| {"event": "start", "run": "specfull_B_s2", "condition": "B", "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": "parity_a", "p3_objective": "none", "freeze_in_p3": true, "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": 2.1} |
| {"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": 2.1} |
| {"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": 7.6} |
| {"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": 9.283, "lr": 1e-07, "router_trainable": true, "sec_per_step": 0.001, "alpha": 0.03, "t": 8.3} |
| {"event": "probe", "step": 1, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 10167.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.58421, "cps": 79.429688, "mean_p": 0.506587, "boundary_entropy": 0.687329, "bpb": 8.008764}, "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": 8.1039}, "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": 7.984225}, "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": 8.10167}, "fi": {"chunks": 2113.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.112328, "cps": 16.507812, "mean_p": 0.479375, "boundary_entropy": 0.686741, "bpb": 8.102497}, "ru": {"chunks": 23612.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.246016, "cps": 184.46875, "mean_p": 0.507155, "boundary_entropy": 0.690197, "bpb": 8.268146}, "el": {"chunks": 1453.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.070799, "cps": 11.351562, "mean_p": 0.480416, "boundary_entropy": 0.689542, "bpb": 7.887667}, "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": 7.80527}, "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": 7.606905}, "th": {"chunks": 27171.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 1.689214, "cps": 212.273438, "mean_p": 0.503153, "boundary_entropy": 0.691076, "bpb": 7.867808}, "zh": {"chunks": 14701.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.489164, "cps": 114.851562, "mean_p": 0.531829, "boundary_entropy": 0.685237, "bpb": 8.023152}, "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": 8.244535}}, "bpb": {"en": 8.00876, "de": 8.1039, "vi": 7.98423, "tr": 8.10167, "fi": 8.1025, "ru": 8.26815, "el": 7.88767, "hi": 7.80527, "ta": 7.6069, "th": 7.86781, "zh": 8.02315, "my": 8.24454}, "gini_cpc": 0.562633, "gini_cps": 0.489179, "bpic": 3.455587, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 0.913670164877338, "vi": 0.5320381524392923, "tr": 0.6007027229045585, "fi": 0.19227326903692754, "ru": 2.1328232034262644, "el": 0.12118700870438104, "hi": 0.778791887221879, "ta": 0.03268692324671697, "th": 2.8914511127032156, "zh": 4.260737039226829, "my": 0.07127626334949846}, "t": 13.4} |
| {"event": "probe", "step": 2, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 10124.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.581739, "cps": 79.09375, "mean_p": 0.506492, "boundary_entropy": 0.687331, "bpb": 7.991826}, "de": {"chunks": 10929.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.532265, "cps": 85.382812, "mean_p": 0.503777, "boundary_entropy": 0.688251, "bpb": 8.08875}, "vi": {"chunks": 5705.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.309466, "cps": 44.570312, "mean_p": 0.489604, "boundary_entropy": 0.68847, "bpb": 7.974744}, "tr": {"chunks": 6335.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.349942, "cps": 49.492188, "mean_p": 0.495677, "boundary_entropy": 0.687873, "bpb": 8.090158}, "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": 8.092225}, "ru": {"chunks": 23596.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.245172, "cps": 184.34375, "mean_p": 0.507146, "boundary_entropy": 0.690197, "bpb": 8.24157}, "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": 7.88101}, "hi": {"chunks": 7646.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.454443, "cps": 59.734375, "mean_p": 0.494265, "boundary_entropy": 0.690845, "bpb": 7.794775}, "ta": {"chunks": 357.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.019096, "cps": 2.789062, "mean_p": 0.47535, "boundary_entropy": 0.689905, "bpb": 7.59927}, "th": {"chunks": 27109.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 1.685359, "cps": 211.789062, "mean_p": 0.50313, "boundary_entropy": 0.691076, "bpb": 7.852506}, "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": 8.019966}, "my": {"chunks": 787.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.041535, "cps": 6.148438, "mean_p": 0.475637, "boundary_entropy": 0.69006, "bpb": 8.245595}}, "bpb": {"en": 7.99183, "de": 8.08875, "vi": 7.97474, "tr": 8.09016, "fi": 8.09222, "ru": 8.24157, "el": 7.88101, "hi": 7.79478, "ta": 7.59927, "th": 7.85251, "zh": 8.01997, "my": 8.24559}, "gini_cpc": 0.562925, "gini_cps": 0.489397, "bpic": 3.462446, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9149555640097371, "vi": 0.5319667530273108, "tr": 0.6015449032996205, "fi": 0.1927243901211076, "ru": 2.140430628543788, "el": 0.12136669456032148, "hi": 0.7811801594845021, "ta": 0.03282575549677711, "th": 2.897106203253729, "zh": 4.277087464575732, "my": 0.07139755505449963}, "t": 19.2} |
| {"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": 7.929881}, "de": {"chunks": 10794.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.52569, "cps": 84.328125, "mean_p": 0.503497, "boundary_entropy": 0.68825, "bpb": 8.032997}, "vi": {"chunks": 5597.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.303607, "cps": 43.726562, "mean_p": 0.489352, "boundary_entropy": 0.688457, "bpb": 7.940486}, "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": 8.048472}, "fi": {"chunks": 2075.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.110308, "cps": 16.210938, "mean_p": 0.479034, "boundary_entropy": 0.686703, "bpb": 8.056363}, "ru": {"chunks": 23587.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.244697, "cps": 184.273438, "mean_p": 0.507132, "boundary_entropy": 0.690199, "bpb": 8.161051}, "el": {"chunks": 1433.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.069824, "cps": 11.195312, "mean_p": 0.480358, "boundary_entropy": 0.689538, "bpb": 7.861438}, "hi": {"chunks": 7623.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.453076, "cps": 59.554688, "mean_p": 0.49424, "boundary_entropy": 0.690844, "bpb": 7.756266}, "ta": {"chunks": 355.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.018989, "cps": 2.773438, "mean_p": 0.475327, "boundary_entropy": 0.689904, "bpb": 7.573648}, "th": {"chunks": 26841.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 1.668698, "cps": 209.695312, "mean_p": 0.503021, "boundary_entropy": 0.691077, "bpb": 7.796688}, "zh": {"chunks": 14685.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.486454, "cps": 114.726562, "mean_p": 0.531609, "boundary_entropy": 0.685263, "bpb": 8.010274}, "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": 8.237875}}, "bpb": {"en": 7.92988, "de": 8.033, "vi": 7.94049, "tr": 8.04847, "fi": 8.05636, "ru": 8.16105, "el": 7.86144, "hi": 7.75627, "ta": 7.57365, "th": 7.79669, "zh": 8.01027, "my": 8.23788}, "gini_cpc": 0.565547, "gini_cps": 0.49198, "bpic": 3.493682, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9154081646864708, "vi": 0.5286849408261719, "tr": 0.5954219938470291, "fi": 0.19208391230049282, "ru": 2.167445908178152, "el": 0.12158783356550985, "hi": 0.788961156783223, "ta": 0.03306645661641435, "th": 2.9057778030200154, "zh": 4.32977454945707, "my": 0.06212524221309226}, "t": 25.4} |
| {"event": "probe", "step": 8, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 9506.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.546228, "cps": 74.265625, "mean_p": 0.504807, "boundary_entropy": 0.687345, "bpb": 7.710402}, "de": {"chunks": 10309.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.50207, "cps": 80.539062, "mean_p": 0.502492, "boundary_entropy": 0.688246, "bpb": 7.834279}, "vi": {"chunks": 5308.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.287931, "cps": 41.46875, "mean_p": 0.488558, "boundary_entropy": 0.688414, "bpb": 7.819042}, "tr": {"chunks": 5763.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.318345, "cps": 45.023438, "mean_p": 0.494147, "boundary_entropy": 0.687814, "bpb": 7.896803}, "fi": {"chunks": 1999.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.106268, "cps": 15.617188, "mean_p": 0.478139, "boundary_entropy": 0.686602, "bpb": 7.927109}, "ru": {"chunks": 23471.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.238575, "cps": 183.367188, "mean_p": 0.507026, "boundary_entropy": 0.690206, "bpb": 7.863756}, "el": {"chunks": 1384.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.067437, "cps": 10.8125, "mean_p": 0.48006, "boundary_entropy": 0.689512, "bpb": 7.786056}, "hi": {"chunks": 7574.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.450163, "cps": 59.171875, "mean_p": 0.494142, "boundary_entropy": 0.69084, "bpb": 7.597947}, "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": 7.471779}, "th": {"chunks": 25798.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 1.603855, "cps": 201.546875, "mean_p": 0.502639, "boundary_entropy": 0.691078, "bpb": 7.594832}, "zh": {"chunks": 14624.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.476126, "cps": 114.25, "mean_p": 0.530957, "boundary_entropy": 0.68534, "bpb": 7.97657}, "my": {"chunks": 663.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.034991, "cps": 5.179688, "mean_p": 0.475313, "boundary_entropy": 0.690027, "bpb": 8.208432}}, "bpb": {"en": 7.7104, "de": 7.83428, "vi": 7.81904, "tr": 7.8968, "fi": 7.92711, "ru": 7.86376, "el": 7.78606, "hi": 7.59795, "ta": 7.47178, "th": 7.59483, "zh": 7.97657, "my": 8.20843}, "gini_cpc": 0.569907, "gini_cps": 0.494788, "bpic": 3.595138, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9191585740130713, "vi": 0.527125568588688, "tr": 0.5828064887460083, "fi": 0.19454820176693577, "ru": 2.2675072763376223, "el": 0.12345865895082597, "hi": 0.8241315456039494, "ta": 0.034372249629138665, "th": 2.936238192837454, "zh": 4.53313910354727, "my": 0.0640584553977242}, "t": 32.2} |
| {"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": 0.500197, "boundary_entropy": 0.687333, "bpb": 7.094022}, "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": 7.237167}, "vi": {"chunks": 4299.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.233198, "cps": 33.585938, "mean_p": 0.485888, "boundary_entropy": 0.688256, "bpb": 7.450468}, "tr": {"chunks": 4458.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.246258, "cps": 34.828125, "mean_p": 0.490191, "boundary_entropy": 0.687621, "bpb": 7.411876}, "fi": {"chunks": 1742.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.092605, "cps": 13.609375, "mean_p": 0.475269, "boundary_entropy": 0.686257, "bpb": 7.520132}, "ru": {"chunks": 21908.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 1.156095, "cps": 171.15625, "mean_p": 0.505753, "boundary_entropy": 0.690247, "bpb": 6.961047}, "el": {"chunks": 1243.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.060566, "cps": 9.710938, "mean_p": 0.479039, "boundary_entropy": 0.689419, "bpb": 7.495426}, "hi": {"chunks": 7017.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.417058, "cps": 54.820312, "mean_p": 0.493466, "boundary_entropy": 0.690814, "bpb": 7.01073}, "ta": {"chunks": 334.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.017866, "cps": 2.609375, "mean_p": 0.475379, "boundary_entropy": 0.68992, "bpb": 7.122747}, "th": {"chunks": 20368.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 1.266273, "cps": 159.125, "mean_p": 0.500671, "boundary_entropy": 0.691079, "bpb": 6.821586}, "zh": {"chunks": 14343.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.428547, "cps": 112.054688, "mean_p": 0.52831, "boundary_entropy": 0.685635, "bpb": 7.853334}, "my": {"chunks": 711.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.037524, "cps": 5.554688, "mean_p": 0.474641, "boundary_entropy": 0.689957, "bpb": 8.062628}}, "bpb": {"en": 7.09402, "de": 7.23717, "vi": 7.45047, "tr": 7.41188, "fi": 7.52013, "ru": 6.96105, "el": 7.49543, "hi": 7.01073, "ta": 7.12275, "th": 6.82159, "zh": 7.85333, "my": 8.06263}, "gini_cpc": 0.583151, "gini_cps": 0.496872, "bpic": 4.135036, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9522174260669408, "vi": 0.5258279283726122, "tr": 0.555275943448589, "fi": 0.2088120769924926, "ru": 2.6068309219239403, "el": 0.13656821386864612, "hi": 0.9404067756931078, "ta": 0.040284719430256195, "th": 2.855266612498402, "zh": 5.476031046821732, "my": 0.08461075504963517}, "t": 40.5} |
| {"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": 0.494947, "boundary_entropy": 0.687303, "bpb": 6.284901}, "de": {"chunks": 7587.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.369503, "cps": 59.273438, "mean_p": 0.497467, "boundary_entropy": 0.688246, "bpb": 6.375035}, "vi": {"chunks": 2735.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.148359, "cps": 21.367188, "mean_p": 0.481677, "boundary_entropy": 0.688017, "bpb": 6.958903}, "tr": {"chunks": 3602.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.198973, "cps": 28.140625, "mean_p": 0.487619, "boundary_entropy": 0.687527, "bpb": 6.664045}, "fi": {"chunks": 1555.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.082664, "cps": 12.148438, "mean_p": 0.475613, "boundary_entropy": 0.686361, "bpb": 6.667407}, "ru": {"chunks": 11728.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.618892, "cps": 91.625, "mean_p": 0.497704, "boundary_entropy": 0.690281, "bpb": 5.628036}, "el": {"chunks": 1380.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.067242, "cps": 10.78125, "mean_p": 0.479199, "boundary_entropy": 0.689428, "bpb": 6.805854}, "hi": {"chunks": 7739.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.45997, "cps": 60.460938, "mean_p": 0.493589, "boundary_entropy": 0.690807, "bpb": 5.97272}, "ta": {"chunks": 434.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.023215, "cps": 3.390625, "mean_p": 0.480487, "boundary_entropy": 0.690421, "bpb": 6.395951}, "th": {"chunks": 4855.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.301834, "cps": 37.929688, "mean_p": 0.492132, "boundary_entropy": 0.69092, "bpb": 5.491049}, "zh": {"chunks": 12218.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.068744, "cps": 95.453125, "mean_p": 0.517177, "boundary_entropy": 0.686551, "bpb": 7.475834}, "my": {"chunks": 660.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.034832, "cps": 5.15625, "mean_p": 0.473343, "boundary_entropy": 0.689811, "bpb": 7.371202}}, "bpb": {"en": 6.2849, "de": 6.37503, "vi": 6.9589, "tr": 6.66405, "fi": 6.66741, "ru": 5.62804, "el": 6.80585, "hi": 5.97272, "ta": 6.39595, "th": 5.49105, "zh": 7.47583, "my": 7.3712}, "gini_cpc": 0.579022, "gini_cps": 0.435991, "bpic": 6.417206, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.2105527837498145, "vi": 0.48604921108355315, "tr": 0.6518673226297446, "fi": 0.27082238835198325, "ru": 2.0275930635470645, "el": 0.22029482496570113, "hi": 1.5069395374961958, "ta": 0.07605544537067768, "th": 0.9888586635943838, "zh": 6.77755002825389, "my": 0.11411601917486286}, "t": 51.4} |
| {"event": "probe", "step": 64, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4509.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.259093, "cps": 35.226562, "mean_p": 0.495517, "boundary_entropy": 0.687561, "bpb": 5.795579}, "de": {"chunks": 6724.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.327473, "cps": 52.53125, "mean_p": 0.497466, "boundary_entropy": 0.688439, "bpb": 5.915419}, "vi": {"chunks": 2634.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.14288, "cps": 20.578125, "mean_p": 0.484866, "boundary_entropy": 0.688433, "bpb": 6.444676}, "tr": {"chunks": 3983.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.220019, "cps": 31.117188, "mean_p": 0.491604, "boundary_entropy": 0.687945, "bpb": 6.241486}, "fi": {"chunks": 2660.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.141407, "cps": 20.78125, "mean_p": 0.488988, "boundary_entropy": 0.687795, "bpb": 5.966039}, "ru": {"chunks": 7434.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.392296, "cps": 58.078125, "mean_p": 0.494759, "boundary_entropy": 0.690294, "bpb": 5.095865}, "el": {"chunks": 14524.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.707694, "cps": 113.46875, "mean_p": 0.498255, "boundary_entropy": 0.690464, "bpb": 5.331095}, "hi": {"chunks": 11277.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.670253, "cps": 88.101562, "mean_p": 0.496189, "boundary_entropy": 0.690885, "bpb": 5.070042}, "ta": {"chunks": 11502.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.615245, "cps": 89.859375, "mean_p": 0.495333, "boundary_entropy": 0.691224, "bpb": 4.852045}, "th": {"chunks": 6882.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.427852, "cps": 53.765625, "mean_p": 0.494658, "boundary_entropy": 0.69102, "bpb": 4.739559}, "zh": {"chunks": 3678.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.622757, "cps": 28.734375, "mean_p": 0.491351, "boundary_entropy": 0.686742, "bpb": 6.978968}, "my": {"chunks": 9441.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.498258, "cps": 73.757812, "mean_p": 0.495154, "boundary_entropy": 0.69131, "bpb": 4.719446}}, "bpb": {"en": 5.79558, "de": 5.91542, "vi": 6.44468, "tr": 6.24149, "fi": 5.96604, "ru": 5.09586, "el": 5.33109, "hi": 5.07004, "ta": 4.85204, "th": 4.73956, "zh": 6.97897, "my": 4.71945}, "gini_cpc": 0.267373, "gini_cps": 0.293051, "bpic": 4.501936, "worst_lang": "el", "cpc_parity_vs_en": {"en": 1.0, "de": 1.2639188254442915, "vi": 0.5514631707563988, "tr": 0.8491875921507627, "fi": 0.5457749997143911, "ru": 1.514109301186545, "el": 2.7314249994059208, "hi": 2.5869163901024104, "ta": 2.3746071910738564, "th": 1.651343753276785, "zh": 2.403599843996396, "my": 1.9230851153930264}, "t": 66.7} |
| {"event": "probe", "step": 128, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2723.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.156467, "cps": 21.273438, "mean_p": 0.48725, "boundary_entropy": 0.687012, "bpb": 5.34593}, "de": {"chunks": 3888.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.189354, "cps": 30.375, "mean_p": 0.487965, "boundary_entropy": 0.687816, "bpb": 5.491479}, "vi": {"chunks": 3722.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.201899, "cps": 29.078125, "mean_p": 0.487437, "boundary_entropy": 0.688545, "bpb": 5.861366}, "tr": {"chunks": 2791.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.154173, "cps": 21.804688, "mean_p": 0.484608, "boundary_entropy": 0.687371, "bpb": 5.766075}, "fi": {"chunks": 2238.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.118973, "cps": 17.484375, "mean_p": 0.483492, "boundary_entropy": 0.68728, "bpb": 5.459131}, "ru": {"chunks": 4336.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.228813, "cps": 33.875, "mean_p": 0.491556, "boundary_entropy": 0.690201, "bpb": 4.674916}, "el": {"chunks": 9740.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.474589, "cps": 76.09375, "mean_p": 0.495144, "boundary_entropy": 0.690459, "bpb": 4.998426}, "hi": {"chunks": 9325.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.554235, "cps": 72.851562, "mean_p": 0.492505, "boundary_entropy": 0.690709, "bpb": 4.363428}, "ta": {"chunks": 7608.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.406954, "cps": 59.4375, "mean_p": 0.493961, "boundary_entropy": 0.691182, "bpb": 4.298156}, "th": {"chunks": 7067.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.439353, "cps": 55.210938, "mean_p": 0.487697, "boundary_entropy": 0.690464, "bpb": 4.133277}, "zh": {"chunks": 5782.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.979004, "cps": 45.171875, "mean_p": 0.499561, "boundary_entropy": 0.687249, "bpb": 6.654818}, "my": {"chunks": 15007.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.79201, "cps": 117.242188, "mean_p": 0.496821, "boundary_entropy": 0.691323, "bpb": 4.171333}}, "bpb": {"en": 5.34593, "de": 5.49148, "vi": 5.86137, "tr": 5.76608, "fi": 5.45913, "ru": 4.67492, "el": 4.99843, "hi": 4.36343, "ta": 4.29816, "th": 4.13328, "zh": 6.65482, "my": 4.17133}, "gini_cpc": 0.360283, "gini_cps": 0.314085, "bpic": 5.170369, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.2101809938120087, "vi": 1.2903564757556027, "tr": 0.9853392071228114, "fi": 0.7603694816596965, "ru": 1.4623675385192298, "el": 3.0331549052146807, "hi": 3.542176899331489, "ta": 2.6008871757694525, "th": 2.807957336455442, "zh": 6.25692751130174, "my": 5.0618233554291}, "t": 88.9} |
| {"event": "probe", "step": 256, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2738.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.157329, "cps": 21.390625, "mean_p": 0.39577, "boundary_entropy": 0.643049, "bpb": 4.260958}, "de": {"chunks": 4468.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.217601, "cps": 34.90625, "mean_p": 0.408179, "boundary_entropy": 0.64906, "bpb": 4.454674}, "vi": {"chunks": 4935.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.267697, "cps": 38.554688, "mean_p": 0.429125, "boundary_entropy": 0.665101, "bpb": 4.25521}, "tr": {"chunks": 3219.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.177816, "cps": 25.148438, "mean_p": 0.423288, "boundary_entropy": 0.660973, "bpb": 4.548973}, "fi": {"chunks": 3558.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.189145, "cps": 27.796875, "mean_p": 0.424384, "boundary_entropy": 0.662148, "bpb": 4.434563}, "ru": {"chunks": 5583.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.294617, "cps": 43.617188, "mean_p": 0.473974, "boundary_entropy": 0.68524, "bpb": 3.134397}, "el": {"chunks": 7787.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.379428, "cps": 60.835938, "mean_p": 0.47145, "boundary_entropy": 0.685581, "bpb": 3.346784}, "hi": {"chunks": 8171.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.485646, "cps": 63.835938, "mean_p": 0.486422, "boundary_entropy": 0.68905, "bpb": 2.714938}, "ta": {"chunks": 9780.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.523135, "cps": 76.40625, "mean_p": 0.492249, "boundary_entropy": 0.690825, "bpb": 2.666498}, "th": {"chunks": 7834.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.487038, "cps": 61.203125, "mean_p": 0.477604, "boundary_entropy": 0.688556, "bpb": 2.594276}, "zh": {"chunks": 3845.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.651033, "cps": 30.039062, "mean_p": 0.493115, "boundary_entropy": 0.68693, "bpb": 5.995085}, "my": {"chunks": 9396.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.495883, "cps": 73.40625, "mean_p": 0.477145, "boundary_entropy": 0.689575, "bpb": 2.649848}}, "bpb": {"en": 4.26096, "de": 4.45467, "vi": 4.25521, "tr": 4.54897, "fi": 4.43456, "ru": 3.1344, "el": 3.34678, "hi": 2.71494, "ta": 2.6665, "th": 2.59428, "zh": 5.99509, "my": 2.64985}, "gini_cpc": 0.246004, "gini_cps": 0.228913, "bpic": 5.381566, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.3830931604087802, "vi": 1.7015107271853032, "tr": 1.1302150905255364, "fi": 1.2022221817511234, "ru": 1.872617553016184, "el": 2.411681799966664, "hi": 3.0868165068807616, "ta": 3.325095029178358, "th": 3.0956594497865955, "zh": 4.138029456864632, "my": 3.1518846011682737}, "t": 120.8} |
| {"event": "probe", "step": 500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2824.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.162271, "cps": 22.0625, "mean_p": 0.288866, "boundary_entropy": 0.512905, "bpb": 3.383261}, "de": {"chunks": 4428.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.215653, "cps": 34.59375, "mean_p": 0.310936, "boundary_entropy": 0.523804, "bpb": 3.621965}, "vi": {"chunks": 4323.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.2345, "cps": 33.773438, "mean_p": 0.348265, "boundary_entropy": 0.582236, "bpb": 2.981661}, "tr": {"chunks": 4405.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.24333, "cps": 34.414062, "mean_p": 0.327171, "boundary_entropy": 0.547154, "bpb": 3.633986}, "fi": {"chunks": 3118.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.165754, "cps": 24.359375, "mean_p": 0.317919, "boundary_entropy": 0.563422, "bpb": 3.712604}, "ru": {"chunks": 5111.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.26971, "cps": 39.929688, "mean_p": 0.305533, "boundary_entropy": 0.556462, "bpb": 2.59213}, "el": {"chunks": 8258.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.402378, "cps": 64.515625, "mean_p": 0.368119, "boundary_entropy": 0.599765, "bpb": 2.640637}, "hi": {"chunks": 9351.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.55578, "cps": 73.054688, "mean_p": 0.418546, "boundary_entropy": 0.63681, "bpb": 2.04758}, "ta": {"chunks": 10058.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.538005, "cps": 78.578125, "mean_p": 0.479144, "boundary_entropy": 0.681264, "bpb": 1.853228}, "th": {"chunks": 11801.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.733665, "cps": 92.195312, "mean_p": 0.354664, "boundary_entropy": 0.613199, "bpb": 2.006525}, "zh": {"chunks": 9030.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.528954, "cps": 70.546875, "mean_p": 0.548871, "boundary_entropy": 0.598976, "bpb": 4.576202}, "my": {"chunks": 16878.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.890754, "cps": 131.859375, "mean_p": 0.367253, "boundary_entropy": 0.601821, "bpb": 1.999658}}, "bpb": {"en": 3.38326, "de": 3.62197, "vi": 2.98166, "tr": 3.63399, "fi": 3.7126, "ru": 2.59213, "el": 2.64064, "hi": 2.04758, "ta": 1.85323, "th": 2.00652, "zh": 4.5762, "my": 1.99966}, "gini_cpc": 0.39007, "gini_cps": 0.292617, "bpic": 4.283987, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.3289683338463378, "vi": 1.4451120466903467, "tr": 1.499528723660126, "fi": 1.0214653878456685, "ru": 1.6620959622384834, "el": 2.479667564298412, "hi": 3.4250144800037043, "ta": 3.3154737181852414, "th": 4.5212359578760495, "zh": 9.422230741484388, "my": 5.489300858290559}, "t": 174.7} |
| {"event": "train", "step": 500, "phase": 1, "objective": "global", "lm_loss": 2.16829, "bpb": 3.12818, "chunk_loss": 0.034909, "F": 0.23773, "G": 0.36535, "chunks": 243.44, "bytes_per_chunk": 4.206, "boundary_entropy": 0.57078, "router_noise": 0.0, "grad_norm": 2.3227, "lr": 7.51e-05, "router_trainable": true, "sec_per_step": 0.334, "alpha": 0.03, "t": 175.2} |
| {"event": "probe", "step": 1000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3137.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.180256, "cps": 24.507812, "mean_p": 0.230893, "boundary_entropy": 0.281119, "bpb": 2.660751}, "de": {"chunks": 4167.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.202942, "cps": 32.554688, "mean_p": 0.234079, "boundary_entropy": 0.280663, "bpb": 2.985569}, "vi": {"chunks": 4866.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.263954, "cps": 38.015625, "mean_p": 0.249962, "boundary_entropy": 0.342329, "bpb": 2.340097}, "tr": {"chunks": 3768.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.208142, "cps": 29.4375, "mean_p": 0.240643, "boundary_entropy": 0.321499, "bpb": 2.977584}, "fi": {"chunks": 3816.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.20286, "cps": 29.8125, "mean_p": 0.24702, "boundary_entropy": 0.324715, "bpb": 3.228674}, "ru": {"chunks": 6723.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.354776, "cps": 52.523438, "mean_p": 0.303313, "boundary_entropy": 0.444473, "bpb": 1.906212}, "el": {"chunks": 6848.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.333674, "cps": 53.5, "mean_p": 0.310397, "boundary_entropy": 0.435169, "bpb": 2.032731}, "hi": {"chunks": 9007.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.535334, "cps": 70.367188, "mean_p": 0.261784, "boundary_entropy": 0.393389, "bpb": 1.573073}, "ta": {"chunks": 9062.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.484729, "cps": 70.796875, "mean_p": 0.249721, "boundary_entropy": 0.412764, "bpb": 1.317058}, "th": {"chunks": 7501.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.466335, "cps": 58.601562, "mean_p": 0.224569, "boundary_entropy": 0.40073, "bpb": 1.645863}, "zh": {"chunks": 2515.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.425838, "cps": 19.648438, "mean_p": 0.303995, "boundary_entropy": 0.53201, "bpb": 3.74174}, "my": {"chunks": 4742.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.250264, "cps": 37.046875, "mean_p": 0.251908, "boundary_entropy": 0.465308, "bpb": 1.645878}}, "bpb": {"en": 2.66075, "de": 2.98557, "vi": 2.3401, "tr": 2.97758, "fi": 3.22867, "ru": 1.90621, "el": 2.03273, "hi": 1.57307, "ta": 1.31706, "th": 1.64586, "zh": 3.74174, "my": 1.64588}, "gini_cpc": 0.208022, "gini_cps": 0.219104, "bpic": 5.801503, "worst_lang": "hi", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1258504216161762, "vi": 1.4643286654754286, "tr": 1.1547020695569614, "fi": 1.1253978576924117, "ru": 1.9681740365019433, "el": 1.851111197990216, "hi": 2.969851207914358, "ta": 2.689107660230076, "th": 2.5870671517239514, "zh": 2.3624038855036416, "my": 1.3883781655264558}, "t": 271.6} |
| {"event": "train", "step": 1000, "phase": 1, "objective": "global", "lm_loss": 1.77556, "bpb": 2.5616, "chunk_loss": 0.030538, "F": 0.1915, "G": 0.26281, "chunks": 196.09, "bytes_per_chunk": 5.222, "boundary_entropy": 0.36516, "router_noise": 0.0, "grad_norm": 1.4926, "lr": 0.0001501, "router_trainable": true, "sec_per_step": 0.194, "alpha": 0.03, "t": 272.1} |
| {"event": "probe", "step": 1500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3069.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.176349, "cps": 23.976562, "mean_p": 0.202057, "boundary_entropy": 0.177606, "bpb": 2.358918}, "de": {"chunks": 3919.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.190863, "cps": 30.617188, "mean_p": 0.209952, "boundary_entropy": 0.174998, "bpb": 2.654795}, "vi": {"chunks": 4480.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.243016, "cps": 35.0, "mean_p": 0.217087, "boundary_entropy": 0.233376, "bpb": 2.117119}, "tr": {"chunks": 3677.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.203116, "cps": 28.726562, "mean_p": 0.211825, "boundary_entropy": 0.190629, "bpb": 2.666673}, "fi": {"chunks": 3582.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.19042, "cps": 27.984375, "mean_p": 0.211422, "boundary_entropy": 0.197173, "bpb": 2.868189}, "ru": {"chunks": 6254.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.330026, "cps": 48.859375, "mean_p": 0.229684, "boundary_entropy": 0.279226, "bpb": 1.659221}, "el": {"chunks": 6445.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.314038, "cps": 50.351562, "mean_p": 0.248177, "boundary_entropy": 0.33842, "bpb": 1.772382}, "hi": {"chunks": 7105.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.422288, "cps": 55.507812, "mean_p": 0.231452, "boundary_entropy": 0.372299, "bpb": 1.423788}, "ta": {"chunks": 9112.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.487403, "cps": 71.1875, "mean_p": 0.218993, "boundary_entropy": 0.342545, "bpb": 1.147252}, "th": {"chunks": 8652.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.537892, "cps": 67.59375, "mean_p": 0.2309, "boundary_entropy": 0.378204, "bpb": 1.357468}, "zh": {"chunks": 3040.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.514731, "cps": 23.75, "mean_p": 0.317885, "boundary_entropy": 0.480847, "bpb": 3.527473}, "my": {"chunks": 12076.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.637323, "cps": 94.34375, "mean_p": 0.274343, "boundary_entropy": 0.44368, "bpb": 1.294446}}, "bpb": {"en": 2.35892, "de": 2.65479, "vi": 2.11712, "tr": 2.66667, "fi": 2.86819, "ru": 1.65922, "el": 1.77238, "hi": 1.42379, "ta": 1.14725, "th": 1.35747, "zh": 3.52747, "my": 1.29445}, "gini_cpc": 0.244073, "gini_cps": 0.251788, "bpic": 5.374256, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0823060548789234, "vi": 1.3780408881592638, "tr": 1.1517820612570546, "fi": 1.0797940494977747, "ru": 1.8714399420195658, "el": 1.7807760589423334, "hi": 2.3946179912531824, "ta": 2.763856389952724, "th": 3.0501603924167524, "zh": 2.918820398693245, "my": 3.6139901125702676}, "t": 360.9} |
| {"event": "train", "step": 1500, "phase": 1, "objective": "global", "lm_loss": 1.65222, "bpb": 2.38365, "chunk_loss": 0.030247, "F": 0.183, "G": 0.2236, "chunks": 187.39, "bytes_per_chunk": 5.465, "boundary_entropy": 0.24583, "router_noise": 0.0, "grad_norm": 0.9049, "lr": 0.0002251, "router_trainable": true, "sec_per_step": 0.179, "alpha": 0.03, "t": 361.4} |
| {"event": "probe", "step": 2000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3150.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.181003, "cps": 24.609375, "mean_p": 0.198119, "boundary_entropy": 0.142662, "bpb": 2.236078}, "de": {"chunks": 4101.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.199727, "cps": 32.039062, "mean_p": 0.21478, "boundary_entropy": 0.147485, "bpb": 2.487177}, "vi": {"chunks": 4392.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.238242, "cps": 34.3125, "mean_p": 0.200591, "boundary_entropy": 0.178091, "bpb": 2.04802}, "tr": {"chunks": 3640.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.201072, "cps": 28.4375, "mean_p": 0.204081, "boundary_entropy": 0.153121, "bpb": 2.456998}, "fi": {"chunks": 3632.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.193079, "cps": 28.375, "mean_p": 0.209962, "boundary_entropy": 0.165607, "bpb": 2.673537}, "ru": {"chunks": 5838.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.308074, "cps": 45.609375, "mean_p": 0.203045, "boundary_entropy": 0.233229, "bpb": 1.541616}, "el": {"chunks": 6335.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.308678, "cps": 49.492188, "mean_p": 0.216193, "boundary_entropy": 0.269025, "bpb": 1.629004}, "hi": {"chunks": 5680.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.337593, "cps": 44.375, "mean_p": 0.234678, "boundary_entropy": 0.399388, "bpb": 1.371033}, "ta": {"chunks": 8566.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.458197, "cps": 66.921875, "mean_p": 0.211099, "boundary_entropy": 0.325725, "bpb": 1.082428}, "th": {"chunks": 8376.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.520734, "cps": 65.4375, "mean_p": 0.198218, "boundary_entropy": 0.306462, "bpb": 1.268327}, "zh": {"chunks": 2662.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.450728, "cps": 20.796875, "mean_p": 0.252421, "boundary_entropy": 0.384191, "bpb": 3.414815}, "my": {"chunks": 11076.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.584547, "cps": 86.53125, "mean_p": 0.242327, "boundary_entropy": 0.372472, "bpb": 1.247354}}, "bpb": {"en": 2.23608, "de": 2.48718, "vi": 2.04802, "tr": 2.457, "fi": 2.67354, "ru": 1.54162, "el": 1.629, "hi": 1.37103, "ta": 1.08243, "th": 1.26833, "zh": 3.41482, "my": 1.24735}, "gini_cpc": 0.224866, "gini_cps": 0.240284, "bpic": 5.690028, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1034456032449507, "vi": 1.3162329419969778, "tr": 1.110872967648088, "fi": 1.0667128402581065, "ru": 1.7020348284960423, "el": 1.705372833731261, "hi": 1.8651202150994128, "ta": 2.5314314242412665, "th": 2.876929170922333, "zh": 2.4901652879235, "my": 3.2294840013269397}, "t": 449.2} |
| {"event": "train", "step": 2000, "phase": 1, "objective": "global", "lm_loss": 1.52958, "bpb": 2.20671, "chunk_loss": 0.030143, "F": 0.17992, "G": 0.21113, "chunks": 184.23, "bytes_per_chunk": 5.558, "boundary_entropy": 0.21255, "router_noise": 0.0, "grad_norm": 0.8503, "lr": 0.0003, "router_trainable": true, "sec_per_step": 0.177, "alpha": 0.03, "t": 449.8} |
| {"event": "probe", "step": 2500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3141.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.180486, "cps": 24.539062, "mean_p": 0.191539, "boundary_entropy": 0.120368, "bpb": 2.124137}, "de": {"chunks": 3989.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.194273, "cps": 31.164062, "mean_p": 0.203657, "boundary_entropy": 0.121369, "bpb": 2.345196}, "vi": {"chunks": 4308.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.233686, "cps": 33.65625, "mean_p": 0.192292, "boundary_entropy": 0.157507, "bpb": 1.931713}, "tr": {"chunks": 3488.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.192675, "cps": 27.25, "mean_p": 0.190827, "boundary_entropy": 0.12343, "bpb": 2.299602}, "fi": {"chunks": 3571.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.189836, "cps": 27.898438, "mean_p": 0.198572, "boundary_entropy": 0.12495, "bpb": 2.493613}, "ru": {"chunks": 5995.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.316359, "cps": 46.835938, "mean_p": 0.198423, "boundary_entropy": 0.195435, "bpb": 1.431925}, "el": {"chunks": 6234.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.303757, "cps": 48.703125, "mean_p": 0.19344, "boundary_entropy": 0.208467, "bpb": 1.505871}, "hi": {"chunks": 6386.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.379554, "cps": 49.890625, "mean_p": 0.221605, "boundary_entropy": 0.35098, "bpb": 1.254558}, "ta": {"chunks": 8804.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.470928, "cps": 68.78125, "mean_p": 0.206796, "boundary_entropy": 0.292897, "bpb": 1.032695}, "th": {"chunks": 7445.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.462854, "cps": 58.164062, "mean_p": 0.20787, "boundary_entropy": 0.324173, "bpb": 1.254483}, "zh": {"chunks": 2852.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.482899, "cps": 22.28125, "mean_p": 0.263903, "boundary_entropy": 0.376604, "bpb": 3.2484}, "my": {"chunks": 8494.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.44828, "cps": 66.359375, "mean_p": 0.251191, "boundary_entropy": 0.423078, "bpb": 1.157957}}, "bpb": {"en": 2.12414, "de": 2.3452, "vi": 1.93171, "tr": 2.2996, "fi": 2.49361, "ru": 1.43193, "el": 1.50587, "hi": 1.25456, "ta": 1.0327, "th": 1.25448, "zh": 3.2484, "my": 1.15796}, "gini_cpc": 0.204635, "gini_cps": 0.211009, "bpic": 5.931058, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0763854360565057, "vi": 1.2947583976225614, "tr": 1.0675349656040696, "fi": 1.051802383301404, "ru": 1.7528153059501579, "el": 1.6829923293831912, "hi": 2.102955220533868, "ta": 2.60922029676098, "th": 2.5644829774012763, "zh": 2.6755450158176926, "my": 2.483733896761438}, "t": 536.8} |
| {"event": "train", "step": 2500, "phase": 1, "objective": "global", "lm_loss": 1.31093, "bpb": 1.89128, "chunk_loss": 0.030139, "F": 0.17558, "G": 0.20778, "chunks": 179.8, "bytes_per_chunk": 5.695, "boundary_entropy": 0.21413, "router_noise": 0.0, "grad_norm": 0.5004, "lr": 0.0003, "router_trainable": true, "sec_per_step": 0.175, "alpha": 0.03, "t": 537.3} |
| {"event": "probe", "step": 3000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3098.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.178015, "cps": 24.203125, "mean_p": 0.187939, "boundary_entropy": 0.112516, "bpb": 2.046091}, "de": {"chunks": 3880.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.188964, "cps": 30.3125, "mean_p": 0.197744, "boundary_entropy": 0.116459, "bpb": 2.255632}, "vi": {"chunks": 4177.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.22658, "cps": 32.632812, "mean_p": 0.190202, "boundary_entropy": 0.160895, "bpb": 1.884873}, "tr": {"chunks": 3472.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.191791, "cps": 27.125, "mean_p": 0.191252, "boundary_entropy": 0.118358, "bpb": 2.204967}, "fi": {"chunks": 3544.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.1884, "cps": 27.6875, "mean_p": 0.196649, "boundary_entropy": 0.1205, "bpb": 2.413817}, "ru": {"chunks": 5830.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.307652, "cps": 45.546875, "mean_p": 0.193251, "boundary_entropy": 0.179963, "bpb": 1.366996}, "el": {"chunks": 6350.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.309409, "cps": 49.609375, "mean_p": 0.206847, "boundary_entropy": 0.22755, "bpb": 1.419283}, "hi": {"chunks": 7786.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.462764, "cps": 60.828125, "mean_p": 0.227333, "boundary_entropy": 0.324664, "bpb": 1.189588}, "ta": {"chunks": 8761.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.468628, "cps": 68.445312, "mean_p": 0.206556, "boundary_entropy": 0.287805, "bpb": 0.984882}, "th": {"chunks": 8016.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.498353, "cps": 62.625, "mean_p": 0.204925, "boundary_entropy": 0.303604, "bpb": 1.130734}, "zh": {"chunks": 2924.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.49509, "cps": 22.84375, "mean_p": 0.263593, "boundary_entropy": 0.349087, "bpb": 3.116328}, "my": {"chunks": 10533.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.55589, "cps": 82.289062, "mean_p": 0.319481, "boundary_entropy": 0.512274, "bpb": 1.110565}}, "bpb": {"en": 2.04609, "de": 2.25563, "vi": 1.88487, "tr": 2.20497, "fi": 2.41382, "ru": 1.367, "el": 1.41928, "hi": 1.18959, "ta": 0.98488, "th": 1.13073, "zh": 3.11633, "my": 1.11057}, "gini_cpc": 0.228897, "gini_cps": 0.239803, "bpic": 5.613213, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0615049536690326, "vi": 1.2728113520836297, "tr": 1.0773873495638453, "fi": 1.0583383573729352, "ru": 1.7282320196364662, "el": 1.7381033111317863, "hi": 2.5995730936989494, "ta": 2.632515382759948, "th": 2.7994927689066293, "zh": 2.781164213084832, "my": 3.122708280608421}, "t": 625.4} |
| {"event": "train", "step": 3000, "phase": 1, "objective": "global", "lm_loss": 1.33374, "bpb": 1.92419, "chunk_loss": 0.030147, "F": 0.18057, "G": 0.2046, "chunks": 184.91, "bytes_per_chunk": 5.538, "boundary_entropy": 0.17073, "router_noise": 0.0, "grad_norm": 0.4508, "lr": 0.0002999, "router_trainable": true, "sec_per_step": 0.177, "alpha": 0.03, "t": 626.0} |
| {"event": "probe", "step": 3500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3148.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.180888, "cps": 24.59375, "mean_p": 0.190412, "boundary_entropy": 0.107194, "bpb": 1.97753}, "de": {"chunks": 3975.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.193591, "cps": 31.054688, "mean_p": 0.200881, "boundary_entropy": 0.104695, "bpb": 2.160045}, "vi": {"chunks": 4167.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.226037, "cps": 32.554688, "mean_p": 0.189649, "boundary_entropy": 0.144943, "bpb": 1.811529}, "tr": {"chunks": 3387.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.187096, "cps": 26.460938, "mean_p": 0.184874, "boundary_entropy": 0.107462, "bpb": 2.102795}, "fi": {"chunks": 3526.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.187444, "cps": 27.546875, "mean_p": 0.195864, "boundary_entropy": 0.111462, "bpb": 2.314712}, "ru": {"chunks": 5917.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.312243, "cps": 46.226562, "mean_p": 0.190936, "boundary_entropy": 0.173847, "bpb": 1.316819}, "el": {"chunks": 6745.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.328656, "cps": 52.695312, "mean_p": 0.207581, "boundary_entropy": 0.193191, "bpb": 1.352171}, "hi": {"chunks": 7837.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.465795, "cps": 61.226562, "mean_p": 0.233166, "boundary_entropy": 0.330766, "bpb": 1.140805}, "ta": {"chunks": 8450.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.451993, "cps": 66.015625, "mean_p": 0.190182, "boundary_entropy": 0.261408, "bpb": 0.949869}, "th": {"chunks": 7989.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.496674, "cps": 62.414062, "mean_p": 0.215884, "boundary_entropy": 0.323665, "bpb": 1.083375}, "zh": {"chunks": 2756.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.466644, "cps": 21.53125, "mean_p": 0.24469, "boundary_entropy": 0.311985, "bpb": 3.023648}, "my": {"chunks": 8322.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.439202, "cps": 65.015625, "mean_p": 0.298452, "boundary_entropy": 0.515548, "bpb": 1.083448}}, "bpb": {"en": 1.97753, "de": 2.16004, "vi": 1.81153, "tr": 2.10279, "fi": 2.31471, "ru": 1.31682, "el": 1.35217, "hi": 1.1408, "ta": 0.94987, "th": 1.08338, "zh": 3.02365, "my": 1.08345}, "gini_cpc": 0.209292, "gini_cps": 0.217748, "bpic": 5.795633, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.070222611247608, "vi": 1.249596370226928, "tr": 1.0343179024782725, "fi": 1.036238731872866, "ru": 1.726162793816403, "el": 1.8168978264136169, "hi": 2.5750411308578665, "ta": 2.4987375080413767, "th": 2.745748483576173, "zh": 2.5797354229390477, "my": 2.428028230274577}, "t": 714.2} |
| {"event": "train", "step": 3500, "phase": 1, "objective": "global", "lm_loss": 1.22474, "bpb": 1.76692, "chunk_loss": 0.03013, "F": 0.18079, "G": 0.20321, "chunks": 185.12, "bytes_per_chunk": 5.531, "boundary_entropy": 0.17136, "router_noise": 0.0, "grad_norm": 0.3981, "lr": 0.0002997, "router_trainable": true, "sec_per_step": 0.178, "alpha": 0.03, "t": 714.8} |
| {"event": "probe", "step": 4000, "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": 0.184931, "boundary_entropy": 0.09049, "bpb": 1.891703}, "de": {"chunks": 3906.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.19023, "cps": 30.515625, "mean_p": 0.196092, "boundary_entropy": 0.098338, "bpb": 2.057883}, "vi": {"chunks": 3989.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.216382, "cps": 31.164062, "mean_p": 0.176872, "boundary_entropy": 0.132108, "bpb": 1.740817}, "tr": {"chunks": 3184.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.175882, "cps": 24.875, "mean_p": 0.171552, "boundary_entropy": 0.097798, "bpb": 1.988563}, "fi": {"chunks": 3418.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.181702, "cps": 26.703125, "mean_p": 0.186444, "boundary_entropy": 0.096733, "bpb": 2.198381}, "ru": {"chunks": 5730.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.302375, "cps": 44.765625, "mean_p": 0.178812, "boundary_entropy": 0.150324, "bpb": 1.288163}, "el": {"chunks": 6557.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.319495, "cps": 51.226562, "mean_p": 0.197487, "boundary_entropy": 0.169391, "bpb": 1.310051}, "hi": {"chunks": 7770.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.461813, "cps": 60.703125, "mean_p": 0.217401, "boundary_entropy": 0.305187, "bpb": 1.100346}, "ta": {"chunks": 7779.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.416101, "cps": 60.773438, "mean_p": 0.187789, "boundary_entropy": 0.278064, "bpb": 0.96339}, "th": {"chunks": 8028.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.499099, "cps": 62.71875, "mean_p": 0.206037, "boundary_entropy": 0.310565, "bpb": 1.074241}, "zh": {"chunks": 2699.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.456993, "cps": 21.085938, "mean_p": 0.225858, "boundary_entropy": 0.275153, "bpb": 2.953162}, "my": {"chunks": 7590.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.40057, "cps": 59.296875, "mean_p": 0.371073, "boundary_entropy": 0.5928, "bpb": 1.063853}}, "bpb": {"en": 1.8917, "de": 2.05788, "vi": 1.74082, "tr": 1.98856, "fi": 2.19838, "ru": 1.28816, "el": 1.31005, "hi": 1.10035, "ta": 0.96339, "th": 1.07424, "zh": 2.95316, "my": 1.06385}, "gini_cpc": 0.213634, "gini_cps": 0.213667, "bpic": 6.020661, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.069999667239118, "vi": 1.2170956359009775, "tr": 0.989296148902454, "fi": 1.022030804364045, "ru": 1.7007842227615624, "el": 1.7970830557268729, "hi": 2.5975849340764996, "ta": 2.3404647945544377, "th": 2.80730829813095, "zh": 2.570474221103691, "my": 2.253109043100847}, "t": 801.7} |
| {"event": "eval", "step": 4000, "phase": 1, "bpb": {"en": 1.89361, "de": 2.0597, "vi": 1.75149, "tr": 1.9814, "fi": 2.20877, "ru": 1.28507, "el": 1.30829, "hi": 1.10737, "ta": 0.96379, "th": 1.0698, "zh": 2.96403, "my": 1.07107}, "bytes_per_chunk": {"en": 5.461, "de": 5.331, "vi": 6.066, "tr": 5.878, "fi": 5.526, "ru": 5.784, "el": 5.91, "hi": 5.609, "ta": 6.288, "th": 5.693, "zh": 6.073, "my": 7.351}, "bpb_mean": 1.6387, "t": 807.0} |
| {"event": "train", "step": 4000, "phase": 1, "objective": "global", "lm_loss": 1.18272, "bpb": 1.7063, "chunk_loss": 0.030076, "F": 0.17517, "G": 0.19543, "chunks": 179.38, "bytes_per_chunk": 5.709, "boundary_entropy": 0.16385, "router_noise": 0.0, "grad_norm": 0.3692, "lr": 0.0002995, "router_trainable": true, "sec_per_step": 0.185, "alpha": 0.03, "t": 807.2} |
| {"event": "probe", "step": 4500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3116.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.17905, "cps": 24.34375, "mean_p": 0.184384, "boundary_entropy": 0.083883, "bpb": 1.849839}, "de": {"chunks": 3943.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.192032, "cps": 30.804688, "mean_p": 0.195534, "boundary_entropy": 0.089927, "bpb": 2.004008}, "vi": {"chunks": 4069.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.220721, "cps": 31.789062, "mean_p": 0.182326, "boundary_entropy": 0.129168, "bpb": 1.677961}, "tr": {"chunks": 3344.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.184721, "cps": 26.125, "mean_p": 0.180299, "boundary_entropy": 0.093534, "bpb": 1.920432}, "fi": {"chunks": 3414.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.18149, "cps": 26.671875, "mean_p": 0.185674, "boundary_entropy": 0.089349, "bpb": 2.146748}, "ru": {"chunks": 5855.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.308971, "cps": 45.742188, "mean_p": 0.181424, "boundary_entropy": 0.150558, "bpb": 1.25415}, "el": {"chunks": 6561.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.31969, "cps": 51.257812, "mean_p": 0.195215, "boundary_entropy": 0.152005, "bpb": 1.275804}, "hi": {"chunks": 7891.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.469004, "cps": 61.648438, "mean_p": 0.213157, "boundary_entropy": 0.282639, "bpb": 1.066424}, "ta": {"chunks": 8416.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.450174, "cps": 65.75, "mean_p": 0.18593, "boundary_entropy": 0.23308, "bpb": 0.904951}, "th": {"chunks": 7968.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.495368, "cps": 62.25, "mean_p": 0.210796, "boundary_entropy": 0.312977, "bpb": 1.031511}, "zh": {"chunks": 2867.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.485439, "cps": 22.398438, "mean_p": 0.237454, "boundary_entropy": 0.272922, "bpb": 2.864163}, "my": {"chunks": 8460.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.446485, "cps": 66.09375, "mean_p": 0.357198, "boundary_entropy": 0.577726, "bpb": 1.015012}}, "bpb": {"en": 1.84984, "de": 2.00401, "vi": 1.67796, "tr": 1.92043, "fi": 2.14675, "ru": 1.25415, "el": 1.2758, "hi": 1.06642, "ta": 0.90495, "th": 1.03151, "zh": 2.86416, "my": 1.01501}, "gini_cpc": 0.215768, "gini_cps": 0.219531, "bpic": 5.823334, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.072509236673211, "vi": 1.2327392361114737, "tr": 1.031673769204134, "fi": 1.0136273132188722, "ru": 1.7256167842542194, "el": 1.7854835931282709, "hi": 2.619410968075427, "ta": 2.5142411407222416, "th": 2.7666545226877592, "zh": 2.7111960400576076, "my": 2.4936394396609542}, "t": 893.2} |
| {"event": "train", "step": 4500, "phase": 1, "objective": "global", "lm_loss": 1.30253, "bpb": 1.87915, "chunk_loss": 0.030105, "F": 0.17772, "G": 0.19318, "chunks": 181.98, "bytes_per_chunk": 5.627, "boundary_entropy": 0.12826, "router_noise": 0.0, "grad_norm": 0.3761, "lr": 0.0002992, "router_trainable": true, "sec_per_step": 0.173, "alpha": 0.03, "t": 893.7} |
| {"event": "probe", "step": 5000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3123.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.179452, "cps": 24.398438, "mean_p": 0.185431, "boundary_entropy": 0.078641, "bpb": 1.816798}, "de": {"chunks": 4036.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.196562, "cps": 31.53125, "mean_p": 0.199663, "boundary_entropy": 0.085692, "bpb": 1.969073}, "vi": {"chunks": 4064.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.22045, "cps": 31.75, "mean_p": 0.182316, "boundary_entropy": 0.126925, "bpb": 1.64724}, "tr": {"chunks": 3427.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.189306, "cps": 26.773438, "mean_p": 0.184061, "boundary_entropy": 0.089071, "bpb": 1.892953}, "fi": {"chunks": 3471.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.18452, "cps": 27.117188, "mean_p": 0.189281, "boundary_entropy": 0.088295, "bpb": 2.101597}, "ru": {"chunks": 5831.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.307704, "cps": 45.554688, "mean_p": 0.180457, "boundary_entropy": 0.145127, "bpb": 1.216107}, "el": {"chunks": 6376.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.310676, "cps": 49.8125, "mean_p": 0.189692, "boundary_entropy": 0.147613, "bpb": 1.242813}, "hi": {"chunks": 7529.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.447489, "cps": 58.820312, "mean_p": 0.205059, "boundary_entropy": 0.281232, "bpb": 1.064782}, "ta": {"chunks": 8255.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.441562, "cps": 64.492188, "mean_p": 0.184701, "boundary_entropy": 0.241405, "bpb": 0.875335}, "th": {"chunks": 7579.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.471184, "cps": 59.210938, "mean_p": 0.2191, "boundary_entropy": 0.344339, "bpb": 0.997919}, "zh": {"chunks": 2975.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.503725, "cps": 23.242188, "mean_p": 0.227459, "boundary_entropy": 0.243113, "bpb": 2.753206}, "my": {"chunks": 8425.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.444638, "cps": 65.820312, "mean_p": 0.348146, "boundary_entropy": 0.565495, "bpb": 1.00595}}, "bpb": {"en": 1.8168, "de": 1.96907, "vi": 1.64724, "tr": 1.89295, "fi": 2.1016, "ru": 1.21611, "el": 1.24281, "hi": 1.06478, "ta": 0.87533, "th": 0.99792, "zh": 2.75321, "my": 1.00595}, "gini_cpc": 0.210852, "gini_cps": 0.21032, "bpic": 5.896069, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0953448896233344, "vi": 1.2284647332958676, "tr": 1.054910679486255, "fi": 1.0282408799682938, "ru": 1.714691373983307, "el": 1.7312492537249857, "hi": 2.493643470602761, "ta": 2.4606154357617442, "th": 2.6256871441552248, "zh": 2.807021010886859, "my": 2.4777567586257105}, "t": 978.9} |
| {"event": "train", "step": 5000, "phase": 1, "objective": "global", "lm_loss": 1.18927, "bpb": 1.71575, "chunk_loss": 0.030075, "F": 0.17915, "G": 0.20301, "chunks": 183.45, "bytes_per_chunk": 5.582, "boundary_entropy": 0.16394, "router_noise": 0.0, "grad_norm": 0.3625, "lr": 0.0002989, "router_trainable": true, "sec_per_step": 0.171, "alpha": 0.03, "t": 979.4} |
| {"event": "probe", "step": 5500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3130.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.179854, "cps": 24.453125, "mean_p": 0.18422, "boundary_entropy": 0.073985, "bpb": 1.79464}, "de": {"chunks": 3926.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.191204, "cps": 30.671875, "mean_p": 0.19317, "boundary_entropy": 0.079704, "bpb": 1.942195}, "vi": {"chunks": 3964.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.215026, "cps": 30.96875, "mean_p": 0.176517, "boundary_entropy": 0.125193, "bpb": 1.609989}, "tr": {"chunks": 3296.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.182069, "cps": 25.75, "mean_p": 0.175296, "boundary_entropy": 0.079897, "bpb": 1.85128}, "fi": {"chunks": 3430.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.18234, "cps": 26.796875, "mean_p": 0.183567, "boundary_entropy": 0.07724, "bpb": 2.061818}, "ru": {"chunks": 5681.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.299789, "cps": 44.382812, "mean_p": 0.17156, "boundary_entropy": 0.132678, "bpb": 1.192733}, "el": {"chunks": 6574.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.320324, "cps": 51.359375, "mean_p": 0.194774, "boundary_entropy": 0.141938, "bpb": 1.224336}, "hi": {"chunks": 7737.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.459851, "cps": 60.445312, "mean_p": 0.211982, "boundary_entropy": 0.271946, "bpb": 1.01778}, "ta": {"chunks": 9171.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.490559, "cps": 71.648438, "mean_p": 0.18944, "boundary_entropy": 0.207732, "bpb": 0.856487}, "th": {"chunks": 7895.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.49083, "cps": 61.679688, "mean_p": 0.215522, "boundary_entropy": 0.320867, "bpb": 0.982561}, "zh": {"chunks": 2925.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.495259, "cps": 22.851562, "mean_p": 0.213747, "boundary_entropy": 0.201007, "bpb": 2.69204}, "my": {"chunks": 8372.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.441841, "cps": 65.40625, "mean_p": 0.384034, "boundary_entropy": 0.597479, "bpb": 0.982602}}, "bpb": {"en": 1.79464, "de": 1.94219, "vi": 1.60999, "tr": 1.85128, "fi": 2.06182, "ru": 1.19273, "el": 1.22434, "hi": 1.01778, "ta": 0.85649, "th": 0.98256, "zh": 2.69204, "my": 0.9826}, "gini_cpc": 0.220524, "gini_cps": 0.226009, "bpic": 5.805979, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0631086963726593, "vi": 1.195556999768637, "tr": 1.0123167766131997, "fi": 1.01382271911802, "ru": 1.6668455410656933, "el": 1.7810193472639364, "hi": 2.556803231916601, "ta": 2.727539235431387, "th": 2.7290459728220986, "zh": 2.7536720116760023, "my": 2.456663233696257}, "t": 1064.3} |
| {"event": "train", "step": 5500, "phase": 1, "objective": "global", "lm_loss": 1.11672, "bpb": 1.61109, "chunk_loss": 0.030104, "F": 0.17931, "G": 0.19433, "chunks": 183.61, "bytes_per_chunk": 5.577, "boundary_entropy": 0.13906, "router_noise": 0.0, "grad_norm": 0.3272, "lr": 0.0002985, "router_trainable": true, "sec_per_step": 0.171, "alpha": 0.03, "t": 1064.9} |
| {"event": "probe", "step": 6000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3102.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.178245, "cps": 24.234375, "mean_p": 0.1821, "boundary_entropy": 0.070562, "bpb": 1.774429}, "de": {"chunks": 3865.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.188234, "cps": 30.195312, "mean_p": 0.190638, "boundary_entropy": 0.076167, "bpb": 1.92326}, "vi": {"chunks": 4049.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.219637, "cps": 31.632812, "mean_p": 0.181112, "boundary_entropy": 0.116889, "bpb": 1.571882}, "tr": {"chunks": 3295.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.182014, "cps": 25.742188, "mean_p": 0.175193, "boundary_entropy": 0.07869, "bpb": 1.826078}, "fi": {"chunks": 3456.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.183722, "cps": 27.0, "mean_p": 0.185502, "boundary_entropy": 0.075135, "bpb": 2.051146}, "ru": {"chunks": 5779.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.30496, "cps": 45.148438, "mean_p": 0.176707, "boundary_entropy": 0.127746, "bpb": 1.169965}, "el": {"chunks": 6693.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.326122, "cps": 52.289062, "mean_p": 0.196721, "boundary_entropy": 0.142871, "bpb": 1.195215}, "hi": {"chunks": 7592.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.451233, "cps": 59.3125, "mean_p": 0.216276, "boundary_entropy": 0.285345, "bpb": 0.993197}, "ta": {"chunks": 8561.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.45793, "cps": 66.882812, "mean_p": 0.182913, "boundary_entropy": 0.212297, "bpb": 0.839626}, "th": {"chunks": 7875.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.489587, "cps": 61.523438, "mean_p": 0.216497, "boundary_entropy": 0.324678, "bpb": 0.952955}, "zh": {"chunks": 2872.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.486285, "cps": 22.4375, "mean_p": 0.205907, "boundary_entropy": 0.185708, "bpb": 2.660243}, "my": {"chunks": 11328.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.597847, "cps": 88.5, "mean_p": 0.409905, "boundary_entropy": 0.603885, "bpb": 0.964309}}, "bpb": {"en": 1.77443, "de": 1.92326, "vi": 1.57188, "tr": 1.82608, "fi": 2.05115, "ru": 1.16997, "el": 1.19521, "hi": 0.9932, "ta": 0.83963, "th": 0.95296, "zh": 2.66024, "my": 0.96431}, "gini_cpc": 0.237061, "gini_cps": 0.249635, "bpic": 5.605343, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0560376884462348, "vi": 1.23221633435265, "tr": 1.0211444804717813, "fi": 1.0307282448322606, "ru": 1.7109046508423367, "el": 1.8296259299021118, "hi": 2.5315321850543624, "ta": 2.569102041682035, "th": 2.7467037720908447, "zh": 2.7281818771324575, "my": 3.3540705316270243}, "t": 1150.1} |
| {"event": "train", "step": 6000, "phase": 1, "objective": "global", "lm_loss": 1.13501, "bpb": 1.63747, "chunk_loss": 0.030148, "F": 0.17854, "G": 0.20234, "chunks": 182.83, "bytes_per_chunk": 5.601, "boundary_entropy": 0.15288, "router_noise": 0.0, "grad_norm": 0.3179, "lr": 0.0002981, "router_trainable": true, "sec_per_step": 0.171, "alpha": 0.03, "t": 1150.6} |
| {"event": "probe", "step": 6500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3115.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.178992, "cps": 24.335938, "mean_p": 0.181947, "boundary_entropy": 0.069486, "bpb": 1.757498}, "de": {"chunks": 3920.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.190912, "cps": 30.625, "mean_p": 0.192148, "boundary_entropy": 0.072273, "bpb": 1.899921}, "vi": {"chunks": 3925.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.21291, "cps": 30.664062, "mean_p": 0.175467, "boundary_entropy": 0.1232, "bpb": 1.555641}, "tr": {"chunks": 3390.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.187262, "cps": 26.484375, "mean_p": 0.180773, "boundary_entropy": 0.079954, "bpb": 1.809413}, "fi": {"chunks": 3440.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.182872, "cps": 26.875, "mean_p": 0.184201, "boundary_entropy": 0.071683, "bpb": 2.010286}, "ru": {"chunks": 5702.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.300897, "cps": 44.546875, "mean_p": 0.172837, "boundary_entropy": 0.124785, "bpb": 1.15652}, "el": {"chunks": 6452.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.314379, "cps": 50.40625, "mean_p": 0.192442, "boundary_entropy": 0.140437, "bpb": 1.179072}, "hi": {"chunks": 8493.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.504785, "cps": 66.351562, "mean_p": 0.217464, "boundary_entropy": 0.258069, "bpb": 0.968156}, "ta": {"chunks": 8910.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.476598, "cps": 69.609375, "mean_p": 0.185253, "boundary_entropy": 0.199519, "bpb": 0.82971}, "th": {"chunks": 8559.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.532111, "cps": 66.867188, "mean_p": 0.211197, "boundary_entropy": 0.300881, "bpb": 0.939535}, "zh": {"chunks": 2885.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.488486, "cps": 22.539062, "mean_p": 0.20371, "boundary_entropy": 0.180318, "bpb": 2.628559}, "my": {"chunks": 10033.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.529502, "cps": 78.382812, "mean_p": 0.404244, "boundary_entropy": 0.608132, "bpb": 0.948703}}, "bpb": {"en": 1.7575, "de": 1.89992, "vi": 1.55564, "tr": 1.80941, "fi": 2.01029, "ru": 1.15652, "el": 1.17907, "hi": 0.96816, "ta": 0.82971, "th": 0.93954, "zh": 2.62856, "my": 0.9487}, "gini_cpc": 0.234144, "gini_cps": 0.244106, "bpic": 5.576267, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0665954558214594, "vi": 1.189494910973047, "tr": 1.0462012051778973, "fi": 1.0216746749984833, "ru": 1.6810633033623161, "el": 1.7563844471875858, "hi": 2.820149427946547, "ta": 2.6626758770617465, "th": 2.9728160040634855, "zh": 2.7290936822416905, "my": 2.9582406830419727}, "t": 1235.6} |
| {"event": "train", "step": 6500, "phase": 1, "objective": "global", "lm_loss": 1.14274, "bpb": 1.64862, "chunk_loss": 0.030107, "F": 0.17836, "G": 0.18848, "chunks": 182.64, "bytes_per_chunk": 5.607, "boundary_entropy": 0.12388, "router_noise": 0.0, "grad_norm": 0.3038, "lr": 0.0002976, "router_trainable": true, "sec_per_step": 0.171, "alpha": 0.03, "t": 1236.0} |
| {"event": "probe", "step": 7000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3091.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.177613, "cps": 24.148438, "mean_p": 0.179665, "boundary_entropy": 0.066583, "bpb": 1.74752}, "de": {"chunks": 3872.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.188574, "cps": 30.25, "mean_p": 0.190488, "boundary_entropy": 0.070409, "bpb": 1.883939}, "vi": {"chunks": 3961.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.214863, "cps": 30.945312, "mean_p": 0.177471, "boundary_entropy": 0.115171, "bpb": 1.530593}, "tr": {"chunks": 3320.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.183395, "cps": 25.9375, "mean_p": 0.174497, "boundary_entropy": 0.070876, "bpb": 1.787924}, "fi": {"chunks": 3415.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.181543, "cps": 26.679688, "mean_p": 0.181637, "boundary_entropy": 0.068634, "bpb": 1.99339}, "ru": {"chunks": 5673.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.299367, "cps": 44.320312, "mean_p": 0.17044, "boundary_entropy": 0.11303, "bpb": 1.14423}, "el": {"chunks": 6504.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.316913, "cps": 50.8125, "mean_p": 0.188388, "boundary_entropy": 0.123212, "bpb": 1.166934}, "hi": {"chunks": 7527.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.44737, "cps": 58.804688, "mean_p": 0.198681, "boundary_entropy": 0.251942, "bpb": 0.964819}, "ta": {"chunks": 8691.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.464884, "cps": 67.898438, "mean_p": 0.180347, "boundary_entropy": 0.202122, "bpb": 0.824577}, "th": {"chunks": 7429.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.461859, "cps": 58.039062, "mean_p": 0.208985, "boundary_entropy": 0.318622, "bpb": 0.937577}, "zh": {"chunks": 2988.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.505926, "cps": 23.34375, "mean_p": 0.211423, "boundary_entropy": 0.174015, "bpb": 2.608863}, "my": {"chunks": 7234.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.381782, "cps": 56.515625, "mean_p": 0.375483, "boundary_entropy": 0.598173, "bpb": 0.948899}}, "bpb": {"en": 1.74752, "de": 1.88394, "vi": 1.53059, "tr": 1.78792, "fi": 1.99339, "ru": 1.14423, "el": 1.16693, "hi": 0.96482, "ta": 0.82458, "th": 0.93758, "zh": 2.60886, "my": 0.9489}, "gini_cpc": 0.21624, "gini_cps": 0.209459, "bpic": 6.024347, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0617152529224287, "vi": 1.2097254450636101, "tr": 1.0325537003622447, "fi": 1.022124832794595, "ru": 1.6854997187340215, "el": 1.7842873749540742, "hi": 2.5187899875158815, "ta": 2.6173957660685514, "th": 2.6003655764773583, "zh": 2.848474038925152, "my": 2.1495138288876836}, "t": 1320.6} |
| {"event": "train", "step": 7000, "phase": 1, "objective": "global", "lm_loss": 1.24051, "bpb": 1.78967, "chunk_loss": 0.030121, "F": 0.17902, "G": 0.19118, "chunks": 183.31, "bytes_per_chunk": 5.586, "boundary_entropy": 0.11116, "router_noise": 0.0, "grad_norm": 0.3102, "lr": 0.000297, "router_trainable": true, "sec_per_step": 0.17, "alpha": 0.03, "t": 1321.1} |
| {"event": "probe", "step": 7500, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3117.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.179107, "cps": 24.351562, "mean_p": 0.18156, "boundary_entropy": 0.060119, "bpb": 1.733767}, "de": {"chunks": 3907.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.190279, "cps": 30.523438, "mean_p": 0.191617, "boundary_entropy": 0.06294, "bpb": 1.863535}, "vi": {"chunks": 4006.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.217304, "cps": 31.296875, "mean_p": 0.1774, "boundary_entropy": 0.107785, "bpb": 1.515181}, "tr": {"chunks": 3287.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.181572, "cps": 25.679688, "mean_p": 0.172698, "boundary_entropy": 0.068349, "bpb": 1.766053}, "fi": {"chunks": 3409.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.181224, "cps": 26.632812, "mean_p": 0.181596, "boundary_entropy": 0.065228, "bpb": 1.981275}, "ru": {"chunks": 5702.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.300897, "cps": 44.546875, "mean_p": 0.169393, "boundary_entropy": 0.115775, "bpb": 1.136142}, "el": {"chunks": 6304.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.307168, "cps": 49.25, "mean_p": 0.183103, "boundary_entropy": 0.119065, "bpb": 1.153502}, "hi": {"chunks": 7493.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.445349, "cps": 58.539062, "mean_p": 0.192558, "boundary_entropy": 0.235655, "bpb": 0.948579}, "ta": {"chunks": 8758.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.468468, "cps": 68.421875, "mean_p": 0.180641, "boundary_entropy": 0.176491, "bpb": 0.80733}, "th": {"chunks": 8563.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.532359, "cps": 66.898438, "mean_p": 0.205854, "boundary_entropy": 0.279054, "bpb": 0.927806}, "zh": {"chunks": 2870.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.485946, "cps": 22.421875, "mean_p": 0.196614, "boundary_entropy": 0.146671, "bpb": 2.579525}, "my": {"chunks": 8951.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.472398, "cps": 69.929688, "mean_p": 0.404156, "boundary_entropy": 0.605017, "bpb": 0.927158}}, "bpb": {"en": 1.73377, "de": 1.86353, "vi": 1.51518, "tr": 1.76605, "fi": 1.98128, "ru": 1.13614, "el": 1.1535, "hi": 0.94858, "ta": 0.80733, "th": 0.92781, "zh": 2.57952, "my": 0.92716}, "gini_cpc": 0.226006, "gini_cps": 0.230912, "bpic": 5.782708, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0623761735316848, "vi": 1.213263467903382, "tr": 1.01376307969736, "fi": 1.011818080558091, "ru": 1.6799846615250607, "el": 1.714994285385204, "hi": 2.486497217721349, "ta": 2.6155726608767815, "th": 2.972296955867574, "zh": 2.713162288754165, "my": 2.637518405572985}, "t": 1406.0} |
| {"event": "train", "step": 7500, "phase": 1, "objective": "global", "lm_loss": 1.08802, "bpb": 1.56969, "chunk_loss": 0.030099, "F": 0.17792, "G": 0.1868, "chunks": 182.19, "bytes_per_chunk": 5.621, "boundary_entropy": 0.11683, "router_noise": 0.0, "grad_norm": 0.2852, "lr": 0.0002964, "router_trainable": true, "sec_per_step": 0.171, "alpha": 0.03, "t": 1406.5} |
| {"event": "probe", "step": 8000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3126.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.179624, "cps": 24.421875, "mean_p": 0.182167, "boundary_entropy": 0.058659, "bpb": 1.726018}, "de": {"chunks": 3899.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.189889, "cps": 30.460938, "mean_p": 0.191019, "boundary_entropy": 0.063124, "bpb": 1.857406}, "vi": {"chunks": 3966.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.215134, "cps": 30.984375, "mean_p": 0.177043, "boundary_entropy": 0.111434, "bpb": 1.506171}, "tr": {"chunks": 3327.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.183782, "cps": 25.992188, "mean_p": 0.177014, "boundary_entropy": 0.075771, "bpb": 1.767251}, "fi": {"chunks": 3399.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180692, "cps": 26.554688, "mean_p": 0.180426, "boundary_entropy": 0.063216, "bpb": 1.959687}, "ru": {"chunks": 5692.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.300369, "cps": 44.46875, "mean_p": 0.170251, "boundary_entropy": 0.099373, "bpb": 1.123269}, "el": {"chunks": 6441.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.313843, "cps": 50.320312, "mean_p": 0.186763, "boundary_entropy": 0.124127, "bpb": 1.139891}, "hi": {"chunks": 6901.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.410163, "cps": 53.914062, "mean_p": 0.180545, "boundary_entropy": 0.230522, "bpb": 0.94589}, "ta": {"chunks": 8434.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.451137, "cps": 65.890625, "mean_p": 0.176897, "boundary_entropy": 0.185761, "bpb": 0.800226}, "th": {"chunks": 7423.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.461486, "cps": 57.992188, "mean_p": 0.198021, "boundary_entropy": 0.299411, "bpb": 0.920214}, "zh": {"chunks": 2955.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.500339, "cps": 23.085938, "mean_p": 0.202663, "boundary_entropy": 0.146016, "bpb": 2.571462}, "my": {"chunks": 10127.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.534463, "cps": 79.117188, "mean_p": 0.400208, "boundary_entropy": 0.607209, "bpb": 0.934218}}, "bpb": {"en": 1.72602, "de": 1.85741, "vi": 1.50617, "tr": 1.76725, "fi": 1.95969, "ru": 1.12327, "el": 1.13989, "hi": 0.94589, "ta": 0.80023, "th": 0.92021, "zh": 2.57146, "my": 0.93422}, "gini_cpc": 0.225899, "gini_cps": 0.231134, "bpic": 5.842305, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0571484431269433, "vi": 1.1976908024094617, "tr": 1.0231454940661049, "fi": 1.0059454431601715, "ru": 1.6722100284109613, "el": 1.7472200329895313, "hi": 2.283453126203181, "ta": 2.511558357490353, "th": 2.569174139065676, "zh": 2.7854745133775274, "my": 2.975449477446508}, "t": 1491.4} |
| {"event": "eval", "step": 8000, "phase": 1, "bpb": {"en": 1.72367, "de": 1.86167, "vi": 1.52197, "tr": 1.76144, "fi": 1.96816, "ru": 1.12184, "el": 1.13806, "hi": 0.95155, "ta": 0.80469, "th": 0.91834, "zh": 2.57631, "my": 0.94001}, "bytes_per_chunk": {"en": 5.455, "de": 5.401, "vi": 6.092, "tr": 5.729, "fi": 5.589, "ru": 5.832, "el": 5.955, "hi": 6.23, "ta": 5.956, "th": 6.374, "zh": 5.668, "my": 5.419}, "bpb_mean": 1.44064, "t": 1496.7} |
| {"event": "train", "step": 8000, "phase": 1, "objective": "global", "lm_loss": 1.09, "bpb": 1.57254, "chunk_loss": 0.030184, "F": 0.18002, "G": 0.19515, "chunks": 184.34, "bytes_per_chunk": 5.555, "boundary_entropy": 0.12525, "router_noise": 0.0, "grad_norm": 0.3285, "lr": 0.0002957, "router_trainable": true, "sec_per_step": 0.181, "alpha": 0.03, "t": 1496.9} |
| {"event": "train", "step": 8500, "phase": 1, "objective": "global", "lm_loss": 1.09827, "bpb": 1.58447, "chunk_loss": 0.030059, "F": 0.17648, "G": 0.19386, "chunks": 180.72, "bytes_per_chunk": 5.666, "boundary_entropy": 0.12944, "router_noise": 0.0, "grad_norm": 0.2876, "lr": 0.0002949, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 1578.6} |
| {"event": "train", "step": 9000, "phase": 1, "objective": "global", "lm_loss": 1.10903, "bpb": 1.59999, "chunk_loss": 0.030064, "F": 0.17592, "G": 0.18994, "chunks": 180.14, "bytes_per_chunk": 5.684, "boundary_entropy": 0.10581, "router_noise": 0.0, "grad_norm": 0.2925, "lr": 0.0002941, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 1660.4} |
| {"event": "train", "step": 9500, "phase": 1, "objective": "global", "lm_loss": 1.04148, "bpb": 1.50254, "chunk_loss": 0.030084, "F": 0.17766, "G": 0.19242, "chunks": 181.92, "bytes_per_chunk": 5.629, "boundary_entropy": 0.12289, "router_noise": 0.0, "grad_norm": 0.275, "lr": 0.0002933, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 1742.3} |
| {"event": "train", "step": 10000, "phase": 1, "objective": "global", "lm_loss": 1.07198, "bpb": 1.54654, "chunk_loss": 0.030079, "F": 0.17581, "G": 0.18714, "chunks": 180.03, "bytes_per_chunk": 5.688, "boundary_entropy": 0.10751, "router_noise": 0.0, "grad_norm": 0.2806, "lr": 0.0002924, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 1824.2} |
| {"event": "train", "step": 10500, "phase": 1, "objective": "global", "lm_loss": 1.02785, "bpb": 1.48287, "chunk_loss": 0.03005, "F": 0.17181, "G": 0.18797, "chunks": 175.94, "bytes_per_chunk": 5.82, "boundary_entropy": 0.12433, "router_noise": 0.0, "grad_norm": 0.2842, "lr": 0.0002914, "router_trainable": true, "sec_per_step": 0.165, "alpha": 0.03, "t": 1906.8} |
| {"event": "train", "step": 11000, "phase": 1, "objective": "global", "lm_loss": 0.99686, "bpb": 1.43817, "chunk_loss": 0.030104, "F": 0.17313, "G": 0.19266, "chunks": 177.28, "bytes_per_chunk": 5.776, "boundary_entropy": 0.13019, "router_noise": 0.0, "grad_norm": 0.2644, "lr": 0.0002903, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 1988.7} |
| {"event": "train", "step": 11500, "phase": 1, "objective": "global", "lm_loss": 0.99717, "bpb": 1.43861, "chunk_loss": 0.030094, "F": 0.17902, "G": 0.18528, "chunks": 183.31, "bytes_per_chunk": 5.586, "boundary_entropy": 0.08874, "router_noise": 0.0, "grad_norm": 0.2623, "lr": 0.0002893, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 2070.6} |
| {"event": "probe", "step": 12000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3091.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.177613, "cps": 24.148438, "mean_p": 0.178184, "boundary_entropy": 0.046105, "bpb": 1.668824}, "de": {"chunks": 3828.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.186432, "cps": 29.90625, "mean_p": 0.186354, "boundary_entropy": 0.046639, "bpb": 1.78994}, "vi": {"chunks": 4056.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.220016, "cps": 31.6875, "mean_p": 0.178581, "boundary_entropy": 0.09277, "bpb": 1.429731}, "tr": {"chunks": 3345.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.184776, "cps": 26.132812, "mean_p": 0.173262, "boundary_entropy": 0.055682, "bpb": 1.692062}, "fi": {"chunks": 3384.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179895, "cps": 26.4375, "mean_p": 0.176805, "boundary_entropy": 0.046517, "bpb": 1.87607}, "ru": {"chunks": 5701.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.300844, "cps": 44.539062, "mean_p": 0.168074, "boundary_entropy": 0.084182, "bpb": 1.070867}, "el": {"chunks": 6439.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.313746, "cps": 50.304688, "mean_p": 0.185036, "boundary_entropy": 0.100504, "bpb": 1.080803}, "hi": {"chunks": 7493.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.445349, "cps": 58.539062, "mean_p": 0.18581, "boundary_entropy": 0.210739, "bpb": 0.878982}, "ta": {"chunks": 8917.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.476972, "cps": 69.664062, "mean_p": 0.178341, "boundary_entropy": 0.152308, "bpb": 0.753206}, "th": {"chunks": 7782.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.483805, "cps": 60.796875, "mean_p": 0.19914, "boundary_entropy": 0.278967, "bpb": 0.860339}, "zh": {"chunks": 2925.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.495259, "cps": 22.851562, "mean_p": 0.19449, "boundary_entropy": 0.101925, "bpb": 2.471587}, "my": {"chunks": 9727.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.513352, "cps": 75.992188, "mean_p": 0.411092, "boundary_entropy": 0.604544, "bpb": 0.852079}}, "bpb": {"en": 1.66882, "de": 1.78994, "vi": 1.42973, "tr": 1.69206, "fi": 1.87607, "ru": 1.07087, "el": 1.0808, "hi": 0.87898, "ta": 0.75321, "th": 0.86034, "zh": 2.47159, "my": 0.85208}, "gini_cpc": 0.225533, "gini_cps": 0.235347, "bpic": 5.754873, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.049650306866492, "vi": 1.2387393095627375, "tr": 1.0403289541300327, "fi": 1.0128463936096368, "ru": 1.6938187725194176, "el": 1.7664554746816241, "hi": 2.5074124321052875, "ta": 2.6854582954819093, "th": 2.7239258199147667, "zh": 2.7884158513574526, "my": 2.8902849064957836}, "t": 2155.0} |
| {"event": "eval", "step": 12000, "phase": 1, "bpb": {"en": 1.66734, "de": 1.79708, "vi": 1.44818, "tr": 1.68378, "fi": 1.88881, "ru": 1.06945, "el": 1.07865, "hi": 0.8856, "ta": 0.75732, "th": 0.85901, "zh": 2.47558, "my": 0.8575}, "bytes_per_chunk": {"en": 5.491, "de": 5.518, "vi": 6.015, "tr": 5.687, "fi": 5.611, "ru": 5.851, "el": 5.934, "hi": 5.709, "ta": 5.643, "th": 5.865, "zh": 5.654, "my": 5.555}, "bpb_mean": 1.37236, "t": 2160.3} |
| {"event": "train", "step": 12000, "phase": 1, "objective": "global", "lm_loss": 1.07584, "bpb": 1.55211, "chunk_loss": 0.030088, "F": 0.17691, "G": 0.18238, "chunks": 181.16, "bytes_per_chunk": 5.653, "boundary_entropy": 0.07963, "router_noise": 0.0, "grad_norm": 0.2731, "lr": 0.0002881, "router_trainable": true, "sec_per_step": 0.18, "alpha": 0.03, "t": 2160.5} |
| {"event": "train", "step": 12500, "phase": 1, "objective": "global", "lm_loss": 1.07744, "bpb": 1.55442, "chunk_loss": 0.030052, "F": 0.17638, "G": 0.18354, "chunks": 180.61, "bytes_per_chunk": 5.67, "boundary_entropy": 0.08945, "router_noise": 0.0, "grad_norm": 0.2749, "lr": 0.0002869, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 2242.2} |
| {"event": "train", "step": 13000, "phase": 1, "objective": "global", "lm_loss": 1.04896, "bpb": 1.51332, "chunk_loss": 0.030075, "F": 0.17708, "G": 0.18616, "chunks": 181.33, "bytes_per_chunk": 5.647, "boundary_entropy": 0.09206, "router_noise": 0.0, "grad_norm": 0.2781, "lr": 0.0002857, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 2323.9} |
| {"event": "train", "step": 13500, "phase": 1, "objective": "global", "lm_loss": 1.0158, "bpb": 1.46548, "chunk_loss": 0.030061, "F": 0.17667, "G": 0.18199, "chunks": 180.91, "bytes_per_chunk": 5.66, "boundary_entropy": 0.08611, "router_noise": 0.0, "grad_norm": 0.2821, "lr": 0.0002844, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 2405.3} |
| {"event": "train", "step": 14000, "phase": 1, "objective": "global", "lm_loss": 1.05159, "bpb": 1.51713, "chunk_loss": 0.030104, "F": 0.1758, "G": 0.19071, "chunks": 180.02, "bytes_per_chunk": 5.688, "boundary_entropy": 0.10112, "router_noise": 0.0, "grad_norm": 0.2819, "lr": 0.000283, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 2486.3} |
| {"event": "train", "step": 14500, "phase": 1, "objective": "global", "lm_loss": 1.09626, "bpb": 1.58157, "chunk_loss": 0.029991, "F": 0.16945, "G": 0.18335, "chunks": 173.52, "bytes_per_chunk": 5.901, "boundary_entropy": 0.0997, "router_noise": 0.0, "grad_norm": 0.2718, "lr": 0.0002816, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 2567.7} |
| {"event": "train", "step": 15000, "phase": 1, "objective": "global", "lm_loss": 0.97512, "bpb": 1.40679, "chunk_loss": 0.030054, "F": 0.17323, "G": 0.17698, "chunks": 177.39, "bytes_per_chunk": 5.773, "boundary_entropy": 0.07338, "router_noise": 0.0, "grad_norm": 0.2661, "lr": 0.0002801, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 2649.3} |
| {"event": "train", "step": 15500, "phase": 1, "objective": "global", "lm_loss": 1.0175, "bpb": 1.46795, "chunk_loss": 0.030066, "F": 0.17845, "G": 0.18543, "chunks": 182.73, "bytes_per_chunk": 5.604, "boundary_entropy": 0.0947, "router_noise": 0.0, "grad_norm": 0.2671, "lr": 0.0002786, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 2730.4} |
| {"event": "probe", "step": 16000, "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": 0.176646, "boundary_entropy": 0.040492, "bpb": 1.633813}, "de": {"chunks": 3797.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.184922, "cps": 29.664062, "mean_p": 0.183409, "boundary_entropy": 0.042257, "bpb": 1.752959}, "vi": {"chunks": 4017.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.217901, "cps": 31.382812, "mean_p": 0.177874, "boundary_entropy": 0.088186, "bpb": 1.392387}, "tr": {"chunks": 3343.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.184666, "cps": 26.117188, "mean_p": 0.172204, "boundary_entropy": 0.050095, "bpb": 1.649384}, "fi": {"chunks": 3396.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180533, "cps": 26.53125, "mean_p": 0.177182, "boundary_entropy": 0.042147, "bpb": 1.822939}, "ru": {"chunks": 5723.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.302005, "cps": 44.710938, "mean_p": 0.166181, "boundary_entropy": 0.071744, "bpb": 1.044191}, "el": {"chunks": 6407.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.312186, "cps": 50.054688, "mean_p": 0.183361, "boundary_entropy": 0.094114, "bpb": 1.050833}, "hi": {"chunks": 7320.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.435067, "cps": 57.1875, "mean_p": 0.176395, "boundary_entropy": 0.190697, "bpb": 0.857966}, "ta": {"chunks": 8653.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.462851, "cps": 67.601562, "mean_p": 0.170329, "boundary_entropy": 0.141913, "bpb": 0.731647}, "th": {"chunks": 7344.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.456574, "cps": 57.375, "mean_p": 0.184225, "boundary_entropy": 0.261171, "bpb": 0.845062}, "zh": {"chunks": 2911.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.492889, "cps": 22.742188, "mean_p": 0.189819, "boundary_entropy": 0.077139, "bpb": 2.429911}, "my": {"chunks": 8337.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.439994, "cps": 65.132812, "mean_p": 0.393389, "boundary_entropy": 0.59339, "bpb": 0.828182}}, "bpb": {"en": 1.63381, "de": 1.75296, "vi": 1.39239, "tr": 1.64938, "fi": 1.82294, "ru": 1.04419, "el": 1.05083, "hi": 0.85797, "ta": 0.73165, "th": 0.84506, "zh": 2.42991, "my": 0.82818}, "gini_cpc": 0.213614, "gini_cps": 0.217241, "bpic": 5.966466, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.046567369837727, "vi": 1.23321185179305, "tr": 1.0451167910452788, "fi": 1.0217268308242065, "ru": 1.7092025484265394, "el": 1.7668223278344461, "hi": 2.4622662269416153, "ta": 2.619511047038793, "th": 2.583988657816326, "zh": 2.7895089739248218, "my": 2.490149523471169}, "t": 2815.0} |
| {"event": "eval", "step": 16000, "phase": 1, "bpb": {"en": 1.63195, "de": 1.76034, "vi": 1.41077, "tr": 1.64267, "fi": 1.83645, "ru": 1.04319, "el": 1.04826, "hi": 0.86302, "ta": 0.73635, "th": 0.84289, "zh": 2.43176, "my": 0.83374}, "bytes_per_chunk": {"en": 5.493, "de": 5.559, "vi": 6.042, "tr": 5.723, "fi": 5.613, "ru": 5.838, "el": 5.96, "hi": 5.841, "ta": 5.794, "th": 6.33, "zh": 5.651, "my": 6.623}, "bpb_mean": 1.34012, "t": 2820.3} |
| {"event": "train", "step": 16000, "phase": 1, "objective": "global", "lm_loss": 0.98252, "bpb": 1.41747, "chunk_loss": 0.03006, "F": 0.17619, "G": 0.18494, "chunks": 180.42, "bytes_per_chunk": 5.676, "boundary_entropy": 0.10055, "router_noise": 0.0, "grad_norm": 0.2561, "lr": 0.000277, "router_trainable": true, "sec_per_step": 0.18, "alpha": 0.03, "t": 2820.5} |
| {"event": "train", "step": 16500, "phase": 1, "objective": "global", "lm_loss": 1.01289, "bpb": 1.46129, "chunk_loss": 0.030177, "F": 0.18059, "G": 0.18442, "chunks": 184.92, "bytes_per_chunk": 5.537, "boundary_entropy": 0.08052, "router_noise": 0.0, "grad_norm": 0.2596, "lr": 0.0002754, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 2901.8} |
| {"event": "train", "step": 17000, "phase": 1, "objective": "global", "lm_loss": 0.97988, "bpb": 1.41367, "chunk_loss": 0.03005, "F": 0.17232, "G": 0.18297, "chunks": 176.45, "bytes_per_chunk": 5.803, "boundary_entropy": 0.09185, "router_noise": 0.0, "grad_norm": 0.2502, "lr": 0.0002737, "router_trainable": true, "sec_per_step": 0.167, "alpha": 0.03, "t": 2985.3} |
| {"event": "train", "step": 17500, "phase": 1, "objective": "global", "lm_loss": 0.97967, "bpb": 1.41336, "chunk_loss": 0.030058, "F": 0.17403, "G": 0.17786, "chunks": 178.2, "bytes_per_chunk": 5.746, "boundary_entropy": 0.07748, "router_noise": 0.0, "grad_norm": 0.2662, "lr": 0.000272, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 3067.2} |
| {"event": "train", "step": 18000, "phase": 1, "objective": "global", "lm_loss": 1.04387, "bpb": 1.50599, "chunk_loss": 0.030057, "F": 0.1752, "G": 0.17936, "chunks": 179.41, "bytes_per_chunk": 5.708, "boundary_entropy": 0.07091, "router_noise": 0.0, "grad_norm": 0.266, "lr": 0.0002703, "router_trainable": true, "sec_per_step": 0.165, "alpha": 0.03, "t": 3149.5} |
| {"event": "train", "step": 18500, "phase": 1, "objective": "global", "lm_loss": 1.0913, "bpb": 1.57441, "chunk_loss": 0.03014, "F": 0.17645, "G": 0.18082, "chunks": 180.69, "bytes_per_chunk": 5.667, "boundary_entropy": 0.07157, "router_noise": 0.0, "grad_norm": 0.4326, "lr": 0.0002685, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 3231.3} |
| {"event": "train", "step": 19000, "phase": 1, "objective": "global", "lm_loss": 1.00101, "bpb": 1.44416, "chunk_loss": 0.030062, "F": 0.17294, "G": 0.17958, "chunks": 177.09, "bytes_per_chunk": 5.782, "boundary_entropy": 0.07661, "router_noise": 0.0, "grad_norm": 0.2621, "lr": 0.0002666, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 3313.3} |
| {"event": "train", "step": 19500, "phase": 1, "objective": "global", "lm_loss": 0.96954, "bpb": 1.39876, "chunk_loss": 0.030046, "F": 0.17473, "G": 0.18503, "chunks": 178.92, "bytes_per_chunk": 5.723, "boundary_entropy": 0.08989, "router_noise": 0.0, "grad_norm": 0.2628, "lr": 0.0002647, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 3394.8} |
| {"event": "probe", "step": 20000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3081.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.177038, "cps": 24.070312, "mean_p": 0.176306, "boundary_entropy": 0.036651, "bpb": 1.609755}, "de": {"chunks": 3808.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.185458, "cps": 29.75, "mean_p": 0.183813, "boundary_entropy": 0.036587, "bpb": 1.721742}, "vi": {"chunks": 4023.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.218226, "cps": 31.429688, "mean_p": 0.174956, "boundary_entropy": 0.080646, "bpb": 1.364261}, "tr": {"chunks": 3345.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.184776, "cps": 26.132812, "mean_p": 0.17198, "boundary_entropy": 0.042738, "bpb": 1.620428}, "fi": {"chunks": 3377.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179523, "cps": 26.382812, "mean_p": 0.175068, "boundary_entropy": 0.035048, "bpb": 1.788538}, "ru": {"chunks": 5701.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.300844, "cps": 44.539062, "mean_p": 0.167166, "boundary_entropy": 0.05812, "bpb": 1.024491}, "el": {"chunks": 6469.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.315207, "cps": 50.539062, "mean_p": 0.184075, "boundary_entropy": 0.089976, "bpb": 1.032023}, "hi": {"chunks": 7753.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.460802, "cps": 60.570312, "mean_p": 0.182252, "boundary_entropy": 0.181537, "bpb": 0.841562}, "ta": {"chunks": 8409.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.449799, "cps": 65.695312, "mean_p": 0.167464, "boundary_entropy": 0.149526, "bpb": 0.72138}, "th": {"chunks": 8291.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.515449, "cps": 64.773438, "mean_p": 0.186134, "boundary_entropy": 0.22137, "bpb": 0.818478}, "zh": {"chunks": 2898.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.490687, "cps": 22.640625, "mean_p": 0.188577, "boundary_entropy": 0.067899, "bpb": 2.392752}, "my": {"chunks": 8354.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.440891, "cps": 65.265625, "mean_p": 0.392791, "boundary_entropy": 0.598609, "bpb": 0.803148}}, "bpb": {"en": 1.60975, "de": 1.72174, "vi": 1.36426, "tr": 1.62043, "fi": 1.78854, "ru": 1.02449, "el": 1.03202, "hi": 0.84156, "ta": 0.72138, "th": 0.81848, "zh": 2.39275, "my": 0.80315}, "gini_cpc": 0.222422, "gini_cps": 0.22165, "bpic": 5.858447, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0475552902680088, "vi": 1.2326486729737287, "tr": 1.0437055492424312, "fi": 1.0140318569947342, "ru": 1.6993163991747873, "el": 1.7804456785155809, "hi": 2.6028379662919883, "ta": 2.540687815710244, "th": 2.9115099040243546, "zh": 2.7716434461216637, "my": 2.490367939807178}, "t": 3479.2} |
| {"event": "eval", "step": 20000, "phase": 1, "bpb": {"en": 1.60742, "de": 1.72867, "vi": 1.38116, "tr": 1.61345, "fi": 1.80343, "ru": 1.02367, "el": 1.02961, "hi": 0.84777, "ta": 0.72845, "th": 0.81775, "zh": 2.39617, "my": 0.80853}, "bytes_per_chunk": {"en": 5.5, "de": 5.536, "vi": 6.056, "tr": 5.745, "fi": 5.64, "ru": 5.858, "el": 5.918, "hi": 5.522, "ta": 5.93, "th": 5.644, "zh": 5.678, "my": 6.564}, "bpb_mean": 1.31551, "t": 3484.5} |
| {"event": "train", "step": 20000, "phase": 1, "objective": "global", "lm_loss": 0.97845, "bpb": 1.4116, "chunk_loss": 0.030031, "F": 0.17619, "G": 0.18725, "chunks": 180.42, "bytes_per_chunk": 5.676, "boundary_entropy": 0.09266, "router_noise": 0.0, "grad_norm": 0.2619, "lr": 0.0002628, "router_trainable": true, "sec_per_step": 0.18, "alpha": 0.03, "t": 3484.7} |
| {"event": "train", "step": 20500, "phase": 1, "objective": "global", "lm_loss": 0.96951, "bpb": 1.39871, "chunk_loss": 0.030065, "F": 0.17659, "G": 0.18006, "chunks": 180.83, "bytes_per_chunk": 5.663, "boundary_entropy": 0.06413, "router_noise": 0.0, "grad_norm": 0.2536, "lr": 0.0002608, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 3566.0} |
| {"event": "train", "step": 21000, "phase": 1, "objective": "global", "lm_loss": 0.9931, "bpb": 1.43275, "chunk_loss": 0.030084, "F": 0.17842, "G": 0.18112, "chunks": 182.7, "bytes_per_chunk": 5.605, "boundary_entropy": 0.07609, "router_noise": 0.0, "grad_norm": 0.2628, "lr": 0.0002587, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 3647.3} |
| {"event": "train", "step": 21500, "phase": 1, "objective": "global", "lm_loss": 0.9832, "bpb": 1.41845, "chunk_loss": 0.030078, "F": 0.177, "G": 0.18647, "chunks": 181.25, "bytes_per_chunk": 5.65, "boundary_entropy": 0.08927, "router_noise": 0.0, "grad_norm": 0.2407, "lr": 0.0002567, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 3728.8} |
| {"event": "train", "step": 22000, "phase": 1, "objective": "global", "lm_loss": 1.04082, "bpb": 1.50158, "chunk_loss": 0.030065, "F": 0.17581, "G": 0.18276, "chunks": 180.03, "bytes_per_chunk": 5.688, "boundary_entropy": 0.07179, "router_noise": 0.0, "grad_norm": 0.2927, "lr": 0.0002545, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 3810.3} |
| {"event": "train", "step": 22500, "phase": 1, "objective": "global", "lm_loss": 0.9355, "bpb": 1.34965, "chunk_loss": 0.030064, "F": 0.17613, "G": 0.1785, "chunks": 180.36, "bytes_per_chunk": 5.678, "boundary_entropy": 0.06551, "router_noise": 0.0, "grad_norm": 0.2473, "lr": 0.0002524, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 3891.4} |
| {"event": "train", "step": 23000, "phase": 1, "objective": "global", "lm_loss": 0.98809, "bpb": 1.42551, "chunk_loss": 0.030078, "F": 0.17838, "G": 0.18392, "chunks": 182.66, "bytes_per_chunk": 5.606, "boundary_entropy": 0.08091, "router_noise": 0.0, "grad_norm": 0.2472, "lr": 0.0002502, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 3972.9} |
| {"event": "train", "step": 23500, "phase": 1, "objective": "global", "lm_loss": 1.0734, "bpb": 1.54858, "chunk_loss": 0.030068, "F": 0.17598, "G": 0.18584, "chunks": 180.2, "bytes_per_chunk": 5.682, "boundary_entropy": 0.08094, "router_noise": 0.0, "grad_norm": 0.276, "lr": 0.000248, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 4054.9} |
| {"event": "probe", "step": 24000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3090.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.177556, "cps": 24.140625, "mean_p": 0.176774, "boundary_entropy": 0.035727, "bpb": 1.58898}, "de": {"chunks": 3772.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.183704, "cps": 29.46875, "mean_p": 0.182167, "boundary_entropy": 0.033933, "bpb": 1.699818}, "vi": {"chunks": 4054.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.219908, "cps": 31.671875, "mean_p": 0.174414, "boundary_entropy": 0.070344, "bpb": 1.336928}, "tr": {"chunks": 3379.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.186654, "cps": 26.398438, "mean_p": 0.173138, "boundary_entropy": 0.040764, "bpb": 1.59797}, "fi": {"chunks": 3393.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180373, "cps": 26.507812, "mean_p": 0.175814, "boundary_entropy": 0.032926, "bpb": 1.758911}, "ru": {"chunks": 5660.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.298681, "cps": 44.21875, "mean_p": 0.1655, "boundary_entropy": 0.051936, "bpb": 1.009673}, "el": {"chunks": 6212.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.302685, "cps": 48.53125, "mean_p": 0.173963, "boundary_entropy": 0.078343, "bpb": 1.015638}, "hi": {"chunks": 7646.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.454443, "cps": 59.734375, "mean_p": 0.180596, "boundary_entropy": 0.172403, "bpb": 0.824861}, "ta": {"chunks": 9062.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.484729, "cps": 70.796875, "mean_p": 0.171484, "boundary_entropy": 0.135268, "bpb": 0.704821}, "th": {"chunks": 7281.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.452658, "cps": 56.882812, "mean_p": 0.166418, "boundary_entropy": 0.208301, "bpb": 0.812769}, "zh": {"chunks": 2877.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.487132, "cps": 22.476562, "mean_p": 0.185548, "boundary_entropy": 0.058087, "bpb": 2.369417}, "my": {"chunks": 7763.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.4097, "cps": 60.648438, "mean_p": 0.362038, "boundary_entropy": 0.572921, "bpb": 0.784808}}, "bpb": {"en": 1.58898, "de": 1.69982, "vi": 1.33693, "tr": 1.59797, "fi": 1.75891, "ru": 1.00967, "el": 1.01564, "hi": 0.82486, "ta": 0.70482, "th": 0.81277, "zh": 2.36942, "my": 0.78481}, "gini_cpc": 0.215879, "gini_cps": 0.21853, "bpic": 5.978922, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0346296442525151, "vi": 1.238529180194912, "tr": 1.0512433974093227, "fi": 1.0158687877832275, "ru": 1.682181520096319, "el": 1.7047325070758879, "hi": 2.559439461042427, "ta": 2.730009945029692, "th": 2.549386058892232, "zh": 2.7435448235955535, "my": 2.307447618520591}, "t": 4140.3} |
| {"event": "eval", "step": 24000, "phase": 1, "bpb": {"en": 1.58587, "de": 1.70798, "vi": 1.35594, "tr": 1.59119, "fi": 1.77577, "ru": 1.00904, "el": 1.01268, "hi": 0.83059, "ta": 0.70995, "th": 0.81319, "zh": 2.37181, "my": 0.79081}, "bytes_per_chunk": {"en": 5.474, "de": 5.576, "vi": 5.998, "tr": 5.714, "fi": 5.622, "ru": 5.869, "el": 6.127, "hi": 5.614, "ta": 5.58, "th": 6.435, "zh": 5.693, "my": 6.918}, "bpb_mean": 1.29624, "t": 4145.5} |
| {"event": "train", "step": 24000, "phase": 1, "objective": "global", "lm_loss": 0.93689, "bpb": 1.35165, "chunk_loss": 0.029994, "F": 0.17491, "G": 0.19251, "chunks": 179.11, "bytes_per_chunk": 5.717, "boundary_entropy": 0.09562, "router_noise": 0.0, "grad_norm": 0.2597, "lr": 0.0002457, "router_trainable": true, "sec_per_step": 0.181, "alpha": 0.03, "t": 4145.6} |
| {"event": "train", "step": 24500, "phase": 1, "objective": "global", "lm_loss": 1.05872, "bpb": 1.52742, "chunk_loss": 0.030062, "F": 0.17567, "G": 0.17789, "chunks": 179.89, "bytes_per_chunk": 5.692, "boundary_entropy": 0.05968, "router_noise": 0.0, "grad_norm": 0.2719, "lr": 0.0002434, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 4227.4} |
| {"event": "train", "step": 25000, "phase": 1, "objective": "global", "lm_loss": 0.99485, "bpb": 1.43527, "chunk_loss": 0.030071, "F": 0.17776, "G": 0.17841, "chunks": 182.03, "bytes_per_chunk": 5.625, "boundary_entropy": 0.05986, "router_noise": 0.0, "grad_norm": 0.2714, "lr": 0.000241, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 4308.9} |
| {"event": "train", "step": 25500, "phase": 1, "objective": "global", "lm_loss": 0.95986, "bpb": 1.38479, "chunk_loss": 0.030046, "F": 0.17355, "G": 0.17886, "chunks": 177.72, "bytes_per_chunk": 5.762, "boundary_entropy": 0.07771, "router_noise": 0.0, "grad_norm": 0.2577, "lr": 0.0002387, "router_trainable": true, "sec_per_step": 0.164, "alpha": 0.03, "t": 4390.7} |
| {"event": "train", "step": 26000, "phase": 1, "objective": "global", "lm_loss": 1.00542, "bpb": 1.45052, "chunk_loss": 0.030074, "F": 0.17397, "G": 0.18725, "chunks": 178.14, "bytes_per_chunk": 5.748, "boundary_entropy": 0.09061, "router_noise": 0.0, "grad_norm": 0.2772, "lr": 0.0002363, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 4472.2} |
| {"event": "train", "step": 26500, "phase": 1, "objective": "global", "lm_loss": 0.97162, "bpb": 1.40175, "chunk_loss": 0.030108, "F": 0.18202, "G": 0.18354, "chunks": 186.39, "bytes_per_chunk": 5.494, "boundary_entropy": 0.06846, "router_noise": 0.0, "grad_norm": 0.2649, "lr": 0.0002338, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 4553.5} |
| {"event": "train", "step": 27000, "phase": 1, "objective": "global", "lm_loss": 0.97415, "bpb": 1.40541, "chunk_loss": 0.030079, "F": 0.17616, "G": 0.1778, "chunks": 180.39, "bytes_per_chunk": 5.677, "boundary_entropy": 0.06346, "router_noise": 0.0, "grad_norm": 0.2743, "lr": 0.0002313, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 4634.6} |
| {"event": "train", "step": 27500, "phase": 1, "objective": "global", "lm_loss": 0.97625, "bpb": 1.40843, "chunk_loss": 0.030054, "F": 0.17535, "G": 0.18415, "chunks": 179.56, "bytes_per_chunk": 5.703, "boundary_entropy": 0.0749, "router_noise": 0.0, "grad_norm": 0.2601, "lr": 0.0002288, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 4716.1} |
| {"event": "probe", "step": 28000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3093.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.177728, "cps": 24.164062, "mean_p": 0.176975, "boundary_entropy": 0.036988, "bpb": 1.572867}, "de": {"chunks": 3809.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.185506, "cps": 29.757812, "mean_p": 0.184433, "boundary_entropy": 0.034924, "bpb": 1.683863}, "vi": {"chunks": 4037.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.218986, "cps": 31.539062, "mean_p": 0.175367, "boundary_entropy": 0.068846, "bpb": 1.321466}, "tr": {"chunks": 3346.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.184831, "cps": 26.140625, "mean_p": 0.172044, "boundary_entropy": 0.041319, "bpb": 1.57632}, "fi": {"chunks": 3397.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180586, "cps": 26.539062, "mean_p": 0.175941, "boundary_entropy": 0.032399, "bpb": 1.732146}, "ru": {"chunks": 5713.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.301478, "cps": 44.632812, "mean_p": 0.166403, "boundary_entropy": 0.054094, "bpb": 0.994143}, "el": {"chunks": 6394.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.311553, "cps": 49.953125, "mean_p": 0.179649, "boundary_entropy": 0.077733, "bpb": 0.997999}, "hi": {"chunks": 7603.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.451887, "cps": 59.398438, "mean_p": 0.176803, "boundary_entropy": 0.168982, "bpb": 0.813065}, "ta": {"chunks": 8866.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.474244, "cps": 69.265625, "mean_p": 0.170811, "boundary_entropy": 0.137281, "bpb": 0.694377}, "th": {"chunks": 7504.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.466522, "cps": 58.625, "mean_p": 0.171466, "boundary_entropy": 0.212944, "bpb": 0.799677}, "zh": {"chunks": 2872.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.486285, "cps": 22.4375, "mean_p": 0.18367, "boundary_entropy": 0.052966, "bpb": 2.345265}, "my": {"chunks": 8355.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.440944, "cps": 65.273438, "mean_p": 0.384071, "boundary_entropy": 0.583161, "bpb": 0.77377}}, "bpb": {"en": 1.57287, "de": 1.68386, "vi": 1.32147, "tr": 1.57632, "fi": 1.73215, "ru": 0.99414, "el": 0.998, "hi": 0.81307, "ta": 0.69438, "th": 0.79968, "zh": 2.34527, "my": 0.77377}, "gini_cpc": 0.215106, "gini_cps": 0.221906, "bpic": 5.905322, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.043765086251589, "vi": 1.2321392935095574, "tr": 1.039967064206848, "fi": 1.016079907844333, "ru": 1.6962865164082312, "el": 1.7529760676826684, "hi": 2.5425770216887194, "ta": 2.668372508923038, "th": 2.6249193255881473, "zh": 2.736120330703163, "my": 2.4810029368241877}, "t": 4800.5} |
| {"event": "eval", "step": 28000, "phase": 1, "bpb": {"en": 1.57018, "de": 1.69145, "vi": 1.33964, "tr": 1.5698, "fi": 1.74773, "ru": 0.992, "el": 0.99615, "hi": 0.81905, "ta": 0.69969, "th": 0.80022, "zh": 2.34668, "my": 0.77973}, "bytes_per_chunk": {"en": 5.457, "de": 5.516, "vi": 6.017, "tr": 5.745, "fi": 5.608, "ru": 5.832, "el": 5.942, "hi": 5.616, "ta": 5.671, "th": 6.172, "zh": 5.693, "my": 6.499}, "bpb_mean": 1.27936, "t": 4805.7} |
| {"event": "train", "step": 28000, "phase": 1, "objective": "global", "lm_loss": 0.94354, "bpb": 1.36124, "chunk_loss": 0.030091, "F": 0.17828, "G": 0.18341, "chunks": 182.56, "bytes_per_chunk": 5.609, "boundary_entropy": 0.08206, "router_noise": 0.0, "grad_norm": 0.267, "lr": 0.0002263, "router_trainable": true, "sec_per_step": 0.18, "alpha": 0.03, "t": 4805.9} |
| {"event": "train", "step": 28500, "phase": 1, "objective": "global", "lm_loss": 0.9784, "bpb": 1.41153, "chunk_loss": 0.030062, "F": 0.17563, "G": 0.17723, "chunks": 179.84, "bytes_per_chunk": 5.694, "boundary_entropy": 0.0693, "router_noise": 0.0, "grad_norm": 0.2631, "lr": 0.0002238, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 4887.0} |
| {"event": "train", "step": 29000, "phase": 1, "objective": "global", "lm_loss": 0.91949, "bpb": 1.32655, "chunk_loss": 0.030138, "F": 0.17824, "G": 0.18679, "chunks": 182.52, "bytes_per_chunk": 5.61, "boundary_entropy": 0.08269, "router_noise": 0.0, "grad_norm": 0.2496, "lr": 0.0002212, "router_trainable": true, "sec_per_step": 0.162, "alpha": 0.03, "t": 4968.1} |
| {"event": "train", "step": 29500, "phase": 1, "objective": "global", "lm_loss": 0.97563, "bpb": 1.40754, "chunk_loss": 0.030087, "F": 0.17551, "G": 0.18456, "chunks": 179.72, "bytes_per_chunk": 5.698, "boundary_entropy": 0.08056, "router_noise": 0.0, "grad_norm": 0.265, "lr": 0.0002186, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 5049.8} |
| {"event": "train", "step": 30000, "phase": 1, "objective": "global", "lm_loss": 0.94513, "bpb": 1.36353, "chunk_loss": 0.030132, "F": 0.17915, "G": 0.19353, "chunks": 183.45, "bytes_per_chunk": 5.582, "boundary_entropy": 0.10045, "router_noise": 0.0, "grad_norm": 0.2473, "lr": 0.0002159, "router_trainable": true, "sec_per_step": 0.166, "alpha": 0.03, "t": 5132.7} |
| {"event": "train", "step": 30500, "phase": 1, "objective": "global", "lm_loss": 1.00958, "bpb": 1.45652, "chunk_loss": 0.030051, "F": 0.1747, "G": 0.17912, "chunks": 178.89, "bytes_per_chunk": 5.724, "boundary_entropy": 0.06399, "router_noise": 0.0, "grad_norm": 0.2739, "lr": 0.0002133, "router_trainable": true, "sec_per_step": 0.168, "alpha": 0.03, "t": 5216.6} |
| {"event": "phase_transition", "step": 30520, "phase": 2, "objective": "parity_a", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 5219.7} |
| {"event": "train", "step": 31000, "phase": 2, "objective": "parity_a", "lm_loss": 0.98256, "bpb": 1.41754, "chunk_loss": 0.030402, "F": 0.15353, "G": 0.15538, "chunks": 157.22, "bytes_per_chunk": 6.513, "boundary_entropy": 0.06032, "router_noise": 0.0, "grad_norm": 0.29, "lr": 0.0002106, "router_trainable": true, "sec_per_step": 0.163, "n_target_mean": 9.1793, "alpha": 0.03, "t": 5298.2} |
| {"event": "train", "step": 31500, "phase": 2, "objective": "parity_a", "lm_loss": 0.9597, "bpb": 1.38455, "chunk_loss": 0.030583, "F": 0.14627, "G": 0.14561, "chunks": 149.78, "bytes_per_chunk": 6.837, "boundary_entropy": 0.06162, "router_noise": 0.0, "grad_norm": 0.2654, "lr": 0.0002079, "router_trainable": true, "sec_per_step": 0.163, "n_target_mean": 10.48, "alpha": 0.03, "t": 5379.5} |
| {"event": "probe", "step": 32000, "phase": 2, "objective": "parity_a", "router_trainable": true, "per_lang": {"en": {"chunks": 3085.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.177268, "cps": 24.101562, "mean_p": 0.173703, "boundary_entropy": 0.033435, "bpb": 1.557787}, "de": {"chunks": 3776.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.183899, "cps": 29.5, "mean_p": 0.179605, "boundary_entropy": 0.029791, "bpb": 1.660159}, "vi": {"chunks": 3860.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.209384, "cps": 30.15625, "mean_p": 0.157429, "boundary_entropy": 0.039078, "bpb": 1.315775}, "tr": {"chunks": 3296.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.182069, "cps": 25.75, "mean_p": 0.167291, "boundary_entropy": 0.030805, "bpb": 1.563589}, "fi": {"chunks": 3383.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179842, "cps": 26.429688, "mean_p": 0.172904, "boundary_entropy": 0.02542, "bpb": 1.710966}, "ru": {"chunks": 4967.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.262111, "cps": 38.804688, "mean_p": 0.125124, "boundary_entropy": 0.070857, "bpb": 0.999103}, "el": {"chunks": 4048.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.197242, "cps": 31.625, "mean_p": 0.10735, "boundary_entropy": 0.048246, "bpb": 1.028119}, "hi": {"chunks": 3741.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.222348, "cps": 29.226562, "mean_p": 0.097219, "boundary_entropy": 0.118353, "bpb": 0.868102}, "ta": {"chunks": 2946.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.157582, "cps": 23.015625, "mean_p": 0.072307, "boundary_entropy": 0.105962, "bpb": 0.777663}, "th": {"chunks": 3254.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.2023, "cps": 25.421875, "mean_p": 0.084774, "boundary_entropy": 0.117091, "bpb": 0.850251}, "zh": {"chunks": 2881.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.487809, "cps": 22.507812, "mean_p": 0.182564, "boundary_entropy": 0.044453, "bpb": 2.323793}, "my": {"chunks": 2981.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.157325, "cps": 23.289062, "mean_p": 0.085472, "boundary_entropy": 0.168443, "bpb": 0.822123}}, "bpb": {"en": 1.55779, "de": 1.66016, "vi": 1.31578, "tr": 1.56359, "fi": 1.71097, "ru": 0.9991, "el": 1.02812, "hi": 0.8681, "ta": 0.77766, "th": 0.85025, "zh": 2.32379, "my": 0.82212}, "gini_cpc": 0.16269, "gini_cps": 0.08689, "bpic": 9.090459, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0374054621011313, "vi": 1.1811719216714385, "tr": 1.0270831477469418, "fi": 1.0145163865378606, "ru": 1.4786108799493678, "el": 1.1126757179812727, "hi": 1.2543004761090548, "ta": 0.888947729113991, "th": 1.141209649506849, "zh": 2.751812046206341, "my": 0.8874983448687853}, "t": 5463.5} |
| {"event": "eval", "step": 32000, "phase": 2, "bpb": {"en": 1.55531, "de": 1.67019, "vi": 1.3339, "tr": 1.55798, "fi": 1.72669, "ru": 0.99714, "el": 1.02594, "hi": 0.87223, "ta": 0.78366, "th": 0.84995, "zh": 2.32794, "my": 0.8281}, "bytes_per_chunk": {"en": 5.494, "de": 5.588, "vi": 6.304, "tr": 5.839, "fi": 5.638, "ru": 6.718, "el": 9.336, "hi": 11.8, "ta": 16.906, "th": 13.994, "zh": 5.687, "my": 19.723}, "bpb_mean": 1.29409, "t": 5468.5} |
| {"event": "train", "step": 32000, "phase": 2, "objective": "parity_a", "lm_loss": 0.97272, "bpb": 1.40334, "chunk_loss": 0.030342, "F": 0.15379, "G": 0.15331, "chunks": 157.48, "bytes_per_chunk": 6.502, "boundary_entropy": 0.05487, "router_noise": 0.0, "grad_norm": 0.2765, "lr": 0.0002052, "router_trainable": true, "sec_per_step": 0.178, "n_target_mean": 9.1998, "alpha": 0.03, "t": 5468.6} |
| {"event": "train", "step": 32500, "phase": 2, "objective": "parity_a", "lm_loss": 0.97092, "bpb": 1.40074, "chunk_loss": 0.03033, "F": 0.15399, "G": 0.15173, "chunks": 157.69, "bytes_per_chunk": 6.494, "boundary_entropy": 0.05001, "router_noise": 0.0, "grad_norm": 0.2796, "lr": 0.0002025, "router_trainable": true, "sec_per_step": 0.162, "n_target_mean": 8.8252, "alpha": 0.03, "t": 5549.6} |
| {"event": "train", "step": 33000, "phase": 2, "objective": "parity_a", "lm_loss": 0.91066, "bpb": 1.31381, "chunk_loss": 0.030318, "F": 0.14676, "G": 0.14402, "chunks": 150.28, "bytes_per_chunk": 6.814, "boundary_entropy": 0.05149, "router_noise": 0.0, "grad_norm": 0.2556, "lr": 0.0001997, "router_trainable": true, "sec_per_step": 0.162, "n_target_mean": 9.6296, "alpha": 0.03, "t": 5630.4} |
| {"event": "train", "step": 33500, "phase": 2, "objective": "parity_a", "lm_loss": 0.93628, "bpb": 1.35077, "chunk_loss": 0.030208, "F": 0.14687, "G": 0.1448, "chunks": 150.39, "bytes_per_chunk": 6.809, "boundary_entropy": 0.04944, "router_noise": 0.0, "grad_norm": 0.2642, "lr": 0.0001969, "router_trainable": true, "sec_per_step": 0.161, "n_target_mean": 9.4667, "alpha": 0.03, "t": 5710.9} |
| {"event": "train", "step": 34000, "phase": 2, "objective": "parity_a", "lm_loss": 1.05329, "bpb": 1.51958, "chunk_loss": 0.030414, "F": 0.15891, "G": 0.15724, "chunks": 162.72, "bytes_per_chunk": 6.293, "boundary_entropy": 0.04291, "router_noise": 0.0, "grad_norm": 0.2839, "lr": 0.0001942, "router_trainable": true, "sec_per_step": 0.161, "n_target_mean": 8.7212, "alpha": 0.03, "t": 5791.6} |
| {"event": "train", "step": 34500, "phase": 2, "objective": "parity_a", "lm_loss": 0.94913, "bpb": 1.36931, "chunk_loss": 0.030216, "F": 0.14357, "G": 0.14195, "chunks": 147.02, "bytes_per_chunk": 6.965, "boundary_entropy": 0.04726, "router_noise": 0.0, "grad_norm": 0.2632, "lr": 0.0001914, "router_trainable": true, "sec_per_step": 0.162, "n_target_mean": 9.3194, "alpha": 0.03, "t": 5872.5} |
| {"event": "train", "step": 35000, "phase": 2, "objective": "parity_a", "lm_loss": 0.92263, "bpb": 1.33108, "chunk_loss": 0.030245, "F": 0.14507, "G": 0.14381, "chunks": 148.55, "bytes_per_chunk": 6.893, "boundary_entropy": 0.04953, "router_noise": 0.0, "grad_norm": 0.2766, "lr": 0.0001886, "router_trainable": true, "sec_per_step": 0.161, "n_target_mean": 9.7215, "alpha": 0.03, "t": 5953.1} |
| {"event": "train", "step": 35500, "phase": 2, "objective": "parity_a", "lm_loss": 0.93034, "bpb": 1.34219, "chunk_loss": 0.030203, "F": 0.15445, "G": 0.15288, "chunks": 158.16, "bytes_per_chunk": 6.475, "boundary_entropy": 0.04475, "router_noise": 0.0, "grad_norm": 0.2764, "lr": 0.0001858, "router_trainable": true, "sec_per_step": 0.161, "n_target_mean": 8.4729, "alpha": 0.03, "t": 6033.7} |
| {"event": "probe", "step": 36000, "phase": 2, "objective": "parity_a", "router_trainable": true, "per_lang": {"en": {"chunks": 3070.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.176406, "cps": 23.984375, "mean_p": 0.173098, "boundary_entropy": 0.028265, "bpb": 1.54373}, "de": {"chunks": 3733.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.181805, "cps": 29.164062, "mean_p": 0.175875, "boundary_entropy": 0.030483, "bpb": 1.649558}, "vi": {"chunks": 3772.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.204611, "cps": 29.46875, "mean_p": 0.15156, "boundary_entropy": 0.037198, "bpb": 1.293485}, "tr": {"chunks": 3248.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.179418, "cps": 25.375, "mean_p": 0.163537, "boundary_entropy": 0.025957, "bpb": 1.545534}, "fi": {"chunks": 3386.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180001, "cps": 26.453125, "mean_p": 0.172343, "boundary_entropy": 0.022249, "bpb": 1.700747}, "ru": {"chunks": 4368.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.230501, "cps": 34.125, "mean_p": 0.116853, "boundary_entropy": 0.054415, "bpb": 0.995571}, "el": {"chunks": 3758.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.183112, "cps": 29.359375, "mean_p": 0.097462, "boundary_entropy": 0.04006, "bpb": 1.008822}, "hi": {"chunks": 3449.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.204993, "cps": 26.945312, "mean_p": 0.084562, "boundary_entropy": 0.089916, "bpb": 0.848549}, "ta": {"chunks": 3044.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.162824, "cps": 23.78125, "mean_p": 0.067245, "boundary_entropy": 0.088774, "bpb": 0.745141}, "th": {"chunks": 3308.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.205657, "cps": 25.84375, "mean_p": 0.079417, "boundary_entropy": 0.094724, "bpb": 0.829767}, "zh": {"chunks": 2885.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.488486, "cps": 22.539062, "mean_p": 0.188839, "boundary_entropy": 0.063451, "bpb": 2.300107}, "my": {"chunks": 2330.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.122968, "cps": 18.203125, "mean_p": 0.054927, "boundary_entropy": 0.090064, "bpb": 0.804982}}, "bpb": {"en": 1.54373, "de": 1.64956, "vi": 1.29348, "tr": 1.54553, "fi": 1.70075, "ru": 0.99557, "el": 1.00882, "hi": 0.84855, "ta": 0.74514, "th": 0.82977, "zh": 2.30011, "my": 0.80498}, "gini_cpc": 0.166534, "gini_cps": 0.081359, "bpic": 9.511065, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.030602822405055, "vi": 1.1598832768358587, "tr": 1.017070865393664, "fi": 1.0203773625418768, "ru": 1.306649660945571, "el": 1.038010329556368, "hi": 1.1620474611709928, "ta": 0.9230068483185858, "th": 1.1658164535109188, "zh": 2.769096684098653, "my": 0.6970730454478878}, "t": 6117.5} |
| {"event": "eval", "step": 36000, "phase": 2, "bpb": {"en": 1.53987, "de": 1.65819, "vi": 1.31108, "tr": 1.53916, "fi": 1.71691, "ru": 0.99227, "el": 1.00772, "hi": 0.85323, "ta": 0.75098, "th": 0.82971, "zh": 2.30124, "my": 0.81061}, "bytes_per_chunk": {"en": 5.514, "de": 5.654, "vi": 6.441, "tr": 5.943, "fi": 5.643, "ru": 7.667, "el": 10.098, "hi": 12.64, "ta": 16.591, "th": 14.382, "zh": 5.682, "my": 24.72}, "bpb_mean": 1.27591, "t": 6122.4} |
| {"event": "train", "step": 36000, "phase": 2, "objective": "parity_a", "lm_loss": 0.94934, "bpb": 1.36961, "chunk_loss": 0.030271, "F": 0.15562, "G": 0.15366, "chunks": 159.36, "bytes_per_chunk": 6.426, "boundary_entropy": 0.04249, "router_noise": 0.0, "grad_norm": 0.2661, "lr": 0.0001829, "router_trainable": true, "sec_per_step": 0.178, "n_target_mean": 8.5226, "alpha": 0.03, "t": 6122.6} |
| {"event": "train", "step": 36500, "phase": 2, "objective": "parity_a", "lm_loss": 0.97289, "bpb": 1.40358, "chunk_loss": 0.030205, "F": 0.14299, "G": 0.14163, "chunks": 146.42, "bytes_per_chunk": 6.993, "boundary_entropy": 0.04847, "router_noise": 0.0, "grad_norm": 0.2789, "lr": 0.0001801, "router_trainable": true, "sec_per_step": 0.163, "n_target_mean": 9.5, "alpha": 0.03, "t": 6204.2} |
| {"event": "train", "step": 37000, "phase": 2, "objective": "parity_a", "lm_loss": 1.01816, "bpb": 1.4689, "chunk_loss": 0.030178, "F": 0.14331, "G": 0.14443, "chunks": 146.75, "bytes_per_chunk": 6.978, "boundary_entropy": 0.0519, "router_noise": 0.0, "grad_norm": 0.2889, "lr": 0.0001773, "router_trainable": true, "sec_per_step": 0.162, "n_target_mean": 9.1291, "alpha": 0.03, "t": 6285.3} |
| {"event": "train", "step": 37500, "phase": 2, "objective": "parity_a", "lm_loss": 0.91447, "bpb": 1.3193, "chunk_loss": 0.030291, "F": 0.13869, "G": 0.13774, "chunks": 142.02, "bytes_per_chunk": 7.21, "boundary_entropy": 0.05344, "router_noise": 0.0, "grad_norm": 0.2821, "lr": 0.0001744, "router_trainable": true, "sec_per_step": 0.162, "n_target_mean": 10.2608, "alpha": 0.03, "t": 6366.3} |
| {"event": "train", "step": 38000, "phase": 2, "objective": "parity_a", "lm_loss": 0.99649, "bpb": 1.43763, "chunk_loss": 0.03021, "F": 0.151, "G": 0.15077, "chunks": 154.62, "bytes_per_chunk": 6.622, "boundary_entropy": 0.04715, "router_noise": 0.0, "grad_norm": 0.2757, "lr": 0.0001716, "router_trainable": true, "sec_per_step": 0.162, "n_target_mean": 8.8116, "alpha": 0.03, "t": 6447.3} |
| {"event": "train", "step": 38500, "phase": 2, "objective": "parity_a", "lm_loss": 0.88848, "bpb": 1.2818, "chunk_loss": 0.03023, "F": 0.13792, "G": 0.1367, "chunks": 141.23, "bytes_per_chunk": 7.25, "boundary_entropy": 0.04847, "router_noise": 0.0, "grad_norm": 0.2729, "lr": 0.0001687, "router_trainable": true, "sec_per_step": 0.161, "n_target_mean": 10.0461, "alpha": 0.03, "t": 6527.8} |
| {"event": "train", "step": 39000, "phase": 2, "objective": "parity_a", "lm_loss": 0.96324, "bpb": 1.38965, "chunk_loss": 0.030215, "F": 0.14935, "G": 0.14844, "chunks": 152.94, "bytes_per_chunk": 6.696, "boundary_entropy": 0.0463, "router_noise": 0.0, "grad_norm": 0.2783, "lr": 0.0001659, "router_trainable": true, "sec_per_step": 0.162, "n_target_mean": 8.8585, "alpha": 0.03, "t": 6608.6} |
| {"event": "train", "step": 39500, "phase": 2, "objective": "parity_a", "lm_loss": 0.94354, "bpb": 1.36124, "chunk_loss": 0.030313, "F": 0.14598, "G": 0.1446, "chunks": 149.48, "bytes_per_chunk": 6.85, "boundary_entropy": 0.04859, "router_noise": 0.0, "grad_norm": 0.2678, "lr": 0.000163, "router_trainable": true, "sec_per_step": 0.161, "n_target_mean": 9.6138, "alpha": 0.03, "t": 6689.1} |
| {"event": "probe", "step": 40000, "phase": 2, "objective": "parity_a", "router_trainable": true, "per_lang": {"en": {"chunks": 3069.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.176349, "cps": 23.976562, "mean_p": 0.172931, "boundary_entropy": 0.028901, "bpb": 1.531747}, "de": {"chunks": 3723.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.181318, "cps": 29.085938, "mean_p": 0.175326, "boundary_entropy": 0.031493, "bpb": 1.630868}, "vi": {"chunks": 3728.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.202224, "cps": 29.125, "mean_p": 0.149396, "boundary_entropy": 0.037284, "bpb": 1.281997}, "tr": {"chunks": 3224.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.178092, "cps": 25.1875, "mean_p": 0.163031, "boundary_entropy": 0.027915, "bpb": 1.530191}, "fi": {"chunks": 3385.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179948, "cps": 26.445312, "mean_p": 0.172436, "boundary_entropy": 0.022489, "bpb": 1.674807}, "ru": {"chunks": 4157.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.219367, "cps": 32.476562, "mean_p": 0.112587, "boundary_entropy": 0.047896, "bpb": 0.98224}, "el": {"chunks": 3799.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.185109, "cps": 29.679688, "mean_p": 0.0988, "boundary_entropy": 0.039292, "bpb": 0.990789}, "hi": {"chunks": 3239.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.192511, "cps": 25.304688, "mean_p": 0.080423, "boundary_entropy": 0.087999, "bpb": 0.836188}, "ta": {"chunks": 3005.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.160738, "cps": 23.476562, "mean_p": 0.06563, "boundary_entropy": 0.081257, "bpb": 0.728107}, "th": {"chunks": 2683.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.166801, "cps": 20.960938, "mean_p": 0.0682, "boundary_entropy": 0.088387, "bpb": 0.817289}, "zh": {"chunks": 2882.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.487978, "cps": 22.515625, "mean_p": 0.205687, "boundary_entropy": 0.114766, "bpb": 2.278094}, "my": {"chunks": 3434.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.181233, "cps": 26.828125, "mean_p": 0.068771, "boundary_entropy": 0.077799, "bpb": 0.777044}}, "bpb": {"en": 1.53175, "de": 1.63087, "vi": 1.282, "tr": 1.53019, "fi": 1.67481, "ru": 0.98224, "el": 0.99079, "hi": 0.83619, "ta": 0.72811, "th": 0.81729, "zh": 2.27809, "my": 0.77704}, "gini_cpc": 0.144045, "gini_cps": 0.069377, "bpic": 9.51649, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0281769436882449, "vi": 1.1467268819325303, "tr": 1.0098845160437162, "fi": 1.0204083912758144, "ru": 1.2439360151863343, "el": 1.0496769973501823, "hi": 1.0916492151539843, "ta": 0.9114781005056997, "th": 0.9458599552535998, "zh": 2.7671185490243198, "my": 1.0276947703350694}, "t": 6772.2} |
| {"event": "eval", "step": 40000, "phase": 2, "bpb": {"en": 1.52867, "de": 1.64022, "vi": 1.29963, "tr": 1.52474, "fi": 1.69048, "ru": 0.98007, "el": 0.98802, "hi": 0.84122, "ta": 0.73393, "th": 0.81872, "zh": 2.2816, "my": 0.78327}, "bytes_per_chunk": {"en": 5.496, "de": 5.655, "vi": 6.493, "tr": 5.957, "fi": 5.635, "ru": 8.013, "el": 9.994, "hi": 13.387, "ta": 17.089, "th": 17.136, "zh": 5.678, "my": 17.089}, "bpb_mean": 1.25921, "t": 6777.0} |
| {"event": "train", "step": 40000, "phase": 2, "objective": "parity_a", "lm_loss": 0.94106, "bpb": 1.35766, "chunk_loss": 0.030195, "F": 0.15433, "G": 0.15425, "chunks": 158.03, "bytes_per_chunk": 6.48, "boundary_entropy": 0.04667, "router_noise": 0.0, "grad_norm": 0.2876, "lr": 0.0001601, "router_trainable": true, "sec_per_step": 0.176, "n_target_mean": 8.5405, "alpha": 0.03, "t": 6777.2} |
| {"event": "train", "step": 40500, "phase": 2, "objective": "parity_a", "lm_loss": 0.99075, "bpb": 1.42935, "chunk_loss": 0.030157, "F": 0.14948, "G": 0.15331, "chunks": 153.06, "bytes_per_chunk": 6.69, "boundary_entropy": 0.05704, "router_noise": 0.0, "grad_norm": 0.2966, "lr": 0.0001573, "router_trainable": true, "sec_per_step": 0.161, "n_target_mean": 8.9237, "alpha": 0.03, "t": 6857.7} |
| {"event": "train", "step": 41000, "phase": 2, "objective": "parity_a", "lm_loss": 0.91106, "bpb": 1.31438, "chunk_loss": 0.030257, "F": 0.13699, "G": 0.14032, "chunks": 140.28, "bytes_per_chunk": 7.3, "boundary_entropy": 0.06199, "router_noise": 0.0, "grad_norm": 0.2904, "lr": 0.0001545, "router_trainable": true, "sec_per_step": 0.161, "n_target_mean": 10.1822, "alpha": 0.03, "t": 6938.4} |
| {"event": "train", "step": 41500, "phase": 2, "objective": "parity_a", "lm_loss": 0.96039, "bpb": 1.38554, "chunk_loss": 0.030256, "F": 0.15065, "G": 0.15121, "chunks": 154.27, "bytes_per_chunk": 6.638, "boundary_entropy": 0.05216, "router_noise": 0.0, "grad_norm": 0.2776, "lr": 0.0001516, "router_trainable": true, "sec_per_step": 0.161, "n_target_mean": 9.0581, "alpha": 0.03, "t": 7018.8} |
| {"event": "train", "step": 42000, "phase": 2, "objective": "parity_a", "lm_loss": 0.98152, "bpb": 1.41603, "chunk_loss": 0.030172, "F": 0.1418, "G": 0.14761, "chunks": 145.2, "bytes_per_chunk": 7.052, "boundary_entropy": 0.06389, "router_noise": 0.0, "grad_norm": 0.2921, "lr": 0.0001488, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 9.9519, "alpha": 0.03, "t": 7099.0} |
| {"event": "train", "step": 42500, "phase": 2, "objective": "parity_a", "lm_loss": 0.96138, "bpb": 1.38699, "chunk_loss": 0.030154, "F": 0.15401, "G": 0.15714, "chunks": 157.7, "bytes_per_chunk": 6.493, "boundary_entropy": 0.05594, "router_noise": 0.0, "grad_norm": 0.2916, "lr": 0.0001459, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 8.4679, "alpha": 0.03, "t": 7179.2} |
| {"event": "train", "step": 43000, "phase": 2, "objective": "parity_a", "lm_loss": 0.98455, "bpb": 1.42041, "chunk_loss": 0.030195, "F": 0.15501, "G": 0.15788, "chunks": 158.73, "bytes_per_chunk": 6.451, "boundary_entropy": 0.05458, "router_noise": 0.0, "grad_norm": 0.29, "lr": 0.0001431, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 8.4587, "alpha": 0.03, "t": 7259.3} |
| {"event": "train", "step": 43500, "phase": 2, "objective": "parity_a", "lm_loss": 0.96815, "bpb": 1.39675, "chunk_loss": 0.030257, "F": 0.15031, "G": 0.15404, "chunks": 153.92, "bytes_per_chunk": 6.653, "boundary_entropy": 0.0617, "router_noise": 0.0, "grad_norm": 0.2899, "lr": 0.0001403, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 9.0659, "alpha": 0.03, "t": 7339.3} |
| {"event": "probe", "step": 44000, "phase": 2, "objective": "parity_a", "router_trainable": true, "per_lang": {"en": {"chunks": 3061.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.175889, "cps": 23.914062, "mean_p": 0.172458, "boundary_entropy": 0.030587, "bpb": 1.518229}, "de": {"chunks": 3709.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.180636, "cps": 28.976562, "mean_p": 0.17372, "boundary_entropy": 0.034673, "bpb": 1.620837}, "vi": {"chunks": 3731.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.202387, "cps": 29.148438, "mean_p": 0.148998, "boundary_entropy": 0.038317, "bpb": 1.265749}, "tr": {"chunks": 3215.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.177595, "cps": 25.117188, "mean_p": 0.161846, "boundary_entropy": 0.028682, "bpb": 1.519911}, "fi": {"chunks": 3392.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.18032, "cps": 26.5, "mean_p": 0.172798, "boundary_entropy": 0.023027, "bpb": 1.655196}, "ru": {"chunks": 3871.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.204274, "cps": 30.242188, "mean_p": 0.104402, "boundary_entropy": 0.04902, "bpb": 0.979653}, "el": {"chunks": 3648.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.177752, "cps": 28.5, "mean_p": 0.094228, "boundary_entropy": 0.04113, "bpb": 0.981829}, "hi": {"chunks": 3307.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.196553, "cps": 25.835938, "mean_p": 0.077458, "boundary_entropy": 0.076735, "bpb": 0.821692}, "ta": {"chunks": 2912.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.155764, "cps": 22.75, "mean_p": 0.061713, "boundary_entropy": 0.075432, "bpb": 0.716926}, "th": {"chunks": 2828.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.175816, "cps": 22.09375, "mean_p": 0.068934, "boundary_entropy": 0.085384, "bpb": 0.805286}, "zh": {"chunks": 2889.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.489164, "cps": 22.570312, "mean_p": 0.22423, "boundary_entropy": 0.166443, "bpb": 2.257441}, "my": {"chunks": 2712.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.143129, "cps": 21.1875, "mean_p": 0.055359, "boundary_entropy": 0.06459, "bpb": 0.775685}}, "bpb": {"en": 1.51823, "de": 1.62084, "vi": 1.26575, "tr": 1.51991, "fi": 1.6552, "ru": 0.97965, "el": 0.98183, "hi": 0.82169, "ta": 0.71693, "th": 0.80529, "zh": 2.25744, "my": 0.77568}, "gini_cpc": 0.153974, "gini_cps": 0.066584, "bpic": 9.771636, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0269876398997746, "vi": 1.1506490882832898, "tr": 1.0096973500018036, "fi": 1.0251909193309656, "ru": 1.1613810824579203, "el": 1.0105894848407384, "hi": 1.1174803949218783, "ta": 0.8855777404419839, "th": 0.9995836192426362, "zh": 2.7810890159487642, "my": 0.8137426890300433}, "t": 7422.3} |
| {"event": "eval", "step": 44000, "phase": 2, "bpb": {"en": 1.51514, "de": 1.62932, "vi": 1.28375, "tr": 1.51386, "fi": 1.67211, "ru": 0.97692, "el": 0.98064, "hi": 0.82718, "ta": 0.72305, "th": 0.80494, "zh": 2.26059, "my": 0.7819}, "bytes_per_chunk": {"en": 5.496, "de": 5.687, "vi": 6.489, "tr": 5.985, "fi": 5.628, "ru": 8.489, "el": 10.352, "hi": 13.099, "ta": 17.16, "th": 16.093, "zh": 5.649, "my": 21.018}, "bpb_mean": 1.24745, "t": 7427.1} |
| {"event": "train", "step": 44000, "phase": 2, "objective": "parity_a", "lm_loss": 0.88186, "bpb": 1.27226, "chunk_loss": 0.030215, "F": 0.1389, "G": 0.13665, "chunks": 142.23, "bytes_per_chunk": 7.199, "boundary_entropy": 0.04744, "router_noise": 0.0, "grad_norm": 0.2853, "lr": 0.0001375, "router_trainable": true, "sec_per_step": 0.176, "n_target_mean": 9.8552, "alpha": 0.03, "t": 7427.3} |
| {"event": "train", "step": 44500, "phase": 2, "objective": "parity_a", "lm_loss": 0.95168, "bpb": 1.37299, "chunk_loss": 0.030116, "F": 0.15262, "G": 0.15757, "chunks": 156.28, "bytes_per_chunk": 6.552, "boundary_entropy": 0.05912, "router_noise": 0.0, "grad_norm": 0.2883, "lr": 0.0001347, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 8.4997, "alpha": 0.03, "t": 7507.3} |
| {"event": "train", "step": 45000, "phase": 2, "objective": "parity_a", "lm_loss": 0.96399, "bpb": 1.39075, "chunk_loss": 0.030165, "F": 0.14508, "G": 0.14883, "chunks": 148.56, "bytes_per_chunk": 6.893, "boundary_entropy": 0.05886, "router_noise": 0.0, "grad_norm": 0.2917, "lr": 0.0001319, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 9.0154, "alpha": 0.03, "t": 7587.4} |
| {"event": "train", "step": 45500, "phase": 2, "objective": "parity_a", "lm_loss": 0.92125, "bpb": 1.32909, "chunk_loss": 0.030174, "F": 0.14229, "G": 0.14635, "chunks": 145.7, "bytes_per_chunk": 7.028, "boundary_entropy": 0.06184, "router_noise": 0.0, "grad_norm": 0.303, "lr": 0.0001292, "router_trainable": true, "sec_per_step": 0.161, "n_target_mean": 9.6525, "alpha": 0.03, "t": 7667.7} |
| {"event": "train", "step": 46000, "phase": 2, "objective": "parity_a", "lm_loss": 0.91843, "bpb": 1.32501, "chunk_loss": 0.030231, "F": 0.14877, "G": 0.15725, "chunks": 152.34, "bytes_per_chunk": 6.722, "boundary_entropy": 0.07343, "router_noise": 0.0, "grad_norm": 0.3071, "lr": 0.0001264, "router_trainable": true, "sec_per_step": 0.161, "n_target_mean": 9.1585, "alpha": 0.03, "t": 7748.1} |
| {"event": "train", "step": 46500, "phase": 2, "objective": "parity_a", "lm_loss": 0.91745, "bpb": 1.3236, "chunk_loss": 0.030208, "F": 0.14803, "G": 0.15825, "chunks": 151.58, "bytes_per_chunk": 6.756, "boundary_entropy": 0.07968, "router_noise": 0.0, "grad_norm": 0.3105, "lr": 0.0001237, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 9.2902, "alpha": 0.03, "t": 7828.1} |
| {"event": "train", "step": 47000, "phase": 2, "objective": "parity_a", "lm_loss": 0.90118, "bpb": 1.30013, "chunk_loss": 0.030137, "F": 0.13837, "G": 0.14146, "chunks": 141.69, "bytes_per_chunk": 7.227, "boundary_entropy": 0.05624, "router_noise": 0.0, "grad_norm": 0.2931, "lr": 0.000121, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 9.683, "alpha": 0.03, "t": 7908.1} |
| {"event": "train", "step": 47500, "phase": 2, "objective": "parity_a", "lm_loss": 0.89587, "bpb": 1.29246, "chunk_loss": 0.030163, "F": 0.14009, "G": 0.14401, "chunks": 143.45, "bytes_per_chunk": 7.138, "boundary_entropy": 0.06107, "router_noise": 0.0, "grad_norm": 0.2953, "lr": 0.0001183, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 9.8869, "alpha": 0.03, "t": 7988.0} |
| {"event": "probe", "step": 48000, "phase": 2, "objective": "parity_a", "router_trainable": true, "per_lang": {"en": {"chunks": 3064.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.176062, "cps": 23.9375, "mean_p": 0.173175, "boundary_entropy": 0.030868, "bpb": 1.508066}, "de": {"chunks": 3721.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.18122, "cps": 29.070312, "mean_p": 0.174453, "boundary_entropy": 0.038078, "bpb": 1.606927}, "vi": {"chunks": 3721.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.201844, "cps": 29.070312, "mean_p": 0.148805, "boundary_entropy": 0.038296, "bpb": 1.252651}, "tr": {"chunks": 3200.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.176766, "cps": 25.0, "mean_p": 0.162449, "boundary_entropy": 0.028881, "bpb": 1.50365}, "fi": {"chunks": 3388.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180107, "cps": 26.46875, "mean_p": 0.172778, "boundary_entropy": 0.024198, "bpb": 1.639694}, "ru": {"chunks": 3831.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.202164, "cps": 29.929688, "mean_p": 0.103524, "boundary_entropy": 0.04652, "bpb": 0.968455}, "el": {"chunks": 3647.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.177703, "cps": 28.492188, "mean_p": 0.094964, "boundary_entropy": 0.039539, "bpb": 0.967195}, "hi": {"chunks": 3229.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.191917, "cps": 25.226562, "mean_p": 0.075231, "boundary_entropy": 0.074861, "bpb": 0.811459}, "ta": {"chunks": 3018.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.161434, "cps": 23.578125, "mean_p": 0.062792, "boundary_entropy": 0.071375, "bpb": 0.706759}, "th": {"chunks": 2973.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.184831, "cps": 23.226562, "mean_p": 0.071176, "boundary_entropy": 0.082132, "bpb": 0.792221}, "zh": {"chunks": 2899.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.490857, "cps": 22.648438, "mean_p": 0.253657, "boundary_entropy": 0.236369, "bpb": 2.244293}, "my": {"chunks": 3300.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.174161, "cps": 25.78125, "mean_p": 0.066339, "boundary_entropy": 0.069379, "bpb": 0.757967}}, "bpb": {"en": 1.50807, "de": 1.60693, "vi": 1.25265, "tr": 1.50365, "fi": 1.63969, "ru": 0.96845, "el": 0.96719, "hi": 0.81146, "ta": 0.70676, "th": 0.79222, "zh": 2.24429, "my": 0.75797}, "gini_cpc": 0.138532, "gini_cps": 0.053043, "bpic": 9.596684, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0293015389224427, "vi": 1.146441469959019, "tr": 1.004002481304759, "fi": 1.0229793744646658, "ru": 1.1482548723106705, "el": 1.0093232498875044, "hi": 1.0900547992504626, "ta": 0.9169151006573154, "th": 1.0498063689066977, "zh": 2.7879830685084603, "my": 0.9892041348839287}, "t": 8070.7} |
| {"event": "eval", "step": 48000, "phase": 2, "bpb": {"en": 1.50336, "de": 1.61553, "vi": 1.27126, "tr": 1.49923, "fi": 1.65751, "ru": 0.96662, "el": 0.96577, "hi": 0.81669, "ta": 0.71206, "th": 0.79215, "zh": 2.24742, "my": 0.76339}, "bytes_per_chunk": {"en": 5.497, "de": 5.668, "vi": 6.485, "tr": 6.006, "fi": 5.633, "ru": 8.628, "el": 10.338, "hi": 13.523, "ta": 16.432, "th": 15.133, "zh": 5.63, "my": 17.546}, "bpb_mean": 1.23425, "t": 8075.6} |
| {"event": "train", "step": 48000, "phase": 2, "objective": "parity_a", "lm_loss": 0.9568, "bpb": 1.38037, "chunk_loss": 0.03014, "F": 0.13995, "G": 0.14991, "chunks": 143.31, "bytes_per_chunk": 7.145, "boundary_entropy": 0.0753, "router_noise": 0.0, "grad_norm": 0.3015, "lr": 0.0001157, "router_trainable": true, "sec_per_step": 0.176, "n_target_mean": 9.8721, "alpha": 0.03, "t": 8075.8} |
| {"event": "train", "step": 48500, "phase": 2, "objective": "parity_a", "lm_loss": 0.93861, "bpb": 1.35412, "chunk_loss": 0.030191, "F": 0.15367, "G": 0.1638, "chunks": 157.36, "bytes_per_chunk": 6.507, "boundary_entropy": 0.07312, "router_noise": 0.0, "grad_norm": 0.3004, "lr": 0.000113, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 8.5351, "alpha": 0.03, "t": 8156.0} |
| {"event": "train", "step": 49000, "phase": 2, "objective": "parity_a", "lm_loss": 0.92477, "bpb": 1.33416, "chunk_loss": 0.030134, "F": 0.15027, "G": 0.15726, "chunks": 153.88, "bytes_per_chunk": 6.655, "boundary_entropy": 0.06435, "router_noise": 0.0, "grad_norm": 0.2989, "lr": 0.0001104, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 8.842, "alpha": 0.03, "t": 8236.1} |
| {"event": "train", "step": 49500, "phase": 2, "objective": "parity_a", "lm_loss": 0.89528, "bpb": 1.29162, "chunk_loss": 0.030142, "F": 0.14568, "G": 0.15199, "chunks": 149.17, "bytes_per_chunk": 6.865, "boundary_entropy": 0.06805, "router_noise": 0.0, "grad_norm": 0.2935, "lr": 0.0001078, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 9.0065, "alpha": 0.03, "t": 8316.1} |
| {"event": "train", "step": 50000, "phase": 2, "objective": "parity_a", "lm_loss": 0.89792, "bpb": 1.29542, "chunk_loss": 0.030216, "F": 0.14232, "G": 0.15073, "chunks": 145.73, "bytes_per_chunk": 7.026, "boundary_entropy": 0.07133, "router_noise": 0.0, "grad_norm": 0.3038, "lr": 0.0001052, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 9.7915, "alpha": 0.03, "t": 8396.2} |
| {"event": "train", "step": 50500, "phase": 2, "objective": "parity_a", "lm_loss": 0.90071, "bpb": 1.29945, "chunk_loss": 0.030232, "F": 0.15201, "G": 0.15252, "chunks": 155.66, "bytes_per_chunk": 6.579, "boundary_entropy": 0.04888, "router_noise": 0.0, "grad_norm": 0.2979, "lr": 0.0001027, "router_trainable": true, "sec_per_step": 0.161, "n_target_mean": 8.8817, "alpha": 0.03, "t": 8476.5} |
| {"event": "train", "step": 51000, "phase": 2, "objective": "parity_a", "lm_loss": 0.89114, "bpb": 1.28564, "chunk_loss": 0.030244, "F": 0.14413, "G": 0.14889, "chunks": 147.59, "bytes_per_chunk": 6.938, "boundary_entropy": 0.06218, "router_noise": 0.0, "grad_norm": 0.3017, "lr": 0.0001002, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 9.5685, "alpha": 0.03, "t": 8556.5} |
| {"event": "train", "step": 51500, "phase": 2, "objective": "parity_a", "lm_loss": 0.90744, "bpb": 1.30916, "chunk_loss": 0.030139, "F": 0.14383, "G": 0.14967, "chunks": 147.28, "bytes_per_chunk": 6.953, "boundary_entropy": 0.06324, "router_noise": 0.0, "grad_norm": 0.2872, "lr": 9.77e-05, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 9.2916, "alpha": 0.03, "t": 8636.8} |
| {"event": "probe", "step": 52000, "phase": 2, "objective": "parity_a", "router_trainable": true, "per_lang": {"en": {"chunks": 3081.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.177038, "cps": 24.070312, "mean_p": 0.173992, "boundary_entropy": 0.032012, "bpb": 1.494742}, "de": {"chunks": 3714.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.18088, "cps": 29.015625, "mean_p": 0.174245, "boundary_entropy": 0.037747, "bpb": 1.593736}, "vi": {"chunks": 3712.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.201356, "cps": 29.0, "mean_p": 0.147996, "boundary_entropy": 0.040406, "bpb": 1.237879}, "tr": {"chunks": 3334.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.184168, "cps": 26.046875, "mean_p": 0.168961, "boundary_entropy": 0.028198, "bpb": 1.49783}, "fi": {"chunks": 3393.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180373, "cps": 26.507812, "mean_p": 0.173158, "boundary_entropy": 0.02521, "bpb": 1.622503}, "ru": {"chunks": 3877.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.204591, "cps": 30.289062, "mean_p": 0.105481, "boundary_entropy": 0.045658, "bpb": 0.954034}, "el": {"chunks": 3770.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.183696, "cps": 29.453125, "mean_p": 0.099305, "boundary_entropy": 0.041326, "bpb": 0.955035}, "hi": {"chunks": 3230.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.191976, "cps": 25.234375, "mean_p": 0.075553, "boundary_entropy": 0.070056, "bpb": 0.798928}, "ta": {"chunks": 2920.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.156191, "cps": 22.8125, "mean_p": 0.060709, "boundary_entropy": 0.06942, "bpb": 0.69915}, "th": {"chunks": 2778.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.172707, "cps": 21.703125, "mean_p": 0.067216, "boundary_entropy": 0.077829, "bpb": 0.785821}, "zh": {"chunks": 2938.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.49746, "cps": 22.953125, "mean_p": 0.289927, "boundary_entropy": 0.305043, "bpb": 2.227852}, "my": {"chunks": 2809.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.148248, "cps": 21.945312, "mean_p": 0.05592, "boundary_entropy": 0.058203, "bpb": 0.751366}}, "bpb": {"en": 1.49474, "de": 1.59374, "vi": 1.23788, "tr": 1.49783, "fi": 1.6225, "ru": 0.95403, "el": 0.95503, "hi": 0.79893, "ta": 0.69915, "th": 0.78582, "zh": 2.22785, "my": 0.75137}, "gini_cpc": 0.153521, "gini_cps": 0.065742, "bpic": 9.70222, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0216965199725276, "vi": 1.137358159104768, "tr": 1.0402733336843843, "fi": 1.0188362720708124, "ru": 1.1556305349238116, "el": 1.037607081156862, "hi": 1.084375935911663, "ta": 0.8822462149927356, "th": 0.9755366678783809, "zh": 2.809899394308298, "my": 0.8373765313524494}, "t": 8719.9} |
| {"event": "eval", "step": 52000, "phase": 2, "bpb": {"en": 1.49095, "de": 1.60229, "vi": 1.25708, "tr": 1.49083, "fi": 1.64084, "ru": 0.95174, "el": 0.95316, "hi": 0.80479, "ta": 0.7048, "th": 0.7863, "zh": 2.23093, "my": 0.75699}, "bytes_per_chunk": {"en": 5.486, "de": 5.664, "vi": 6.497, "tr": 5.822, "fi": 5.635, "ru": 8.547, "el": 10.126, "hi": 13.214, "ta": 16.869, "th": 16.399, "zh": 5.587, "my": 20.637}, "bpb_mean": 1.22256, "t": 8724.8} |
| {"event": "train", "step": 52000, "phase": 2, "objective": "parity_a", "lm_loss": 0.96963, "bpb": 1.39888, "chunk_loss": 0.030135, "F": 0.15633, "G": 0.17016, "chunks": 160.08, "bytes_per_chunk": 6.397, "boundary_entropy": 0.07634, "router_noise": 0.0, "grad_norm": 0.3171, "lr": 9.52e-05, "router_trainable": true, "sec_per_step": 0.176, "n_target_mean": 8.2005, "alpha": 0.03, "t": 8725.0} |
| {"event": "train", "step": 52500, "phase": 2, "objective": "parity_a", "lm_loss": 0.9097, "bpb": 1.31241, "chunk_loss": 0.030213, "F": 0.15065, "G": 0.16016, "chunks": 154.27, "bytes_per_chunk": 6.638, "boundary_entropy": 0.07246, "router_noise": 0.0, "grad_norm": 0.3088, "lr": 9.28e-05, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 9.1438, "alpha": 0.03, "t": 8805.2} |
| {"event": "train", "step": 53000, "phase": 2, "objective": "parity_a", "lm_loss": 0.99487, "bpb": 1.4353, "chunk_loss": 0.030228, "F": 0.15509, "G": 0.16648, "chunks": 158.81, "bytes_per_chunk": 6.448, "boundary_entropy": 0.07771, "router_noise": 0.0, "grad_norm": 0.3186, "lr": 9.04e-05, "router_trainable": true, "sec_per_step": 0.16, "n_target_mean": 8.5535, "alpha": 0.03, "t": 8885.2} |
| {"event": "phase_transition", "step": 53410, "phase": 3, "objective": "none", "router_trainable": false, "encoder_trainable": true, "frozen_router_params": 295680, "t": 8950.6} |
| {"event": "train", "step": 53500, "phase": 3, "objective": "none", "lm_loss": 0.93977, "bpb": 1.3558, "chunk_loss": 0.0, "F": 0.14836, "G": 0.16033, "chunks": 151.92, "bytes_per_chunk": 6.74, "boundary_entropy": 0.07627, "router_noise": 0.0, "grad_norm": 0.315, "lr": 8.8e-05, "router_trainable": false, "sec_per_step": 0.16, "alpha": 0.0, "t": 8965.0} |
| {"event": "train", "step": 54000, "phase": 3, "objective": "none", "lm_loss": 0.96672, "bpb": 1.39469, "chunk_loss": 0.0, "F": 0.15897, "G": 0.17416, "chunks": 162.78, "bytes_per_chunk": 6.291, "boundary_entropy": 0.08718, "router_noise": 0.0, "grad_norm": 0.3311, "lr": 8.57e-05, "router_trainable": false, "sec_per_step": 0.16, "alpha": 0.0, "t": 9045.0} |
| {"event": "train", "step": 54500, "phase": 3, "objective": "none", "lm_loss": 0.91487, "bpb": 1.31987, "chunk_loss": 0.0, "F": 0.15584, "G": 0.16084, "chunks": 159.58, "bytes_per_chunk": 6.417, "boundary_entropy": 0.05757, "router_noise": 0.0, "grad_norm": 0.3072, "lr": 8.34e-05, "router_trainable": false, "sec_per_step": 0.16, "alpha": 0.0, "t": 9125.2} |
| {"event": "train", "step": 55000, "phase": 3, "objective": "none", "lm_loss": 0.84639, "bpb": 1.22108, "chunk_loss": 0.0, "F": 0.1445, "G": 0.15921, "chunks": 147.97, "bytes_per_chunk": 6.92, "boundary_entropy": 0.08826, "router_noise": 0.0, "grad_norm": 0.3037, "lr": 8.11e-05, "router_trainable": false, "sec_per_step": 0.16, "alpha": 0.0, "t": 9205.0} |
| {"event": "train", "step": 55500, "phase": 3, "objective": "none", "lm_loss": 0.93949, "bpb": 1.35539, "chunk_loss": 0.0, "F": 0.16319, "G": 0.17455, "chunks": 167.11, "bytes_per_chunk": 6.128, "boundary_entropy": 0.08201, "router_noise": 0.0, "grad_norm": 0.3277, "lr": 7.89e-05, "router_trainable": false, "sec_per_step": 0.16, "alpha": 0.0, "t": 9285.1} |
| {"event": "probe", "step": 56000, "phase": 3, "objective": "none", "router_trainable": false, "per_lang": {"en": {"chunks": 3089.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.177498, "cps": 24.132812, "mean_p": 0.175364, "boundary_entropy": 0.031338, "bpb": 1.485236}, "de": {"chunks": 3734.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.181854, "cps": 29.171875, "mean_p": 0.176448, "boundary_entropy": 0.035363, "bpb": 1.581975}, "vi": {"chunks": 3789.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.205533, "cps": 29.601562, "mean_p": 0.153563, "boundary_entropy": 0.040413, "bpb": 1.227055}, "tr": {"chunks": 3371.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.186212, "cps": 26.335938, "mean_p": 0.171268, "boundary_entropy": 0.028155, "bpb": 1.481208}, "fi": {"chunks": 3393.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180373, "cps": 26.507812, "mean_p": 0.173581, "boundary_entropy": 0.025749, "bpb": 1.604241}, "ru": {"chunks": 4067.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.214617, "cps": 31.773438, "mean_p": 0.113367, "boundary_entropy": 0.043061, "bpb": 0.9435}, "el": {"chunks": 3966.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.193247, "cps": 30.984375, "mean_p": 0.104771, "boundary_entropy": 0.046297, "bpb": 0.94249}, "hi": {"chunks": 6831.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.406003, "cps": 53.367188, "mean_p": 0.152134, "boundary_entropy": 0.114082, "bpb": 0.785112}, "ta": {"chunks": 6457.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.345386, "cps": 50.445312, "mean_p": 0.124746, "boundary_entropy": 0.098241, "bpb": 0.685816}, "th": {"chunks": 3727.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.231707, "cps": 29.117188, "mean_p": 0.086753, "boundary_entropy": 0.095986, "bpb": 0.771157}, "zh": {"chunks": 2902.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.491365, "cps": 22.671875, "mean_p": 0.285632, "boundary_entropy": 0.301775, "bpb": 2.210238}, "my": {"chunks": 11392.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.601224, "cps": 89.0, "mean_p": 0.20774, "boundary_entropy": 0.106081, "bpb": 0.730663}}, "bpb": {"en": 1.48524, "de": 1.58198, "vi": 1.22706, "tr": 1.48121, "fi": 1.60424, "ru": 0.9435, "el": 0.94249, "hi": 0.78511, "ta": 0.68582, "th": 0.77116, "zh": 2.21024, "my": 0.73066}, "gini_cpc": 0.242783, "gini_cps": 0.22452, "bpic": 6.766476, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0245381117656034, "vi": 1.1579443159156049, "tr": 1.049094023628363, "fi": 1.0161976543380296, "ru": 1.2091249142629692, "el": 1.0887247017523172, "hi": 2.28736475159664, "ta": 1.9458597231061987, "th": 1.3054027985231018, "zh": 2.7682810343597115, "my": 3.387215375976083}, "t": 9368.2} |
| {"event": "eval", "step": 56000, "phase": 3, "bpb": {"en": 1.48032, "de": 1.59019, "vi": 1.24447, "tr": 1.47527, "fi": 1.62377, "ru": 0.94042, "el": 0.9412, "hi": 0.79089, "ta": 0.69136, "th": 0.77151, "zh": 2.21329, "my": 0.73718}, "bytes_per_chunk": {"en": 5.456, "de": 5.639, "vi": 6.389, "tr": 5.734, "fi": 5.627, "ru": 8.138, "el": 9.584, "hi": 6.087, "ta": 7.589, "th": 11.635, "zh": 5.614, "my": 4.826}, "bpb_mean": 1.20832, "t": 9373.5} |
| {"event": "train", "step": 56000, "phase": 3, "objective": "none", "lm_loss": 0.91204, "bpb": 1.31579, "chunk_loss": 0.0, "F": 0.15485, "G": 0.1662, "chunks": 158.56, "bytes_per_chunk": 6.458, "boundary_entropy": 0.07851, "router_noise": 0.0, "grad_norm": 0.3163, "lr": 7.68e-05, "router_trainable": false, "sec_per_step": 0.177, "alpha": 0.0, "t": 9373.6} |
| {"event": "train", "step": 56500, "phase": 3, "objective": "none", "lm_loss": 0.92405, "bpb": 1.33312, "chunk_loss": 0.0, "F": 0.16278, "G": 0.17031, "chunks": 166.69, "bytes_per_chunk": 6.143, "boundary_entropy": 0.06938, "router_noise": 0.0, "grad_norm": 0.308, "lr": 7.46e-05, "router_trainable": false, "sec_per_step": 0.161, "alpha": 0.0, "t": 9453.9} |
| {"event": "train", "step": 57000, "phase": 3, "objective": "none", "lm_loss": 0.89641, "bpb": 1.29325, "chunk_loss": 0.0, "F": 0.15823, "G": 0.16651, "chunks": 162.03, "bytes_per_chunk": 6.32, "boundary_entropy": 0.07312, "router_noise": 0.0, "grad_norm": 0.3089, "lr": 7.25e-05, "router_trainable": false, "sec_per_step": 0.163, "alpha": 0.0, "t": 9535.4} |
| {"event": "train", "step": 57500, "phase": 3, "objective": "none", "lm_loss": 0.86615, "bpb": 1.2496, "chunk_loss": 0.0, "F": 0.15506, "G": 0.16561, "chunks": 158.78, "bytes_per_chunk": 6.449, "boundary_entropy": 0.07629, "router_noise": 0.0, "grad_norm": 0.3254, "lr": 7.05e-05, "router_trainable": false, "sec_per_step": 0.165, "alpha": 0.0, "t": 9617.8} |
| {"event": "train", "step": 58000, "phase": 3, "objective": "none", "lm_loss": 0.843, "bpb": 1.21619, "chunk_loss": 0.0, "F": 0.15886, "G": 0.16564, "chunks": 162.67, "bytes_per_chunk": 6.295, "boundary_entropy": 0.0704, "router_noise": 0.0, "grad_norm": 0.3055, "lr": 6.84e-05, "router_trainable": false, "sec_per_step": 0.166, "alpha": 0.0, "t": 9701.0} |
| {"event": "train", "step": 58500, "phase": 3, "objective": "none", "lm_loss": 0.82977, "bpb": 1.19711, "chunk_loss": 0.0, "F": 0.15599, "G": 0.16263, "chunks": 159.73, "bytes_per_chunk": 6.411, "boundary_entropy": 0.07738, "router_noise": 0.0, "grad_norm": 0.3132, "lr": 6.65e-05, "router_trainable": false, "sec_per_step": 0.164, "alpha": 0.0, "t": 9783.0} |
| {"event": "train", "step": 59000, "phase": 3, "objective": "none", "lm_loss": 0.8665, "bpb": 1.25009, "chunk_loss": 0.0, "F": 0.16125, "G": 0.16404, "chunks": 165.12, "bytes_per_chunk": 6.201, "boundary_entropy": 0.05398, "router_noise": 0.0, "grad_norm": 0.3144, "lr": 6.45e-05, "router_trainable": false, "sec_per_step": 0.167, "alpha": 0.0, "t": 9866.3} |
| {"event": "train", "step": 59500, "phase": 3, "objective": "none", "lm_loss": 0.87058, "bpb": 1.25598, "chunk_loss": 0.0, "F": 0.15552, "G": 0.16985, "chunks": 159.25, "bytes_per_chunk": 6.43, "boundary_entropy": 0.08659, "router_noise": 0.0, "grad_norm": 0.3228, "lr": 6.27e-05, "router_trainable": false, "sec_per_step": 0.168, "alpha": 0.0, "t": 9950.0} |
| {"event": "probe", "step": 60000, "phase": 3, "objective": "none", "router_trainable": false, "per_lang": {"en": {"chunks": 3090.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.177556, "cps": 24.140625, "mean_p": 0.175263, "boundary_entropy": 0.029884, "bpb": 1.475013}, "de": {"chunks": 3737.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.182, "cps": 29.195312, "mean_p": 0.176865, "boundary_entropy": 0.033285, "bpb": 1.569582}, "vi": {"chunks": 3773.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.204665, "cps": 29.476562, "mean_p": 0.152597, "boundary_entropy": 0.038985, "bpb": 1.213315}, "tr": {"chunks": 3385.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.186986, "cps": 26.445312, "mean_p": 0.171135, "boundary_entropy": 0.028011, "bpb": 1.467697}, "fi": {"chunks": 3385.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179948, "cps": 26.445312, "mean_p": 0.173091, "boundary_entropy": 0.024734, "bpb": 1.590294}, "ru": {"chunks": 4072.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.214881, "cps": 31.8125, "mean_p": 0.113796, "boundary_entropy": 0.042628, "bpb": 0.935091}, "el": {"chunks": 4448.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.216732, "cps": 34.75, "mean_p": 0.117541, "boundary_entropy": 0.052082, "bpb": 0.931861}, "hi": {"chunks": 6938.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.412363, "cps": 54.203125, "mean_p": 0.154329, "boundary_entropy": 0.119863, "bpb": 0.771366}, "ta": {"chunks": 6266.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.33517, "cps": 48.953125, "mean_p": 0.120511, "boundary_entropy": 0.104975, "bpb": 0.676028}, "th": {"chunks": 5008.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.311346, "cps": 39.125, "mean_p": 0.112336, "boundary_entropy": 0.113634, "bpb": 0.759217}, "zh": {"chunks": 2906.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.492042, "cps": 22.703125, "mean_p": 0.289569, "boundary_entropy": 0.306967, "bpb": 2.198063}, "my": {"chunks": 15001.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.791693, "cps": 117.195312, "mean_p": 0.266733, "boundary_entropy": 0.115773, "bpb": 0.721611}}, "bpb": {"en": 1.47501, "de": 1.56958, "vi": 1.21332, "tr": 1.4677, "fi": 1.59029, "ru": 0.93509, "el": 0.93186, "hi": 0.77137, "ta": 0.67603, "th": 0.75922, "zh": 2.19806, "my": 0.72161}, "gini_cpc": 0.275269, "gini_cps": 0.266616, "bpic": 6.189118, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0250294222088145, "vi": 1.1526814496485949, "tr": 1.0531100622167973, "fi": 1.013473576966173, "ru": 1.2102196377795424, "el": 1.2206455556139004, "hi": 2.3224419278985557, "ta": 1.8876895073445212, "th": 1.7535126195484545, "zh": 2.771199602839304, "my": 4.458846029296327}, "t": 10037.2} |
| {"event": "eval", "step": 60000, "phase": 3, "bpb": {"en": 1.47009, "de": 1.57893, "vi": 1.2316, "tr": 1.46235, "fi": 1.61035, "ru": 0.93217, "el": 0.93075, "hi": 0.77705, "ta": 0.68188, "th": 0.75919, "zh": 2.20132, "my": 0.72772}, "bytes_per_chunk": {"en": 5.457, "de": 5.635, "vi": 6.404, "tr": 5.716, "fi": 5.633, "ru": 8.116, "el": 8.703, "hi": 5.953, "ta": 7.759, "th": 8.916, "zh": 5.616, "my": 3.66}, "bpb_mean": 1.19695, "t": 10042.4} |
| {"event": "train", "step": 60000, "phase": 3, "objective": "none", "lm_loss": 0.89285, "bpb": 1.2881, "chunk_loss": 0.0, "F": 0.15704, "G": 0.17487, "chunks": 160.81, "bytes_per_chunk": 6.368, "boundary_entropy": 0.09054, "router_noise": 0.0, "grad_norm": 0.3258, "lr": 6.08e-05, "router_trainable": false, "sec_per_step": 0.185, "alpha": 0.0, "t": 10042.6} |
| {"event": "train", "step": 60500, "phase": 3, "objective": "none", "lm_loss": 0.85739, "bpb": 1.23696, "chunk_loss": 0.0, "F": 0.16144, "G": 0.16174, "chunks": 165.31, "bytes_per_chunk": 6.194, "boundary_entropy": 0.05061, "router_noise": 0.0, "grad_norm": 0.31, "lr": 5.9e-05, "router_trainable": false, "sec_per_step": 0.172, "alpha": 0.0, "t": 10128.4} |
| {"event": "train", "step": 61000, "phase": 3, "objective": "none", "lm_loss": 0.88958, "bpb": 1.28339, "chunk_loss": 0.0, "F": 0.15854, "G": 0.16731, "chunks": 162.34, "bytes_per_chunk": 6.308, "boundary_entropy": 0.07586, "router_noise": 0.0, "grad_norm": 0.315, "lr": 5.73e-05, "router_trainable": false, "sec_per_step": 0.178, "alpha": 0.0, "t": 10217.4} |
| {"event": "train", "step": 61500, "phase": 3, "objective": "none", "lm_loss": 0.93016, "bpb": 1.34194, "chunk_loss": 0.0, "F": 0.16818, "G": 0.17573, "chunks": 172.22, "bytes_per_chunk": 5.946, "boundary_entropy": 0.07609, "router_noise": 0.0, "grad_norm": 0.3289, "lr": 5.56e-05, "router_trainable": false, "sec_per_step": 0.177, "alpha": 0.0, "t": 10305.8} |
| {"event": "train", "step": 62000, "phase": 3, "objective": "none", "lm_loss": 0.95327, "bpb": 1.37528, "chunk_loss": 0.0, "F": 0.16441, "G": 0.18382, "chunks": 168.36, "bytes_per_chunk": 6.082, "boundary_entropy": 0.09652, "router_noise": 0.0, "grad_norm": 0.3383, "lr": 5.39e-05, "router_trainable": false, "sec_per_step": 0.178, "alpha": 0.0, "t": 10394.6} |
| {"event": "train", "step": 62500, "phase": 3, "objective": "none", "lm_loss": 0.91533, "bpb": 1.32054, "chunk_loss": 0.0, "F": 0.16592, "G": 0.18159, "chunks": 169.91, "bytes_per_chunk": 6.027, "boundary_entropy": 0.08976, "router_noise": 0.0, "grad_norm": 0.3339, "lr": 5.23e-05, "router_trainable": false, "sec_per_step": 0.175, "alpha": 0.0, "t": 10482.0} |
| {"event": "train", "step": 63000, "phase": 3, "objective": "none", "lm_loss": 0.91268, "bpb": 1.31671, "chunk_loss": 0.0, "F": 0.15973, "G": 0.16624, "chunks": 163.56, "bytes_per_chunk": 6.261, "boundary_entropy": 0.06254, "router_noise": 0.0, "grad_norm": 0.3282, "lr": 5.08e-05, "router_trainable": false, "sec_per_step": 0.178, "alpha": 0.0, "t": 10570.8} |
| {"event": "train", "step": 63500, "phase": 3, "objective": "none", "lm_loss": 0.90193, "bpb": 1.30121, "chunk_loss": 0.0, "F": 0.16183, "G": 0.1752, "chunks": 165.72, "bytes_per_chunk": 6.179, "boundary_entropy": 0.09026, "router_noise": 0.0, "grad_norm": 0.3503, "lr": 4.93e-05, "router_trainable": false, "sec_per_step": 0.179, "alpha": 0.0, "t": 10660.4} |
| {"event": "probe", "step": 64000, "phase": 3, "objective": "none", "router_trainable": false, "per_lang": {"en": {"chunks": 3111.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.178762, "cps": 24.304688, "mean_p": 0.177168, "boundary_entropy": 0.029982, "bpb": 1.465852}, "de": {"chunks": 3752.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.18273, "cps": 29.3125, "mean_p": 0.178543, "boundary_entropy": 0.033819, "bpb": 1.560043}, "vi": {"chunks": 3797.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.205967, "cps": 29.664062, "mean_p": 0.154316, "boundary_entropy": 0.038563, "bpb": 1.201621}, "tr": {"chunks": 3405.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.18809, "cps": 26.601562, "mean_p": 0.1728, "boundary_entropy": 0.027725, "bpb": 1.45829}, "fi": {"chunks": 3407.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.181117, "cps": 26.617188, "mean_p": 0.174534, "boundary_entropy": 0.025203, "bpb": 1.576737}, "ru": {"chunks": 4140.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.21847, "cps": 32.34375, "mean_p": 0.115947, "boundary_entropy": 0.042973, "bpb": 0.925866}, "el": {"chunks": 4475.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.218048, "cps": 34.960938, "mean_p": 0.119714, "boundary_entropy": 0.05183, "bpb": 0.92365}, "hi": {"chunks": 7082.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.420921, "cps": 55.328125, "mean_p": 0.157491, "boundary_entropy": 0.120361, "bpb": 0.761774}, "ta": {"chunks": 6334.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.338807, "cps": 49.484375, "mean_p": 0.121391, "boundary_entropy": 0.103291, "bpb": 0.665519}, "th": {"chunks": 5390.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.335095, "cps": 42.109375, "mean_p": 0.119514, "boundary_entropy": 0.115026, "bpb": 0.751659}, "zh": {"chunks": 2920.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.494412, "cps": 22.8125, "mean_p": 0.289584, "boundary_entropy": 0.306506, "bpb": 2.186707}, "my": {"chunks": 14495.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.764988, "cps": 113.242188, "mean_p": 0.258033, "boundary_entropy": 0.120086, "bpb": 0.703855}}, "bpb": {"en": 1.46585, "de": 1.56004, "vi": 1.20162, "tr": 1.45829, "fi": 1.57674, "ru": 0.92587, "el": 0.92365, "hi": 0.76177, "ta": 0.66552, "th": 0.75166, "zh": 2.18671, "my": 0.70386}, "gini_cpc": 0.270001, "gini_cps": 0.262131, "bpic": 6.159418, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.022196834307837, "vi": 1.1521832684306899, "tr": 1.0521815299797137, "fi": 1.0131747530515574, "ru": 1.2221238960569738, "el": 1.2197653821234318, "hi": 2.3546423919725004, "ta": 1.8952944838772323, "th": 1.874527469635306, "zh": 2.7657538008680516, "my": 4.279361278871332}, "t": 10753.7} |
| {"event": "eval", "step": 64000, "phase": 3, "bpb": {"en": 1.46151, "de": 1.56854, "vi": 1.22049, "tr": 1.45372, "fi": 1.59743, "ru": 0.92317, "el": 0.92192, "hi": 0.7671, "ta": 0.67145, "th": 0.75108, "zh": 2.19087, "my": 0.7094}, "bytes_per_chunk": {"en": 5.433, "de": 5.602, "vi": 6.379, "tr": 5.672, "fi": 5.613, "ru": 8.032, "el": 8.519, "hi": 5.878, "ta": 7.771, "th": 8.322, "zh": 5.6, "my": 3.797}, "bpb_mean": 1.18639, "t": 10758.9} |
| {"event": "train", "step": 64000, "phase": 3, "objective": "none", "lm_loss": 0.79825, "bpb": 1.15163, "chunk_loss": 0.0, "F": 0.15642, "G": 0.16097, "chunks": 160.17, "bytes_per_chunk": 6.393, "boundary_entropy": 0.0714, "router_noise": 0.0, "grad_norm": 0.3227, "lr": 4.78e-05, "router_trainable": false, "sec_per_step": 0.197, "alpha": 0.0, "t": 10759.1} |
| {"event": "train", "step": 64500, "phase": 3, "objective": "none", "lm_loss": 0.92755, "bpb": 1.33817, "chunk_loss": 0.0, "F": 0.17365, "G": 0.17859, "chunks": 177.81, "bytes_per_chunk": 5.759, "boundary_entropy": 0.06306, "router_noise": 0.0, "grad_norm": 0.3391, "lr": 4.65e-05, "router_trainable": false, "sec_per_step": 0.176, "alpha": 0.0, "t": 10847.2} |
| {"event": "train", "step": 65000, "phase": 3, "objective": "none", "lm_loss": 0.88243, "bpb": 1.27308, "chunk_loss": 0.0, "F": 0.16565, "G": 0.17562, "chunks": 169.62, "bytes_per_chunk": 6.037, "boundary_entropy": 0.068, "router_noise": 0.0, "grad_norm": 0.3298, "lr": 4.51e-05, "router_trainable": false, "sec_per_step": 0.179, "alpha": 0.0, "t": 10936.5} |
| {"event": "train", "step": 65500, "phase": 3, "objective": "none", "lm_loss": 0.82379, "bpb": 1.18848, "chunk_loss": 0.0, "F": 0.15738, "G": 0.16382, "chunks": 161.16, "bytes_per_chunk": 6.354, "boundary_entropy": 0.07241, "router_noise": 0.0, "grad_norm": 0.3273, "lr": 4.38e-05, "router_trainable": false, "sec_per_step": 0.178, "alpha": 0.0, "t": 11025.7} |
| {"event": "train", "step": 66000, "phase": 3, "objective": "none", "lm_loss": 0.85962, "bpb": 1.24017, "chunk_loss": 0.0, "F": 0.15851, "G": 0.16617, "chunks": 162.31, "bytes_per_chunk": 6.309, "boundary_entropy": 0.07384, "router_noise": 0.0, "grad_norm": 0.3264, "lr": 4.26e-05, "router_trainable": false, "sec_per_step": 0.177, "alpha": 0.0, "t": 11114.4} |
| {"event": "train", "step": 66500, "phase": 3, "objective": "none", "lm_loss": 0.89627, "bpb": 1.29304, "chunk_loss": 0.0, "F": 0.15929, "G": 0.17118, "chunks": 163.11, "bytes_per_chunk": 6.278, "boundary_entropy": 0.08033, "router_noise": 0.0, "grad_norm": 0.3373, "lr": 4.14e-05, "router_trainable": false, "sec_per_step": 0.177, "alpha": 0.0, "t": 11202.9} |
| {"event": "train", "step": 67000, "phase": 3, "objective": "none", "lm_loss": 0.90618, "bpb": 1.30733, "chunk_loss": 0.0, "F": 0.16542, "G": 0.17368, "chunks": 169.39, "bytes_per_chunk": 6.045, "boundary_entropy": 0.07231, "router_noise": 0.0, "grad_norm": 0.3362, "lr": 4.03e-05, "router_trainable": false, "sec_per_step": 0.18, "alpha": 0.0, "t": 11293.0} |
| {"event": "train", "step": 67500, "phase": 3, "objective": "none", "lm_loss": 0.85446, "bpb": 1.23273, "chunk_loss": 0.0, "F": 0.16006, "G": 0.16428, "chunks": 163.91, "bytes_per_chunk": 6.247, "boundary_entropy": 0.06489, "router_noise": 0.0, "grad_norm": 0.3221, "lr": 3.92e-05, "router_trainable": false, "sec_per_step": 0.18, "alpha": 0.0, "t": 11382.8} |
| {"event": "probe", "step": 68000, "phase": 3, "objective": "none", "router_trainable": false, "per_lang": {"en": {"chunks": 3098.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.178015, "cps": 24.203125, "mean_p": 0.176188, "boundary_entropy": 0.029388, "bpb": 1.461092}, "de": {"chunks": 3753.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.182779, "cps": 29.320312, "mean_p": 0.178421, "boundary_entropy": 0.033501, "bpb": 1.553367}, "vi": {"chunks": 3789.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.205533, "cps": 29.601562, "mean_p": 0.153152, "boundary_entropy": 0.039418, "bpb": 1.193672}, "tr": {"chunks": 3404.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.188035, "cps": 26.59375, "mean_p": 0.172494, "boundary_entropy": 0.027965, "bpb": 1.45099}, "fi": {"chunks": 3402.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180852, "cps": 26.578125, "mean_p": 0.174297, "boundary_entropy": 0.025197, "bpb": 1.568251}, "ru": {"chunks": 4211.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.222216, "cps": 32.898438, "mean_p": 0.118737, "boundary_entropy": 0.041875, "bpb": 0.918124}, "el": {"chunks": 4474.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.217999, "cps": 34.953125, "mean_p": 0.118406, "boundary_entropy": 0.051556, "bpb": 0.918501}, "hi": {"chunks": 7150.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.424963, "cps": 55.859375, "mean_p": 0.157917, "boundary_entropy": 0.119633, "bpb": 0.753794}, "ta": {"chunks": 6131.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.327949, "cps": 47.898438, "mean_p": 0.117603, "boundary_entropy": 0.100669, "bpb": 0.659896}, "th": {"chunks": 6011.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.373702, "cps": 46.960938, "mean_p": 0.131196, "boundary_entropy": 0.120767, "bpb": 0.744641}, "zh": {"chunks": 2918.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.494074, "cps": 22.796875, "mean_p": 0.289929, "boundary_entropy": 0.307092, "bpb": 2.173963}, "my": {"chunks": 15327.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.808898, "cps": 119.742188, "mean_p": 0.270295, "boundary_entropy": 0.123066, "bpb": 0.69584}}, "bpb": {"en": 1.46109, "de": 1.55337, "vi": 1.19367, "tr": 1.45099, "fi": 1.56825, "ru": 0.91812, "el": 0.9185, "hi": 0.75379, "ta": 0.6599, "th": 0.74464, "zh": 2.17396, "my": 0.69584}, "gini_cpc": 0.27932, "gini_cps": 0.27184, "bpic": 6.027848, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.026759817298938, "vi": 1.154580371808684, "tr": 1.056286445252111, "fi": 1.0159331523088955, "ru": 1.2482993198437666, "el": 1.2246101124415136, "hi": 2.3872267685522077, "ta": 1.8422499499716058, "th": 2.0992703385600984, "zh": 2.7754573097748083, "my": 4.5439808047930565}, "t": 11475.5} |
| {"event": "eval", "step": 68000, "phase": 3, "bpb": {"en": 1.45559, "de": 1.56174, "vi": 1.2135, "tr": 1.44614, "fi": 1.58859, "ru": 0.91576, "el": 0.91644, "hi": 0.75987, "ta": 0.66599, "th": 0.74426, "zh": 2.17769, "my": 0.70138}, "bytes_per_chunk": {"en": 5.441, "de": 5.59, "vi": 6.384, "tr": 5.684, "fi": 5.611, "ru": 7.876, "el": 8.399, "hi": 5.786, "ta": 8.068, "th": 7.491, "zh": 5.581, "my": 3.578}, "bpb_mean": 1.17891, "t": 11480.8} |
| {"event": "train", "step": 68000, "phase": 3, "objective": "none", "lm_loss": 0.84666, "bpb": 1.22148, "chunk_loss": 0.0, "F": 0.16208, "G": 0.16982, "chunks": 165.97, "bytes_per_chunk": 6.17, "boundary_entropy": 0.06965, "router_noise": 0.0, "grad_norm": 0.3405, "lr": 3.82e-05, "router_trainable": false, "sec_per_step": 0.196, "alpha": 0.0, "t": 11480.9} |
| {"event": "train", "step": 68500, "phase": 3, "objective": "none", "lm_loss": 0.91269, "bpb": 1.31673, "chunk_loss": 0.0, "F": 0.17169, "G": 0.18131, "chunks": 175.81, "bytes_per_chunk": 5.824, "boundary_entropy": 0.0781, "router_noise": 0.0, "grad_norm": 0.3535, "lr": 3.73e-05, "router_trainable": false, "sec_per_step": 0.183, "alpha": 0.0, "t": 11572.5} |
| {"event": "train", "step": 69000, "phase": 3, "objective": "none", "lm_loss": 0.97393, "bpb": 1.40509, "chunk_loss": 0.0, "F": 0.17293, "G": 0.1778, "chunks": 177.08, "bytes_per_chunk": 5.783, "boundary_entropy": 0.0574, "router_noise": 0.0, "grad_norm": 0.3439, "lr": 3.64e-05, "router_trainable": false, "sec_per_step": 0.18, "alpha": 0.0, "t": 11662.7} |
| {"event": "train", "step": 69500, "phase": 3, "objective": "none", "lm_loss": 0.89787, "bpb": 1.29535, "chunk_loss": 0.0, "F": 0.16658, "G": 0.17246, "chunks": 170.58, "bytes_per_chunk": 6.003, "boundary_entropy": 0.06758, "router_noise": 0.0, "grad_norm": 0.3527, "lr": 3.55e-05, "router_trainable": false, "sec_per_step": 0.181, "alpha": 0.0, "t": 11753.4} |
| {"event": "train", "step": 70000, "phase": 3, "objective": "none", "lm_loss": 0.95917, "bpb": 1.38379, "chunk_loss": 0.0, "F": 0.17339, "G": 0.1837, "chunks": 177.55, "bytes_per_chunk": 5.767, "boundary_entropy": 0.07754, "router_noise": 0.0, "grad_norm": 0.347, "lr": 3.48e-05, "router_trainable": false, "sec_per_step": 0.181, "alpha": 0.0, "t": 11843.9} |
| {"event": "train", "step": 70500, "phase": 3, "objective": "none", "lm_loss": 0.87252, "bpb": 1.25878, "chunk_loss": 0.0, "F": 0.16945, "G": 0.17391, "chunks": 173.52, "bytes_per_chunk": 5.901, "boundary_entropy": 0.06227, "router_noise": 0.0, "grad_norm": 0.3386, "lr": 3.4e-05, "router_trainable": false, "sec_per_step": 0.182, "alpha": 0.0, "t": 11935.0} |
| {"event": "train", "step": 71000, "phase": 3, "objective": "none", "lm_loss": 0.9041, "bpb": 1.30434, "chunk_loss": 0.0, "F": 0.16649, "G": 0.17634, "chunks": 170.48, "bytes_per_chunk": 6.006, "boundary_entropy": 0.07405, "router_noise": 0.0, "grad_norm": 0.353, "lr": 3.34e-05, "router_trainable": false, "sec_per_step": 0.181, "alpha": 0.0, "t": 12025.5} |
| {"event": "train", "step": 71500, "phase": 3, "objective": "none", "lm_loss": 0.95472, "bpb": 1.37737, "chunk_loss": 0.0, "F": 0.16884, "G": 0.1756, "chunks": 172.89, "bytes_per_chunk": 5.923, "boundary_entropy": 0.06502, "router_noise": 0.0, "grad_norm": 0.3616, "lr": 3.28e-05, "router_trainable": false, "sec_per_step": 0.181, "alpha": 0.0, "t": 12116.2} |
| {"event": "probe", "step": 72000, "phase": 3, "objective": "none", "router_trainable": false, "per_lang": {"en": {"chunks": 3096.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.1779, "cps": 24.1875, "mean_p": 0.176777, "boundary_entropy": 0.028867, "bpb": 1.455279}, "de": {"chunks": 3747.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.182487, "cps": 29.273438, "mean_p": 0.17845, "boundary_entropy": 0.032355, "bpb": 1.547241}, "vi": {"chunks": 3785.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.205316, "cps": 29.570312, "mean_p": 0.153883, "boundary_entropy": 0.037043, "bpb": 1.188707}, "tr": {"chunks": 3401.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.187869, "cps": 26.570312, "mean_p": 0.1727, "boundary_entropy": 0.027712, "bpb": 1.44367}, "fi": {"chunks": 3399.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180692, "cps": 26.554688, "mean_p": 0.174162, "boundary_entropy": 0.025267, "bpb": 1.559533}, "ru": {"chunks": 4145.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.218734, "cps": 32.382812, "mean_p": 0.116978, "boundary_entropy": 0.0412, "bpb": 0.913264}, "el": {"chunks": 4741.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.231009, "cps": 37.039062, "mean_p": 0.125822, "boundary_entropy": 0.051527, "bpb": 0.911103}, "hi": {"chunks": 7206.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.428291, "cps": 56.296875, "mean_p": 0.159017, "boundary_entropy": 0.119327, "bpb": 0.747812}, "ta": {"chunks": 6451.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.345066, "cps": 50.398438, "mean_p": 0.126042, "boundary_entropy": 0.102012, "bpb": 0.654198}, "th": {"chunks": 6055.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.376438, "cps": 47.304688, "mean_p": 0.131548, "boundary_entropy": 0.120454, "bpb": 0.740063}, "zh": {"chunks": 2918.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.494074, "cps": 22.796875, "mean_p": 0.290607, "boundary_entropy": 0.308895, "bpb": 2.166648}, "my": {"chunks": 15692.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.828161, "cps": 122.59375, "mean_p": 0.27583, "boundary_entropy": 0.122253, "bpb": 0.68757}}, "bpb": {"en": 1.45528, "de": 1.54724, "vi": 1.18871, "tr": 1.44367, "fi": 1.55953, "ru": 0.91326, "el": 0.9111, "hi": 0.74781, "ta": 0.6542, "th": 0.74006, "zh": 2.16665, "my": 0.68757}, "gini_cpc": 0.282421, "gini_cps": 0.277812, "bpic": 5.937573, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0257805359268517, "vi": 1.1541065612890882, "tr": 1.0560372772168902, "fi": 1.0156929765241267, "ru": 1.2295281851465505, "el": 1.2985308693177289, "hi": 2.407478143681537, "ta": 1.9396561178941432, "th": 2.1160028602892096, "zh": 2.7772502408534745, "my": 4.655197292303866}, "t": 12210.0} |
| {"event": "eval", "step": 72000, "phase": 3, "bpb": {"en": 1.45001, "de": 1.55606, "vi": 1.2084, "tr": 1.4386, "fi": 1.57981, "ru": 0.91163, "el": 0.90872, "hi": 0.75387, "ta": 0.6602, "th": 0.73976, "zh": 2.17068, "my": 0.69358}, "bytes_per_chunk": {"en": 5.442, "de": 5.616, "vi": 6.385, "tr": 5.665, "fi": 5.616, "ru": 7.976, "el": 7.98, "hi": 5.737, "ta": 7.649, "th": 7.491, "zh": 5.597, "my": 3.503}, "bpb_mean": 1.17261, "t": 12215.3} |
| {"event": "train", "step": 72000, "phase": 3, "objective": "none", "lm_loss": 0.80111, "bpb": 1.15576, "chunk_loss": 0.0, "F": 0.15558, "G": 0.16249, "chunks": 159.31, "bytes_per_chunk": 6.428, "boundary_entropy": 0.07453, "router_noise": 0.0, "grad_norm": 0.339, "lr": 3.22e-05, "router_trainable": false, "sec_per_step": 0.199, "alpha": 0.0, "t": 12215.5} |
| {"event": "train", "step": 72500, "phase": 3, "objective": "none", "lm_loss": 0.856, "bpb": 1.23494, "chunk_loss": 0.0, "F": 0.16127, "G": 0.16512, "chunks": 165.14, "bytes_per_chunk": 6.201, "boundary_entropy": 0.05479, "router_noise": 0.0, "grad_norm": 0.3493, "lr": 3.17e-05, "router_trainable": false, "sec_per_step": 0.181, "alpha": 0.0, "t": 12306.0} |
| {"event": "train", "step": 73000, "phase": 3, "objective": "none", "lm_loss": 0.8315, "bpb": 1.1996, "chunk_loss": 0.0, "F": 0.16489, "G": 0.17548, "chunks": 168.84, "bytes_per_chunk": 6.065, "boundary_entropy": 0.08906, "router_noise": 0.0, "grad_norm": 0.3545, "lr": 3.13e-05, "router_trainable": false, "sec_per_step": 0.181, "alpha": 0.0, "t": 12396.5} |
| {"event": "train", "step": 73500, "phase": 3, "objective": "none", "lm_loss": 0.88648, "bpb": 1.27892, "chunk_loss": 0.0, "F": 0.16982, "G": 0.18492, "chunks": 173.89, "bytes_per_chunk": 5.889, "boundary_entropy": 0.092, "router_noise": 0.0, "grad_norm": 0.3648, "lr": 3.09e-05, "router_trainable": false, "sec_per_step": 0.181, "alpha": 0.0, "t": 12487.2} |
| {"event": "train", "step": 74000, "phase": 3, "objective": "none", "lm_loss": 0.8868, "bpb": 1.27939, "chunk_loss": 0.0, "F": 0.16779, "G": 0.18178, "chunks": 171.81, "bytes_per_chunk": 5.96, "boundary_entropy": 0.08103, "router_noise": 0.0, "grad_norm": 0.3544, "lr": 3.06e-05, "router_trainable": false, "sec_per_step": 0.183, "alpha": 0.0, "t": 12578.9} |
| {"event": "train", "step": 74500, "phase": 3, "objective": "none", "lm_loss": 0.86162, "bpb": 1.24305, "chunk_loss": 0.0, "F": 0.1646, "G": 0.16614, "chunks": 168.55, "bytes_per_chunk": 6.075, "boundary_entropy": 0.05513, "router_noise": 0.0, "grad_norm": 0.3516, "lr": 3.04e-05, "router_trainable": false, "sec_per_step": 0.184, "alpha": 0.0, "t": 12671.1} |
| {"event": "train", "step": 75000, "phase": 3, "objective": "none", "lm_loss": 0.93102, "bpb": 1.34318, "chunk_loss": 0.0, "F": 0.16599, "G": 0.17872, "chunks": 169.97, "bytes_per_chunk": 6.025, "boundary_entropy": 0.07936, "router_noise": 0.0, "grad_norm": 0.3675, "lr": 3.02e-05, "router_trainable": false, "sec_per_step": 0.184, "alpha": 0.0, "t": 12763.2} |
| {"event": "train", "step": 75500, "phase": 3, "objective": "none", "lm_loss": 0.91117, "bpb": 1.31454, "chunk_loss": 0.0, "F": 0.16486, "G": 0.17338, "chunks": 168.81, "bytes_per_chunk": 6.066, "boundary_entropy": 0.06914, "router_noise": 0.0, "grad_norm": 0.3696, "lr": 3.01e-05, "router_trainable": false, "sec_per_step": 0.18, "alpha": 0.0, "t": 12853.1} |
| {"event": "probe", "step": 76000, "phase": 3, "objective": "none", "router_trainable": false, "per_lang": {"en": {"chunks": 3105.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.178418, "cps": 24.257812, "mean_p": 0.176953, "boundary_entropy": 0.028679, "bpb": 1.452499}, "de": {"chunks": 3744.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.182341, "cps": 29.25, "mean_p": 0.178421, "boundary_entropy": 0.032308, "bpb": 1.543267}, "vi": {"chunks": 3796.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.205913, "cps": 29.65625, "mean_p": 0.154212, "boundary_entropy": 0.037266, "bpb": 1.185062}, "tr": {"chunks": 3415.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.188643, "cps": 26.679688, "mean_p": 0.17303, "boundary_entropy": 0.027873, "bpb": 1.440358}, "fi": {"chunks": 3409.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.181224, "cps": 26.632812, "mean_p": 0.174559, "boundary_entropy": 0.024559, "bpb": 1.553541}, "ru": {"chunks": 4181.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.220633, "cps": 32.664062, "mean_p": 0.118354, "boundary_entropy": 0.040885, "bpb": 0.911924}, "el": {"chunks": 4925.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.239975, "cps": 38.476562, "mean_p": 0.130401, "boundary_entropy": 0.050632, "bpb": 0.907372}, "hi": {"chunks": 7208.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.42841, "cps": 56.3125, "mean_p": 0.158291, "boundary_entropy": 0.120058, "bpb": 0.745891}, "ta": {"chunks": 6525.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.349024, "cps": 50.976562, "mean_p": 0.125141, "boundary_entropy": 0.10622, "bpb": 0.650744}, "th": {"chunks": 5927.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.36848, "cps": 46.304688, "mean_p": 0.128887, "boundary_entropy": 0.118784, "bpb": 0.736623}, "zh": {"chunks": 2925.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.495259, "cps": 22.851562, "mean_p": 0.291315, "boundary_entropy": 0.309855, "bpb": 2.162155}, "my": {"chunks": 14919.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.787365, "cps": 116.554688, "mean_p": 0.262905, "boundary_entropy": 0.126528, "bpb": 0.680135}}, "bpb": {"en": 1.4525, "de": 1.54327, "vi": 1.18506, "tr": 1.44036, "fi": 1.55354, "ru": 0.91192, "el": 0.90737, "hi": 0.74589, "ta": 0.65074, "th": 0.73662, "zh": 2.16215, "my": 0.68014}, "gini_cpc": 0.274354, "gini_cps": 0.269408, "bpic": 5.989185, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0219883580831703, "vi": 1.1541056774749774, "tr": 1.0573108095089798, "fi": 1.015728488598895, "ru": 1.2366120352312848, "el": 1.345017408299537, "hi": 2.4011661964956104, "ta": 1.9562194030024302, "th": 2.0652678178577273, "zh": 2.7758432839117186, "my": 4.413050056141445}, "t": 12946.9} |
| {"event": "eval", "step": 76000, "phase": 3, "bpb": {"en": 1.44719, "de": 1.55206, "vi": 1.20467, "tr": 1.43497, "fi": 1.57523, "ru": 0.90927, "el": 0.90541, "hi": 0.75137, "ta": 0.65675, "th": 0.73601, "zh": 2.1666, "my": 0.68634}, "bytes_per_chunk": {"en": 5.441, "de": 5.606, "vi": 6.376, "tr": 5.666, "fi": 5.606, "ru": 7.89, "el": 7.681, "hi": 5.781, "ta": 7.604, "th": 7.672, "zh": 5.587, "my": 3.676}, "bpb_mean": 1.16882, "t": 12952.5} |
| {"event": "train", "step": 76000, "phase": 3, "objective": "none", "lm_loss": 0.79863, "bpb": 1.15218, "chunk_loss": 0.0, "F": 0.1718, "G": 0.17687, "chunks": 175.92, "bytes_per_chunk": 5.821, "boundary_entropy": 0.07799, "router_noise": 0.0, "grad_norm": 0.3568, "lr": 3e-05, "router_trainable": false, "sec_per_step": 0.199, "alpha": 0.0, "t": 12952.7} |
| {"event": "probe", "step": 76300, "phase": 3, "objective": "none", "router_trainable": false, "per_lang": {"en": {"chunks": 3115.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.178992, "cps": 24.335938, "mean_p": 0.177854, "boundary_entropy": 0.029047, "bpb": 1.452139}, "de": {"chunks": 3777.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.183948, "cps": 29.507812, "mean_p": 0.179414, "boundary_entropy": 0.03258, "bpb": 1.542992}, "vi": {"chunks": 3796.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.205913, "cps": 29.65625, "mean_p": 0.154227, "boundary_entropy": 0.036882, "bpb": 1.18454}, "tr": {"chunks": 3418.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.188808, "cps": 26.703125, "mean_p": 0.173053, "boundary_entropy": 0.027726, "bpb": 1.439087}, "fi": {"chunks": 3401.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180798, "cps": 26.570312, "mean_p": 0.174466, "boundary_entropy": 0.024073, "bpb": 1.554182}, "ru": {"chunks": 4114.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.217098, "cps": 32.140625, "mean_p": 0.115904, "boundary_entropy": 0.040733, "bpb": 0.912153}, "el": {"chunks": 4805.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.234128, "cps": 37.539062, "mean_p": 0.127319, "boundary_entropy": 0.052052, "bpb": 0.907882}, "hi": {"chunks": 7430.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.441605, "cps": 58.046875, "mean_p": 0.164417, "boundary_entropy": 0.117483, "bpb": 0.743868}, "ta": {"chunks": 7035.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.376304, "cps": 54.960938, "mean_p": 0.135833, "boundary_entropy": 0.109554, "bpb": 0.651572}, "th": {"chunks": 6013.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.373827, "cps": 46.976562, "mean_p": 0.129468, "boundary_entropy": 0.120429, "bpb": 0.736427}, "zh": {"chunks": 2928.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.495767, "cps": 22.875, "mean_p": 0.288747, "boundary_entropy": 0.305772, "bpb": 2.160891}, "my": {"chunks": 15706.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.8289, "cps": 122.703125, "mean_p": 0.275462, "boundary_entropy": 0.120954, "bpb": 0.680435}}, "bpb": {"en": 1.45214, "de": 1.54299, "vi": 1.18454, "tr": 1.43909, "fi": 1.55418, "ru": 0.91215, "el": 0.90788, "hi": 0.74387, "ta": 0.65157, "th": 0.73643, "zh": 2.16089, "my": 0.68044}, "gini_cpc": 0.282584, "gini_cps": 0.281272, "bpic": 5.855855, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0276864889381765, "vi": 1.15040068332578, "tr": 1.054842395073172, "fi": 1.01009173536914, "ru": 1.2128892371154978, "el": 1.3080327446894526, "hi": 2.4671741728061742, "ta": 2.10234846185515, "th": 2.088508310834646, "zh": 2.769769948562797, "my": 4.630930745326146}, "t": 13010.8} |
| {"event": "eval", "step": 76300, "phase": 3, "bpb": {"en": 1.4468, "de": 1.55227, "vi": 1.20398, "tr": 1.43412, "fi": 1.57561, "ru": 0.90957, "el": 0.90585, "hi": 0.74971, "ta": 0.65783, "th": 0.73567, "zh": 2.16492, "my": 0.68656}, "bytes_per_chunk": {"en": 5.428, "de": 5.596, "vi": 6.376, "tr": 5.661, "fi": 5.612, "ru": 8.03, "el": 7.854, "hi": 5.617, "ta": 7.029, "th": 7.552, "zh": 5.59, "my": 3.495}, "bpb_mean": 1.16857, "t": 13016.1} |
| {"event": "probe", "step": 76300, "phase": 3, "objective": "none", "router_trainable": false, "per_lang": {"en": {"chunks": 3115.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.178992, "cps": 24.335938, "mean_p": 0.177854, "boundary_entropy": 0.029047, "bpb": 1.452139}, "de": {"chunks": 3777.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.183948, "cps": 29.507812, "mean_p": 0.179414, "boundary_entropy": 0.03258, "bpb": 1.542992}, "vi": {"chunks": 3796.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.205913, "cps": 29.65625, "mean_p": 0.154227, "boundary_entropy": 0.036882, "bpb": 1.18454}, "tr": {"chunks": 3418.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.188808, "cps": 26.703125, "mean_p": 0.173053, "boundary_entropy": 0.027726, "bpb": 1.439087}, "fi": {"chunks": 3401.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180798, "cps": 26.570312, "mean_p": 0.174466, "boundary_entropy": 0.024073, "bpb": 1.554182}, "ru": {"chunks": 4114.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.217098, "cps": 32.140625, "mean_p": 0.115904, "boundary_entropy": 0.040733, "bpb": 0.912153}, "el": {"chunks": 4805.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.234128, "cps": 37.539062, "mean_p": 0.127319, "boundary_entropy": 0.052052, "bpb": 0.907882}, "hi": {"chunks": 7430.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.441605, "cps": 58.046875, "mean_p": 0.164417, "boundary_entropy": 0.117483, "bpb": 0.743868}, "ta": {"chunks": 7035.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.376304, "cps": 54.960938, "mean_p": 0.135833, "boundary_entropy": 0.109554, "bpb": 0.651572}, "th": {"chunks": 6013.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.373827, "cps": 46.976562, "mean_p": 0.129468, "boundary_entropy": 0.120429, "bpb": 0.736427}, "zh": {"chunks": 2928.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.495767, "cps": 22.875, "mean_p": 0.288747, "boundary_entropy": 0.305772, "bpb": 2.160891}, "my": {"chunks": 15706.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.8289, "cps": 122.703125, "mean_p": 0.275462, "boundary_entropy": 0.120954, "bpb": 0.680435}}, "bpb": {"en": 1.45214, "de": 1.54299, "vi": 1.18454, "tr": 1.43909, "fi": 1.55418, "ru": 0.91215, "el": 0.90788, "hi": 0.74387, "ta": 0.65157, "th": 0.73643, "zh": 2.16089, "my": 0.68044}, "gini_cpc": 0.282584, "gini_cps": 0.281272, "bpic": 5.855855, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0276864889381765, "vi": 1.15040068332578, "tr": 1.054842395073172, "fi": 1.01009173536914, "ru": 1.2128892371154978, "el": 1.3080327446894526, "hi": 2.4671741728061742, "ta": 2.10234846185515, "th": 2.088508310834646, "zh": 2.769769948562797, "my": 4.630930745326146}, "t": 13019.2} |
| {"event": "reps_dumped", "dirs": ["/local/scratch/sas245/beetle-hnet/scratch/reps/specfull_B_s2/th0.30", "/local/scratch/sas245/beetle-hnet/scratch/reps/specfull_B_s2/th0.40", "/local/scratch/sas245/beetle-hnet/scratch/reps/specfull_B_s2/th0.50", "/local/scratch/sas245/beetle-hnet/scratch/reps/specfull_B_s2/th0.60", "/local/scratch/sas245/beetle-hnet/scratch/reps/specfull_B_s2/th0.70"], "t": 13022.6} |
| {"event": "final", "run": "specfull_B_s2", "condition": "B", "step": 76300, "params": 82470528, "per_lang": {"en": {"cpc": 0.1789921277940585, "cps": 24.3359375, "bpb": 1.4437584902361158, "bytes_per_chunk": 5.421771345546301}, "de": {"cpc": 0.18394779136024936, "cps": 29.5078125, "bpb": 1.5484230630866367, "bytes_per_chunk": 5.574922376759814}, "vi": {"cpc": 0.20591266612422024, "cps": 29.65625, "bpb": 1.2020494130585777, "bytes_per_chunk": 6.381693141013937}, "tr": {"cpc": 0.18880848478152792, "cps": 26.703125, "bpb": 1.4326551075648464, "bytes_per_chunk": 5.667396321458877}, "fi": {"cpc": 0.18079846898091542, "cps": 26.5703125, "bpb": 1.5804008830020422, "bytes_per_chunk": 5.609098009008141}, "ru": {"cpc": 0.2170976253298153, "cps": 32.140625, "bpb": 0.9120210147013733, "bytes_per_chunk": 8.020744570995847}, "el": {"cpc": 0.2341275641962676, "cps": 37.5390625, "bpb": 0.9042780064197701, "bytes_per_chunk": 7.813764940832812}, "hi": {"cpc": 0.4416047548291233, "cps": 58.046875, "bpb": 0.7510825710417928, "bytes_per_chunk": 5.623175348842198}, "ta": {"cpc": 0.3763038245520193, "cps": 54.9609375, "bpb": 0.6558157294864648, "bytes_per_chunk": 7.02582313763853}, "th": {"cpc": 0.3738265464718682, "cps": 46.9765625, "bpb": 0.7376263420941898, "bytes_per_chunk": 7.557794739838981}, "zh": {"cpc": 0.495767016593295, "cps": 22.875, "bpb": 2.164724977743153, "bytes_per_chunk": 5.597032213296325}, "my": {"cpc": 0.828900147772852, "cps": 122.703125, "bpb": 0.6876772326980799, "bytes_per_chunk": 3.4925291605880746}}, "gini_cpc": 0.28258356994305767, "gini_cps": 0.28127193383990967, "bpic": 5.855854618694498, "seed": 2, "mean_boundary_entropy": 0.09001854062080383, "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": "parity_a", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0}, {"step": 53410, "phase": 3, "objective": "none", "router_trainable": false, "encoder_trainable": true, "frozen_router_params": 295680}], "t": 13033.6} |
|
|
