812, "mean_p": 0.182524, "boundary_entropy": 0.056769, "bpb": 1.027403}, "hi": {"chunks": 7230.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.429718, "cps": 56.484375, "mean_p": 0.169117, "boundary_entropy": 0.139591, "bpb": 0.841113}, "ta": {"chunks": 8927.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.477507, "cps": 69.742188, "mean_p": 0.164661, "boundary_entropy": 0.117786, "bpb": 0.714995}, "th": {"chunks": 7891.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.490581, "cps": 61.648438, "mean_p": 0.159239, "boundary_entropy": 0.185384, "bpb": 0.821335}, "zh": {"chunks": 2776.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.47003, "cps": 21.6875, "mean_p": 0.181215, "boundary_entropy": 0.06834, "bpb": 2.408905}, "my": {"chunks": 7772.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.410175, "cps": 60.71875, "mean_p": 0.363649, "boundary_entropy": 0.572696, "bpb": 0.817014}}, "bpb": {"en": 1.61135, "de": 1.72262, "vi": 1.36085, "tr": 1.61987, "fi": 1.7921, "ru": 1.02241, "el": 1.0274, "hi": 0.84111, "ta": 0.715, "th": 0.82133, "zh": 2.4089, "my": 0.81701}, "gini_cpc": 0.218018, "gini_cps": 0.223842, "bpic": 5.956003, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.049027606442561, "vi": 1.2732337850354032, "tr": 1.0415499377968387, "fi": 1.037718318107219, "ru": 1.7445254519171391, "el": 1.8901846209649575, "hi": 2.486162506718093, "ta": 2.762653090910643, "th": 2.8382932633483025, "zh": 2.7193950796521387, "my": 2.3730981684475765}, "t": 3687.7}
{"event": "eval", "step": 20000, "phase": 1, "bpb": {"en": 1.6083, "de": 1.7295, "vi": 1.37768, "tr": 1.61236, "fi": 1.80718, "ru": 1.02174, "el": 1.02479, "hi": 0.84682, "ta": 0.71982, "th": 0.82068, "zh": 2.41028, "my": 0.82276}, "bytes_per_chunk": {"en": 5.631, "de": 5.636, "vi": 5.922, "tr": 5.903, "fi": 5.768, "ru": 5.816, "el": 5.765, "hi": 5.856, "ta": 5.596, "th": 5.979, "zh": 5.767, "my": 7.069}, "bpb_mean": 1.31683, "t": 3693.0}
{"event": "train", "step": 20000, "phase": 1, "objective": "global", "lm_loss": 0.97915, "bpb": 1.41261, "chunk_loss": 0.030022, "F": 0.17406, "G": 0.18256, "chunks": 178.23, "bytes_per_chunk": 5.745, "boundary_entropy": 0.08013, "router_noise": 0.0, "grad_norm": 0.2612, "lr": 0.0002628, "router_trainable": true, "sec_per_step": 0.189, "alpha": 0.03, "t": 3693.2}
{"event": "train", "step": 20500, "phase": 1, "objective": "global", "lm_loss": 0.97361, "bpb": 1.40463, "chunk_loss": 0.03005, "F": 0.17439, "G": 0.17666, "chunks": 178.58, "bytes_per_chunk": 5.734, "boundary_entropy": 0.05429, "router_noise": 0.0, "grad_norm": 0.27, "lr": 0.0002608, "router_trainable": true, "sec_per_step": 0.166, "alpha": 0.03, "t": 3776.3}
{"event": "train", "step": 21000, "phase": 1, "objective": "global", "lm_loss": 0.99726, "bpb": 1.43874, "chunk_loss": 0.030051, "F": 0.17686, "G": 0.17702, "chunks": 181.11, "bytes_per_chunk": 5.654, "boundary_entropy": 0.06486, "router_noise": 0.0, "grad_norm": 0.2646, "lr": 0.0002587, "router_trainable": true, "sec_per_step": 0.171, "alpha": 0.03, "t": 3862.0}
{"event": "train", "step": 21500, "phase": 1, "objective": "global", "lm_loss": 0.98345, "bpb": 1.41882, "chunk_loss": 0.03015, "F": 0.17743, "G": 0.18477, "chunks": 181.69, "bytes_per_chunk": 5.636, "boundary_entropy": 0.07736, "router_noise": 0.0, "grad_norm": 0.2418, "lr": 0.0002567, "router_trainable": true, "sec_per_step": 0.166, "alpha": 0.03, "t": 3944.8}
{"event": "train", "step": 22000, "phase": 1, "objective": "global", "lm_loss": 1.04221, "bpb": 1.5036, "chunk_loss": 0.030079, "F": 0.17633, "G": 0.18162, "chunks": 180.56, "bytes_per_chunk": 5.671, "boundary_entropy": 0.06475, "router_noise": 0.0, "grad_norm": 0.2898, "lr": 0.0002545, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 4026.8}
{"event": "train", "step": 22500, "phase": 1, "objective": "global", "lm_loss": 0.9373, "bpb": 1.35224, "chunk_loss": 0.030058, "F": 0.17665, "G": 0.17704, "chunks": 180.89, "bytes_per_chunk": 5.661, "boundary_entropy": 0.05213, "router_noise": 0.0, "grad_norm": 0.2515, "lr": 0.0002524, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 4108.0}
{"event": "train", "step": 23000, "phase": 1, "objective": "global", "lm_loss": 0.98624, "bpb": 1.42284, "chunk_loss": 0.030057, "F": 0.17679, "G": 0.18011, "chunks": 181.03, "bytes_per_chunk": 5.656, "boundary_entropy": 0.06844, "router_noise": 0.0, "grad_norm": 0.2527, "lr": 0.0002502, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 4189.2}
{"event": "train", "step": 23500, "phase": 1, "objective": "global", "lm_loss": 1.0746, "bpb": 1.55032, "chunk_loss": 0.030088, "F": 0.17679, "G": 0.18351, "chunks": 181.03, "bytes_per_chunk": 5.656, "boundary_entropy": 0.06945, "router_noise": 0.0, "grad_norm": 0.2707, "lr": 0.000248, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 4270.6}
{"event": "probe", "step": 24000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3009.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.172901, "cps": 23.507812, "mean_p": 0.173926, "boundary_entropy": 0.030307, "bpb": 1.590031}, "de": {"chunks": 3745.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.182389, "cps": 29.257812, "mean_p": 0.181921, "boundary_entropy": 0.034973, "bpb": 1.698329}, "vi": {"chunks": 4019.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.218009, "cps": 31.398438, "mean_p": 0.170904, "boundary_entropy": 0.050317, "bpb": 1.335828}, "tr": {"chunks": 3351.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.185107, "cps": 26.179688, "mean_p": 0.173121, "boundary_entropy": 0.039636, "bpb": 1.603484}, "fi": {"chunks": 3364.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.178832, "cps": 26.28125, "mean_p": 0.175371, "boundary_entropy": 0.034452, "bpb": 1.767447}, "ru": {"chunks": 5710.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.301319, "cps": 44.609375, "mean_p": 0.166347, "boundary_entropy": 0.049463, "bpb": 1.007675}, "el": {"chunks": 6662.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.324611, "cps": 52.046875, "mean_p": 0.179215, "boundary_entropy": 0.052064, "bpb": 1.012451}, "hi": {"chunks": 7463.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.443566, "cps": 58.304688, "mean_p": 0.171695, "boundary_entropy": 0.122952, "bpb": 0.821422}, "ta": {"chunks": 8260.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.441829, "cps": 64.53125, "mean_p": 0.155019, "boundary_entropy": 0.107263, "bpb": 0.710359}, "th": {"chunks": 8369.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.520298, "cps": 65.382812, "mean_p": 0.164908, "boundary_entropy": 0.174241, "bpb": 0.805286}, "zh": {"chunks": 2722.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.460887, "cps": 21.265625, "mean_p": 0.175175, "boundary_entropy": 0.057865, "bpb": 2.383003}, "my": {"chunks": 8590.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.453346, "cps": 67.109375, "mean_p": 0.330278, "boundary_entropy": 0.53102, "bpb": 0.790529}}, "bpb": {"en": 1.59003, "de": 1.69833, "vi": 1.33583, "tr": 1.60348, "fi": 1.76745, "ru": 1.00767, "el": 1.01245, "hi": 0.82142, "ta": 0.71036, "th": 0.80529, "zh": 2.383, "my": 0.79053}, "gini_cpc": 0.219833, "gini_cps": 0.227394, "bpic": 5.880439, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0548758439045112, "vi": 1.260888827956056, "tr": 1.0705964744004743, "fi": 1.0342988365189902, "ru": 1.7427248597216267, "el": 1.8774384831666868, "hi": 2.56543078046692, "ta": 2.5553859949970894, "th": 3.009223433212951, "zh": 2.665610010469776, "my": 2.621994161067568}, "t": 4354.9}
{"event": "eval", "step": 24000, "phase": 1, "bpb": {"en": 1.58737, "de": 1.70746, "vi": 1.35432, "tr": 1.59573, "fi": 1.78428, "ru": 1.00646, "el": 1.01003, "hi": 0.82822, "ta": 0.71471, "th": 0.80341, "zh": 2.38391, "my": 0.79617}, "bytes_per_chunk": {"en": 5.61, "de": 5.593, "vi": 5.941, "tr": 5.744, "fi": 5.756, "ru": 5.837, "el": 5.815, "hi": 5.687, "ta": 6.022, "th": 5.641, "zh": 5.829, "my": 6.308}, "bpb_mean": 1.29767, "t": 4360.1}
{"event": "train", "step": 24000, "phase": 1, "objective": "global", "lm_loss": 0.93825, "bpb": 1.3536, "chunk_loss": 0.030033, "F": 0.17453, "G": 0.18883, "chunks": 178.72, "bytes_per_chunk": 5.73, "boundary_entropy": 0.08633, "router_noise": 0.0, "grad_norm": 0.2575, "lr": 0.0002457, "router_trainable": true, "sec_per_step": 0.179, "alpha": 0.03, "t": 4360.3}
{"event": "train", "step": 24500, "phase": 1, "objective": "global", "lm_loss": 1.06144, "bpb": 1.53133, "chunk_loss": 0.03005, "F": 0.17426, "G": 0.17549, "chunks": 178.44, "bytes_per_chunk": 5.739, "boundary_entropy": 0.0511, "router_noise": 0.0, "grad_norm": 0.2989, "lr": 0.0002434, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 4441.4}
{"event": "train", "step": 25000, "phase": 1, "objective": "global", "lm_loss": 0.99531, "bpb": 1.43592, "chunk_loss": 0.030054, "F": 0.17464, "G": 0.17585, "chunks": 178.83, "bytes_per_chunk": 5.726, "boundary_entropy": 0.04869, "router_noise": 0.0, "grad_norm": 0.272, "lr": 0.000241, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 4522.3}
{"event": "train", "step": 25500, "phase": 1, "objective": "global", "lm_loss": 0.96168, "bpb": 1.38741, "chunk_loss": 0.030055, "F": 0.1741, "G": 0.1763, "chunks": 178.28, "bytes_per_chunk": 5.744, "boundary_entropy": 0.06846, "router_noise": 0.0, "grad_norm": 0.2676, "lr": 0.0002387, "router_trainable": true, "sec_per_step": 0.161, "alpha": 0.03, "t": 4603.0}
{"event": "train", "step": 26000, "phase": 1, "objective": "global", "lm_loss": 1.0086, "bpb": 1.45511, "chunk_loss": 0.030059, "F": 0.17305, "G": 0.18311, "chunks": 177.2, "bytes_per_chunk": 5.779, "boundary_entropy": 0.07491, "router_noise": 0.0, "grad_norm": 0.2955, "lr": 0.0002363, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 4683.8}
{"event": "train", "step": 26500, "phase": 1, "objective": "global", "lm_loss": 0.9711, "bpb": 1.40099, "chunk_loss": 0.030068, "F": 0.1792, "G": 0.17918, "chunks": 183.5, "bytes_per_chunk": 5.58, "boundary_entropy": 0.0564, "router_noise": 0.0, "grad_norm": 0.2696, "lr": 0.0002338, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 4765.0}
{"event": "train", "step": 27000, "phase": 1, "objective": "global", "lm_loss": 0.97208, "bpb": 1.40242, "chunk_loss": 0.030064, "F": 0.17546, "G": 0.17629, "chunks": 179.67, "bytes_per_chunk": 5.699, "boundary_entropy": 0.0527, "router_noise": 0.0, "grad_norm": 0.2713, "lr": 0.0002313, "router_trainable": true, "sec_per_step": 0.166, "alpha": 0.03, "t": 4847.8}
{"event": "train", "step": 27500, "phase": 1, "objective": "global", "lm_loss": 0.97676, "bpb": 1.40917, "chunk_loss": 0.030077, "F": 0.17441, "G": 0.18191, "chunks": 178.59, "bytes_per_chunk": 5.734, "boundary_entropy": 0.06475, "router_noise": 0.0, "grad_norm": 0.2644, "lr": 0.0002288, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 4929.5}
{"event": "probe", "step": 28000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3026.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.173878, "cps": 23.640625, "mean_p": 0.174917, "boundary_entropy": 0.03026, "bpb": 1.575171}, "de": {"chunks": 3749.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.182584, "cps": 29.289062, "mean_p": 0.181737, "boundary_entropy": 0.033982, "bpb": 1.688662}, "vi": {"chunks": 4065.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.220504, "cps": 31.757812, "mean_p": 0.171853, "boundary_entropy": 0.048861, "bpb": 1.321049}, "tr": {"chunks": 3372.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.186267, "cps": 26.34375, "mean_p": 0.172413, "boundary_entropy": 0.034855, "bpb": 1.581215}, "fi": {"chunks": 3359.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.178566, "cps": 26.242188, "mean_p": 0.173973, "boundary_entropy": 0.033217, "bpb": 1.738326}, "ru": {"chunks": 5712.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.301425, "cps": 44.625, "mean_p": 0.16736, "boundary_entropy": 0.04604, "bpb": 0.995265}, "el": {"chunks": 6680.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.325488, "cps": 52.1875, "mean_p": 0.179558, "boundary_entropy": 0.047954, "bpb": 0.996301}, "hi": {"chunks": 7334.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.435899, "cps": 57.296875, "mean_p": 0.167559, "boundary_entropy": 0.120222, "bpb": 0.812518}, "ta": {"chunks": 8915.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.476865, "cps": 69.648438, "mean_p": 0.164775, "boundary_entropy": 0.107533, "bpb": 0.692852}, "th": {"chunks": 8048.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.500342, "cps": 62.875, "mean_p": 0.160521, "boundary_entropy": 0.173294, "bpb": 0.795605}, "zh": {"chunks": 2733.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.46275, "cps": 21.351562, "mean_p": 0.177168, "boundary_entropy": 0.057698, "bpb": 2.364364}, "my": {"chunks": 8537.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.450549, "cps": 66.695312, "mean_p": 0.338313, "boundary_entropy": 0.537554, "bpb": 0.778717}}, "bpb": {"en": 1.57517, "de": 1.68866, "vi": 1.32105, "tr": 1.58121, "fi": 1.73833, "ru": 0.99526, "el": 0.9963, "hi": 0.81252, "ta": 0.69285, "th": 0.79561, "zh": 2.36436, "my": 0.77872}, "gini_cpc": 0.218834, "gini_cps": 0.22976, "bpic": 5.85657, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0500699484964944, "vi": 1.2681557771351837, "tr": 1.0712533943700802, "fi": 1.0269595003949785, "ru": 1.7335412528534584, "el": 1.8719352127587319, "hi": 2.506923198549667, "ta": 2.742528027134717, "th": 2.8775448342116743, "zh": 2.661346275612487, "my": 2.591177129852446}, "t": 5013.8}
{"event": "eval", "step": 28000, "phase": 1, "bpb": {"en": 1.5715, "de": 1.69555, "vi": 1.33876, "tr": 1.57314, "fi": 1.75406, "ru": 0.99351, "el": 0.99431, "hi": 0.81924, "ta": 0.69703, "th": 0.79452, "zh": 2.36421, "my": 0.7847}, "bytes_per_chunk": {"en": 5.584, "de": 5.615, "vi": 5.901, "tr": 5.731, "fi": 5.776, "ru": 5.826, "el": 5.793, "hi": 5.748, "ta": 5.642, "th": 5.778, "zh": 5.806, "my": 6.354}, "bpb_mean": 1.28171, "t": 5019.1}
{"event": "train", "step": 28000, "phase": 1, "objective": "global", "lm_loss": 0.94164, "bpb": 1.35849, "chunk_loss": 0.030063, "F": 0.17554, "G": 0.17838, "chunks": 179.75, "bytes_per_chunk": 5.697, "boundary_entropy": 0.06716, "router_noise": 0.0, "grad_norm": 0.2757, "lr": 0.0002263, "router_trainable": true, "sec_per_step": 0.18, "alpha": 0.03, "t": 5019.3}
{"event": "train", "step": 28500, "phase": 1, "objective": "global", "lm_loss": 0.98076, "bpb": 1.41494, "chunk_loss": 0.030051, "F": 0.17386, "G": 0.1743, "chunks": 178.03, "bytes_per_chunk": 5.752, "boundary_entropy": 0.05753, "router_noise": 0.0, "grad_norm": 0.2654, "lr": 0.0002238, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 5100.6}
{"event": "train", "step": 29000, "phase": 1, "objective": "global", "lm_loss": 0.92024, "bpb": 1.32763, "chunk_loss": 0.030087, "F": 0.17447, "G": 0.18025, "chunks": 178.66, "bytes_per_chunk": 5.732, "boundary_entropy": 0.06621, "router_noise": 0.0, "grad_norm": 0.2574, "lr": 0.0002212, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 5181.9}
{"event": "train", "step": 29500, "phase": 1, "objective": "global", "lm_loss": 0.97634, "bpb": 1.40856, "chunk_loss": 0.030091, "F": 0.17651, "G": 0.18271, "chunks": 180.75, "bytes_per_chunk": 5.665, "boundary_entropy": 0.06778, "router_noise": 0.0, "grad_norm": 0.2644, "lr": 0.0002186, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 5264.1}
{"event": "train", "step": 30000, "phase": 1, "objective": "global", "lm_loss": 0.94685, "bpb": 1.36602, "chunk_loss": 0.030123, "F": 0.17747, "G": 0.18844, "chunks": 181.73, "bytes_per_chunk": 5.635, "boundary_entropy": 0.08528, "router_noise": 0.0, "grad_norm": 0.2558, "lr": 0.0002159, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 5346.1}
{"event": "train", "step": 30500, "phase": 1, "objective": "global", "lm_loss": 1.00931, "bpb": 1.45612, "chunk_loss": 0.030049, "F": 0.17299, "G": 0.17639, "chunks": 177.14, "bytes_per_chunk": 5.781, "boundary_entropy": 0.05288, "router_noise": 0.0, "grad_norm": 0.2732, "lr": 0.0002133, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 5427.9}
{"event": "phase_transition", "step": 30520, "phase": 2, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 5431.1}
{"event": "train", "step": 31000, "phase": 2, "objective": "global", "lm_loss": 0.96795, "bpb": 1.39646, "chunk_loss": 0.030049, "F": 0.17595, "G": 0.18171, "chunks": 180.17, "bytes_per_chunk": 5.683, "boundary_entropy": 0.06781, "router_noise": 0.0, "grad_norm": 0.2956, "lr": 0.0002106, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 5509.3}
{"event": "train", "step": 31500, "phase": 2, "objective": "global", "lm_loss": 0.94363, "bpb": 1.36137, "chunk_loss": 0.030045, "F": 0.17368, "G": 0.185, "chunks": 177.84, "bytes_per_chunk": 5.758, "boundary_entropy": 0.07644, "router_noise": 0.0, "grad_norm": 0.2602, "lr": 0.0002079, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 5590.5}
{"event": "probe", "step": 32000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3006.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.172729, "cps": 23.484375, "mean_p": 0.172707, "boundary_entropy": 0.02828, "bpb": 1.558392}, "de": {"chunks": 3728.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.181561, "cps": 29.125, "mean_p": 0.180283, "boundary_entropy": 0.031373, "bpb": 1.6627}, "vi": {"chunks": 4071.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.22083, "cps": 31.804688, "mean_p": 0.171493, "boundary_entropy": 0.044434, "bpb": 1.300875}, "tr": {"chunks": 3360.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.185605, "cps": 26.25, "mean_p": 0.171699, "boundary_entropy": 0.034209, "bpb": 1.565335}, "fi": {"chunks": 3364.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.178832, "cps": 26.28125, "mean_p": 0.174684, "boundary_entropy": 0.030613, "bpb": 1.715138}, "ru": {"chunks": 5709.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.301266, "cps": 44.601562, "mean_p": 0.166484, "boundary_entropy": 0.041454, "bpb": 0.980537}, "el": {"chunks": 6701.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.326512, "cps": 52.351562, "mean_p": 0.179782, "boundary_entropy": 0.049076, "bpb": 0.981247}, "hi": {"chunks": 7267.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.431917, "cps": 56.773438, "mean_p": 0.16735, "boundary_entropy": 0.12658, "bpb": 0.7998}, "ta": {"chunks": 8634.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.461835, "cps": 67.453125, "mean_p": 0.162087, "boundary_entropy": 0.1126, "bpb": 0.68543}, "th": {"chunks": 8306.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.516382, "cps": 64.890625, "mean_p": 0.160784, "boundary_entropy": 0.16903, "bpb": 0.784157}, "zh": {"chunks": 2717.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.460041, "cps": 21.226562, "mean_p": 0.175067, "boundary_entropy": 0.053368, "bpb": 2.331961}, "my": {"chunks": 8736.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.461051, "cps": 68.25, "mean_p": 0.341217, "boundary_entropy": 0.540932, "bpb": 0.76475}}, "bpb": {"en": 1.55839, "de": 1.6627, "vi": 1.30087, "tr": 1.56534, "fi": 1.71514, "ru": 0.98054, "el": 0.98125, "hi": 0.7998, "ta": 0.68543, "th": 0.78416, "zh": 2.33196, "my": 0.76475}, "gini_cpc": 0.220473, "gini_cps": 0.230637, "bpic": 5.850409, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.05113534680019, "vi": 1.278477544504157, "tr": 1.0745431746020195, "fi": 1.0353310708867736, "ru": 1.7441585957372852, "el": 1.8903138516612927, "hi": 2.5005482052240247, "ta": 2.67375567132651, "th": 2.989551267375104, "zh": 2.663368995162401, "my": 2.6692201410984238}, "t": 5674.4}
{"event": "eval", "step": 32000, "phase": 2, "bpb": {"en": 1.55489, "de": 1.67139, "vi": 1.31902, "tr": 1.55882, "fi": 1.73054, "ru": 0.97887, "el": 0.98034, "hi": 0.80519, "ta": 0.69069, "th": 0.78184, "zh": 2.33479, "my": 0.77116}, "bytes_per_chunk": {"en": 5.617, "de": 5.606, "vi": 5.898, "tr": 5.738, "fi": 5.767, "ru": 5.832, "el": 5.774, "hi": 5.815, "ta": 5.774, "th": 5.614, "zh": 5.814, "my": 6.246}, "bpb_mean": 1.26479, "t": 5679.8}
{"event": "train", "step": 32000, "phase": 2, "objective": "global", "lm_loss": 0.96111, "bpb": 1.38658, "chunk_loss": 0.030045, "F": 0.17667, "G": 0.17898, "chunks": 180.91, "bytes_per_chunk": 5.66, "boundary_entropy": 0.06195, "router_noise": 0.0, "grad_norm": 0.273, "lr": 0.0002052, "router_trainable": true, "sec_per_step": 0.179, "alpha": 0.03, "t": 5680.0}
{"event": "train", "step": 32500, "phase": 2, "objective": "global", "lm_loss": 0.96384, "bpb": 1.39052, "chunk_loss": 0.030061, "F": 0.17479, "G": 0.17507, "chunks": 178.98, "bytes_per_chunk": 5.721, "boundary_entropy": 0.05083, "router_noise": 0.0, "grad_norm": 0.2773, "lr": 0.0002025, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 5761.2}
{"event": "train", "step": 33000, "phase": 2, "objective": "global", "lm_loss": 0.89976, "bpb": 1.29808, "chunk_loss": 0.03007, "F": 0.17416, "G": 0.17943, "chunks": 178.34, "bytes_per_chunk": 5.742, "boundary_entropy": 0.06367, "router_noise": 0.0, "grad_norm": 0.2455, "lr": 0.0001997, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 5842.0}
{"event": "train", "step": 33500, "phase": 2, "objective": "global", "lm_loss": 0.92455, "bpb": 1.33385, "chunk_loss": 0.030023, "F": 0.17256, "G": 0.17905, "chunks": 176.7, "bytes_per_chunk": 5.795, "boundary_entropy": 0.06236, "router_noise": 0.0, "grad_norm": 0.2554, "lr": 0.0001969, "router_trainable": true, "sec_per_step": 0.16, "alpha": 0.03, "t": 5921.8}
{"event": "train", "step": 34000, "phase": 2, "objective": "global", "lm_loss": 1.04357, "bpb": 1.50555, "chunk_loss": 0.030036, "F": 0.17529, "G": 0.17887, "chunks": 179.5, "bytes_per_chunk": 5.705, "boundary_entropy": 0.0524, "router_noise": 0.0, "grad_norm": 0.2888, "lr": 0.0001942, "router_trainable": true, "sec_per_step": 0.16, "alpha": 0.03, "t": 6001.5}
{"event": "train", "step": 34500, "phase": 2, "objective": "global", "lm_loss": 0.93741, "bpb": 1.35239, "chunk_loss": 0.030057, "F": 0.17252, "G": 0.17188, "chunks": 176.66, "bytes_per_chunk": 5.797, "boundary_entropy": 0.05304, "router_noise": 0.0, "grad_norm": 0.2614, "lr": 0.0001914, "router_trainable": true, "sec_per_step": 0.16, "alpha": 0.03, "t": 6081.7}
{"event": "train", "step": 35000, "phase": 2, "objective": "global", "lm_loss": 0.9114, "bpb": 1.31487, "chunk_loss": 0.030038, "F": 0.17595, "G": 0.17752, "chunks": 180.17, "bytes_per_chunk": 5.683, "boundary_entropy": 0.06139, "router_noise": 0.0, "grad_norm": 0.2835, "lr": 0.0001886, "router_trainable": true, "sec_per_step": 0.159, "alpha": 0.03, "t": 6161.4}
{"event": "train", "step": 35500, "phase": 2, "objective": "global", "lm_loss": 0.92558, "bpb": 1.33533, "chunk_loss": 0.030043, "F": 0.1754, "G": 0.17483, "chunks": 179.61, "bytes_per_chunk": 5.701, "boundary_entropy": 0.04726, "router_noise": 0.0, "grad_norm": 0.2775, "lr": 0.0001858, "router_trainable": true, "sec_per_step": 0.16, "alpha": 0.03, "t": 6241.3}
{"event": "probe", "step": 36000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3004.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.172614, "cps": 23.46875, "mean_p": 0.172401, "boundary_entropy": 0.027248, "bpb": 1.545194}, "de": {"chunks": 3683.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.17937, "cps": 28.773438, "mean_p": 0.17789, "boundary_entropy": 0.029992, "bpb": 1.649567}, "vi": {"chunks": 4039.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.219094, "cps": 31.554688, "mean_p": 0.1701, "boundary_entropy": 0.04363, "bpb": 1.284298}, "tr": {"chunks": 3345.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.184776, "cps": 26.132812, "mean_p": 0.170863, "boundary_entropy": 0.033329, "bpb": 1.549907}, "fi": {"chunks": 3379.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179629, "cps": 26.398438, "mean_p": 0.174861, "boundary_entropy": 0.032194, "bpb": 1.701491}, "ru": {"chunks": 5713.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.301478, "cps": 44.632812, "mean_p": 0.166808, "boundary_entropy": 0.039978, "bpb": 0.97203}, "el": {"chunks": 6699.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.326414, "cps": 52.335938, "mean_p": 0.180392, "boundary_entropy": 0.045294, "bpb": 0.967915}, "hi": {"chunks": 7251.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.430966, "cps": 56.648438, "mean_p": 0.166312, "boundary_entropy": 0.116463, "bpb": 0.796148}, "ta": {"chunks": 8543.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.456967, "cps": 66.742188, "mean_p": 0.160093, "boundary_entropy": 0.105789, "bpb": 0.679435}, "th": {"chunks": 7970.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.495493, "cps": 62.265625, "mean_p": 0.156148, "boundary_entropy": 0.164357, "bpb": 0.775499}, "zh": {"chunks": 2724.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.461226, "cps": 21.28125, "mean_p": 0.174555, "boundary_entropy": 0.048403, "bpb": 2.312522}, "my": {"chunks": 9751.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.514619, "cps": 76.179688, "mean_p": 0.365866, "boundary_entropy": 0.55777, "bpb": 0.751005}}, "bpb": {"en": 1.54519, "de": 1.64957, "vi": 1.2843, "tr": 1.54991, "fi": 1.70149, "ru": 0.97203, "el": 0.96792, "hi": 0.79615, "ta": 0.67944, "th": 0.7755, "zh": 2.31252, "my": 0.75101}, "gini_cpc": 0.225185, "gini_cps": 0.239034, "bpic": 5.805979, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0391386623708527, "vi": 1.2692725945010668, "tr": 1.07045832131023, "fi": 1.0406399639670276, "ru": 1.7465426748504194, "el": 1.8910078186449644, "hi": 2.496703810471625, "ta": 2.647336385577009, "th": 2.87052575635346, "zh": 2.6720086053001686, "my": 2.9813294644506083}, "t": 6324.1}
{"event": "eval", "step": 36000, "phase": 2, "bpb": {"en": 1.54114, "de": 1.65854, "vi": 1.30137, "tr": 1.54274, "fi": 1.71762, "ru": 0.96947, "el": 0.96699, "hi": 0.80007, "ta": 0.68338, "th": 0.77323, "zh": 2.3129, "my": 0.75758}, "bytes_per_chunk": {"en": 5.618, "de": 5.684, "vi": 5.93, "tr": 5.78, "fi": 5.764, "ru": 5.829, "el": 5.76, "hi": 5.851, "ta": 5.849, "th": 5.92, "zh": 5.801, "my": 5.602}, "bpb_mean": 1.25209, "t": 6329.3}
{"event": "train", "step": 36000, "phase": 2, "objective": "global", "lm_loss": 0.94177, "bpb": 1.35868, "chunk_loss": 0.03007, "F": 0.1768, "G": 0.17703, "chunks": 181.05, "bytes_per_chunk": 5.656, "boundary_entropy": 0.0429, "router_noise": 0.0, "grad_norm": 0.2702, "lr": 0.0001829, "router_trainable": true, "sec_per_step": 0.176, "alpha": 0.03, "t": 6329.5}
{"event": "train", "step": 36500, "phase": 2, "objective": "global", "lm_loss": 0.9605, "bpb": 1.3857, "chunk_loss": 0.030057, "F": 0.17351, "G": 0.17574, "chunks": 177.67, "bytes_per_chunk": 5.763, "boundary_entropy": 0.05676, "router_noise": 0.0, "grad_norm": 0.2825, "lr": 0.0001801, "router_trainable": true, "sec_per_step": 0.159, "alpha": 0.03, "t": 6409.0}
{"event": "train", "step": 37000, "phase": 2, "objective": "global", "lm_loss": 1.01115, "bpb": 1.45879, "chunk_loss": 0.030033, "F": 0.16882, "G": 0.16902, "chunks": 172.88, "bytes_per_chunk": 5.923, "boundary_entropy": 0.05381, "router_noise": 0.0, "grad_norm": 0.2873, "lr": 0.0001773, "router_trainable": true, "sec_per_step": 0.16, "alpha": 0.03, "t": 6489.2}
{"event": "train", "step": 37500, "phase": 2, "objective": "global", "lm_loss": 0.90195, "bpb": 1.30124, "chunk_loss": 0.030055, "F": 0.17722, "G": 0.18203, "chunks": 181.47, "bytes_per_chunk": 5.643, "boundary_entropy": 0.07127, "router_noise": 0.0, "grad_norm": 0.2808, "lr": 0.0001744, "router_trainable": true, "sec_per_step": 0.167, "alpha": 0.03, "t": 6572.5}
{"event": "train", "step": 38000, "phase": 2, "objective": "global", "lm_loss": 0.99079, "bpb": 1.42941, "chunk_loss": 0.030054, "F": 0.17517, "G": 0.17492, "chunks": 179.38, "bytes_per_chunk": 5.709, "boundary_entropy": 0.04658, "router_noise": 0.0, "grad_norm": 0.2623, "lr": 0.0001716, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 6654.7}
{"event": "train", "step": 38500, "phase": 2, "objective": "global", "lm_loss": 0.87649, "bpb": 1.26451, "chunk_loss": 0.03005, "F": 0.17291, "G": 0.17829, "chunks": 177.06, "bytes_per_chunk": 5.783, "boundary_entropy": 0.06549, "router_noise": 0.0, "grad_norm": 0.2606, "lr": 0.0001687, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 6736.9}
{"event": "train", "step": 39000, "phase": 2, "objective": "global", "lm_loss": 0.95556, "bpb": 1.37858, "chunk_loss": 0.030063, "F": 0.17365, "G": 0.17577, "chunks": 177.81, "bytes_per_chunk": 5.759, "boundary_entropy": 0.05224, "router_noise": 0.0, "grad_norm": 0.2681, "lr": 0.0001659, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 6818.5}
{"event": "train", "step": 39500, "phase": 2, "objective": "global", "lm_loss": 0.93391, "bpb": 1.34735, "chunk_loss": 0.030054, "F": 0.17433, "G": 0.17295, "chunks": 178.52, "bytes_per_chunk": 5.736, "boundary_entropy": 0.05234, "router_noise": 0.0, "grad_norm": 0.269, "lr": 0.000163, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 6900.2}
{"event": "probe", "step": 40000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3005.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.172671, "cps": 23.476562, "mean_p": 0.172723, "boundary_entropy": 0.025955, "bpb": 1.532955}, "de": {"chunks": 3695.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.179954, "cps": 28.867188, "mean_p": 0.178303, "boundary_entropy": 0.027665, "bpb": 1.631847}, "vi": {"chunks": 4021.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.218118, "cps": 31.414062, "mean_p": 0.169583, "boundary_entropy": 0.042624, "bpb": 1.273008}, "tr": {"chunks": 3377.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.186544, "cps": 26.382812, "mean_p": 0.171682, "boundary_entropy": 0.032294, "bpb": 1.530324}, "fi": {"chunks": 3345.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.177821, "cps": 26.132812, "mean_p": 0.172807, "boundary_entropy": 0.027933, "bpb": 1.67753}, "ru": {"chunks": 5703.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.30095, "cps": 44.554688, "mean_p": 0.165913, "boundary_entropy": 0.039132, "bpb": 0.958073}, "el": {"chunks": 6681.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.325537, "cps": 52.195312, "mean_p": 0.179346, "boundary_entropy": 0.042502, "bpb": 0.95815}, "hi": {"chunks": 7605.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.452006, "cps": 59.414062, "mean_p": 0.168816, "boundary_entropy": 0.114535, "bpb": 0.782879}, "ta": {"chunks": 8745.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.467772, "cps": 68.320312, "mean_p": 0.162503, "boundary_entropy": 0.103884, "bpb": 0.66657}, "th": {"chunks": 8287.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.5152, "cps": 64.742188, "mean_p": 0.161143, "boundary_entropy": 0.163704, "bpb": 0.763506}, "zh": {"chunks": 2714.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.459533, "cps": 21.203125, "mean_p": 0.173673, "boundary_entropy": 0.044835, "bpb": 2.287917}, "my": {"chunks": 8619.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.454877, "cps": 67.335938, "mean_p": 0.34974, "boundary_entropy": 0.544336, "bpb": 0.740504}}, "bpb": {"en": 1.53295, "de": 1.63185, "vi": 1.27301, "tr": 1.53032, "fi": 1.67753, "ru": 0.95807, "el": 0.95815, "hi": 0.78288, "ta": 0.66657, "th": 0.76351, "zh": 2.28792, "my": 0.7405}, "gini_cpc": 0.221648, "gini_cps": 0.232079, "bpic": 5.832804, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.042177467982428, "vi": 1.2631955149337488, "tr": 1.0803392456715837, "fi": 1.029826071781026, "ru": 1.7429053424591379, "el": 1.8852991445091878, "hi": 2.6177236057783833, "ta": 2.7090310835244282, "th": 2.9837052431006863, "zh": 2.661313546166706, "my": 2.634348020346077}, "t": 6985.2}
{"event": "eval", "step": 40000, "phase": 2, "bpb": {"en": 1.52954, "de": 1.64066, "vi": 1.29004, "tr": 1.52295, "fi": 1.69204, "ru": 0.95556, "el": 0.95736, "hi": 0.7881, "ta": 0.67102, "th": 0.76312, "zh": 2.28959, "my": 0.7463}, "bytes_per_chunk": {"en": 5.609, "de": 5.658, "vi": 5.939, "tr": 5.712, "fi": 5.797, "ru": 5.83, "el": 5.783, "hi": 5.587, "ta": 5.691, "th": 5.619, "zh": 5.811, "my": 6.331}, "bpb_mean": 1.23719, "t": 6990.4}
{"event": "train", "step": 40000, "phase": 2, "objective": "global", "lm_loss": 0.93598, "bpb": 1.35033, "chunk_loss": 0.030053, "F": 0.17311, "G": 0.17586, "chunks": 177.27, "bytes_per_chunk": 5.777, "boundary_entropy": 0.04807, "router_noise": 0.0, "grad_norm": 0.2991, "lr": 0.0001601, "router_trainable": true, "sec_per_step": 0.181, "alpha": 0.03, "t": 6990.6}
{"event": "train", "step": 40500, "phase": 2, "objective": "global", "lm_loss": 0.98007, "bpb": 1.41394, "chunk_loss": 0.030046, "F": 0.1747, "G": 0.18009, "chunks": 178.89, "bytes_per_chunk": 5.724, "boundary_entropy": 0.06275, "router_noise": 0.0, "grad_norm": 0.2947, "lr": 0.0001573, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 7071.9}
{"event": "train", "step": 41000, "phase": 2, "objective": "global", "lm_loss": 0.89418, "bpb": 1.29003, "chunk_loss": 0.030099, "F": 0.17613, "G": 0.1864, "chunks": 180.36, "bytes_per_chunk": 5.678, "boundary_entropy": 0.08345, "router_noise": 0.0, "grad_norm": 0.2778, "lr": 0.0001545, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 7153.2}
{"event": "train", "step": 41500, "phase": 2, "objective": "global", "lm_loss": 0.95298, "bpb": 1.37486, "chunk_loss": 0.030076, "F": 0.17892, "G": 0.1774, "chunks": 183.22, "bytes_per_chunk": 5.589, "boundary_entropy": 0.05043, "router_noise": 0.0, "grad_norm": 0.2783, "lr": 0.0001516, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 7235.0}
{"event": "train", "step": 42000, "phase": 2, "objective": "global", "lm_loss": 0.96789, "bpb": 1.39637, "chunk_loss": 0.030031, "F": 0.17361, "G": 0.18828, "chunks": 177.78, "bytes_per_chunk": 5.76, "boundary_entropy": 0.0835, "router_noise": 0.0, "grad_norm": 0.2966, "lr": 0.0001488, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 7316.6}
{"event": "train", "step": 42500, "phase": 2, "objective": "global", "lm_loss": 0.95645, "bpb": 1.37986, "chunk_loss": 0.030065, "F": 0.17424, "G": 0.18002, "chunks": 178.42, "bytes_per_chunk": 5.739, "boundary_entropy": 0.05671, "router_noise": 0.0, "grad_norm": 0.3085, "lr": 0.0001459, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 7397.5}
{"event": "train", "step": 43000, "phase": 2, "objective": "global", "lm_loss": 0.97476, "bpb": 1.40628, "chunk_loss": 0.03007, "F": 0.177, "G": 0.18046, "chunks": 181.25, "bytes_per_chunk": 5.65, "boundary_entropy": 0.04854, "router_noise": 0.0, "grad_norm": 0.2852, "lr": 0.0001431, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 7478.3}
{"event": "train", "step": 43500, "phase": 2, "objective": "global", "lm_loss": 0.95941, "bpb": 1.38413, "chunk_loss": 0.030063, "F": 0.17566, "G": 0.17477, "chunks": 179.88, "bytes_per_chunk": 5.693, "boundary_entropy": 0.05297, "router_noise": 0.0, "grad_norm": 0.2861, "lr": 0.0001403, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 7559.4}
{"event": "probe", "step": 44000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3012.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.173074, "cps": 23.53125, "mean_p": 0.173338, "boundary_entropy": 0.026842, "bpb": 1.519243}, "de": {"chunks": 3717.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.181026, "cps": 29.039062, "mean_p": 0.17954, "boundary_entropy": 0.027558, "bpb": 1.620775}, "vi": {"chunks": 4053.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.219854, "cps": 31.664062, "mean_p": 0.170911, "boundary_entropy": 0.045172, "bpb": 1.255969}, "tr": {"chunks": 3387.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.187096, "cps": 26.460938, "mean_p": 0.172688, "boundary_entropy": 0.030832, "bpb": 1.519815}, "fi": {"chunks": 3381.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179735, "cps": 26.414062, "mean_p": 0.174463, "boundary_entropy": 0.027364, "bpb": 1.656904}, "ru": {"chunks": 5711.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.301372, "cps": 44.617188, "mean_p": 0.166101, "boundary_entropy": 0.037372, "bpb": 0.950045}, "el": {"chunks": 6673.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.325147, "cps": 52.132812, "mean_p": 0.177838, "boundary_entropy": 0.045126, "bpb": 0.947117}, "hi": {"chunks": 7417.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.440832, "cps": 57.945312, "mean_p": 0.167533, "boundary_entropy": 0.115126, "bpb": 0.772509}, "ta": {"chunks": 8335.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.445841, "cps": 65.117188, "mean_p": 0.153438, "boundary_entropy": 0.108458, "bpb": 0.664894}, "th": {"chunks": 7793.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.484489, "cps": 60.882812, "mean_p": 0.153644, "boundary_entropy": 0.156577, "bpb": 0.755336}, "zh": {"chunks": 2721.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.460718, "cps": 21.257812, "mean_p": 0.172688, "boundary_entropy": 0.041115, "bpb": 2.269432}, "my": {"chunks": 9655.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.509552, "cps": 75.429688, "mean_p": 0.362262, "boundary_entropy": 0.55174, "bpb": 0.730837}}, "bpb": {"en": 1.51924, "de": 1.62078, "vi": 1.25597, "tr": 1.51982, "fi": 1.6569, "ru": 0.95005, "el": 0.94712, "hi": 0.77251, "ta": 0.66489, "th": 0.75534, "zh": 2.26943, "my": 0.73084}, "gini_cpc": 0.221349, "gini_cps": 0.235038, "bpic": 5.827667, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0459461040559914, "vi": 1.270289225522645, "tr": 1.0810201716472783, "fi": 1.0384902895461905, "ru": 1.7412939797538078, "el": 1.8786653803826743, "hi": 2.547078668979357, "ta": 2.5760203369458776, "th": 2.7993213964162003, "zh": 2.661976712182762, "my": 2.9441372676909}, "t": 7644.0}
{"event": "eval", "step": 44000, "phase": 2, "bpb": {"en": 1.51527, "de": 1.62908, "vi": 1.27353, "tr": 1.51236, "fi": 1.67257, "ru": 0.94738, "el": 0.94621, "hi": 0.77819, "ta": 0.66981, "th": 0.7545, "zh": 2.2714, "my": 0.73621}, "bytes_per_chunk": {"en": 5.588, "de": 5.646, "vi": 5.932, "tr": 5.705, "fi": 5.765, "ru": 5.821, "el": 5.782, "hi": 5.725, "ta": 5.984, "th": 5.929, "zh": 5.817, "my": 5.65}, "bpb_mean": 1.22554, "t": 7649.2}
{"event": "train", "step": 44000, "phase": 2, "objective": "global", "lm_loss": 0.87096, "bpb": 1.25654, "chunk_loss": 0.030042, "F": 0.17361, "G": 0.17842, "chunks": 177.78, "bytes_per_chunk": 5.76, "boundary_entropy": 0.06077, "router_noise": 0.0, "grad_norm": 0.2708, "lr": 0.0001375, "router_trainable": true, "sec_per_step": 0.18, "alpha": 0.03, "t": 7649.4}
{"event": "train", "step": 44500, "phase": 2, "objective": "global", "lm_loss": 0.94925, "bpb": 1.36947, "chunk_loss": 0.030092, "F": 0.17451, "G": 0.18285, "chunks": 178.7, "bytes_per_chunk": 5.73, "boundary_entropy": 0.06098, "router_noise": 0.0, "grad_norm": 0.291, "lr": 0.0001347, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 7730.4}
{"event": "train", "step": 45000, "phase": 2, "objective": "global", "lm_loss": 0.95732, "bpb": 1.38113, "chunk_loss": 0.030022, "F": 0.17329, "G": 0.17764, "chunks": 177.45, "bytes_per_chunk": 5.771, "boundary_entropy": 0.05756, "router_noise": 0.0, "grad_norm": 0.2796, "lr": 0.0001319, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 7812.1}
{"event": "train", "step": 45500, "phase": 2, "objective": "global", "lm_loss": 0.91289, "bpb": 1.31702, "chunk_loss": 0.030033, "F": 0.17441, "G": 0.17464, "chunks": 178.59, "bytes_per_chunk": 5.734, "boundary_entropy": 0.06048, "router_noise": 0.0, "grad_norm": 0.2843, "lr": 0.0001292, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 7893.5}
{"event": "train", "step": 46000, "phase": 2, "objective": "global", "lm_loss": 0.91014, "bpb": 1.31305, "chunk_loss": 0.030054, "F": 0.17603, "G": 0.17479, "chunks": 180.25, "bytes_per_chunk": 5.681, "boundary_entropy": 0.0465, "router_noise": 0.0, "grad_norm": 0.2995, "lr": 0.0001264, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 7974.8}
{"event": "train", "step": 46500, "phase": 2, "objective": "global", "lm_loss": 0.90777, "bpb": 1.30964, "chunk_loss": 0.030074, "F": 0.17841, "G": 0.17778, "chunks": 182.69, "bytes_per_chunk": 5.605, "boundary_entropy": 0.05017, "router_noise": 0.0, "grad_norm": 0.2937, "lr": 0.0001237, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 8056.7}
{"event": "train", "step": 47000, "phase": 2, "objective": "global", "lm_loss": 0.89125, "bpb": 1.2858, "chunk_loss": 0.030027, "F": 0.17117, "G": 0.1729, "chunks": 175.28, "bytes_per_chunk": 5.842, "boundary_entropy": 0.0566, "router_noise": 0.0, "grad_norm": 0.2866, "lr": 0.000121, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 8138.6}
{"event": "train", "step": 47500, "phase": 2, "objective": "global", "lm_loss": 0.88852, "bpb": 1.28187, "chunk_loss": 0.030038, "F": 0.17384, "G": 0.18093, "chunks": 178.02, "bytes_per_chunk": 5.752, "boundary_entropy": 0.07241, "router_noise": 0.0, "grad_norm": 0.2864, "lr": 0.0001183, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 8220.7}
{"event": "probe", "step": 48000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3019.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.173476, "cps": 23.585938, "mean_p": 0.173217, "boundary_entropy": 0.025257, "bpb": 1.507932}, "de": {"chunks": 3709.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.180636, "cps": 28.976562, "mean_p": 0.178699, "boundary_entropy": 0.026567, "bpb": 1.605414}, "vi": {"chunks": 4049.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.219637, "cps": 31.632812, "mean_p": 0.170556, "boundary_entropy": 0.041202, "bpb": 1.244112}, "tr": {"chunks": 3443.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.190189, "cps": 26.898438, "mean_p": 0.174731, "boundary_entropy": 0.029602, "bpb": 1.500269}, "fi": {"chunks": 3395.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.18048, "cps": 26.523438, "mean_p": 0.175103, "boundary_entropy": 0.027515, "bpb": 1.640979}, "ru": {"chunks": 5707.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.301161, "cps": 44.585938, "mean_p": 0.165258, "boundary_entropy": 0.037395, "bpb": 0.9408}, "el": {"chunks": 6684.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.325683, "cps": 52.21875, "mean_p": 0.177887, "boundary_entropy": 0.043395, "bpb": 0.933096}, "hi": {"chunks": 7457.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.44321, "cps": 58.257812, "mean_p": 0.165812, "boundary_entropy": 0.109447, "bpb": 0.760314}, "ta": {"chunks": 8598.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.459909, "cps": 67.171875, "mean_p": 0.158198, "boundary_entropy": 0.098762, "bpb": 0.652622}, "th": {"chunks": 8060.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.501088, "cps": 62.96875, "mean_p": 0.15751, "boundary_entropy": 0.152893, "bpb": 0.745755}, "zh": {"chunks": 2730.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.462242, "cps": 21.328125, "mean_p": 0.173355, "boundary_entropy": 0.039305, "bpb": 2.252922}, "my": {"chunks": 9534.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.503167, "cps": 74.484375, "mean_p": 0.334813, "boundary_entropy": 0.524275, "bpb": 0.718768}}, "bpb": {"en": 1.50793, "de": 1.60541, "vi": 1.24411, "tr": 1.50027, "fi": 1.64098, "ru": 0.9408, "el": 0.9331, "hi": 0.76031, "ta": 0.65262, "th": 0.74576, "zh": 2.25292, "my": 0.71877}, "gini_cpc": 0.222606, "gini_cps": 0.236453, "bpic": 5.78114, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0412749803687347, "vi": 1.2660931000867572, "tr": 1.0963456012087325, "fi": 1.0403725917874618, "ru": 1.736039751756903, "el": 1.877399090949543, "hi": 2.5548774748534173, "ta": 2.65114192976443, "th": 2.888517371536584, "zh": 2.6645888826684865, "my": 2.9004993915461736}, "t": 8306.4}
{"event": "eval", "step": 48000, "phase": 2, "bpb": {"en": 1.50372, "de": 1.61378, "vi": 1.26272, "tr": 1.49344, "fi": 1.65789, "ru": 0.93802, "el": 0.93252, "hi": 0.76583, "ta": 0.65731, "th": 0.7446, "zh": 2.25568, "my": 0.72421}, "bytes_per_chunk": {"en": 5.588, "de": 5.644, "vi": 5.94, "tr": 5.637, "fi": 5.755, "ru": 5.834, "el": 5.768, "hi": 5.676, "ta": 5.804, "th": 5.798, "zh": 5.812, "my": 5.781}, "bpb_mean": 1.21248, "t": 8311.9}
{"event": "train", "step": 48000, "phase": 2, "objective": "global", "lm_loss": 0.94574, "bpb": 1.36441, "chunk_loss": 0.030062, "F": 0.17491, "G": 0.18108, "chunks": 179.11, "bytes_per_chunk": 5.717, "boundary_entropy": 0.07021, "router_noise": 0.0, "grad_norm": 0.2917, "lr": 0.0001157, "router_trainable": true, "sec_per_step": 0.183, "alpha": 0.03, "t": 8312.1}
{"event": "train", "step": 48500, "phase": 2, "objective": "global", "lm_loss": 0.93595, "bpb": 1.35029, "chunk_loss": 0.030049, "F": 0.17432, "G": 0.17657, "chunks": 178.5, "bytes_per_chunk": 5.737, "boundary_entropy": 0.0465, "router_noise": 0.0, "grad_norm": 0.2984, "lr": 0.000113, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 8393.3}
{"event": "train", "step": 49000, "phase": 2, "objective": "global", "lm_loss": 0.91919, "bpb": 1.32611, "chunk_loss": 0.030034, "F": 0.17488, "G": 0.17747, "chunks": 179.08, "bytes_per_chunk": 5.718, "boundary_entropy": 0.04797, "router_noise": 0.0, "grad_norm": 0.3308, "lr": 0.0001104, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 8474.2}
{"event": "train", "step": 49500, "phase": 2, "objective": "global", "lm_loss": 0.88834, "bpb": 1.28161, "chunk_loss": 0.030055, "F": 0.1748, "G": 0.17406, "chunks": 179.0, "bytes_per_chunk": 5.721, "boundary_entropy": 0.04213, "router_noise": 0.0, "grad_norm": 0.2917, "lr": 0.0001078, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 8555.4}
{"event": "train", "step": 50000, "phase": 2, "objective": "global", "lm_loss": 0.88915, "bpb": 1.28277, "chunk_loss": 0.030072, "F": 0.17799, "G": 0.18112, "chunks": 182.27, "bytes_per_chunk": 5.618, "boundary_entropy": 0.06044, "router_noise": 0.0, "grad_norm": 0.323, "lr": 0.0001052, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 8637.1}
{"event": "train", "step": 50500, "phase": 2, "objective": "global", "lm_loss": 0.89416, "bpb": 1.29, "chunk_loss": 0.030054, "F": 0.17572, "G": 0.1812, "chunks": 179.94, "bytes_per_chunk": 5.691, "boundary_entropy": 0.05555, "router_noise": 0.0, "grad_norm": 0.294, "lr": 0.0001027, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 8719.0}
{"event": "train", "step": 51000, "phase": 2, "objective": "global", "lm_loss": 0.8812, "bpb": 1.27131, "chunk_loss": 0.030065, "F": 0.17671, "G": 0.18143, "chunks": 180.95, "bytes_per_chunk": 5.659, "boundary_entropy": 0.05674, "router_noise": 0.0, "grad_norm": 0.2893, "lr": 0.0001002, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 8800.4}
{"event": "train", "step": 51500, "phase": 2, "objective": "global", "lm_loss": 0.89808, "bpb": 1.29565, "chunk_loss": 0.030064, "F": 0.17413, "G": 0.17505, "chunks": 178.31, "bytes_per_chunk": 5.743, "boundary_entropy": 0.05524, "router_noise": 0.0, "grad_norm": 0.2774, "lr": 9.77e-05, "router_trainable": true, "sec_per_step": 0.166, "alpha": 0.03, "t": 8883.4}
{"event": "probe", "step": 52000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3026.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.173878, "cps": 23.640625, "mean_p": 0.172951, "boundary_entropy": 0.026594, "bpb": 1.495753}, "de": {"chunks": 3713.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.180831, "cps": 29.007812, "mean_p": 0.178542, "boundary_entropy": 0.024387, "bpb": 1.593821}, "vi": {"chunks": 4058.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.220125, "cps": 31.703125, "mean_p": 0.170824, "boundary_entropy": 0.04062, "bpb": 1.231363}, "tr": {"chunks": 3403.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.18798, "cps": 26.585938, "mean_p": 0.172605, "boundary_entropy": 0.027652, "bpb": 1.489282}, "fi": {"chunks": 3378.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179576, "cps": 26.390625, "mean_p": 0.173676, "boundary_entropy": 0.025551, "bpb": 1.624005}, "ru": {"chunks": 5712.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.301425, "cps": 44.625, "mean_p": 0.165438, "boundary_entropy": 0.036282, "bpb": 0.929554}, "el": {"chunks": 6694.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.326171, "cps": 52.296875, "mean_p": 0.177948, "boundary_entropy": 0.041895, "bpb": 0.925179}, "hi": {"chunks": 7679.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.456404, "cps": 59.992188, "mean_p": 0.169262, "boundary_entropy": 0.108612, "bpb": 0.750556}, "ta": {"chunks": 8622.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.461193, "cps": 67.359375, "mean_p": 0.15927, "boundary_entropy": 0.098604, "bpb": 0.644124}, "th": {"chunks": 7930.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.493006, "cps": 61.953125, "mean_p": 0.154835, "boundary_entropy": 0.153173, "bpb": 0.737609}, "zh": {"chunks": 2724.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.461226, "cps": 21.28125, "mean_p": 0.172336, "boundary_entropy": 0.036072, "bpb": 2.23782}, "my": {"chunks": 8188.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.43213, "cps": 63.96875, "mean_p": 0.347295, "boundary_entropy": 0.548623, "bpb": 0.713154}}, "bpb": {"en": 1.49575, "de": 1.59382, "vi": 1.23136, "tr": 1.48928, "fi": 1.62401, "ru": 0.92955, "el": 0.92518, "hi": 0.75056, "ta": 0.64412, "th": 0.73761, "zh": 2.23782, "my": 0.71315}, "gini_cpc": 0.215342, "gini_cps": 0.224148, "bpic": 5.892809, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.039986588094821, "vi": 1.265971991049096, "tr": 1.0811018093242535, "fi": 1.0327684407068287, "ru": 1.7335412528534584, "el": 1.8758584302705017, "hi": 2.62485181915229, "ta": 2.652392220970895, "th": 2.8353541917617515, "zh": 2.652582237383247, "my": 2.485247550571844}, "t": 8969.0}
{"event": "eval", "step": 52000, "phase": 2, "bpb": {"en": 1.49168, "de": 1.60196, "vi": 1.24993, "tr": 1.48099, "fi": 1.64155, "ru": 0.92785, "el": 0.92471, "hi": 0.75699, "ta": 0.64893, "th": 0.73695, "zh": 2.24007, "my": 0.71975}, "bytes_per_chunk": {"en": 5.588, "de": 5.665, "vi": 5.92, "tr": 5.688, "fi": 5.784, "ru": 5.83, "el": 5.782, "hi": 5.547, "ta": 5.813, "th": 5.865, "zh": 5.816, "my": 6.705}, "bpb_mean": 1.20178, "t": 8974.4}
{"event": "train", "step": 52000, "phase": 2, "objective": "global", "lm_loss": 0.96118, "bpb": 1.38668, "chunk_loss": 0.030044, "F": 0.17628, "G": 0.17765, "chunks": 180.52, "bytes_per_chunk": 5.673, "boundary_entropy": 0.05151, "router_noise": 0.0, "grad_norm": 0.3192, "lr": 9.52e-05, "router_trainable": true, "sec_per_step": 0.182, "alpha": 0.03, "t": 8974.6}
{"event": "train", "step": 52500, "phase": 2, "objective": "global", "lm_loss": 0.90068, "bpb": 1.29941, "chunk_loss": 0.03008, "F": 0.17661, "G": 0.18595, "chunks": 180.84, "bytes_per_chunk": 5.662, "boundary_entropy": 0.06535, "router_noise": 0.0, "grad_norm": 0.2952, "lr": 9.28e-05, "router_trainable": true, "sec_per_step": 0.166, "alpha": 0.03, "t": 9057.7}
{"event": "train", "step": 53000, "phase": 2, "objective": "global", "lm_loss": 0.98846, "bpb": 1.42605, "chunk_loss": 0.030088, "F": 0.17746, "G": 0.17709, "chunks": 181.72, "bytes_per_chunk": 5.635, "boundary_entropy": 0.04082, "router_noise": 0.0, "grad_norm": 0.314, "lr": 9.04e-05, "router_trainable": true, "sec_per_step": 0.166, "alpha": 0.03, "t": 9140.7}
{"event": "phase_transition", "step": 53410, "phase": 3, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 9207.1}
{"event": "train", "step": 53500, "phase": 3, "objective": "global", "lm_loss": 0.93036, "bpb": 1.34223, "chunk_loss": 0.030054, "F": 0.17325, "G": 0.17603, "chunks": 177.41, "bytes_per_chunk": 5.772, "boundary_entropy": 0.04989, "router_noise": 0.0, "grad_norm": 0.3082, "lr": 8.8e-05, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 9221.7}
{"event": "train", "step": 54000, "phase": 3, "objective": "global", "lm_loss": 0.96088, "bpb": 1.38626, "chunk_loss": 0.030065, "F": 0.17632, "G": 0.18387, "chunks": 180.55, "bytes_per_chunk": 5.672, "boundary_entropy": 0.06519, "router_noise": 0.0, "grad_norm": 0.3108, "lr": 8.57e-05, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 9303.2}
{"event": "train", "step": 54500, "phase": 3, "objective": "global", "lm_loss": 0.90515, "bpb": 1.30585, "chunk_loss": 0.030065, "F": 0.17619, "G": 0.17805, "chunks": 180.42, "bytes_per_chunk": 5.676, "boundary_entropy": 0.04203, "router_noise": 0.0, "grad_norm": 0.2936, "lr": 8.34e-05, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 9384.9}
{"event": "train", "step": 55000, "phase": 3, "objective": "global", "lm_loss": 0.83411, "bpb": 1.20337, "chunk_loss": 0.030058, "F": 0.17517, "G": 0.17566, "chunks": 179.38, "bytes_per_chunk": 5.709, "boundary_entropy": 0.05506, "router_noise": 0.0, "grad_norm": 0.2962, "lr": 8.11e-05, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 9465.7}
{"event": "train", "step": 55500, "phase": 3, "objective": "global", "lm_loss": 0.93186, "bpb": 1.34439, "chunk_loss": 0.030069, "F": 0.17549, "G": 0.18492, "chunks": 179.7, "bytes_per_chunk": 5.698, "boundary_entropy": 0.07395, "router_noise": 0.0, "grad_norm": 0.3119, "lr": 7.89e-05, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 9547.2}
{"event": "probe", "step": 56000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3024.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.173763, "cps": 23.625, "mean_p": 0.173322, "boundary_entropy": 0.023664, "bpb": 1.485871}, "de": {"chunks": 3724.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.181367, "cps": 29.09375, "mean_p": 0.179328, "boundary_entropy": 0.024136, "bpb": 1.583016}, "vi": {"chunks": 4075.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.221047, "cps": 31.835938, "mean_p": 0.171411, "boundary_entropy": 0.037975, "bpb": 1.220458}, "tr": {"chunks": 3402.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.187925, "cps": 26.578125, "mean_p": 0.172733, "boundary_entropy": 0.027216, "bpb": 1.475486}, "fi": {"chunks": 3389.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180161, "cps": 26.476562, "mean_p": 0.174709, "boundary_entropy": 0.024809, "bpb": 1.605894}, "ru": {"chunks": 5714.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.30153, "cps": 44.640625, "mean_p": 0.165647, "boundary_entropy": 0.035346, "bpb": 0.923069}, "el": {"chunks": 6705.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.326707, "cps": 52.382812, "mean_p": 0.179064, "boundary_entropy": 0.038678, "bpb": 0.912368}, "hi": {"chunks": 7465.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.443685, "cps": 58.320312, "mean_p": 0.166827, "boundary_entropy": 0.108876, "bpb": 0.743286}, "ta": {"chunks": 8440.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.451458, "cps": 65.9375, "mean_p": 0.154774, "boundary_entropy": 0.099781, "bpb": 0.639634}, "th": {"chunks": 7976.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.495866, "cps": 62.3125, "mean_p": 0.155779, "boundary_entropy": 0.153122, "bpb": 0.726658}, "zh": {"chunks": 2721.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.460718, "cps": 21.257812, "mean_p": 0.172256, "boundary_entropy": 0.035048, "bpb": 2.220673}, "my": {"chunks": 8986.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.474245, "cps": 70.203125, "mean_p": 0.337976, "boundary_entropy": 0.528691, "bpb": 0.697886}}, "bpb": {"en": 1.48587, "de": 1.58302, "vi": 1.22046, "tr": 1.47549, "fi": 1.60589, "ru": 0.92307, "el": 0.91237, "hi": 0.74329, "ta": 0.63963, "th": 0.72666, "zh": 2.22067, "my": 0.69789}, "gini_cpc": 0.217096, "gini_cps": 0.22943, "bpic": 5.848448, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0437574744178748, "vi": 1.2721162621996633, "tr": 1.08149892283047, "fi": 1.0368167835773228, "ru": 1.7352951585207523, "el": 1.8801836441344553, "hi": 2.553389525696361, "ta": 2.5981206247320183, "th": 2.8536875033921976, "zh": 2.6514133125312433, "my": 2.729262899129112}, "t": 9631.9}
{"event": "eval", "step": 56000, "phase": 3, "bpb": {"en": 1.48058, "de": 1.59175, "vi": 1.23779, "tr": 1.46763, "fi": 1.62457, "ru": 0.92003, "el": 0.91233, "hi": 0.7496, "ta": 0.64413, "th": 0.7265, "zh": 2.22266, "my": 0.70458}, "bytes_per_chunk": {"en": 5.583, "de": 5.651, "vi": 5.899, "tr": 5.675, "fi": 5.764, "ru": 5.826, "el": 5.755, "hi": 5.676, "ta": 5.89, "th": 5.814, "zh": 5.815, "my": 6.116}, "bpb_mean": 1.19018, "t": 9637.1}
{"event": "train", "step": 56000, "phase": 3, "objective": "global", "lm_loss": 0.90903, "bpb": 1.31146, "chunk_loss": 0.030033, "F": 0.1723, "G": 0.17092, "chunks": 176.44, "bytes_per_chunk": 5.804, "boundary_entropy": 0.04515, "router_noise": 0.0, "grad_norm": 0.3093, "lr": 7.68e-05, "router_trainable": true, "sec_per_step": 0.18, "alpha": 0.03, "t": 9637.3}
{"event": "train", "step": 56500, "phase": 3, "objective": "global", "lm_loss": 0.91772, "bpb": 1.32398, "chunk_loss": 0.030057, "F": 0.17639, "G": 0.1763, "chunks": 180.62, "bytes_per_chunk": 5.669, "boundary_entropy": 0.05346, "router_noise": 0.0, "grad_norm": 0.3048, "lr": 7.46e-05, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 9718.9}
{"event": "train", "step": 57000, "phase": 3, "objective": "global", "lm_loss": 0.89038, "bpb": 1.28454, "chunk_loss": 0.030036, "F": 0.17574, "G": 0.17665, "chunks": 179.95, "bytes_per_chunk": 5.69, "boundary_entropy": 0.05177, "router_noise": 0.0, "grad_norm": 0.3037, "lr": 7.25e-05, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 9799.8}
{"event": "train", "step": 57500, "phase": 3, "objective": "global", "lm_loss": 0.86112, "bpb": 1.24233, "chunk_loss": 0.030044, "F": 0.17247, "G": 0.17169, "chunks": 176.61, "bytes_per_chunk": 5.798, "boundary_entropy": 0.04181, "router_noise": 0.0, "grad_norm": 0.3215, "lr": 7.05e-05, "router_trainable": true, "sec_per_step": 0.165, "alpha": 0.03, "t": 9882.2}
{"event": "train", "step": 58000, "phase": 3, "objective": "global", "lm_loss": 0.83737, "bpb": 1.20808, "chunk_loss": 0.030055, "F": 0.17599, "G": 0.17368, "chunks": 180.22, "bytes_per_chunk": 5.682, "boundary_entropy": 0.04724, "router_noise": 0.0, "grad_norm": 0.2985, "lr": 6.84e-05, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 9963.5}
{"event": "train", "step": 58500, "phase": 3, "objective": "global", "lm_loss": 0.81966, "bpb": 1.18252, "chunk_loss": 0.030026, "F": 0.17229, "G": 0.17417, "chunks": 176.42, "bytes_per_chunk": 5.804, "boundary_entropy": 0.07144, "router_noise": 0.0, "grad_norm": 0.3049, "lr": 6.65e-05, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 10044.8}
{"event": "train", "step": 59000, "phase": 3, "objective": "global", "lm_loss": 0.86166, "bpb": 1.24312, "chunk_loss": 0.030048, "F": 0.17529, "G": 0.17423, "chunks": 179.5, "bytes_per_chunk": 5.705, "boundary_entropy": 0.03489, "router_noise": 0.0, "grad_norm": 0.3108, "lr": 6.45e-05, "router_trainable": true, "sec_per_step": 0.165, "alpha": 0.03, "t": 10127.1}
{"event": "train", "step": 59500, "phase": 3, "objective": "global", "lm_loss": 0.86336, "bpb": 1.24557, "chunk_loss": 0.030027, "F": 0.17108, "G": 0.17199, "chunks": 175.19, "bytes_per_chunk": 5.845, "boundary_entropy": 0.04954, "router_noise": 0.0, "grad_norm": 0.3161, "lr": 6.27e-05, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 10208.5}
{"event": "probe", "step": 60000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3011.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.173016, "cps": 23.523438, "mean_p": 0.172299, "boundary_entropy": 0.022929, "bpb": 1.476854}, "de": {"chunks": 3686.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.179516, "cps": 28.796875, "mean_p": 0.177087, "boundary_entropy": 0.022853, "bpb": 1.571906}, "vi": {"chunks": 4079.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.221264, "cps": 31.867188, "mean_p": 0.171112, "boundary_entropy": 0.037518, "bpb": 1.20772}, "tr": {"chunks": 3390.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.187262, "cps": 26.484375, "mean_p": 0.171526, "boundary_entropy": 0.025405, "bpb": 1.465355}, "fi": {"chunks": 3359.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.178566, "cps": 26.242188, "mean_p": 0.172644, "boundary_entropy": 0.022795, "bpb": 1.591632}, "ru": {"chunks": 5704.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.301003, "cps": 44.5625, "mean_p": 0.164618, "boundary_entropy": 0.032591, "bpb": 0.915447}, "el": {"chunks": 6681.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.325537, "cps": 52.195312, "mean_p": 0.17775, "boundary_entropy": 0.037539, "bpb": 0.904301}, "hi": {"chunks": 7278.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.432571, "cps": 56.859375, "mean_p": 0.160664, "boundary_entropy": 0.10987, "bpb": 0.737222}, "ta": {"chunks": 8630.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.461621, "cps": 67.421875, "mean_p": 0.157084, "boundary_entropy": 0.096767, "bpb": 0.629497}, "th": {"chunks": 7993.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.496923, "cps": 62.445312, "mean_p": 0.154284, "boundary_entropy": 0.149036, "bpb": 0.720018}, "zh": {"chunks": 2720.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.460549, "cps": 21.25, "mean_p": 0.171518, "boundary_entropy": 0.031676, "bpb": 2.20653}, "my": {"chunks": 8861.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.467648, "cps": 69.226562, "mean_p": 0.320587, "boundary_entropy": 0.510984, "bpb": 0.690348}}, "bpb": {"en": 1.47685, "de": 1.57191, "vi": 1.20772, "tr": 1.46536, "fi": 1.59163, "ru": 0.91545, "el": 0.9043, "hi": 0.73722, "ta": 0.6295, "th": 0.72002, "zh": 2.20653, "my": 0.69035}, "gini_cpc": 0.217999, "gini_cps": 0.230163, "bpic": 5.868929, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0375673295062537, "vi": 1.2788627216884982, "tr": 1.0823370156523247, "fi": 1.0320755390884109, "ru": 1.7397372694140831, "el": 1.8815423212388274, "hi": 2.500174624692127, "ta": 2.6680790387106064, "th": 2.8721168997274216, "zh": 2.661882162964266, "my": 2.702917095833327}, "t": 10293.1}
{"event": "eval", "step": 60000, "phase": 3, "bpb": {"en": 1.47204, "de": 1.58032, "vi": 1.22603, "tr": 1.45754, "fi": 1.61046, "ru": 0.91236, "el": 0.90403, "hi": 0.74378, "ta": 0.63458, "th": 0.72042, "zh": 2.20898, "my": 0.69636}, "bytes_per_chunk": {"en": 5.595, "de": 5.684, "vi": 5.892, "tr": 5.697, "fi": 5.797, "ru": 5.832, "el": 5.774, "hi": 5.8, "ta": 5.768, "th": 5.834, "zh": 5.816, "my": 6.181}, "bpb_mean": 1.18057, "t": 10299.3}
{"event": "train", "step": 60000, "phase": 3, "objective": "global", "lm_loss": 0.88593, "bpb": 1.27813, "chunk_loss": 0.030049, "F": 0.17281, "G": 0.17208, "chunks": 176.95, "bytes_per_chunk": 5.787, "boundary_entropy": 0.03687, "router_noise": 0.0, "grad_norm": 0.321, "lr": 6.08e-05, "router_trainable": true, "sec_per_step": 0.182, "alpha": 0.03, "t": 10299.4}
{"event": "train", "step": 60500, "phase": 3, "objective": "global", "lm_loss": 0.85181, "bpb": 1.2289, "chunk_loss": 0.030044, "F": 0.17543, "G": 0.17373, "chunks": 179.64, "bytes_per_chunk": 5.7, "boundary_entropy": 0.04162, "router_noise": 0.0, "grad_norm": 0.3093, "lr": 5.9e-05, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 10380.7}
{"event": "train", "step": 61000, "phase": 3, "objective": "global", "lm_loss": 0.88193, "bpb": 1.27236, "chunk_loss": 0.030039, "F": 0.1725, "G": 0.17277, "chunks": 176.64, "bytes_per_chunk": 5.797, "boundary_entropy": 0.05211, "router_noise": 0.0, "grad_norm": 0.3067, "lr": 5.73e-05, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 10462.2}
{"event": "train", "step": 61500, "phase": 3, "objective": "global", "lm_loss": 0.9223, "bpb": 1.3306, "chunk_loss": 0.030057, "F": 0.17735, "G": 0.18043, "chunks": 181.61, "bytes_per_chunk": 5.638, "boundary_entropy": 0.05759, "router_noise": 0.0, "grad_norm": 0.3252, "lr": 5.56e-05, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 10543.5}
{"event": "train", "step": 62000, "phase": 3, "objective": "global", "lm_loss": 0.94763, "bpb": 1.36714, "chunk_loss": 0.030052, "F": 0.17607, "G": 0.17655, "chunks": 180.3, "bytes_per_chunk": 5.68, "boundary_entropy": 0.04389, "router_noise": 0.0, "grad_norm": 0.3359, "lr": 5.39e-05, "router_trainable": true, "sec_per_step": 0.159, "alpha": 0.03, "t": 10622.8}
{"event": "train", "step": 62500, "phase": 3, "objective": "global", "lm_loss": 0.90945, "bpb": 1.31206, "chunk_loss": 0.03007, "F": 0.1778, "G": 0.18112, "chunks": 182.06, "bytes_per_chunk": 5.624, "boundary_entropy": 0.05424, "router_noise": 0.0, "grad_norm": 0.336, "lr": 5.23e-05, "router_trainable": true, "sec_per_step": 0.159, "alpha": 0.03, "t": 10702.4}
{"event": "train", "step": 63000, "phase": 3, "objective": "global", "lm_loss": 0.90835, "bpb": 1.31047, "chunk_loss": 0.030052, "F": 0.17265, "G": 0.17154, "chunks": 176.8, "bytes_per_chunk": 5.792, "boundary_entropy": 0.03363, "router_noise": 0.0, "grad_norm": 0.3181, "lr": 5.08e-05, "router_trainable": true, "sec_per_step": 0.159, "alpha": 0.03, "t": 10782.1}
{"event": "train", "step": 63500, "phase": 3, "objective": "global", "lm_loss": 0.89434, "bpb": 1.29025, "chunk_loss": 0.030105, "F": 0.1727, "G": 0.17255, "chunks": 176.84, "bytes_per_chunk": 5.79, "boundary_entropy": 0.05463, "router_noise": 0.0, "grad_norm": 0.3382, "lr": 4.93e-05, "router_trainable": true, "sec_per_step": 0.161, "alpha": 0.03, "t": 10862.8}
{"event": "probe", "step": 64000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3018.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.173418, "cps": 23.578125, "mean_p": 0.172601, "boundary_entropy": 0.023411, "bpb": 1.468626}, "de": {"chunks": 3708.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.180587, "cps": 28.96875, "mean_p": 0.178316, "boundary_entropy": 0.02267, "bpb": 1.561755}, "vi": {"chunks": 4127.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.223868, "cps": 32.242188, "mean_p": 0.173087, "boundary_entropy": 0.037118, "bpb": 1.195199}, "tr": {"chunks": 3406.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.188146, "cps": 26.609375, "mean_p": 0.172408, "boundary_entropy": 0.025972, "bpb": 1.455355}, "fi": {"chunks": 3368.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179044, "cps": 26.3125, "mean_p": 0.173233, "boundary_entropy": 0.022811, "bpb": 1.57966}, "ru": {"chunks": 5711.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.301372, "cps": 44.617188, "mean_p": 0.165421, "boundary_entropy": 0.031097, "bpb": 0.907052}, "el": {"chunks": 6700.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.326463, "cps": 52.34375, "mean_p": 0.178102, "boundary_entropy": 0.037919, "bpb": 0.89738}, "hi": {"chunks": 7444.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.442437, "cps": 58.15625, "mean_p": 0.164613, "boundary_entropy": 0.103417, "bpb": 0.729263}, "ta": {"chunks": 8575.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.458679, "cps": 66.992188, "mean_p": 0.157756, "boundary_entropy": 0.094364, "bpb": 0.624385}, "th": {"chunks": 8136.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.505813, "cps": 63.5625, "mean_p": 0.157672, "boundary_entropy": 0.145067, "bpb": 0.714259}, "zh": {"chunks": 2724.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.461226, "cps": 21.28125, "mean_p": 0.17203, "boundary_entropy": 0.032036, "bpb": 2.196127}, "my": {"chunks": 9617.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.507547, "cps": 75.132812, "mean_p": 0.339275, "boundary_entropy": 0.521635, "bpb": 0.679612}}, "bpb": {"en": 1.46863, "de": 1.56175, "vi": 1.1952, "tr": 1.45536, "fi": 1.57966, "ru": 0.90705, "el": 0.89738, "hi": 0.72926, "ta": 0.62439, "th": 0.71426, "zh": 2.19613, "my": 0.67961}, "gini_cpc": 0.224009, "gini_cps": 0.237926, "bpic": 5.768194, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0413391660943607, "vi": 1.2909107330521887, "tr": 1.0849231518034257, "fi": 1.0324406228621572, "ru": 1.7378321626966433, "el": 1.8825167205605233, "hi": 2.5512685550884884, "ta": 2.6449261106269577, "th": 2.916720133040736, "zh": 2.6596136018295913, "my": 2.926719658825598}, "t": 10945.9}
{"event": "eval", "step": 64000, "phase": 3, "bpb": {"en": 1.46357, "de": 1.57069, "vi": 1.21429, "tr": 1.44863, "fi": 1.59971, "ru": 0.90437, "el": 0.89677, "hi": 0.73611, "ta": 0.62909, "th": 0.71361, "zh": 2.19877, "my": 0.68594}, "bytes_per_chunk": {"en": 5.588, "de": 5.653, "vi": 5.824, "tr": 5.678, "fi": 5.784, "ru": 5.826, "el": 5.765, "hi": 5.702, "ta": 5.812, "th": 5.753, "zh": 5.808, "my": 5.644}, "bpb_mean": 1.17179, "t": 10951.3}
{"event": "train", "step": 64000, "phase": 3, "objective": "global", "lm_loss": 0.78739, "bpb": 1.13596, "chunk_loss": 0.030056, "F": 0.17607, "G": 0.17595, "chunks": 180.3, "bytes_per_chunk": 5.68, "boundary_entropy": 0.05402, "router_noise": 0.0, "grad_norm": 0.3126, "lr": 4.78e-05, "router_trainable": true, "sec_per_step": 0.177, "alpha": 0.03, "t": 10951.4}
{"event": "train", "step": 64500, "phase": 3, "objective": "global", "lm_loss": 0.92156, "bpb": 1.32953, "chunk_loss": 0.030116, "F": 0.17918, "G": 0.18745, "chunks": 183.48, "bytes_per_chunk": 5.581, "boundary_entropy": 0.06074, "router_noise": 0.0, "grad_norm": 0.3277, "lr": 4.65e-05, "router_trainable": true, "sec_per_step": 0.16, "alpha": 0.03, "t": 11031.3}
{"event": "train", "step": 65000, "phase": 3, "objective": "global", "lm_loss": 0.87907, "bpb": 1.26824, "chunk_loss": 0.030048, "F": 0.17468, "G": 0.17369, "chunks": 178.88, "bytes_per_chunk": 5.725, "boundary_entropy": 0.03428, "router_noise": 0.0, "grad_norm": 0.3191, "lr": 4.51e-05, "router_trainable": true, "sec_per_step": 0.16, "alpha": 0.03, "t": 11111.3}
{"event": "train", "step": 65500, "phase": 3, "objective": "global", "lm_loss": 0.8171, "bpb": 1.17882, "chunk_loss": 0.030046, "F": 0.17426, "G": 0.17422, "chunks": 178.44, "bytes_per_chunk": 5.739, "boundary_entropy": 0.05432, "router_noise": 0.0, "grad_norm": 0.3143, "lr": 4.38e-05, "router_trainable": true, "sec_per_step": 0.16, "alpha": 0.03, "t": 11191.1}
{"event": "train", "step": 66000, "phase": 3, "objective": "global", "lm_loss": 0.85479, "bpb": 1.23321, "chunk_loss": 0.030042, "F": 0.17241, "G": 0.17252, "chunks": 176.55, "bytes_per_chunk": 5.8, "boundary_entropy": 0.0529, "router_noise": 0.0, "grad_norm": 0.3275, "lr": 4.26e-05, "router_trainable": true, "sec_per_step": 0.16, "alpha": 0.03, "t": 11271.0}
{"event": "train", "step": 66500, "phase": 3, "objective": "global", "lm_loss": 0.88988, "bpb": 1.28382, "chunk_loss": 0.030081, "F": 0.17644, "G": 0.17411, "chunks": 180.67, "bytes_per_chunk": 5.668, "boundary_entropy": 0.04126, "router_noise": 0.0, "grad_norm": 0.3332, "lr": 4.14e-05, "router_trainable": true, "sec_per_step": 0.16, "alpha": 0.03, "t": 11350.7}
{"event": "train", "step": 67000, "phase": 3, "objective": "global", "lm_loss": 0.89711, "bpb": 1.29426, "chunk_loss": 0.030057, "F": 0.17709, "G": 0.17704, "chunks": 181.34, "bytes_per_chunk": 5.647, "boundary_entropy": 0.04833, "router_noise": 0.0, "grad_norm": 0.3244, "lr": 4.03e-05, "router_trainable": true, "sec_per_step": 0.159, "alpha": 0.03, "t": 11430.0}
{"event": "train", "step": 67500, "phase": 3, "objective": "global", "lm_loss": 0.84997, "bpb": 1.22625, "chunk_loss": 0.030053, "F": 0.17592, "G": 0.1732, "chunks": 180.14, "bytes_per_chunk": 5.684, "boundary_entropy": 0.04137, "router_noise": 0.0, "grad_norm": 0.3307, "lr": 3.92e-05, "router_trainable": true, "sec_per_step": 0.159, "alpha": 0.03, "t": 11509.7}
{"event": "probe", "step": 68000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3024.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.173763, "cps": 23.625, "mean_p": 0.1731, "boundary_entropy": 0.023186, "bpb": 1.462593}, "de": {"chunks": 3720.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.181172, "cps": 29.0625, "mean_p": 0.1787, "boundary_entropy": 0.022463, "bpb": 1.553749}, "vi": {"chunks": 4125.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.223759, "cps": 32.226562, "mean_p": 0.172849, "boundary_entropy": 0.037046, "bpb": 1.188535}, "tr": {"chunks": 3411.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.188422, "cps": 26.648438, "mean_p": 0.172906, "boundary_entropy": 0.025396, "bpb": 1.44735}, "fi": {"chunks": 3372.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179257, "cps": 26.34375, "mean_p": 0.173697, "boundary_entropy": 0.022389, "bpb": 1.569787}, "ru": {"chunks": 5710.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.301319, "cps": 44.609375, "mean_p": 0.165234, "boundary_entropy": 0.030192, "bpb": 0.901044}, "el": {"chunks": 6735.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.328168, "cps": 52.617188, "mean_p": 0.179133, "boundary_entropy": 0.037153, "bpb": 0.891759}, "hi": {"chunks": 7414.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.440654, "cps": 57.921875, "mean_p": 0.164318, "boundary_entropy": 0.098767, "bpb": 0.723662}, "ta": {"chunks": 8647.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.46253, "cps": 67.554688, "mean_p": 0.158599, "boundary_entropy": 0.092437, "bpb": 0.618166}, "th": {"chunks": 8162.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.507429, "cps": 63.765625, "mean_p": 0.157746, "boundary_entropy": 0.141409, "bpb": 0.70967}, "zh": {"chunks": 2726.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.461565, "cps": 21.296875, "mean_p": 0.172172, "boundary_entropy": 0.031082, "bpb": 2.183896}, "my": {"chunks": 10150.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.535677, "cps": 79.296875, "mean_p": 0.302691, "boundary_entropy": 0.470248, "bpb": 0.674704}}, "bpb": {"en": 1.46259, "de": 1.55375, "vi": 1.18853, "tr": 1.44735, "fi": 1.56979, "ru": 0.90104, "el": 0.89176, "hi": 0.72366, "ta": 0.61817, "th": 0.70967, "zh": 2.1839, "my": 0.6747}, "gini_cpc": 0.229002, "gini_cps": 0.243623, "bpic": 5.711367, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0426363600522273, "vi": 1.2877250506929108, "tr": 1.0843600310919264, "fi": 1.031615873184636, "ru": 1.7340803911714202, "el": 1.8885960989180548, "hi": 2.5359450694591854, "ta": 2.6618423035613463, "th": 2.9202353814803304, "zh": 2.6562854428372544, "my": 3.0827975101447236}, "t": 11592.1}
{"event": "eval", "step": 68000, "phase": 3, "bpb": {"en": 1.45749, "de": 1.56272, "vi": 1.20767, "tr": 1.44012, "fi": 1.58928, "ru": 0.89846, "el": 0.89095, "hi": 0.7303, "ta": 0.62395, "th": 0.70955, "zh": 2.18706, "my": 0.6806}, "bytes_per_chunk": {"en": 5.576, "de": 5.655, "vi": 5.835, "tr": 5.673, "fi": 5.775, "ru": 5.831, "el": 5.734, "hi": 5.692, "ta": 5.775, "th": 5.695, "zh": 5.799, "my": 5.407}, "bpb_mean": 1.16485, "t": 11597.3}
{"event": "train", "step": 68000, "phase": 3, "objective": "global", "lm_loss": 0.84353, "bpb": 1.21696, "chunk_loss": 0.030047, "F": 0.17473, "G": 0.17516, "chunks": 178.92, "bytes_per_chunk": 5.723, "boundary_entropy": 0.04446, "router_noise": 0.0, "grad_norm": 0.3318, "lr": 3.82e-05, "router_trainable": true, "sec_per_step": 0.176, "alpha": 0.03, "t": 11597.5}
{"event": "train", "step": 68500, "phase": 3, "objective": "global", "lm_loss": 0.90619, "bpb": 1.30736, "chunk_loss": 0.030062, "F": 0.17412, "G": 0.17878, "chunks": 178.3, "bytes_per_chunk": 5.743, "boundary_entropy": 0.05565, "router_noise": 0.0, "grad_norm": 0.3474, "lr": 3.73e-05, "router_trainable": true, "sec_per_step": 0.159, "alpha": 0.03, "t": 11676.8}
{"event": "train", "step": 69000, "phase": 3, "objective": "global", "lm_loss": 0.96828, "bpb": 1.39693, "chunk_loss": 0.030057, "F": 0.17729, "G": 0.17848, "chunks": 181.55, "bytes_per_chunk": 5.64, "boundary_entropy": 0.03995, "router_noise": 0.0, "grad_norm": 0.3412, "lr": 3.64e-05, "router_trainable": true, "sec_per_step": 0.158, "alpha": 0.03, "t": 11756.1}
{"event": "train", "step": 69500, "phase": 3, "objective": "global", "lm_loss": 0.89181, "bpb": 1.28661, "chunk_loss": 0.030072, "F": 0.17816, "G": 0.1775, "chunks": 182.44, "bytes_per_chunk": 5.613, "boundary_entropy": 0.04654, "router_noise": 0.0, "grad_norm": 0.344, "lr": 3.55e-05, "router_trainable": true, "sec_per_step": 0.159, "alpha": 0.03, "t": 11835.4}
{"event": "train", "step": 70000, "phase": 3, "objective": "global", "lm_loss": 0.95518, "bpb": 1.37803, "chunk_loss": 0.030092, "F": 0.17479, "G": 0.18369, "chunks": 178.98, "bytes_per_chunk": 5.721, "boundary_entropy": 0.06645, "router_noise": 0.0, "grad_norm": 0.3497, "lr": 3.48e-05, "router_trainable": true, "sec_per_step": 0.158, "alpha": 0.03, "t": 11914.5}
{"event": "train", "step": 70500, "phase": 3, "objective": "global", "lm_loss": 0.8674, "bpb": 1.2514, "chunk_loss": 0.030067, "F": 0.17628, "G": 0.18065, "chunks": 180.52, "bytes_per_chunk": 5.673, "boundary_entropy": 0.05488, "router_noise": 0.0, "grad_norm": 0.3268, "lr": 3.4e-05, "router_trainable": true, "sec_per_step": 0.159, "alpha": 0.03, "t": 11994.0}
{"event": "train", "step": 71000, "phase": 3, "objective": "global", "lm_loss": 0.89765, "bpb": 1.29503, "chunk_loss": 0.030064, "F": 0.17697, "G": 0.17723, "chunks": 181.22, "bytes_per_chunk": 5.651, "boundary_entropy": 0.04324, "router_noise": 0.0, "grad_norm": 0.3452, "lr": 3.34e-05, "router_trainable": true, "sec_per_step": 0.159, "alpha": 0.03, "t": 12073.6}
{"event": "train", "step": 71500, "phase": 3, "objective": "global", "lm_loss": 0.9517, "bpb": 1.37302, "chunk_loss": 0.030052, "F": 0.17502, "G": 0.17528, "chunks": 179.22, "bytes_per_chunk": 5.714, "boundary_entropy": 0.04281, "router_noise": 0.0, "grad_norm": 0.3549, "lr": 3.28e-05, "router_trainable": true, "sec_per_step": 0.16, "alpha": 0.03, "t": 12153.4}
{"event": "probe", "step": 72000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3026.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.173878, "cps": 23.640625, "mean_p": 0.172977, "boundary_entropy": 0.023074, "bpb": 1.457333}, "de": {"chunks": 3721.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.18122, "cps": 29.070312, "mean_p": 0.178991, "boundary_entropy": 0.021593, "bpb": 1.549116}, "vi": {"chunks": 4277.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.232004, "cps": 33.414062, "mean_p": 0.178084, "boundary_entropy": 0.035679, "bpb": 1.184269}, "tr": {"chunks": 3417.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.188753, "cps": 26.695312, "mean_p": 0.173163, "boundary_entropy": 0.024602, "bpb": 1.441887}, "fi": {"chunks": 3370.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.17915, "cps": 26.328125, "mean_p": 0.173782, "boundary_entropy": 0.022497, "bpb": 1.562701}, "ru": {"chunks": 5712.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.301425, "cps": 44.625, "mean_p": 0.164949, "boundary_entropy": 0.030232, "bpb": 0.897173}, "el": {"chunks": 6710.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.32695, "cps": 52.421875, "mean_p": 0.177821, "boundary_entropy": 0.037257, "bpb": 0.886244}, "hi": {"chunks": 7378.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.438514, "cps": 57.640625, "mean_p": 0.162085, "boundary_entropy": 0.100779, "bpb": 0.720538}, "ta": {"chunks": 8610.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.460551, "cps": 67.265625, "mean_p": 0.157368, "boundary_entropy": 0.094204, "bpb": 0.616015}, "th": {"chunks": 7874.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.489524, "cps": 61.515625, "mean_p": 0.151907, "boundary_entropy": 0.143392, "bpb": 0.705853}, "zh": {"chunks": 2730.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.462242, "cps": 21.328125, "mean_p": 0.172174, "boundary_entropy": 0.029628, "bpb": 2.176332}, "my": {"chunks": 9580.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.505594, "cps": 74.84375, "mean_p": 0.258503, "boundary_entropy": 0.405807, "bpb": 0.668436}}, "bpb": {"en": 1.45733, "de": 1.54912, "vi": 1.18427, "tr": 1.44189, "fi": 1.5627, "ru": 0.89717, "el": 0.88624, "hi": 0.72054, "ta": 0.61601, "th": 0.70585, "zh": 2.17633, "my": 0.66844}, "gini_cpc": 0.220261, "gini_cps": 0.235099, "bpic": 5.779399, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0422273348507485, "vi": 1.3342932985995524, "tr": 1.0855494805938801, "fi": 1.0303225711018391, "ru": 1.7335412528534584, "el": 1.88034210742681, "hi": 2.5219633704526103, "ta": 2.6487006521177694, "th": 2.8153315139889066, "zh": 2.658424929536074, "my": 2.907751775095049}, "t": 12236.4}
{"event": "eval", "step": 72000, "phase": 3, "bpb": {"en": 1.45233, "de": 1.55779, "vi": 1.20341, "tr": 1.43459, "fi": 1.58233, "ru": 0.89439, "el": 0.88613, "hi": 0.72691, "ta": 0.62097, "th": 0.70588, "zh": 2.17917, "my": 0.67487}, "bytes_per_chunk": {"en": 5.571, "de": 5.638, "vi": 5.645, "tr": 5.666, "fi": 5.773, "ru": 5.825, "el": 5.761, "hi": 5.757, "ta": 5.79, "th": 5.924, "zh": 5.788, "my": 5.727}, "bpb_mean": 1.1599, "t": 12241.8}
{"event": "train", "step": 72000, "phase": 3, "objective": "global", "lm_loss": 0.79241, "bpb": 1.14321, "chunk_loss": 0.03004, "F": 0.17345, "G": 0.17258, "chunks": 177.61, "bytes_per_chunk": 5.765, "boundary_entropy": 0.05299, "router_noise": 0.0, "grad_norm": 0.3292, "lr": 3.22e-05, "router_trainable": true, "sec_per_step": 0.177, "alpha": 0.03, "t": 12241.9}
{"event": "train", "step": 72500, "phase": 3, "objective": "global", "lm_loss": 0.85169, "bpb": 1.22872, "chunk_loss": 0.03004, "F": 0.17484, "G": 0.17305, "chunks": 179.03, "bytes_per_chunk": 5.72, "boundary_entropy": 0.03745, "router_noise": 0.0, "grad_norm": 0.3673, "lr": 3.17e-05, "router_trainable": true, "sec_per_step": 0.16, "alpha": 0.03, "t": 12321.8}
{"event": "train", "step": 73000, "phase": 3, "objective": "global", "lm_loss": 0.82509, "bpb": 1.19036, "chunk_loss": 0.03004, "F": 0.17387, "G": 0.17453, "chunks": 178.05, "bytes_per_chunk": 5.751, "boundary_entropy": 0.06318, "router_noise": 0.0, "grad_norm": 0.345, "lr": 3.13e-05, "router_trainable": true, "sec_per_step": 0.159, "alpha": 0.03, "t": 12401.5}
{"event": "train", "step": 73500, "phase": 3, "objective": "global", "lm_loss": 0.88164, "bpb": 1.27194, "chunk_loss": 0.030051, "F": 0.17418, "G": 0.17448, "chunks": 178.36, "bytes_per_chunk": 5.741, "boundary_entropy": 0.04881, "router_noise": 0.0, "grad_norm": 0.3569, "lr": 3.09e-05, "router_trainable": true, "sec_per_step": 0.16, "alpha": 0.03, "t": 12481.3}
{"event": "train", "step": 74000, "phase": 3, "objective": "global", "lm_loss": 0.88054, "bpb": 1.27036, "chunk_loss": 0.030062, "F": 0.17786, "G": 0.17663, "chunks": 182.12, "bytes_per_chunk": 5.623, "boundary_entropy": 0.03312, "router_noise": 0.0, "grad_norm": 0.349, "lr": 3.06e-05, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 12562.5}
{"event": "train", "step": 74500, "phase": 3, "objective": "global", "lm_loss": 0.855, "bpb": 1.2335, "chunk_loss": 0.030041, "F": 0.17148, "G": 0.17189, "chunks": 175.59, "bytes_per_chunk": 5.832, "boundary_entropy": 0.05204, "router_noise": 0.0, "grad_norm": 0.34, "lr": 3.04e-05, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 12643.8}
{"event": "train", "step": 75000, "phase": 3, "objective": "global", "lm_loss": 0.92589, "bpb": 1.33577, "chunk_loss": 0.030038, "F": 0.17213, "G": 0.17469, "chunks": 176.27, "bytes_per_chunk": 5.809, "boundary_entropy": 0.04591, "router_noise": 0.0, "grad_norm": 0.3609, "lr": 3.02e-05, "router_trainable": true, "sec_per_step": 0.159, "alpha": 0.03, "t": 12723.2}
{"event": "train", "step": 75500, "phase": 3, "objective": "global", "lm_loss": 0.90335, "bpb": 1.30325, "chunk_loss": 0.030063, "F": 0.17674, "G": 0.1751, "chunks": 180.98, "bytes_per_chunk": 5.658, "boundary_entropy": 0.0402, "router_noise": 0.0, "grad_norm": 0.3574, "lr": 3.01e-05, "router_trainable": true, "sec_per_step": 0.159, "alpha": 0.03, "t": 12802.6}
{"event": "probe", "step": 76000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3026.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.173878, "cps": 23.640625, "mean_p": 0.173026, "boundary_entropy": 0.022874, "bpb": 1.453792}, "de": {"chunks": 3725.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.181415, "cps": 29.101562, "mean_p": 0.179061, "boundary_entropy": 0.021219, "bpb": 1.545866}, "vi": {"chunks": 4119.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.223434, "cps": 32.179688, "mean_p": 0.172502, "boundary_entropy": 0.033653, "bpb": 1.179837}, "tr": {"chunks": 3417.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.188753, "cps": 26.695312, "mean_p": 0.173038, "boundary_entropy": 0.02435, "bpb": 1.438606}, "fi": {"chunks": 3379.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179629, "cps": 26.398438, "mean_p": 0.174083, "boundary_entropy": 0.021607, "bpb": 1.557269}, "ru": {"chunks": 5715.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.301583, "cps": 44.648438, "mean_p": 0.164902, "boundary_entropy": 0.029432, "bpb": 0.895443}, "el": {"chunks": 6722.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.327535, "cps": 52.515625, "mean_p": 0.178374, "boundary_entropy": 0.036336, "bpb": 0.883815}, "hi": {"chunks": 7436.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.441961, "cps": 58.09375, "mean_p": 0.163752, "boundary_entropy": 0.098742, "bpb": 0.717989}, "ta": {"chunks": 8528.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.456165, "cps": 66.625, "mean_p": 0.155802, "boundary_entropy": 0.092, "bpb": 0.614086}, "th": {"chunks": 8027.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.499036, "cps": 62.710938, "mean_p": 0.155134, "boundary_entropy": 0.139891, "bpb": 0.702326}, "zh": {"chunks": 2728.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.461903, "cps": 21.3125, "mean_p": 0.171835, "boundary_entropy": 0.028754, "bpb": 2.173114}, "my": {"chunks": 10241.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.540479, "cps": 80.007812, "mean_p": 0.227864, "boundary_entropy": 0.329327, "bpb": 0.664403}}, "bpb": {"en": 1.45379, "de": 1.54587, "vi": 1.17984, "tr": 1.43861, "fi": 1.55727, "ru": 0.89544, "el": 0.88381, "hi": 0.71799, "ta": 0.61409, "th": 0.70233, "zh": 2.17311, "my": 0.6644}, "gini_cpc": 0.228243, "gini_cps": 0.24278, "bpic": 5.722694, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0433477082287121, "vi": 1.2850021269421457, "tr": 1.0855494805938801, "fi": 1.0330741744074523, "ru": 1.734451726200545, "el": 1.8837048652940411, "hi": 2.541789051597399, "ta": 2.62347493162141, "th": 2.870036330046857, "zh": 2.6564773654851317, "my": 3.1083805771136115}, "t": 12885.0}
{"event": "eval", "step": 76000, "phase": 3, "bpb": {"en": 1.44887, "de": 1.55399, "vi": 1.19938, "tr": 1.43164, "fi": 1.57762, "ru": 0.89188, "el": 0.88343, "hi": 0.72431, "ta": 0.6189, "th": 0.70212, "zh": 2.17633, "my": 0.67037}, "bytes_per_chunk": {"en": 5.57, "de": 5.644, "vi": 5.83, "tr": 5.665, "fi": 5.774, "ru": 5.827, "el": 5.743, "hi": 5.681, "ta": 5.838, "th": 5.791, "zh": 5.8, "my": 5.374}, "bpb_mean": 1.15657, "t": 12890.3}
{"event": "train", "step": 76000, "phase": 3, "objective": "global", "lm_loss": 0.79154, "bpb": 1.14195, "chunk_loss": 0.030088, "F": 0.1787, "G": 0.17981, "chunks": 182.98, "bytes_per_chunk": 5.596, "boundary_entropy": 0.06971, "router_noise": 0.0, "grad_norm": 0.3391, "lr": 3e-05, "router_trainable": true, "sec_per_step": 0.176, "alpha": 0.03, "t": 12890.5}
{"event": "probe", "step": 76300, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3044.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.174912, "cps": 23.78125, "mean_p": 0.173544, "boundary_entropy": 0.023919, "bpb": 1.453581}, "de": {"chunks": 3725.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.181415, "cps": 29.101562, "mean_p": 0.179152, "boundary_entropy": 0.021626, "bpb": 1.545093}, "vi": {"chunks": 4161.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.225712, "cps": 32.507812, "mean_p": 0.174537, "boundary_entropy": 0.035025, "bpb": 1.179661}, "tr": {"chunks": 3407.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.188201, "cps": 26.617188, "mean_p": 0.172703, "boundary_entropy": 0.024447, "bpb": 1.438158}, "fi": {"chunks": 3398.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180639, "cps": 26.546875, "mean_p": 0.174838, "boundary_entropy": 0.02206, "bpb": 1.557058}, "ru": {"chunks": 5714.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.30153, "cps": 44.640625, "mean_p": 0.165041, "boundary_entropy": 0.029823, "bpb": 0.895074}, "el": {"chunks": 6728.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.327827, "cps": 52.5625, "mean_p": 0.178831, "boundary_entropy": 0.03729, "bpb": 0.883054}, "hi": {"chunks": 7318.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.434948, "cps": 57.171875, "mean_p": 0.162048, "boundary_entropy": 0.097931, "bpb": 0.718267}, "ta": {"chunks": 8575.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.458679, "cps": 66.992188, "mean_p": 0.156582, "boundary_entropy": 0.092511, "bpb": 0.613976}, "th": {"chunks": 8016.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.498353, "cps": 62.625, "mean_p": 0.155118, "boundary_entropy": 0.139253, "bpb": 0.702728}, "zh": {"chunks": 2726.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.461565, "cps": 21.296875, "mean_p": 0.171972, "boundary_entropy": 0.028995, "bpb": 2.172661}, "my": {"chunks": 10413.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.549557, "cps": 81.351562, "mean_p": 0.22977, "boundary_entropy": 0.328927, "bpb": 0.665495}}, "bpb": {"en": 1.45358, "de": 1.54509, "vi": 1.17966, "tr": 1.43816, "fi": 1.55706, "ru": 0.89507, "el": 0.88305, "hi": 0.71827, "ta": 0.61398, "th": 0.70273, "zh": 2.17266, "my": 0.66549}, "gini_cpc": 0.229036, "gini_cps": 0.243931, "bpic": 5.708903, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.037178109428411, "vi": 1.2904287964211365, "tr": 1.0759722091059751, "fi": 1.0327399162748387, "ru": 1.7238937448642428, "el": 1.8742374425206603, "hi": 2.4866622669397622, "ta": 2.62233475751385, "th": 2.8491552556086526, "zh": 2.6388328446582974, "my": 3.1418971508439864}, "t": 12941.1}
{"event": "eval", "step": 76300, "phase": 3, "bpb": {"en": 1.4488, "de": 1.5538, "vi": 1.19908, "tr": 1.43109, "fi": 1.57741, "ru": 0.89176, "el": 0.88287, "hi": 0.7249, "ta": 0.61897, "th": 0.70256, "zh": 2.17554, "my": 0.67115}, "bytes_per_chunk": {"en": 5.556, "de": 5.646, "vi": 5.766, "tr": 5.68, "fi": 5.76, "ru": 5.829, "el": 5.731, "hi": 5.783, "ta": 5.825, "th": 5.793, "zh": 5.798, "my": 5.292}, "bpb_mean": 1.15649, "t": 12946.3}
{"event": "probe", "step": 76300, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3044.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.174912, "cps": 23.78125, "mean_p": 0.173544, "boundary_entropy": 0.023919, "bpb": 1.453581}, "de": {"chunks": 3725.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.181415, "cps": 29.101562, "mean_p": 0.179152, "boundary_entropy": 0.021626, "bpb": 1.545093}, "vi": {"chunks": 4161.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.225712, "cps": 32.507812, "mean_p": 0.174537, "boundary_entropy": 0.035025, "bpb": 1.179661}, "tr": {"chunks": 3407.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.188201, "cps": 26.617188, "mean_p": 0.172703, "boundary_entropy": 0.024447, "bpb": 1.438158}, "fi": {"chunks": 3398.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180639, "cps": 26.546875, "mean_p": 0.174838, "boundary_entropy": 0.02206, "bpb": 1.557058}, "ru": {"chunks": 5714.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.30153, "cps": 44.640625, "mean_p": 0.165041, "boundary_entropy": 0.029823, "bpb": 0.895074}, "el": {"chunks": 6728.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.327827, "cps": 52.5625, "mean_p": 0.178831, "boundary_entropy": 0.03729, "bpb": 0.883054}, "hi": {"chunks": 7318.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.434948, "cps": 57.171875, "mean_p": 0.162048, "boundary_entropy": 0.097931, "bpb": 0.718267}, "ta": {"chunks": 8575.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.458679, "cps": 66.992188, "mean_p": 0.156582, "boundary_entropy": 0.092511, "bpb": 0.613976}, "th": {"chunks": 8016.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.498353, "cps": 62.625, "mean_p": 0.155118, "boundary_entropy": 0.139253, "bpb": 0.702728}, "zh": {"chunks": 2726.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.461565, "cps": 21.296875, "mean_p": 0.171972, "boundary_entropy": 0.028995, "bpb": 2.172661}, "my": {"chunks": 10413.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.549557, "cps": 81.351562, "mean_p": 0.22977, "boundary_entropy": 0.328927, "bpb": 0.665495}}, "bpb": {"en": 1.45358, "de": 1.54509, "vi": 1.17966, "tr": 1.43816, "fi": 1.55706, "ru": 0.89507, "el": 0.88305, "hi": 0.71827, "ta": 0.61398, "th": 0.70273, "zh": 2.17266, "my": 0.66549}, "gini_cpc": 0.229036, "gini_cps": 0.243931, "bpic": 5.708903, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.037178109428411, "vi": 1.2904287964211365, "tr": 1.0759722091059751, "fi": 1.0327399162748387, "ru": 1.7238937448642428, "el": 1.8742374425206603, "hi": 2.4866622669397622, "ta": 2.62233475751385, "th": 2.8491552556086526, "zh": 2.6388328446582974, "my": 3.1418971508439864}, "t": 12949.5}
{"event": "reps_dumped", "dirs": ["/local/scratch/sas245/beetle-hnet/scratch/reps/specfull_A_s2/th0.30", "/local/scratch/sas245/beetle-hnet/scratch/reps/specfull_A_s2/th0.40", "/local/scratch/sas245/beetle-hnet/scratch/reps/specfull_A_s2/th0.50", "/local/scratch/sas245/beetle-hnet/scratch/reps/specfull_A_s2/th0.60", "/local/scratch/sas245/beetle-hnet/scratch/reps/specfull_A_s2/th0.70"], "t": 12952.8}
{"event": "final", "run": "specfull_A_s2", "condition": "A", "step": 76300, "params": 82470528, "per_lang": {"en": {"cpc": 0.17491237143021318, "cps": 23.78125, "bpb": 1.4459496566972647, "bytes_per_chunk": 5.552104457775824}, "de": {"cpc": 0.1814152827156285, "cps": 29.1015625, "bpb": 1.5502144021312767, "bytes_per_chunk": 5.632720592185133}, "vi": {"cpc": 0.2257119609438568, "cps": 32.5078125, "bpb": 1.1977475303956786, "bytes_per_chunk": 5.7664125999494065}, "tr": {"cpc": 0.18820085068773132, "cps": 26.6171875, "bpb": 1.4290408491613016, "bytes_per_chunk": 5.68389327956115}, "fi": {"cpc": 0.18063898782627186, "cps": 26.546875, "bpb": 1.5827007424985364, "bytes_per_chunk": 5.751704542337924}, "ru": {"cpc": 0.3015303430079156, "cps": 44.640625, "bpb": 0.8939889669747939, "bytes_per_chunk": 5.815535811342939}, "el": {"cpc": 0.32782731569458656, "cps": 52.5625, "bpb": 0.8818085429524961, "bytes_per_chunk": 5.717301694619528}, "hi": {"cpc": 0.4349479940564636, "cps": 57.171875, "bpb": 0.7264628568609015, "bytes_per_chunk": 5.7822799887506}, "ta": {"cpc": 0.4586787911206205, "cps": 66.9921875, "bpb": 0.6170430255202892, "bytes_per_chunk": 5.838786555894603}, "th": {"cpc": 0.49835250233136463, "cps": 62.625, "bpb": 0.7054821638980503, "bytes_per_chunk": 5.787290408751228}, "zh": {"cpc": 0.4615645106671182, "cps": 21.296875, "bpb": 2.1749229147532927, "bytes_per_chunk": 5.809832505000471}, "my": {"cpc": 0.5495566814439519, "cps": 81.3515625, "bpb": 0.6715674592452543, "bytes_per_chunk": 5.296721170700168}}, "gini_cpc": 0.2290363327631324, "gini_cps": 0.24393082930457433, "bpic": 5.708902937895128, "seed": 2, "mean_boundary_entropy": 0.039606429636478424, "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": 12962.8}