| {"event": "start", "run": "emactl_noema_s0", "condition": "A", "params": 22516992, "langs": ["en", "de", "vi", "tr", "fi", "ru", "el", "hi", "ta", "th", "zh", "my"], "premium": {"en": 1.0, "de": 1.184, "vi": 1.39, "tr": 1.12, "fi": 1.104, "ru": 1.971, "el": 2.174, "hi": 2.552, "ta": 3.192, "th": 2.762, "zh": 0.915, "my": 3.524}, "schedule": {"tau1": 0.4, "tau2": 0.7, "p1_objective": "global", "p2_objective": "global", "p3_objective": "global", "freeze_in_p3": false, "freeze_scope": "router"}, "corpus": "/local/scratch/sas245/beetle-hnet/scratch/data/tier1_corpus", "premium_kind": "flores_per_sent", "doc_filtered": {"my": 37511}, "weights": {"en": 0.2946, "de": 0.0909, "vi": 0.0576, "tr": 0.0612, "fi": 0.0424, "ru": 0.1476, "el": 0.0522, "hi": 0.042, "ta": 0.0313, "th": 0.0542, "zh": 0.1038, "my": 0.0222}, "t": 1.5} |
| {"event": "probe_schedule", "n": 40, "steps": [0, 1, 2, 4, 8, 16, 32, 64, 128, 200, 256, 400, 600, 800, 1000, 1200, 1400, 1600, 1800, 2000, 2200, 2400, 2600, 2800, 3000, 3200, 3400, 3600, 3800, 4000, 4200, 4400, 4600, 4800, 5000, 5200, 5400, 5600, 5800, 6000], "t": 1.5} |
| {"event": "probe", "step": 0, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 6384.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.366833, "cps": 49.875, "mean_p": 0.496725, "boundary_entropy": 0.686995, "bpb": 8.125775}, "de": {"chunks": 8628.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.420202, "cps": 67.40625, "mean_p": 0.498889, "boundary_entropy": 0.687821, "bpb": 8.121816}, "vi": {"chunks": 13115.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.711418, "cps": 102.460938, "mean_p": 0.505745, "boundary_entropy": 0.688385, "bpb": 8.087584}, "tr": {"chunks": 5290.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.292217, "cps": 41.328125, "mean_p": 0.489567, "boundary_entropy": 0.687288, "bpb": 8.14748}, "fi": {"chunks": 8895.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.472862, "cps": 69.492188, "mean_p": 0.500688, "boundary_entropy": 0.687512, "bpb": 8.208588}, "ru": {"chunks": 10744.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.566966, "cps": 83.9375, "mean_p": 0.493655, "boundary_entropy": 0.689893, "bpb": 8.082947}, "el": {"chunks": 15039.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.732788, "cps": 117.492188, "mean_p": 0.497482, "boundary_entropy": 0.690234, "bpb": 8.196862}, "hi": {"chunks": 2563.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.152333, "cps": 20.023438, "mean_p": 0.484638, "boundary_entropy": 0.690256, "bpb": 8.272031}, "ta": {"chunks": 38089.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.03739, "cps": 297.570312, "mean_p": 0.512976, "boundary_entropy": 0.690378, "bpb": 8.008816}, "th": {"chunks": 34006.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.114144, "cps": 265.671875, "mean_p": 0.511916, "boundary_entropy": 0.690552, "bpb": 8.270247}, "zh": {"chunks": 11821.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.001524, "cps": 92.351562, "mean_p": 0.52004, "boundary_entropy": 0.685674, "bpb": 8.131836}, "my": {"chunks": 48201.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.543857, "cps": 376.570312, "mean_p": 0.521563, "boundary_entropy": 0.690206, "bpb": 8.106955}}, "bpb": {"en": 8.12577, "de": 8.12182, "vi": 8.08758, "tr": 8.14748, "fi": 8.20859, "ru": 8.08295, "el": 8.19686, "hi": 8.27203, "ta": 8.00882, "th": 8.27025, "zh": 8.13184, "my": 8.10696}, "gini_cpc": 0.424496, "gini_cps": 0.425905, "bpic": 1.892645, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1454838515955792, "vi": 1.939350894526797, "tr": 0.7965927738475616, "fi": 1.2890367847879758, "ru": 1.5455676790921895, "el": 1.997603802482379, "hi": 0.41526447061189375, "ta": 5.553993192060118, "th": 5.7632270176529286, "zh": 5.456221801328583, "my": 6.934639901884223}, "t": 3.3} |
| {"event": "train", "step": 0, "phase": 1, "objective": "global", "lm_loss": 5.63738, "bpb": 8.13302, "chunk_loss": 0.05247, "F": 0.47223, "G": 0.4995, "chunks": 483.56, "bytes_per_chunk": 2.118, "boundary_entropy": 0.69158, "router_noise": 0.0, "grad_norm": 7.4901, "lr": 1e-06, "router_trainable": true, "sec_per_step": 0.005, "alpha": 0.03, "t": 3.6} |
| {"event": "probe", "step": 1, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 6342.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.36442, "cps": 49.546875, "mean_p": 0.49658, "boundary_entropy": 0.68699, "bpb": 8.105601}, "de": {"chunks": 8605.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.419081, "cps": 67.226562, "mean_p": 0.498815, "boundary_entropy": 0.68782, "bpb": 8.104426}, "vi": {"chunks": 13055.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.708164, "cps": 101.992188, "mean_p": 0.505612, "boundary_entropy": 0.688386, "bpb": 8.073795}, "tr": {"chunks": 5262.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.29067, "cps": 41.109375, "mean_p": 0.489476, "boundary_entropy": 0.687283, "bpb": 8.134076}, "fi": {"chunks": 8850.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.470469, "cps": 69.140625, "mean_p": 0.500541, "boundary_entropy": 0.68751, "bpb": 8.191988}, "ru": {"chunks": 10640.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.561478, "cps": 83.125, "mean_p": 0.493511, "boundary_entropy": 0.689889, "bpb": 8.047788}, "el": {"chunks": 14970.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.729426, "cps": 116.953125, "mean_p": 0.497406, "boundary_entropy": 0.690233, "bpb": 8.183408}, "hi": {"chunks": 2509.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.149123, "cps": 19.601562, "mean_p": 0.484458, "boundary_entropy": 0.690244, "bpb": 8.262003}, "ta": {"chunks": 37841.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.024124, "cps": 295.632812, "mean_p": 0.512671, "boundary_entropy": 0.690392, "bpb": 7.987595}, "th": {"chunks": 33875.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.105999, "cps": 264.648438, "mean_p": 0.511786, "boundary_entropy": 0.690558, "bpb": 8.255146}, "zh": {"chunks": 11819.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.001185, "cps": 92.335938, "mean_p": 0.520016, "boundary_entropy": 0.685676, "bpb": 8.129458}, "my": {"chunks": 48279.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.547973, "cps": 377.179688, "mean_p": 0.521729, "boundary_entropy": 0.690191, "bpb": 8.087067}}, "bpb": {"en": 8.1056, "de": 8.10443, "vi": 8.0738, "tr": 8.13408, "fi": 8.19199, "ru": 8.04779, "el": 8.18341, "hi": 8.262, "ta": 7.98759, "th": 8.25515, "zh": 8.12946, "my": 8.08707}, "gini_cpc": 0.425536, "gini_cps": 0.426894, "bpic": 1.899464, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1499960536102647, "vi": 1.9432631565712355, "tr": 0.7976239296308818, "fi": 1.2910089996960754, "ru": 1.5407433294308828, "el": 2.0016071222858116, "hi": 0.40920739258969774, "ta": 5.554372700444799, "th": 5.779045597698367, "zh": 5.49142646655054, "my": 6.991860784552214}, "t": 5.0} |
| {"event": "probe", "step": 2, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 6283.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.36103, "cps": 49.085938, "mean_p": 0.496333, "boundary_entropy": 0.686981, "bpb": 8.060648}, "de": {"chunks": 8563.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.417036, "cps": 66.898438, "mean_p": 0.498715, "boundary_entropy": 0.687817, "bpb": 8.065145}, "vi": {"chunks": 12903.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.699919, "cps": 100.804688, "mean_p": 0.505286, "boundary_entropy": 0.688389, "bpb": 8.042932}, "tr": {"chunks": 5223.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.288516, "cps": 40.804688, "mean_p": 0.489331, "boundary_entropy": 0.687276, "bpb": 8.10201}, "fi": {"chunks": 8787.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.46712, "cps": 68.648438, "mean_p": 0.500292, "boundary_entropy": 0.687505, "bpb": 8.154015}, "ru": {"chunks": 10493.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.55372, "cps": 81.976562, "mean_p": 0.493304, "boundary_entropy": 0.689883, "bpb": 7.989191}, "el": {"chunks": 14968.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.729328, "cps": 116.9375, "mean_p": 0.497402, "boundary_entropy": 0.690233, "bpb": 8.16804}, "hi": {"chunks": 2421.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.143893, "cps": 18.914062, "mean_p": 0.484203, "boundary_entropy": 0.690226, "bpb": 8.23532}, "ta": {"chunks": 37435.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.002407, "cps": 292.460938, "mean_p": 0.512171, "boundary_entropy": 0.690416, "bpb": 7.950514}, "th": {"chunks": 33647.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.091825, "cps": 262.867188, "mean_p": 0.511603, "boundary_entropy": 0.690565, "bpb": 8.233771}, "zh": {"chunks": 11802.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.998307, "cps": 92.203125, "mean_p": 0.519959, "boundary_entropy": 0.685679, "bpb": 8.123267}, "my": {"chunks": 48413.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.555045, "cps": 378.226562, "mean_p": 0.521971, "boundary_entropy": 0.690169, "bpb": 8.054636}}, "bpb": {"en": 8.06065, "de": 8.06514, "vi": 8.04293, "tr": 8.10201, "fi": 8.15402, "ru": 7.98919, "el": 8.16804, "hi": 8.23532, "ta": 7.95051, "th": 8.23377, "zh": 8.12327, "my": 8.05464}, "gini_cpc": 0.427097, "gini_cps": 0.428428, "bpic": 1.909947, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.15512929311766, "vi": 1.9386732406028797, "tr": 0.7991467450359587, "fi": 1.2938555809092809, "ru": 1.5337250788134165, "el": 2.0201331240891682, "hi": 0.39856281449104747, "ta": 5.546377539003795, "th": 5.794051397431014, "zh": 5.535020429079937, "my": 7.077105662343485}, "t": 6.6} |
| {"event": "probe", "step": 4, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 6012.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.345458, "cps": 46.96875, "mean_p": 0.495302, "boundary_entropy": 0.686943, "bpb": 7.90879}, "de": {"chunks": 8396.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.408903, "cps": 65.59375, "mean_p": 0.498178, "boundary_entropy": 0.687805, "bpb": 7.930161}, "vi": {"chunks": 12514.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.678817, "cps": 97.765625, "mean_p": 0.504182, "boundary_entropy": 0.688394, "bpb": 7.941891}, "tr": {"chunks": 5038.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.278296, "cps": 39.359375, "mean_p": 0.488686, "boundary_entropy": 0.687243, "bpb": 7.994273}, "fi": {"chunks": 8462.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.449843, "cps": 66.109375, "mean_p": 0.499189, "boundary_entropy": 0.687484, "bpb": 8.023488}, "ru": {"chunks": 9652.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.50934, "cps": 75.40625, "mean_p": 0.492215, "boundary_entropy": 0.689848, "bpb": 7.791359}, "el": {"chunks": 14758.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.719096, "cps": 115.296875, "mean_p": 0.497203, "boundary_entropy": 0.690228, "bpb": 8.098143}, "hi": {"chunks": 2128.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.126478, "cps": 16.625, "mean_p": 0.483035, "boundary_entropy": 0.690144, "bpb": 8.13706}, "ta": {"chunks": 35789.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 1.914362, "cps": 279.601562, "mean_p": 0.51029, "boundary_entropy": 0.690497, "bpb": 7.818992}, "th": {"chunks": 32667.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.030898, "cps": 255.210938, "mean_p": 0.5108, "boundary_entropy": 0.690594, "bpb": 8.143608}, "zh": {"chunks": 11755.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.990349, "cps": 91.835938, "mean_p": 0.519678, "boundary_entropy": 0.685697, "bpb": 8.09786}, "my": {"chunks": 48595.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.564651, "cps": 379.648438, "mean_p": 0.522457, "boundary_entropy": 0.690123, "bpb": 7.97682}}, "bpb": {"en": 7.90879, "de": 7.93016, "vi": 7.94189, "tr": 7.99427, "fi": 8.02349, "ru": 7.79136, "el": 8.09814, "hi": 8.13706, "ta": 7.81899, "th": 8.14361, "zh": 8.09786, "my": 7.97682}, "gini_cpc": 0.434209, "gini_cps": 0.434426, "bpic": 1.960407, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.183655092775536, "vi": 1.964980102176986, "tr": 0.805587576435264, "fi": 1.3021658028899938, "ru": 1.4743929555137747, "el": 2.0815737817353646, "hi": 0.3661185207593728, "ta": 5.541524306295139, "th": 5.878862945541928, "zh": 5.761483720672438, "my": 7.423921288729275}, "t": 8.1} |
| {"event": "probe", "step": 8, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 5162.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.296616, "cps": 40.328125, "mean_p": 0.491991, "boundary_entropy": 0.68679, "bpb": 7.479601}, "de": {"chunks": 7854.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.382506, "cps": 61.359375, "mean_p": 0.496409, "boundary_entropy": 0.687763, "bpb": 7.540113}, "vi": {"chunks": 11068.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.60038, "cps": 86.46875, "mean_p": 0.500656, "boundary_entropy": 0.688379, "bpb": 7.633181}, "tr": {"chunks": 4419.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.244103, "cps": 34.523438, "mean_p": 0.486556, "boundary_entropy": 0.687117, "bpb": 7.665735}, "fi": {"chunks": 7407.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.393759, "cps": 57.867188, "mean_p": 0.49566, "boundary_entropy": 0.687385, "bpb": 7.63762}, "ru": {"chunks": 6535.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.344855, "cps": 51.054688, "mean_p": 0.487698, "boundary_entropy": 0.689645, "bpb": 7.249508}, "el": {"chunks": 13148.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.640647, "cps": 102.71875, "mean_p": 0.495748, "boundary_entropy": 0.690195, "bpb": 7.831379}, "hi": {"chunks": 1173.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.069718, "cps": 9.164062, "mean_p": 0.477956, "boundary_entropy": 0.689721, "bpb": 7.800783}, "ta": {"chunks": 28548.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 1.527039, "cps": 223.03125, "mean_p": 0.503463, "boundary_entropy": 0.690678, "bpb": 7.392556}, "th": {"chunks": 28026.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 1.742369, "cps": 218.953125, "mean_p": 0.506761, "boundary_entropy": 0.69071, "bpb": 7.702815}, "zh": {"chunks": 11559.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.957162, "cps": 90.304688, "mean_p": 0.51868, "boundary_entropy": 0.68576, "bpb": 8.013397}, "my": {"chunks": 49641.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.619854, "cps": 387.820312, "mean_p": 0.52523, "boundary_entropy": 0.689847, "bpb": 7.651141}}, "bpb": {"en": 7.4796, "de": 7.54011, "vi": 7.63318, "tr": 7.66574, "fi": 7.63762, "ru": 7.24951, "el": 7.83138, "hi": 7.80078, "ta": 7.39256, "th": 7.70281, "zh": 8.0134, "my": 7.65114}, "gini_cpc": 0.465857, "gini_cps": 0.457914, "bpic": 2.198814, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.2895690748186177, "vi": 2.0241007626300855, "tr": 0.8229615980004042, "fi": 1.3275062706494678, "ru": 1.1626326033864276, "el": 2.1598568604944286, "hi": 0.23504394037350326, "ta": 5.1482110043781795, "th": 5.874165420843382, "zh": 6.598313425901568, "my": 8.832492260267031}, "t": 9.9} |
| {"event": "probe", "step": 16, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2560.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.147101, "cps": 20.0, "mean_p": 0.480118, "boundary_entropy": 0.685967, "bpb": 6.886538}, "de": {"chunks": 4862.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.23679, "cps": 37.984375, "mean_p": 0.487282, "boundary_entropy": 0.687439, "bpb": 6.972519}, "vi": {"chunks": 7619.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.41329, "cps": 59.523438, "mean_p": 0.491752, "boundary_entropy": 0.688149, "bpb": 7.197694}, "tr": {"chunks": 2858.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.157874, "cps": 22.328125, "mean_p": 0.480187, "boundary_entropy": 0.686636, "bpb": 7.184829}, "fi": {"chunks": 4150.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.220616, "cps": 32.421875, "mean_p": 0.48437, "boundary_entropy": 0.686812, "bpb": 7.078677}, "ru": {"chunks": 710.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.037467, "cps": 5.546875, "mean_p": 0.468102, "boundary_entropy": 0.687786, "bpb": 6.5298}, "el": {"chunks": 4387.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.21376, "cps": 34.273438, "mean_p": 0.483302, "boundary_entropy": 0.689587, "bpb": 7.204587}, "hi": {"chunks": 370.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.021991, "cps": 2.890625, "mean_p": 0.465342, "boundary_entropy": 0.688226, "bpb": 7.150344}, "ta": {"chunks": 14378.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.769083, "cps": 112.328125, "mean_p": 0.489901, "boundary_entropy": 0.690513, "bpb": 6.708036}, "th": {"chunks": 15782.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.981163, "cps": 123.296875, "mean_p": 0.493983, "boundary_entropy": 0.69066, "bpb": 6.554778}, "zh": {"chunks": 10987.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.860312, "cps": 85.835938, "mean_p": 0.515843, "boundary_entropy": 0.685925, "bpb": 7.796886}, "my": {"chunks": 52122.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.750792, "cps": 407.203125, "mean_p": 0.533897, "boundary_entropy": 0.688773, "bpb": 6.828077}}, "bpb": {"en": 6.88654, "de": 6.97252, "vi": 7.19769, "tr": 7.18483, "fi": 7.07868, "ru": 6.5298, "el": 7.20459, "hi": 7.15034, "ta": 6.70804, "th": 6.55478, "zh": 7.79689, "my": 6.82808}, "gini_cpc": 0.596777, "gini_cps": 0.578972, "bpic": 3.17739, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.609706516643939, "vi": 2.809564368897478, "tr": 1.0732374727255152, "fi": 1.4997551715086919, "ru": 0.254702547823219, "el": 1.4531516659662331, "hi": 0.14949642459138188, "ta": 5.228259853904788, "th": 6.669989168868511, "zh": 12.646485102544023, "my": 18.70001051397245}, "t": 11.7} |
| {"event": "probe", "step": 32, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 1690.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.09711, "cps": 13.203125, "mean_p": 0.483923, "boundary_entropy": 0.68679, "bpb": 6.388108}, "de": {"chunks": 4653.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.226611, "cps": 36.351562, "mean_p": 0.490999, "boundary_entropy": 0.688091, "bpb": 6.470697}, "vi": {"chunks": 5530.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.299973, "cps": 43.203125, "mean_p": 0.488662, "boundary_entropy": 0.688302, "bpb": 6.933085}, "tr": {"chunks": 4615.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.25493, "cps": 36.054688, "mean_p": 0.491425, "boundary_entropy": 0.687794, "bpb": 6.739156}, "fi": {"chunks": 3409.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.181224, "cps": 26.632812, "mean_p": 0.487801, "boundary_entropy": 0.687577, "bpb": 6.544353}, "ru": {"chunks": 3759.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.198364, "cps": 29.367188, "mean_p": 0.485706, "boundary_entropy": 0.689702, "bpb": 5.838464}, "el": {"chunks": 222.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.010817, "cps": 1.734375, "mean_p": 0.446247, "boundary_entropy": 0.684196, "bpb": 6.772799}, "hi": {"chunks": 480.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.028529, "cps": 3.75, "mean_p": 0.477553, "boundary_entropy": 0.689776, "bpb": 6.161415}, "ta": {"chunks": 3281.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.175501, "cps": 25.632812, "mean_p": 0.477943, "boundary_entropy": 0.689842, "bpb": 5.971626}, "th": {"chunks": 4063.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.252596, "cps": 31.742188, "mean_p": 0.487607, "boundary_entropy": 0.690588, "bpb": 5.602296}, "zh": {"chunks": 8908.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.508297, "cps": 69.59375, "mean_p": 0.506229, "boundary_entropy": 0.686322, "bpb": 7.385994}, "my": {"chunks": 53217.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 2.808581, "cps": 415.757812, "mean_p": 0.541304, "boundary_entropy": 0.687619, "bpb": 5.835393}}, "bpb": {"en": 6.38811, "de": 6.4707, "vi": 6.93308, "tr": 6.73916, "fi": 6.54435, "ru": 5.83846, "el": 6.7728, "hi": 6.16142, "ta": 5.97163, "th": 5.6023, "zh": 7.38599, "my": 5.83539}, "gini_cpc": 0.6365, "gini_cps": 0.618567, "bpic": 4.090304, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 2.333555105549531, "vi": 3.089010645110038, "tr": 2.6251768725115685, "fi": 1.8661757142601003, "ru": 2.042680894911867, "el": 0.11139085690264668, "hi": 0.2937809156211259, "ta": 1.8072497630129247, "th": 2.6011366759062895, "zh": 15.531885536121125, "my": 28.921740690497693}, "t": 14.5} |
| {"event": "train", "step": 50, "phase": 1, "objective": "global", "lm_loss": 4.23245, "bpb": 6.10613, "chunk_loss": 0.03993, "F": 0.29965, "G": 0.4925, "chunks": 306.84, "bytes_per_chunk": 3.337, "boundary_entropy": 0.69175, "router_noise": 0.0, "grad_norm": 1.5857, "lr": 5.1e-05, "router_trainable": true, "sec_per_step": 0.252, "alpha": 0.03, "t": 16.3} |
| {"event": "probe", "step": 64, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2395.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.13762, "cps": 18.710938, "mean_p": 0.489337, "boundary_entropy": 0.687283, "bpb": 5.811397}, "de": {"chunks": 3929.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.191351, "cps": 30.695312, "mean_p": 0.490824, "boundary_entropy": 0.68819, "bpb": 5.928303}, "vi": {"chunks": 3594.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.194955, "cps": 28.078125, "mean_p": 0.483333, "boundary_entropy": 0.688037, "bpb": 6.283128}, "tr": {"chunks": 3833.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.211733, "cps": 29.945312, "mean_p": 0.491023, "boundary_entropy": 0.687911, "bpb": 6.223208}, "fi": {"chunks": 3001.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.159534, "cps": 23.445312, "mean_p": 0.49022, "boundary_entropy": 0.687912, "bpb": 5.977091}, "ru": {"chunks": 7034.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.371187, "cps": 54.953125, "mean_p": 0.493895, "boundary_entropy": 0.690246, "bpb": 5.384123}, "el": {"chunks": 650.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.031672, "cps": 5.078125, "mean_p": 0.479769, "boundary_entropy": 0.689615, "bpb": 5.746216}, "hi": {"chunks": 9688.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.57581, "cps": 75.6875, "mean_p": 0.492239, "boundary_entropy": 0.690666, "bpb": 5.166845}, "ta": {"chunks": 3250.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.173843, "cps": 25.390625, "mean_p": 0.479049, "boundary_entropy": 0.689845, "bpb": 5.136134}, "th": {"chunks": 12167.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.756419, "cps": 95.054688, "mean_p": 0.491059, "boundary_entropy": 0.690545, "bpb": 4.876215}, "zh": {"chunks": 3535.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.598544, "cps": 27.617188, "mean_p": 0.488607, "boundary_entropy": 0.686325, "bpb": 6.995181}, "my": {"chunks": 37811.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 1.995514, "cps": 295.398438, "mean_p": 0.509259, "boundary_entropy": 0.690907, "bpb": 4.932285}}, "bpb": {"en": 5.8114, "de": 5.9283, "vi": 6.28313, "tr": 6.22321, "fi": 5.97709, "ru": 5.38412, "el": 5.74622, "hi": 5.16685, "ta": 5.13613, "th": 4.87622, "zh": 6.99518, "my": 4.93229}, "gini_cpc": 0.506508, "gini_cps": 0.517006, "bpic": 4.222617, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.390427101055412, "vi": 1.4166205360707058, "tr": 1.5385331927475543, "fi": 1.1592382816033215, "ru": 2.6971913121553808, "el": 0.23013947282731925, "hi": 4.184057648580655, "ta": 1.2632127308999928, "th": 5.496434352211599, "zh": 4.3492520609945515, "my": 14.500179879181479}, "t": 18.2} |
| {"event": "train", "step": 100, "phase": 1, "objective": "global", "lm_loss": 3.26123, "bpb": 4.70497, "chunk_loss": 0.037975, "F": 0.28137, "G": 0.47408, "chunks": 288.12, "bytes_per_chunk": 3.554, "boundary_entropy": 0.68652, "router_noise": 0.0, "grad_norm": 1.9186, "lr": 0.000101, "router_trainable": true, "sec_per_step": 0.085, "alpha": 0.03, "t": 20.5} |
| {"event": "probe", "step": 128, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 5600.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.321784, "cps": 43.75, "mean_p": 0.446424, "boundary_entropy": 0.668261, "bpb": 4.458585}, "de": {"chunks": 6011.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.292748, "cps": 46.960938, "mean_p": 0.443456, "boundary_entropy": 0.668893, "bpb": 4.672317}, "vi": {"chunks": 6964.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.37776, "cps": 54.40625, "mean_p": 0.47308, "boundary_entropy": 0.683195, "bpb": 4.823637}, "tr": {"chunks": 8769.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.484395, "cps": 68.507812, "mean_p": 0.467627, "boundary_entropy": 0.67447, "bpb": 4.916523}, "fi": {"chunks": 7600.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.404019, "cps": 59.375, "mean_p": 0.4669, "boundary_entropy": 0.67582, "bpb": 4.680942}, "ru": {"chunks": 10753.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.567441, "cps": 84.007812, "mean_p": 0.494051, "boundary_entropy": 0.690033, "bpb": 3.603232}, "el": {"chunks": 18425.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.897773, "cps": 143.945312, "mean_p": 0.486402, "boundary_entropy": 0.68757, "bpb": 4.331297}, "hi": {"chunks": 17630.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 1.047845, "cps": 137.734375, "mean_p": 0.410535, "boundary_entropy": 0.62536, "bpb": 3.361787}, "ta": {"chunks": 3760.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.201123, "cps": 29.375, "mean_p": 0.401229, "boundary_entropy": 0.638007, "bpb": 3.131866}, "th": {"chunks": 4736.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.294436, "cps": 37.0, "mean_p": 0.436863, "boundary_entropy": 0.669202, "bpb": 3.0333}, "zh": {"chunks": 3551.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.601253, "cps": 27.742188, "mean_p": 0.467835, "boundary_entropy": 0.680273, "bpb": 6.134738}, "my": {"chunks": 28685.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 1.51388, "cps": 224.101562, "mean_p": 0.497868, "boundary_entropy": 0.691031, "bpb": 3.563022}}, "bpb": {"en": 4.45858, "de": 4.67232, "vi": 4.82364, "tr": 4.91652, "fi": 4.68094, "ru": 3.60323, "el": 4.3313, "hi": 3.36179, "ta": 3.13187, "th": 3.0333, "zh": 6.13474, "my": 3.56302}, "gini_cpc": 0.327248, "gini_cps": 0.362899, "bpic": 3.133315, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9097674910075071, "vi": 1.1739557131233291, "tr": 1.5053435006826021, "fi": 1.2555609559756071, "ru": 1.7634230964945348, "el": 2.7899906289807257, "hi": 3.2563669072383785, "ta": 0.6250265540824514, "th": 0.9150118566543807, "zh": 1.8685009493977072, "my": 4.7046527243734735}, "t": 23.1} |
| {"event": "train", "step": 150, "phase": 1, "objective": "global", "lm_loss": 2.81646, "bpb": 4.06329, "chunk_loss": 0.038109, "F": 0.30618, "G": 0.43661, "chunks": 313.53, "bytes_per_chunk": 3.266, "boundary_entropy": 0.66054, "router_noise": 0.0, "grad_norm": 1.6055, "lr": 0.000151, "router_trainable": true, "sec_per_step": 0.08, "alpha": 0.03, "t": 24.5} |
| {"event": "probe", "step": 200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4992.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.286847, "cps": 39.0, "mean_p": 0.417998, "boundary_entropy": 0.646128, "bpb": 3.823833}, "de": {"chunks": 6198.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.301856, "cps": 48.421875, "mean_p": 0.415807, "boundary_entropy": 0.644081, "bpb": 4.061151}, "vi": {"chunks": 6325.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.343097, "cps": 49.414062, "mean_p": 0.448778, "boundary_entropy": 0.664843, "bpb": 3.621859}, "tr": {"chunks": 7928.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.437938, "cps": 61.9375, "mean_p": 0.450591, "boundary_entropy": 0.656265, "bpb": 4.119477}, "fi": {"chunks": 7087.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.376748, "cps": 55.367188, "mean_p": 0.45024, "boundary_entropy": 0.658795, "bpb": 4.108538}, "ru": {"chunks": 6071.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.320369, "cps": 47.429688, "mean_p": 0.461715, "boundary_entropy": 0.684444, "bpb": 2.808159}, "el": {"chunks": 12516.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.609852, "cps": 97.78125, "mean_p": 0.476318, "boundary_entropy": 0.686112, "bpb": 3.002414}, "hi": {"chunks": 3287.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.195364, "cps": 25.679688, "mean_p": 0.374681, "boundary_entropy": 0.594259, "bpb": 2.387284}, "ta": {"chunks": 2340.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.125167, "cps": 18.28125, "mean_p": 0.363198, "boundary_entropy": 0.610628, "bpb": 2.236199}, "th": {"chunks": 7119.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.442586, "cps": 55.617188, "mean_p": 0.412641, "boundary_entropy": 0.648287, "bpb": 2.283523}, "zh": {"chunks": 3035.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.513884, "cps": 23.710938, "mean_p": 0.387119, "boundary_entropy": 0.600063, "bpb": 4.837088}, "my": {"chunks": 19166.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 1.011505, "cps": 149.734375, "mean_p": 0.483639, "boundary_entropy": 0.689191, "bpb": 2.402121}}, "bpb": {"en": 3.82383, "de": 4.06115, "vi": 3.62186, "tr": 4.11948, "fi": 4.10854, "ru": 2.80816, "el": 3.00241, "hi": 2.38728, "ta": 2.2362, "th": 2.28352, "zh": 4.83709, "my": 2.40212}, "gini_cpc": 0.27121, "gini_cps": 0.306663, "bpic": 4.459251, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0523221413746728, "vi": 1.1960984605474536, "tr": 1.526731385363242, "fi": 1.3134093173286499, "ru": 1.1168646932717678, "el": 2.1260538129828075, "hi": 0.6810738012915761, "ta": 0.43635497459213696, "th": 1.5429344333588388, "zh": 1.791491683001641, "my": 3.526286960974164}, "t": 28.2} |
| {"event": "eval", "step": 200, "phase": 1, "bpb": {"en": 3.82806, "de": 4.06059, "vi": 3.62506, "tr": 4.1282, "fi": 4.11327, "ru": 2.79326, "el": 2.9957, "hi": 2.38251, "ta": 2.2522, "th": 2.3169, "zh": 4.84637, "my": 2.41923}, "bytes_per_chunk": {"en": 3.377, "de": 3.297, "vi": 3.871, "tr": 2.414, "fi": 2.607, "ru": 5.251, "el": 3.006, "hi": 13.233, "ta": 20.851, "th": 6.646, "zh": 4.947, "my": 3.619}, "bpb_mean": 3.31345, "t": 29.7} |
| {"event": "train", "step": 200, "phase": 1, "objective": "global", "lm_loss": 2.44273, "bpb": 3.52412, "chunk_loss": 0.035694, "F": 0.2641, "G": 0.42703, "chunks": 270.44, "bytes_per_chunk": 3.786, "boundary_entropy": 0.65283, "router_noise": 0.0, "grad_norm": 1.4535, "lr": 0.000201, "router_trainable": true, "sec_per_step": 0.105, "alpha": 0.03, "t": 29.8} |
| {"event": "train", "step": 250, "phase": 1, "objective": "global", "lm_loss": 2.52311, "bpb": 3.64009, "chunk_loss": 0.036783, "F": 0.28833, "G": 0.43283, "chunks": 295.25, "bytes_per_chunk": 3.468, "boundary_entropy": 0.65358, "router_noise": 0.0, "grad_norm": 1.5963, "lr": 0.000251, "router_trainable": true, "sec_per_step": 0.053, "alpha": 0.03, "t": 32.4} |
| {"event": "probe", "step": 256, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4090.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.235017, "cps": 31.953125, "mean_p": 0.426009, "boundary_entropy": 0.651686, "bpb": 3.589812}, "de": {"chunks": 4263.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.207617, "cps": 33.304688, "mean_p": 0.416211, "boundary_entropy": 0.649304, "bpb": 3.892463}, "vi": {"chunks": 5984.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.3246, "cps": 46.75, "mean_p": 0.445892, "boundary_entropy": 0.666097, "bpb": 3.187772}, "tr": {"chunks": 6756.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.373198, "cps": 52.78125, "mean_p": 0.437895, "boundary_entropy": 0.655506, "bpb": 3.810644}, "fi": {"chunks": 5630.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.299293, "cps": 43.984375, "mean_p": 0.431569, "boundary_entropy": 0.654213, "bpb": 3.907179}, "ru": {"chunks": 8832.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.466069, "cps": 69.0, "mean_p": 0.487794, "boundary_entropy": 0.68857, "bpb": 2.650545}, "el": {"chunks": 7427.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.361887, "cps": 58.023438, "mean_p": 0.474166, "boundary_entropy": 0.686781, "bpb": 2.749727}, "hi": {"chunks": 13414.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.797266, "cps": 104.796875, "mean_p": 0.386534, "boundary_entropy": 0.591752, "bpb": 2.116339}, "ta": {"chunks": 12976.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.694089, "cps": 101.375, "mean_p": 0.378265, "boundary_entropy": 0.607606, "bpb": 1.953771}, "th": {"chunks": 4504.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.280012, "cps": 35.1875, "mean_p": 0.401931, "boundary_entropy": 0.63309, "bpb": 2.137137}, "zh": {"chunks": 3463.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.586353, "cps": 27.054688, "mean_p": 0.390936, "boundary_entropy": 0.607425, "bpb": 4.444195}, "my": {"chunks": 21703.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 1.145398, "cps": 169.554688, "mean_p": 0.503034, "boundary_entropy": 0.690598, "bpb": 2.175686}}, "bpb": {"en": 3.58981, "de": 3.89246, "vi": 3.18777, "tr": 3.81064, "fi": 3.90718, "ru": 2.65054, "el": 2.74973, "hi": 2.11634, "ta": 1.95377, "th": 2.13714, "zh": 4.4442, "my": 2.17569}, "gini_cpc": 0.29091, "gini_cps": 0.314134, "bpic": 3.874932, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8834129019098245, "vi": 1.3811767374035384, "tr": 1.5879612545813977, "fi": 1.2734952329172222, "ru": 1.983127597396314, "el": 1.539832178269309, "hi": 3.3923764627239272, "ta": 2.953358578807534, "th": 1.1914563295694558, "zh": 2.4949385937967024, "my": 4.873682199590902}, "t": 33.8} |
| {"event": "train", "step": 300, "phase": 1, "objective": "global", "lm_loss": 2.28916, "bpb": 3.30256, "chunk_loss": 0.035906, "F": 0.2757, "G": 0.42614, "chunks": 282.31, "bytes_per_chunk": 3.627, "boundary_entropy": 0.65198, "router_noise": 0.0, "grad_norm": 1.072, "lr": 0.0003, "router_trainable": true, "sec_per_step": 0.076, "alpha": 0.03, "t": 36.2} |
| {"event": "train", "step": 350, "phase": 1, "objective": "global", "lm_loss": 2.02703, "bpb": 2.92438, "chunk_loss": 0.033026, "F": 0.21597, "G": 0.41653, "chunks": 221.16, "bytes_per_chunk": 4.63, "boundary_entropy": 0.64604, "router_noise": 0.0, "grad_norm": 0.9989, "lr": 0.0002999, "router_trainable": true, "sec_per_step": 0.05, "alpha": 0.03, "t": 38.7} |
| {"event": "probe", "step": 400, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4353.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.250129, "cps": 34.007812, "mean_p": 0.433003, "boundary_entropy": 0.658932, "bpb": 3.236493}, "de": {"chunks": 5149.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.250767, "cps": 40.226562, "mean_p": 0.430558, "boundary_entropy": 0.660219, "bpb": 3.50534}, "vi": {"chunks": 6173.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.334852, "cps": 48.226562, "mean_p": 0.443878, "boundary_entropy": 0.668479, "bpb": 2.767514}, "tr": {"chunks": 5751.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.317682, "cps": 44.929688, "mean_p": 0.441839, "boundary_entropy": 0.660897, "bpb": 3.433246}, "fi": {"chunks": 4470.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.237627, "cps": 34.921875, "mean_p": 0.423772, "boundary_entropy": 0.656715, "bpb": 3.563935}, "ru": {"chunks": 5860.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.309235, "cps": 45.78125, "mean_p": 0.401156, "boundary_entropy": 0.636974, "bpb": 2.359477}, "el": {"chunks": 10092.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.491741, "cps": 78.84375, "mean_p": 0.46391, "boundary_entropy": 0.677904, "bpb": 2.535287}, "hi": {"chunks": 5908.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.351144, "cps": 46.15625, "mean_p": 0.371029, "boundary_entropy": 0.617778, "bpb": 1.86406}, "ta": {"chunks": 6811.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.364322, "cps": 53.210938, "mean_p": 0.354591, "boundary_entropy": 0.610086, "bpb": 1.598073}, "th": {"chunks": 3393.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.210942, "cps": 26.507812, "mean_p": 0.389431, "boundary_entropy": 0.638856, "bpb": 1.923968}, "zh": {"chunks": 1880.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.31832, "cps": 14.6875, "mean_p": 0.398958, "boundary_entropy": 0.636381, "bpb": 4.158936}, "my": {"chunks": 4045.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.213479, "cps": 31.601562, "mean_p": 0.400749, "boundary_entropy": 0.643939, "bpb": 1.950481}}, "bpb": {"en": 3.23649, "de": 3.50534, "vi": 2.76751, "tr": 3.43325, "fi": 3.56394, "ru": 2.35948, "el": 2.53529, "hi": 1.86406, "ta": 1.59807, "th": 1.92397, "zh": 4.15894, "my": 1.95048}, "gini_cpc": 0.13414, "gini_cps": 0.193756, "bpic": 6.007373, "worst_lang": "el", "cpc_parity_vs_en": {"en": 1.0, "de": 1.002549760797557, "vi": 1.3387164132290386, "tr": 1.2700717880793393, "fi": 0.950016378601007, "ru": 1.2362999587219443, "el": 1.9659472002180876, "hi": 1.4038505186257775, "ta": 1.4565347947362806, "th": 0.8433313545565887, "zh": 1.2726232687943881, "my": 0.8534746042962518}, "t": 42.1} |
| {"event": "eval", "step": 400, "phase": 1, "bpb": {"en": 3.24643, "de": 3.50177, "vi": 2.77384, "tr": 3.44035, "fi": 3.55437, "ru": 2.34274, "el": 2.52984, "hi": 1.86161, "ta": 1.61361, "th": 1.95336, "zh": 4.16561, "my": 1.96601}, "bytes_per_chunk": {"en": 4.028, "de": 3.773, "vi": 4.084, "tr": 3.307, "fi": 4.28, "ru": 6.675, "el": 3.779, "hi": 6.604, "ta": 7.581, "th": 13.157, "zh": 7.471, "my": 10.533}, "bpb_mean": 2.7458, "t": 43.4} |
| {"event": "train", "step": 400, "phase": 1, "objective": "global", "lm_loss": 2.03241, "bpb": 2.93214, "chunk_loss": 0.032607, "F": 0.20947, "G": 0.41771, "chunks": 214.5, "bytes_per_chunk": 4.774, "boundary_entropy": 0.65311, "router_noise": 0.0, "grad_norm": 1.0281, "lr": 0.0002998, "router_trainable": true, "sec_per_step": 0.094, "alpha": 0.03, "t": 43.4} |
| {"event": "train", "step": 450, "phase": 1, "objective": "global", "lm_loss": 2.11296, "bpb": 3.04836, "chunk_loss": 0.034929, "F": 0.25516, "G": 0.41984, "chunks": 261.28, "bytes_per_chunk": 3.919, "boundary_entropy": 0.65407, "router_noise": 0.0, "grad_norm": 0.7114, "lr": 0.0002995, "router_trainable": true, "sec_per_step": 0.047, "alpha": 0.03, "t": 45.8} |
| {"event": "train", "step": 500, "phase": 1, "objective": "global", "lm_loss": 1.95253, "bpb": 2.8169, "chunk_loss": 0.034125, "F": 0.23663, "G": 0.42072, "chunks": 242.31, "bytes_per_chunk": 4.226, "boundary_entropy": 0.65472, "router_noise": 0.0, "grad_norm": 0.8573, "lr": 0.0002992, "router_trainable": true, "sec_per_step": 0.045, "alpha": 0.03, "t": 48.0} |
| {"event": "train", "step": 550, "phase": 1, "objective": "global", "lm_loss": 1.91468, "bpb": 2.7623, "chunk_loss": 0.035465, "F": 0.26093, "G": 0.42951, "chunks": 267.19, "bytes_per_chunk": 3.833, "boundary_entropy": 0.65919, "router_noise": 0.0, "grad_norm": 0.9139, "lr": 0.0002987, "router_trainable": true, "sec_per_step": 0.045, "alpha": 0.03, "t": 50.2} |
| {"event": "probe", "step": 600, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4807.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.276217, "cps": 37.554688, "mean_p": 0.444719, "boundary_entropy": 0.664044, "bpb": 2.956843}, "de": {"chunks": 5013.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.244144, "cps": 39.164062, "mean_p": 0.440052, "boundary_entropy": 0.666729, "bpb": 3.263209}, "vi": {"chunks": 7461.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.404719, "cps": 58.289062, "mean_p": 0.453099, "boundary_entropy": 0.671118, "bpb": 2.572018}, "tr": {"chunks": 6262.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.34591, "cps": 48.921875, "mean_p": 0.450789, "boundary_entropy": 0.663415, "bpb": 3.153021}, "fi": {"chunks": 4807.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.255542, "cps": 37.554688, "mean_p": 0.437145, "boundary_entropy": 0.662507, "bpb": 3.37761}, "ru": {"chunks": 5783.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.305172, "cps": 45.179688, "mean_p": 0.417871, "boundary_entropy": 0.654647, "bpb": 2.144487}, "el": {"chunks": 7991.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.389368, "cps": 62.429688, "mean_p": 0.447491, "boundary_entropy": 0.664589, "bpb": 2.242023}, "hi": {"chunks": 4736.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.281486, "cps": 37.0, "mean_p": 0.400765, "boundary_entropy": 0.651923, "bpb": 1.643579}, "ta": {"chunks": 5319.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.284515, "cps": 41.554688, "mean_p": 0.374639, "boundary_entropy": 0.63262, "bpb": 1.413208}, "th": {"chunks": 4214.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.261983, "cps": 32.921875, "mean_p": 0.406652, "boundary_entropy": 0.659629, "bpb": 1.747961}, "zh": {"chunks": 4921.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.83322, "cps": 38.445312, "mean_p": 0.433121, "boundary_entropy": 0.65411, "bpb": 3.939034}, "my": {"chunks": 16581.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.875079, "cps": 129.539062, "mean_p": 0.426766, "boundary_entropy": 0.660711, "bpb": 1.72817}}, "bpb": {"en": 2.95684, "de": 3.26321, "vi": 2.57202, "tr": 3.15302, "fi": 3.37761, "ru": 2.14449, "el": 2.24202, "hi": 1.64358, "ta": 1.41321, "th": 1.74796, "zh": 3.93903, "my": 1.72817}, "gini_cpc": 0.243802, "gini_cps": 0.207627, "bpic": 4.926902, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8838840470425102, "vi": 1.4652235695740023, "tr": 1.252311907128845, "fi": 0.9251501780872894, "ru": 1.1048262291194735, "el": 1.4096467148980314, "hi": 1.0190761058894116, "ta": 1.0300410167863316, "th": 0.9484697059043584, "zh": 3.016545778460707, "my": 3.168088764630112}, "t": 53.4} |
| {"event": "eval", "step": 600, "phase": 1, "bpb": {"en": 2.9667, "de": 3.26084, "vi": 2.57629, "tr": 3.16291, "fi": 3.36617, "ru": 2.12642, "el": 2.23366, "hi": 1.63961, "ta": 1.42607, "th": 1.77879, "zh": 3.95009, "my": 1.74011}, "bytes_per_chunk": {"en": 3.438, "de": 4.034, "vi": 3.312, "tr": 2.97, "fi": 4.208, "ru": 6.291, "el": 4.652, "hi": 7.555, "ta": 10.357, "th": 9.712, "zh": 2.667, "my": 3.123}, "bpb_mean": 2.51897, "t": 54.8} |
| {"event": "train", "step": 600, "phase": 1, "objective": "global", "lm_loss": 1.93968, "bpb": 2.79836, "chunk_loss": 0.034892, "F": 0.24808, "G": 0.43146, "chunks": 254.03, "bytes_per_chunk": 4.031, "boundary_entropy": 0.66317, "router_noise": 0.0, "grad_norm": 0.8607, "lr": 0.0002982, "router_trainable": true, "sec_per_step": 0.093, "alpha": 0.03, "t": 54.9} |
| {"event": "train", "step": 650, "phase": 1, "objective": "global", "lm_loss": 1.72901, "bpb": 2.49443, "chunk_loss": 0.033664, "F": 0.22656, "G": 0.42737, "chunks": 232.0, "bytes_per_chunk": 4.414, "boundary_entropy": 0.66265, "router_noise": 0.0, "grad_norm": 0.9101, "lr": 0.0002975, "router_trainable": true, "sec_per_step": 0.045, "alpha": 0.03, "t": 57.1} |
| {"event": "train", "step": 700, "phase": 1, "objective": "global", "lm_loss": 1.75445, "bpb": 2.53113, "chunk_loss": 0.034982, "F": 0.2485, "G": 0.43495, "chunks": 254.47, "bytes_per_chunk": 4.024, "boundary_entropy": 0.6657, "router_noise": 0.0, "grad_norm": 0.7238, "lr": 0.0002967, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 59.3} |
| {"event": "train", "step": 750, "phase": 1, "objective": "global", "lm_loss": 1.68464, "bpb": 2.43041, "chunk_loss": 0.036669, "F": 0.27875, "G": 0.43566, "chunks": 285.44, "bytes_per_chunk": 3.587, "boundary_entropy": 0.6652, "router_noise": 0.0, "grad_norm": 0.7197, "lr": 0.0002959, "router_trainable": true, "sec_per_step": 0.046, "alpha": 0.03, "t": 61.6} |
| {"event": "probe", "step": 800, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 5304.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.304775, "cps": 41.4375, "mean_p": 0.443886, "boundary_entropy": 0.66375, "bpb": 2.740212}, "de": {"chunks": 6401.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.311742, "cps": 50.007812, "mean_p": 0.44666, "boundary_entropy": 0.667255, "bpb": 3.053767}, "vi": {"chunks": 6718.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.364416, "cps": 52.484375, "mean_p": 0.45557, "boundary_entropy": 0.672734, "bpb": 2.449346}, "tr": {"chunks": 6778.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.374413, "cps": 52.953125, "mean_p": 0.45575, "boundary_entropy": 0.665219, "bpb": 2.985894}, "fi": {"chunks": 6238.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.331614, "cps": 48.734375, "mean_p": 0.454564, "boundary_entropy": 0.666811, "bpb": 3.234503}, "ru": {"chunks": 6884.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.363272, "cps": 53.78125, "mean_p": 0.423392, "boundary_entropy": 0.661717, "bpb": 2.012024}, "el": {"chunks": 7868.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.383375, "cps": 61.46875, "mean_p": 0.447606, "boundary_entropy": 0.664725, "bpb": 2.143921}, "hi": {"chunks": 8331.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.495156, "cps": 65.085938, "mean_p": 0.408624, "boundary_entropy": 0.656405, "bpb": 1.573643}, "ta": {"chunks": 7899.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.422519, "cps": 61.710938, "mean_p": 0.395328, "boundary_entropy": 0.648082, "bpb": 1.320935}, "th": {"chunks": 12067.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.750202, "cps": 94.273438, "mean_p": 0.43384, "boundary_entropy": 0.66874, "bpb": 1.627401}, "zh": {"chunks": 2970.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.502878, "cps": 23.203125, "mean_p": 0.423656, "boundary_entropy": 0.654697, "bpb": 3.850274}, "my": {"chunks": 7685.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.405584, "cps": 60.039062, "mean_p": 0.417609, "boundary_entropy": 0.665446, "bpb": 1.517367}}, "bpb": {"en": 2.74021, "de": 3.05377, "vi": 2.44935, "tr": 2.98589, "fi": 3.2345, "ru": 2.01202, "el": 2.14392, "hi": 1.57364, "ta": 1.32093, "th": 1.6274, "zh": 3.85027, "my": 1.51737}, "gini_cpc": 0.136286, "gini_cps": 0.146106, "bpic": 4.507487, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0228595982672632, "vi": 1.1956868758644472, "tr": 1.2284899780366896, "fi": 1.0880631242286032, "ru": 1.1919341205124225, "el": 1.2578941891262931, "hi": 1.62466066710145, "ta": 1.3863320037418962, "th": 2.461494401208512, "zh": 1.6499987358511092, "my": 1.3307641740624734}, "t": 64.8} |
| {"event": "eval", "step": 800, "phase": 1, "bpb": {"en": 2.75049, "de": 3.05397, "vi": 2.45567, "tr": 2.99739, "fi": 3.23211, "ru": 1.99346, "el": 2.13104, "hi": 1.56921, "ta": 1.33155, "th": 1.65418, "zh": 3.86263, "my": 1.52988}, "bytes_per_chunk": {"en": 3.131, "de": 3.271, "vi": 3.548, "tr": 2.895, "fi": 3.061, "ru": 5.114, "el": 4.9, "hi": 5.961, "ta": 6.774, "th": 3.894, "zh": 4.312, "my": 6.797}, "bpb_mean": 2.38013, "t": 66.2} |
| {"event": "train", "step": 800, "phase": 1, "objective": "global", "lm_loss": 1.82804, "bpb": 2.63731, "chunk_loss": 0.036341, "F": 0.2735, "G": 0.43758, "chunks": 280.06, "bytes_per_chunk": 3.656, "boundary_entropy": 0.66674, "router_noise": 0.0, "grad_norm": 0.8818, "lr": 0.0002949, "router_trainable": true, "sec_per_step": 0.093, "alpha": 0.03, "t": 66.3} |
| {"event": "train", "step": 850, "phase": 1, "objective": "global", "lm_loss": 1.86505, "bpb": 2.6907, "chunk_loss": 0.036878, "F": 0.2814, "G": 0.43982, "chunks": 288.16, "bytes_per_chunk": 3.554, "boundary_entropy": 0.66723, "router_noise": 0.0, "grad_norm": 0.8759, "lr": 0.0002938, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 68.5} |
| {"event": "train", "step": 900, "phase": 1, "objective": "global", "lm_loss": 1.68423, "bpb": 2.42983, "chunk_loss": 0.033406, "F": 0.22339, "G": 0.43295, "chunks": 228.75, "bytes_per_chunk": 4.477, "boundary_entropy": 0.66622, "router_noise": 0.0, "grad_norm": 0.92, "lr": 0.0002927, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 70.7} |
| {"event": "train", "step": 950, "phase": 1, "objective": "global", "lm_loss": 1.82189, "bpb": 2.62844, "chunk_loss": 0.033664, "F": 0.22812, "G": 0.43176, "chunks": 233.59, "bytes_per_chunk": 4.384, "boundary_entropy": 0.66539, "router_noise": 0.0, "grad_norm": 0.7946, "lr": 0.0002914, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 72.9} |
| {"event": "probe", "step": 1000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 5675.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.326093, "cps": 44.335938, "mean_p": 0.449647, "boundary_entropy": 0.664408, "bpb": 2.602773}, "de": {"chunks": 5406.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.263283, "cps": 42.234375, "mean_p": 0.44671, "boundary_entropy": 0.668775, "bpb": 2.940353}, "vi": {"chunks": 7672.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.416165, "cps": 59.9375, "mean_p": 0.458604, "boundary_entropy": 0.673201, "bpb": 2.314021}, "tr": {"chunks": 6958.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.384356, "cps": 54.359375, "mean_p": 0.456333, "boundary_entropy": 0.663958, "bpb": 2.883147}, "fi": {"chunks": 5895.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.31338, "cps": 46.054688, "mean_p": 0.451279, "boundary_entropy": 0.667084, "bpb": 3.129919}, "ru": {"chunks": 8288.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.437361, "cps": 64.75, "mean_p": 0.427377, "boundary_entropy": 0.663868, "bpb": 1.932576}, "el": {"chunks": 10684.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.520587, "cps": 83.46875, "mean_p": 0.467414, "boundary_entropy": 0.672442, "bpb": 2.067584}, "hi": {"chunks": 8982.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.533848, "cps": 70.171875, "mean_p": 0.411605, "boundary_entropy": 0.656568, "bpb": 1.531101}, "ta": {"chunks": 5327.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.284942, "cps": 41.617188, "mean_p": 0.392846, "boundary_entropy": 0.649656, "bpb": 1.286325}, "th": {"chunks": 8608.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.535157, "cps": 67.25, "mean_p": 0.435162, "boundary_entropy": 0.671587, "bpb": 1.574903}, "zh": {"chunks": 3330.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.563833, "cps": 26.015625, "mean_p": 0.433204, "boundary_entropy": 0.661103, "bpb": 3.681065}, "my": {"chunks": 7989.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.421628, "cps": 62.414062, "mean_p": 0.447778, "boundary_entropy": 0.681815, "bpb": 1.542278}}, "bpb": {"en": 2.60277, "de": 2.94035, "vi": 2.31402, "tr": 2.88315, "fi": 3.12992, "ru": 1.93258, "el": 2.06758, "hi": 1.5311, "ta": 1.28632, "th": 1.5749, "zh": 3.68106, "my": 1.54228}, "gini_cpc": 0.137286, "gini_cps": 0.153536, "bpic": 4.524972, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8073872530541252, "vi": 1.2762145937207643, "tr": 1.178669721790216, "fi": 0.9610150308060917, "ru": 1.3412161751885905, "el": 1.596435176092322, "hi": 1.6371038587166413, "ta": 0.8738069238046876, "th": 1.6411166339613752, "zh": 1.729055944003783, "my": 1.2929665795281127}, "t": 76.1} |
| {"event": "eval", "step": 1000, "phase": 1, "bpb": {"en": 2.61017, "de": 2.94174, "vi": 2.31745, "tr": 2.89298, "fi": 3.13079, "ru": 1.91553, "el": 2.05664, "hi": 1.52676, "ta": 1.29741, "th": 1.599, "zh": 3.69119, "my": 1.55212}, "bytes_per_chunk": {"en": 2.907, "de": 3.888, "vi": 3.221, "tr": 2.79, "fi": 3.317, "ru": 4.075, "el": 3.648, "hi": 4.723, "ta": 8.897, "th": 5.282, "zh": 4.023, "my": 7.288}, "bpb_mean": 2.29432, "t": 77.5} |
| {"event": "train", "step": 1000, "phase": 1, "objective": "global", "lm_loss": 1.74966, "bpb": 2.52423, "chunk_loss": 0.036914, "F": 0.28128, "G": 0.43967, "chunks": 288.03, "bytes_per_chunk": 3.555, "boundary_entropy": 0.6671, "router_noise": 0.0, "grad_norm": 0.7746, "lr": 0.0002901, "router_trainable": true, "sec_per_step": 0.093, "alpha": 0.03, "t": 77.6} |
| {"event": "train", "step": 1050, "phase": 1, "objective": "global", "lm_loss": 1.62265, "bpb": 2.34099, "chunk_loss": 0.033026, "F": 0.2168, "G": 0.43329, "chunks": 222.0, "bytes_per_chunk": 4.613, "boundary_entropy": 0.66771, "router_noise": 0.0, "grad_norm": 0.8216, "lr": 0.0002886, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 79.7} |
| {"event": "train", "step": 1100, "phase": 1, "objective": "global", "lm_loss": 1.82065, "bpb": 2.62665, "chunk_loss": 0.03514, "F": 0.25229, "G": 0.43922, "chunks": 258.34, "bytes_per_chunk": 3.964, "boundary_entropy": 0.66885, "router_noise": 0.0, "grad_norm": 0.7971, "lr": 0.0002871, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 81.9} |
| {"event": "train", "step": 1150, "phase": 1, "objective": "global", "lm_loss": 1.66579, "bpb": 2.40323, "chunk_loss": 0.034919, "F": 0.2485, "G": 0.43682, "chunks": 254.47, "bytes_per_chunk": 4.024, "boundary_entropy": 0.66811, "router_noise": 0.0, "grad_norm": 0.7565, "lr": 0.0002855, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 84.1} |
| {"event": "probe", "step": 1200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4462.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.256393, "cps": 34.859375, "mean_p": 0.441213, "boundary_entropy": 0.663732, "bpb": 2.507415}, "de": {"chunks": 5137.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.250183, "cps": 40.132812, "mean_p": 0.444658, "boundary_entropy": 0.66917, "bpb": 2.830941}, "vi": {"chunks": 5177.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.280825, "cps": 40.445312, "mean_p": 0.446786, "boundary_entropy": 0.672388, "bpb": 2.279344}, "tr": {"chunks": 6490.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.358504, "cps": 50.703125, "mean_p": 0.456296, "boundary_entropy": 0.666753, "bpb": 2.783138}, "fi": {"chunks": 5176.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.275158, "cps": 40.4375, "mean_p": 0.446264, "boundary_entropy": 0.667682, "bpb": 3.068132}, "ru": {"chunks": 6228.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.328654, "cps": 48.65625, "mean_p": 0.420683, "boundary_entropy": 0.664306, "bpb": 1.864272}, "el": {"chunks": 9236.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.450032, "cps": 72.15625, "mean_p": 0.460331, "boundary_entropy": 0.672707, "bpb": 1.985394}, "hi": {"chunks": 10586.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.629183, "cps": 82.703125, "mean_p": 0.415432, "boundary_entropy": 0.656687, "bpb": 1.511117}, "ta": {"chunks": 13838.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.740198, "cps": 108.109375, "mean_p": 0.409601, "boundary_entropy": 0.651852, "bpb": 1.273369}, "th": {"chunks": 9791.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.608704, "cps": 76.492188, "mean_p": 0.439924, "boundary_entropy": 0.673102, "bpb": 1.505858}, "zh": {"chunks": 3190.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.540129, "cps": 24.921875, "mean_p": 0.436417, "boundary_entropy": 0.665075, "bpb": 3.621797}, "my": {"chunks": 8136.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.429386, "cps": 63.5625, "mean_p": 0.442055, "boundary_entropy": 0.678447, "bpb": 1.414417}}, "bpb": {"en": 2.50741, "de": 2.83094, "vi": 2.27934, "tr": 2.78314, "fi": 3.06813, "ru": 1.86427, "el": 1.98539, "hi": 1.51112, "ta": 1.27337, "th": 1.50586, "zh": 3.6218, "my": 1.41442}, "gini_cpc": 0.207415, "gini_cps": 0.223241, "bpic": 4.388727, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9757795515971787, "vi": 1.0952911467563176, "tr": 1.3982624650968178, "fi": 1.0731907937650853, "ru": 1.281840366436481, "el": 1.755244550434205, "hi": 2.4539819895661767, "ta": 2.886970931256413, "th": 2.3741083723333865, "zh": 2.106647123476221, "my": 1.6747196578581103}, "t": 87.3} |
| {"event": "eval", "step": 1200, "phase": 1, "bpb": {"en": 2.51605, "de": 2.83368, "vi": 2.28487, "tr": 2.7961, "fi": 3.06946, "ru": 1.84523, "el": 1.9743, "hi": 1.50958, "ta": 1.28092, "th": 1.52913, "zh": 3.63326, "my": 1.42339}, "bytes_per_chunk": {"en": 3.819, "de": 4.164, "vi": 4.73, "tr": 3.085, "fi": 4.003, "ru": 5.451, "el": 3.973, "hi": 3.941, "ta": 3.744, "th": 4.84, "zh": 4.4, "my": 6.598}, "bpb_mean": 2.22466, "t": 88.6} |
| {"event": "train", "step": 1200, "phase": 1, "objective": "global", "lm_loss": 1.73073, "bpb": 2.49692, "chunk_loss": 0.034988, "F": 0.25104, "G": 0.43815, "chunks": 257.06, "bytes_per_chunk": 3.983, "boundary_entropy": 0.66814, "router_noise": 0.0, "grad_norm": 0.8084, "lr": 0.0002837, "router_trainable": true, "sec_per_step": 0.091, "alpha": 0.03, "t": 88.7} |
| {"event": "train", "step": 1250, "phase": 1, "objective": "global", "lm_loss": 1.84598, "bpb": 2.66318, "chunk_loss": 0.035119, "F": 0.25159, "G": 0.4395, "chunks": 257.62, "bytes_per_chunk": 3.975, "boundary_entropy": 0.66907, "router_noise": 0.0, "grad_norm": 0.8253, "lr": 0.0002819, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 90.9} |
| {"event": "train", "step": 1300, "phase": 1, "objective": "global", "lm_loss": 1.52877, "bpb": 2.20555, "chunk_loss": 0.03465, "F": 0.24405, "G": 0.43443, "chunks": 249.91, "bytes_per_chunk": 4.098, "boundary_entropy": 0.66787, "router_noise": 0.0, "grad_norm": 0.8211, "lr": 0.00028, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 93.0} |
| {"event": "train", "step": 1350, "phase": 1, "objective": "global", "lm_loss": 1.52243, "bpb": 2.19641, "chunk_loss": 0.03417, "F": 0.2366, "G": 0.43379, "chunks": 242.28, "bytes_per_chunk": 4.226, "boundary_entropy": 0.66716, "router_noise": 0.0, "grad_norm": 0.7728, "lr": 0.000278, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 95.2} |
| {"event": "probe", "step": 1400, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4432.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.254669, "cps": 34.625, "mean_p": 0.444266, "boundary_entropy": 0.665302, "bpb": 2.452102}, "de": {"chunks": 5340.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.260069, "cps": 41.71875, "mean_p": 0.449655, "boundary_entropy": 0.670895, "bpb": 2.763237}, "vi": {"chunks": 4968.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.269487, "cps": 38.8125, "mean_p": 0.444516, "boundary_entropy": 0.672212, "bpb": 2.216856}, "tr": {"chunks": 6631.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.366293, "cps": 51.804688, "mean_p": 0.458583, "boundary_entropy": 0.666582, "bpb": 2.724181}, "fi": {"chunks": 5662.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.300994, "cps": 44.234375, "mean_p": 0.450801, "boundary_entropy": 0.668543, "bpb": 3.001045}, "ru": {"chunks": 5568.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.293826, "cps": 43.5, "mean_p": 0.429033, "boundary_entropy": 0.668764, "bpb": 1.817167}, "el": {"chunks": 6209.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.302539, "cps": 48.507812, "mean_p": 0.453358, "boundary_entropy": 0.673003, "bpb": 1.916831}, "hi": {"chunks": 3345.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.198811, "cps": 26.132812, "mean_p": 0.392703, "boundary_entropy": 0.650166, "bpb": 1.472716}, "ta": {"chunks": 6000.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.320941, "cps": 46.875, "mean_p": 0.391951, "boundary_entropy": 0.647181, "bpb": 1.193313}, "th": {"chunks": 6352.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.394902, "cps": 49.625, "mean_p": 0.412957, "boundary_entropy": 0.664493, "bpb": 1.48444}, "zh": {"chunks": 3306.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.55977, "cps": 25.828125, "mean_p": 0.434418, "boundary_entropy": 0.664745, "bpb": 3.594548}, "my": {"chunks": 10794.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.569664, "cps": 84.328125, "mean_p": 0.441753, "boundary_entropy": 0.677768, "bpb": 1.351876}}, "bpb": {"en": 2.4521, "de": 2.76324, "vi": 2.21686, "tr": 2.72418, "fi": 3.00104, "ru": 1.81717, "el": 1.91683, "hi": 1.47272, "ta": 1.19331, "th": 1.48444, "zh": 3.59455, "my": 1.35188}, "gini_cpc": 0.169764, "gini_cps": 0.162351, "bpic": 5.593904, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0212056720879534, "vi": 1.0581879547517428, "tr": 1.4383111426572097, "fi": 1.1819044016990599, "ru": 1.1537570844803444, "el": 1.1879692058975482, "hi": 0.7806662741858481, "ta": 1.2602309735512982, "th": 1.5506500327120478, "zh": 2.1980308069502836, "my": 2.2368837066141114}, "t": 98.3} |
| {"event": "eval", "step": 1400, "phase": 1, "bpb": {"en": 2.45811, "de": 2.76637, "vi": 2.22113, "tr": 2.73611, "fi": 2.99688, "ru": 1.79739, "el": 1.90645, "hi": 1.47011, "ta": 1.20088, "th": 1.50335, "zh": 3.60066, "my": 1.36079}, "bytes_per_chunk": {"en": 3.866, "de": 3.943, "vi": 4.944, "tr": 2.985, "fi": 3.436, "ru": 5.957, "el": 6.151, "hi": 15.682, "ta": 8.121, "th": 7.179, "zh": 4.49, "my": 5.081}, "bpb_mean": 2.16819, "t": 99.7} |
| {"event": "train", "step": 1400, "phase": 1, "objective": "global", "lm_loss": 1.58303, "bpb": 2.28384, "chunk_loss": 0.033562, "F": 0.22488, "G": 0.43567, "chunks": 230.28, "bytes_per_chunk": 4.447, "boundary_entropy": 0.66858, "router_noise": 0.0, "grad_norm": 0.7704, "lr": 0.0002759, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 99.7} |
| {"event": "train", "step": 1450, "phase": 1, "objective": "global", "lm_loss": 1.67593, "bpb": 2.41785, "chunk_loss": 0.034214, "F": 0.23782, "G": 0.43433, "chunks": 243.53, "bytes_per_chunk": 4.205, "boundary_entropy": 0.66731, "router_noise": 0.0, "grad_norm": 0.7655, "lr": 0.0002738, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 101.9} |
| {"event": "train", "step": 1500, "phase": 1, "objective": "global", "lm_loss": 1.75283, "bpb": 2.5288, "chunk_loss": 0.033741, "F": 0.228, "G": 0.43914, "chunks": 233.47, "bytes_per_chunk": 4.386, "boundary_entropy": 0.67168, "router_noise": 0.0, "grad_norm": 0.7665, "lr": 0.0002715, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 104.1} |
| {"event": "train", "step": 1550, "phase": 1, "objective": "global", "lm_loss": 1.61498, "bpb": 2.32992, "chunk_loss": 0.035674, "F": 0.26041, "G": 0.4407, "chunks": 266.66, "bytes_per_chunk": 3.84, "boundary_entropy": 0.66967, "router_noise": 0.0, "grad_norm": 0.8453, "lr": 0.0002692, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 106.2} |
| {"event": "probe", "step": 1600, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4523.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.259898, "cps": 35.335938, "mean_p": 0.444414, "boundary_entropy": 0.6647, "bpb": 2.383004}, "de": {"chunks": 5392.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.262602, "cps": 42.125, "mean_p": 0.450763, "boundary_entropy": 0.671379, "bpb": 2.681891}, "vi": {"chunks": 5251.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.284839, "cps": 41.023438, "mean_p": 0.451259, "boundary_entropy": 0.67448, "bpb": 2.181846}, "tr": {"chunks": 6305.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.348285, "cps": 49.257812, "mean_p": 0.457722, "boundary_entropy": 0.668041, "bpb": 2.651153}, "fi": {"chunks": 6406.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.340545, "cps": 50.046875, "mean_p": 0.461474, "boundary_entropy": 0.671002, "bpb": 2.898171}, "ru": {"chunks": 5795.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.305805, "cps": 45.273438, "mean_p": 0.426861, "boundary_entropy": 0.666922, "bpb": 1.766176}, "el": {"chunks": 8319.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.40535, "cps": 64.992188, "mean_p": 0.458741, "boundary_entropy": 0.67476, "bpb": 1.867388}, "hi": {"chunks": 8415.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.500149, "cps": 65.742188, "mean_p": 0.415099, "boundary_entropy": 0.660205, "bpb": 1.434282}, "ta": {"chunks": 6103.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.326451, "cps": 47.679688, "mean_p": 0.398257, "boundary_entropy": 0.652471, "bpb": 1.175428}, "th": {"chunks": 8760.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.544607, "cps": 68.4375, "mean_p": 0.435817, "boundary_entropy": 0.671417, "bpb": 1.385486}, "zh": {"chunks": 3387.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.573485, "cps": 26.460938, "mean_p": 0.438862, "boundary_entropy": 0.667798, "bpb": 3.549944}, "my": {"chunks": 11408.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.602069, "cps": 89.125, "mean_p": 0.449691, "boundary_entropy": 0.680397, "bpb": 1.312697}}, "bpb": {"en": 2.383, "de": 2.68189, "vi": 2.18185, "tr": 2.65115, "fi": 2.89817, "ru": 1.76618, "el": 1.86739, "hi": 1.43428, "ta": 1.17543, "th": 1.38549, "zh": 3.54994, "my": 1.3127}, "gini_cpc": 0.168213, "gini_cps": 0.171927, "bpic": 4.793428, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0104038882679962, "vi": 1.0959643028750803, "tr": 1.340084154262631, "fi": 1.3103055584406755, "ru": 1.1766349884524538, "el": 1.5596523775264455, "hi": 1.9244054574620917, "ta": 1.2560745982452057, "th": 2.0954657818763645, "zh": 2.2065780131635586, "my": 2.316560617557129}, "t": 109.3} |
| {"event": "eval", "step": 1600, "phase": 1, "bpb": {"en": 2.3904, "de": 2.68805, "vi": 2.18791, "tr": 2.66167, "fi": 2.90171, "ru": 1.74458, "el": 1.85727, "hi": 1.43243, "ta": 1.18359, "th": 1.40198, "zh": 3.55605, "my": 1.32055}, "bytes_per_chunk": {"en": 3.722, "de": 3.841, "vi": 4.579, "tr": 3.145, "fi": 3.087, "ru": 5.87, "el": 4.527, "hi": 5.447, "ta": 8.302, "th": 5.365, "zh": 4.701, "my": 4.655}, "bpb_mean": 2.11052, "t": 110.7} |
| {"event": "train", "step": 1600, "phase": 1, "objective": "global", "lm_loss": 1.61007, "bpb": 2.32284, "chunk_loss": 0.034338, "F": 0.23776, "G": 0.44046, "chunks": 243.47, "bytes_per_chunk": 4.206, "boundary_entropy": 0.67007, "router_noise": 0.0, "grad_norm": 0.86, "lr": 0.0002668, "router_trainable": true, "sec_per_step": 0.091, "alpha": 0.03, "t": 110.8} |
| {"event": "train", "step": 1650, "phase": 1, "objective": "global", "lm_loss": 1.58499, "bpb": 2.28666, "chunk_loss": 0.035582, "F": 0.25812, "G": 0.43844, "chunks": 264.31, "bytes_per_chunk": 3.874, "boundary_entropy": 0.66842, "router_noise": 0.0, "grad_norm": 0.8377, "lr": 0.0002643, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 112.9} |
| {"event": "train", "step": 1700, "phase": 1, "objective": "global", "lm_loss": 1.54954, "bpb": 2.23551, "chunk_loss": 0.033399, "F": 0.22314, "G": 0.43708, "chunks": 228.5, "bytes_per_chunk": 4.481, "boundary_entropy": 0.6706, "router_noise": 0.0, "grad_norm": 0.7568, "lr": 0.0002618, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 115.1} |
| {"event": "train", "step": 1750, "phase": 1, "objective": "global", "lm_loss": 1.57887, "bpb": 2.27783, "chunk_loss": 0.03371, "F": 0.22897, "G": 0.43872, "chunks": 234.47, "bytes_per_chunk": 4.367, "boundary_entropy": 0.6706, "router_noise": 0.0, "grad_norm": 0.8299, "lr": 0.0002591, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 117.2} |
| {"event": "probe", "step": 1800, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4421.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.254037, "cps": 34.539062, "mean_p": 0.445591, "boundary_entropy": 0.665089, "bpb": 2.342342}, "de": {"chunks": 5345.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.260313, "cps": 41.757812, "mean_p": 0.44973, "boundary_entropy": 0.671346, "bpb": 2.642842}, "vi": {"chunks": 5639.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.305886, "cps": 44.054688, "mean_p": 0.45122, "boundary_entropy": 0.673421, "bpb": 2.144462}, "tr": {"chunks": 6384.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.352649, "cps": 49.875, "mean_p": 0.459724, "boundary_entropy": 0.667526, "bpb": 2.623508}, "fi": {"chunks": 6012.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.3196, "cps": 46.96875, "mean_p": 0.457878, "boundary_entropy": 0.669849, "bpb": 2.848821}, "ru": {"chunks": 7440.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.392612, "cps": 58.125, "mean_p": 0.432156, "boundary_entropy": 0.668851, "bpb": 1.711893}, "el": {"chunks": 7904.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.385129, "cps": 61.75, "mean_p": 0.460702, "boundary_entropy": 0.676629, "bpb": 1.81091}, "hi": {"chunks": 6368.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.378484, "cps": 49.75, "mean_p": 0.415328, "boundary_entropy": 0.662372, "bpb": 1.404596}, "ta": {"chunks": 8480.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.453597, "cps": 66.25, "mean_p": 0.410981, "boundary_entropy": 0.659999, "bpb": 1.128214}, "th": {"chunks": 9987.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.620889, "cps": 78.023438, "mean_p": 0.432614, "boundary_entropy": 0.668438, "bpb": 1.349676}, "zh": {"chunks": 3339.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.565357, "cps": 26.085938, "mean_p": 0.441107, "boundary_entropy": 0.669169, "bpb": 3.493269}, "my": {"chunks": 10066.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.531243, "cps": 78.640625, "mean_p": 0.441618, "boundary_entropy": 0.677867, "bpb": 1.278973}}, "bpb": {"en": 2.34234, "de": 2.64284, "vi": 2.14446, "tr": 2.62351, "fi": 2.84882, "ru": 1.71189, "el": 1.81091, "hi": 1.4046, "ta": 1.12821, "th": 1.34968, "zh": 3.49327, "my": 1.27897}, "gini_cpc": 0.158285, "gini_cps": 0.165531, "bpic": 4.715623, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.024705123500556, "vi": 1.2041000042514745, "tr": 1.3881804766844545, "fi": 1.2580870551144048, "ru": 1.545494011252364, "el": 1.5160366195531745, "hi": 1.4898810183599618, "ta": 1.7855581075974105, "th": 2.4440922274697505, "zh": 2.225494788939866, "my": 2.091207632293279}, "t": 120.4} |
| {"event": "eval", "step": 1800, "phase": 1, "bpb": {"en": 2.3521, "de": 2.64501, "vi": 2.15013, "tr": 2.63525, "fi": 2.8489, "ru": 1.69123, "el": 1.80308, "hi": 1.40183, "ta": 1.13644, "th": 1.37093, "zh": 3.49736, "my": 1.28752}, "bytes_per_chunk": {"en": 3.736, "de": 3.88, "vi": 4.166, "tr": 3.095, "fi": 3.309, "ru": 4.612, "el": 4.722, "hi": 7.431, "ta": 5.72, "th": 4.622, "zh": 4.861, "my": 5.319}, "bpb_mean": 2.06832, "t": 121.7} |
| {"event": "train", "step": 1800, "phase": 1, "objective": "global", "lm_loss": 1.53848, "bpb": 2.21956, "chunk_loss": 0.034047, "F": 0.23331, "G": 0.443, "chunks": 238.91, "bytes_per_chunk": 4.286, "boundary_entropy": 0.67209, "router_noise": 0.0, "grad_norm": 0.741, "lr": 0.0002564, "router_trainable": true, "sec_per_step": 0.091, "alpha": 0.03, "t": 121.8} |
| {"event": "train", "step": 1850, "phase": 1, "objective": "global", "lm_loss": 1.58442, "bpb": 2.28583, "chunk_loss": 0.034895, "F": 0.24741, "G": 0.44292, "chunks": 253.34, "bytes_per_chunk": 4.042, "boundary_entropy": 0.67081, "router_noise": 0.0, "grad_norm": 0.7584, "lr": 0.0002537, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 124.0} |
| {"event": "train", "step": 1900, "phase": 1, "objective": "global", "lm_loss": 1.64149, "bpb": 2.36817, "chunk_loss": 0.034951, "F": 0.24957, "G": 0.43775, "chunks": 255.56, "bytes_per_chunk": 4.007, "boundary_entropy": 0.66877, "router_noise": 0.0, "grad_norm": 0.815, "lr": 0.0002508, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 126.1} |
| {"event": "train", "step": 1950, "phase": 1, "objective": "global", "lm_loss": 1.48714, "bpb": 2.14549, "chunk_loss": 0.033221, "F": 0.22028, "G": 0.43833, "chunks": 225.56, "bytes_per_chunk": 4.54, "boundary_entropy": 0.67154, "router_noise": 0.0, "grad_norm": 0.7488, "lr": 0.0002479, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 128.3} |
| {"event": "probe", "step": 2000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4219.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.242429, "cps": 32.960938, "mean_p": 0.443305, "boundary_entropy": 0.664823, "bpb": 2.299141}, "de": {"chunks": 4933.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.240247, "cps": 38.539062, "mean_p": 0.446086, "boundary_entropy": 0.670097, "bpb": 2.587095}, "vi": {"chunks": 4811.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.260971, "cps": 37.585938, "mean_p": 0.447214, "boundary_entropy": 0.672899, "bpb": 2.108716}, "tr": {"chunks": 6573.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.363089, "cps": 51.351562, "mean_p": 0.457351, "boundary_entropy": 0.665437, "bpb": 2.563149}, "fi": {"chunks": 5636.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.299612, "cps": 44.03125, "mean_p": 0.452326, "boundary_entropy": 0.668565, "bpb": 2.790748}, "ru": {"chunks": 4926.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.259947, "cps": 38.484375, "mean_p": 0.427878, "boundary_entropy": 0.669546, "bpb": 1.683464}, "el": {"chunks": 7398.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.360474, "cps": 57.796875, "mean_p": 0.456404, "boundary_entropy": 0.675167, "bpb": 1.75402}, "hi": {"chunks": 7813.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.464368, "cps": 61.039062, "mean_p": 0.419686, "boundary_entropy": 0.663423, "bpb": 1.370974}, "ta": {"chunks": 8111.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.433859, "cps": 63.367188, "mean_p": 0.413002, "boundary_entropy": 0.659488, "bpb": 1.12271}, "th": {"chunks": 8652.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.537892, "cps": 67.59375, "mean_p": 0.430782, "boundary_entropy": 0.669353, "bpb": 1.317196}, "zh": {"chunks": 3072.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.520149, "cps": 24.0, "mean_p": 0.440375, "boundary_entropy": 0.669744, "bpb": 3.455537}, "my": {"chunks": 7742.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.408592, "cps": 60.484375, "mean_p": 0.437588, "boundary_entropy": 0.678179, "bpb": 1.258184}}, "bpb": {"en": 2.29914, "de": 2.58709, "vi": 2.10872, "tr": 2.56315, "fi": 2.79075, "ru": 1.68346, "el": 1.75402, "hi": 1.37097, "ta": 1.12271, "th": 1.3172, "zh": 3.45554, "my": 1.25818}, "gini_cpc": 0.159366, "gini_cps": 0.157714, "bpic": 5.194232, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9909991982220863, "vi": 1.0764820928069403, "tr": 1.4977098142824112, "fi": 1.2358725773168908, "ru": 1.0722592168485197, "el": 1.486921621544929, "hi": 1.9154787846813415, "ta": 1.7896311347924203, "th": 2.218758531483056, "zh": 2.1455683964636134, "my": 1.6854054181444327}, "t": 131.4} |
| {"event": "eval", "step": 2000, "phase": 1, "bpb": {"en": 2.30787, "de": 2.59484, "vi": 2.11582, "tr": 2.57246, "fi": 2.79749, "ru": 1.66205, "el": 1.74562, "hi": 1.37077, "ta": 1.1303, "th": 1.33238, "zh": 3.46089, "my": 1.26764}, "bytes_per_chunk": {"en": 3.916, "de": 4.189, "vi": 4.969, "tr": 2.951, "fi": 3.577, "ru": 6.956, "el": 5.082, "hi": 5.833, "ta": 6.04, "th": 5.502, "zh": 5.218, "my": 6.982}, "bpb_mean": 2.02985, "t": 132.7} |
| {"event": "train", "step": 2000, "phase": 1, "objective": "global", "lm_loss": 1.61066, "bpb": 2.32369, "chunk_loss": 0.033674, "F": 0.2272, "G": 0.44045, "chunks": 232.66, "bytes_per_chunk": 4.401, "boundary_entropy": 0.67081, "router_noise": 0.0, "grad_norm": 0.8075, "lr": 0.000245, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 132.8} |
| {"event": "train", "step": 2050, "phase": 1, "objective": "global", "lm_loss": 1.62163, "bpb": 2.33952, "chunk_loss": 0.03303, "F": 0.21661, "G": 0.4397, "chunks": 221.81, "bytes_per_chunk": 4.617, "boundary_entropy": 0.67165, "router_noise": 0.0, "grad_norm": 0.7586, "lr": 0.0002419, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 134.9} |
| {"event": "train", "step": 2100, "phase": 1, "objective": "global", "lm_loss": 1.55042, "bpb": 2.23679, "chunk_loss": 0.034345, "F": 0.23944, "G": 0.43888, "chunks": 245.19, "bytes_per_chunk": 4.176, "boundary_entropy": 0.66979, "router_noise": 0.0, "grad_norm": 0.7604, "lr": 0.0002388, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 137.1} |
| {"event": "train", "step": 2150, "phase": 1, "objective": "global", "lm_loss": 1.51227, "bpb": 2.18175, "chunk_loss": 0.034821, "F": 0.24704, "G": 0.43992, "chunks": 252.97, "bytes_per_chunk": 4.048, "boundary_entropy": 0.67, "router_noise": 0.0, "grad_norm": 0.7424, "lr": 0.0002357, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 139.2} |
| {"event": "probe", "step": 2200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4499.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.258519, "cps": 35.148438, "mean_p": 0.444169, "boundary_entropy": 0.664351, "bpb": 2.256467}, "de": {"chunks": 5191.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.252813, "cps": 40.554688, "mean_p": 0.450187, "boundary_entropy": 0.670976, "bpb": 2.534368}, "vi": {"chunks": 5790.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.314076, "cps": 45.234375, "mean_p": 0.451487, "boundary_entropy": 0.6732, "bpb": 2.093933}, "tr": {"chunks": 6167.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.340662, "cps": 48.179688, "mean_p": 0.456084, "boundary_entropy": 0.666984, "bpb": 2.515296}, "fi": {"chunks": 5370.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.285471, "cps": 41.953125, "mean_p": 0.452907, "boundary_entropy": 0.669664, "bpb": 2.742823}, "ru": {"chunks": 5942.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.313562, "cps": 46.421875, "mean_p": 0.429003, "boundary_entropy": 0.669546, "bpb": 1.641668}, "el": {"chunks": 7969.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.388296, "cps": 62.257812, "mean_p": 0.459389, "boundary_entropy": 0.676214, "bpb": 1.721243}, "hi": {"chunks": 7331.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.435721, "cps": 57.273438, "mean_p": 0.419863, "boundary_entropy": 0.664526, "bpb": 1.348939}, "ta": {"chunks": 7756.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.41487, "cps": 60.59375, "mean_p": 0.40524, "boundary_entropy": 0.65539, "bpb": 1.10561}, "th": {"chunks": 9020.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.560771, "cps": 70.46875, "mean_p": 0.427497, "boundary_entropy": 0.667187, "bpb": 1.263417}, "zh": {"chunks": 3196.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.541145, "cps": 24.96875, "mean_p": 0.439302, "boundary_entropy": 0.669412, "bpb": 3.471223}, "my": {"chunks": 6249.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.329797, "cps": 48.820312, "mean_p": 0.438915, "boundary_entropy": 0.678669, "bpb": 1.23975}}, "bpb": {"en": 2.25647, "de": 2.53437, "vi": 2.09393, "tr": 2.5153, "fi": 2.74282, "ru": 1.64167, "el": 1.72124, "hi": 1.34894, "ta": 1.10561, "th": 1.26342, "zh": 3.47122, "my": 1.23975}, "gini_cpc": 0.145091, "gini_cps": 0.139474, "bpic": 5.152806, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9779277021759963, "vi": 1.2149084391052638, "tr": 1.3177454460686095, "fi": 1.1042579365033882, "ru": 1.2129183324819766, "el": 1.5020040673008952, "hi": 1.685451553209612, "ta": 1.6047983085718553, "th": 2.1691700493937693, "zh": 2.093251711803812, "my": 1.2757197398450801}, "t": 142.3} |
| {"event": "eval", "step": 2200, "phase": 1, "bpb": {"en": 2.26556, "de": 2.54026, "vi": 2.09963, "tr": 2.52411, "fi": 2.75144, "ru": 1.61914, "el": 1.71242, "hi": 1.34539, "ta": 1.11232, "th": 1.28187, "zh": 3.47651, "my": 1.24808}, "bytes_per_chunk": {"en": 3.765, "de": 3.935, "vi": 4.075, "tr": 3.179, "fi": 3.719, "ru": 5.552, "el": 4.817, "hi": 6.162, "ta": 6.355, "th": 5.296, "zh": 4.906, "my": 8.321}, "bpb_mean": 1.99806, "t": 143.7} |
| {"event": "train", "step": 2200, "phase": 1, "objective": "global", "lm_loss": 1.45795, "bpb": 2.10337, "chunk_loss": 0.033414, "F": 0.22336, "G": 0.43838, "chunks": 228.72, "bytes_per_chunk": 4.477, "boundary_entropy": 0.67027, "router_noise": 0.0, "grad_norm": 0.7077, "lr": 0.0002325, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 143.7} |
| {"event": "train", "step": 2250, "phase": 1, "objective": "global", "lm_loss": 1.5491, "bpb": 2.23488, "chunk_loss": 0.033914, "F": 0.23218, "G": 0.43761, "chunks": 237.75, "bytes_per_chunk": 4.307, "boundary_entropy": 0.66999, "router_noise": 0.0, "grad_norm": 0.7281, "lr": 0.0002293, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 145.8} |
| {"event": "train", "step": 2300, "phase": 1, "objective": "global", "lm_loss": 1.49017, "bpb": 2.14986, "chunk_loss": 0.034137, "F": 0.23599, "G": 0.43847, "chunks": 241.66, "bytes_per_chunk": 4.237, "boundary_entropy": 0.66963, "router_noise": 0.0, "grad_norm": 0.7952, "lr": 0.000226, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 148.0} |
| {"event": "train", "step": 2350, "phase": 1, "objective": "global", "lm_loss": 1.48974, "bpb": 2.14924, "chunk_loss": 0.034817, "F": 0.24707, "G": 0.44098, "chunks": 253.0, "bytes_per_chunk": 4.047, "boundary_entropy": 0.67061, "router_noise": 0.0, "grad_norm": 0.7153, "lr": 0.0002226, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 150.1} |
| {"event": "probe", "step": 2400, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4293.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.246682, "cps": 33.539062, "mean_p": 0.44152, "boundary_entropy": 0.663734, "bpb": 2.236182}, "de": {"chunks": 5238.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.255102, "cps": 40.921875, "mean_p": 0.448563, "boundary_entropy": 0.670647, "bpb": 2.492797}, "vi": {"chunks": 5403.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.293084, "cps": 42.210938, "mean_p": 0.453151, "boundary_entropy": 0.674309, "bpb": 2.069563}, "tr": {"chunks": 6231.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.344197, "cps": 48.679688, "mean_p": 0.454241, "boundary_entropy": 0.665796, "bpb": 2.479986}, "fi": {"chunks": 5512.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.29302, "cps": 43.0625, "mean_p": 0.455005, "boundary_entropy": 0.670151, "bpb": 2.702867}, "ru": {"chunks": 6553.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.345805, "cps": 51.195312, "mean_p": 0.434993, "boundary_entropy": 0.672136, "bpb": 1.618025}, "el": {"chunks": 7266.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.354042, "cps": 56.765625, "mean_p": 0.460505, "boundary_entropy": 0.677123, "bpb": 1.706215}, "hi": {"chunks": 11645.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.692125, "cps": 90.976562, "mean_p": 0.431399, "boundary_entropy": 0.666276, "bpb": 1.32081}, "ta": {"chunks": 9433.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.504573, "cps": 73.695312, "mean_p": 0.416263, "boundary_entropy": 0.660441, "bpb": 1.073658}, "th": {"chunks": 12721.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.790861, "cps": 99.382812, "mean_p": 0.439955, "boundary_entropy": 0.668932, "bpb": 1.245298}, "zh": {"chunks": 3276.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.55469, "cps": 25.59375, "mean_p": 0.440285, "boundary_entropy": 0.669232, "bpb": 3.399085}, "my": {"chunks": 11096.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.585603, "cps": 86.6875, "mean_p": 0.449579, "boundary_entropy": 0.680552, "bpb": 1.197229}}, "bpb": {"en": 2.23618, "de": 2.4928, "vi": 2.06956, "tr": 2.47999, "fi": 2.70287, "ru": 1.61802, "el": 1.70621, "hi": 1.32081, "ta": 1.07366, "th": 1.2453, "zh": 3.39908, "my": 1.19723}, "gini_cpc": 0.21808, "gini_cps": 0.222558, "bpic": 4.328341, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0341328133542966, "vi": 1.1881056394528, "tr": 1.3953091157505313, "fi": 1.187847142235532, "ru": 1.4018262410366755, "el": 1.4352176946027013, "hi": 2.8057414727800776, "ta": 2.045444012896717, "th": 3.205999269690392, "zh": 2.2486076413071023, "my": 2.373921226463039}, "t": 153.3} |
| {"event": "eval", "step": 2400, "phase": 1, "bpb": {"en": 2.24506, "de": 2.49738, "vi": 2.07517, "tr": 2.48916, "fi": 2.71238, "ru": 1.59502, "el": 1.69885, "hi": 1.31956, "ta": 1.08005, "th": 1.26414, "zh": 3.40649, "my": 1.20611}, "bytes_per_chunk": {"en": 3.918, "de": 3.893, "vi": 4.333, "tr": 3.12, "fi": 3.59, "ru": 5.106, "el": 5.312, "hi": 3.899, "ta": 5.432, "th": 3.7, "zh": 4.846, "my": 4.981}, "bpb_mean": 1.96578, "t": 154.7} |
| {"event": "phase_transition", "step": 2400, "phase": 2, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 154.7} |
| {"event": "train", "step": 2400, "phase": 2, "objective": "global", "lm_loss": 1.37965, "bpb": 1.99041, "chunk_loss": 0.034777, "F": 0.24591, "G": 0.44236, "chunks": 251.81, "bytes_per_chunk": 4.067, "boundary_entropy": 0.67075, "router_noise": 0.0, "grad_norm": 0.7342, "lr": 0.0002192, "router_trainable": true, "sec_per_step": 0.092, "alpha": 0.03, "t": 154.7} |
| {"event": "train", "step": 2450, "phase": 2, "objective": "global", "lm_loss": 1.47072, "bpb": 2.12181, "chunk_loss": 0.033431, "F": 0.22427, "G": 0.43923, "chunks": 229.66, "bytes_per_chunk": 4.459, "boundary_entropy": 0.67142, "router_noise": 0.0, "grad_norm": 0.7135, "lr": 0.0002158, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 156.8} |
| {"event": "train", "step": 2500, "phase": 2, "objective": "global", "lm_loss": 1.50823, "bpb": 2.17592, "chunk_loss": 0.033328, "F": 0.22305, "G": 0.43651, "chunks": 228.41, "bytes_per_chunk": 4.483, "boundary_entropy": 0.66937, "router_noise": 0.0, "grad_norm": 0.7402, "lr": 0.0002123, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 159.0} |
| {"event": "train", "step": 2550, "phase": 2, "objective": "global", "lm_loss": 1.57162, "bpb": 2.26737, "chunk_loss": 0.035301, "F": 0.25415, "G": 0.4446, "chunks": 260.25, "bytes_per_chunk": 3.935, "boundary_entropy": 0.67155, "router_noise": 0.0, "grad_norm": 0.7424, "lr": 0.0002088, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 161.2} |
| {"event": "probe", "step": 2600, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4436.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.254899, "cps": 34.65625, "mean_p": 0.442945, "boundary_entropy": 0.663725, "bpb": 2.206671}, "de": {"chunks": 5196.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.253056, "cps": 40.59375, "mean_p": 0.449239, "boundary_entropy": 0.670756, "bpb": 2.455111}, "vi": {"chunks": 5508.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.298779, "cps": 43.03125, "mean_p": 0.453882, "boundary_entropy": 0.673822, "bpb": 2.054329}, "tr": {"chunks": 6339.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.350163, "cps": 49.523438, "mean_p": 0.460268, "boundary_entropy": 0.668037, "bpb": 2.441096}, "fi": {"chunks": 5833.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.310085, "cps": 45.570312, "mean_p": 0.457472, "boundary_entropy": 0.670349, "bpb": 2.668955}, "ru": {"chunks": 6583.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.347388, "cps": 51.429688, "mean_p": 0.434361, "boundary_entropy": 0.671334, "bpb": 1.59017}, "el": {"chunks": 8162.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.3977, "cps": 63.765625, "mean_p": 0.456216, "boundary_entropy": 0.675685, "bpb": 1.680602}, "hi": {"chunks": 9352.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.55584, "cps": 73.0625, "mean_p": 0.425072, "boundary_entropy": 0.664847, "bpb": 1.308957}, "ta": {"chunks": 6695.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.358117, "cps": 52.304688, "mean_p": 0.414307, "boundary_entropy": 0.661691, "bpb": 1.075184}, "th": {"chunks": 10032.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.623687, "cps": 78.375, "mean_p": 0.438543, "boundary_entropy": 0.671166, "bpb": 1.219163}, "zh": {"chunks": 3719.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.629699, "cps": 29.054688, "mean_p": 0.444853, "boundary_entropy": 0.6704, "bpb": 3.3672}, "my": {"chunks": 7580.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.400042, "cps": 59.21875, "mean_p": 0.453648, "boundary_entropy": 0.682309, "bpb": 1.206029}}, "bpb": {"en": 2.20667, "de": 2.45511, "vi": 2.05433, "tr": 2.4411, "fi": 2.66895, "ru": 1.59017, "el": 1.6806, "hi": 1.30896, "ta": 1.07518, "th": 1.21916, "zh": 3.3672, "my": 1.20603}, "gini_cpc": 0.172095, "gini_cps": 0.154884, "bpic": 4.831384, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 0.992771538411171, "vi": 1.17215048706456, "tr": 1.373734429781054, "fi": 1.216501575469603, "ru": 1.3628473796783809, "el": 1.5602289357819619, "hi": 2.1806301274055175, "ta": 1.4049397318397867, "th": 2.44680320647776, "zh": 2.4703888497186304, "my": 1.569417215699062}, "t": 164.3} |
| {"event": "eval", "step": 2600, "phase": 2, "bpb": {"en": 2.21502, "de": 2.46157, "vi": 2.05797, "tr": 2.44754, "fi": 2.6759, "ru": 1.56869, "el": 1.67196, "hi": 1.30547, "ta": 1.08067, "th": 1.23842, "zh": 3.36972, "my": 1.21455}, "bytes_per_chunk": {"en": 3.806, "de": 3.924, "vi": 4.422, "tr": 3.067, "fi": 3.383, "ru": 5.177, "el": 4.643, "hi": 4.858, "ta": 7.091, "th": 4.575, "zh": 4.409, "my": 6.92}, "bpb_mean": 1.94229, "t": 165.6} |
| {"event": "train", "step": 2600, "phase": 2, "objective": "global", "lm_loss": 1.6027, "bpb": 2.31221, "chunk_loss": 0.034591, "F": 0.24249, "G": 0.4427, "chunks": 248.31, "bytes_per_chunk": 4.124, "boundary_entropy": 0.6711, "router_noise": 0.0, "grad_norm": 0.7992, "lr": 0.0002053, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 165.7} |
| {"event": "train", "step": 2650, "phase": 2, "objective": "global", "lm_loss": 1.50903, "bpb": 2.17708, "chunk_loss": 0.03441, "F": 0.24051, "G": 0.43995, "chunks": 246.28, "bytes_per_chunk": 4.158, "boundary_entropy": 0.67056, "router_noise": 0.0, "grad_norm": 0.7672, "lr": 0.0002017, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 167.8} |
| {"event": "train", "step": 2700, "phase": 2, "objective": "global", "lm_loss": 1.52017, "bpb": 2.19314, "chunk_loss": 0.035322, "F": 0.25488, "G": 0.44469, "chunks": 261.0, "bytes_per_chunk": 3.923, "boundary_entropy": 0.67171, "router_noise": 0.0, "grad_norm": 0.7556, "lr": 0.0001981, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 169.9} |
| {"event": "train", "step": 2750, "phase": 2, "objective": "global", "lm_loss": 1.40444, "bpb": 2.02618, "chunk_loss": 0.033638, "F": 0.22849, "G": 0.43555, "chunks": 233.97, "bytes_per_chunk": 4.377, "boundary_entropy": 0.66783, "router_noise": 0.0, "grad_norm": 0.7815, "lr": 0.0001945, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 172.1} |
| {"event": "probe", "step": 2800, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3997.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.229673, "cps": 31.226562, "mean_p": 0.439454, "boundary_entropy": 0.662925, "bpb": 2.193668}, "de": {"chunks": 4804.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.233965, "cps": 37.53125, "mean_p": 0.445527, "boundary_entropy": 0.669751, "bpb": 2.446594}, "vi": {"chunks": 5608.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.304204, "cps": 43.8125, "mean_p": 0.450736, "boundary_entropy": 0.672782, "bpb": 2.042961}, "tr": {"chunks": 5582.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.308347, "cps": 43.609375, "mean_p": 0.455186, "boundary_entropy": 0.66771, "bpb": 2.431151}, "fi": {"chunks": 4937.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.262453, "cps": 38.570312, "mean_p": 0.44953, "boundary_entropy": 0.669869, "bpb": 2.653311}, "ru": {"chunks": 5575.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.294195, "cps": 43.554688, "mean_p": 0.433936, "boundary_entropy": 0.671944, "bpb": 1.569986}, "el": {"chunks": 8177.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.398431, "cps": 63.882812, "mean_p": 0.458817, "boundary_entropy": 0.677025, "bpb": 1.640184}, "hi": {"chunks": 7437.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.442021, "cps": 58.101562, "mean_p": 0.412971, "boundary_entropy": 0.659493, "bpb": 1.294343}, "ta": {"chunks": 6153.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.329125, "cps": 48.070312, "mean_p": 0.40988, "boundary_entropy": 0.660051, "bpb": 1.079179}, "th": {"chunks": 9779.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.607958, "cps": 76.398438, "mean_p": 0.427882, "boundary_entropy": 0.66708, "bpb": 1.209383}, "zh": {"chunks": 3544.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.600068, "cps": 27.6875, "mean_p": 0.44425, "boundary_entropy": 0.670603, "bpb": 3.360087}, "my": {"chunks": 11657.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.61521, "cps": 91.070312, "mean_p": 0.455922, "boundary_entropy": 0.681474, "bpb": 1.182774}}, "bpb": {"en": 2.19367, "de": 2.44659, "vi": 2.04296, "tr": 2.43115, "fi": 2.65331, "ru": 1.56999, "el": 1.64018, "hi": 1.29434, "ta": 1.07918, "th": 1.20938, "zh": 3.36009, "my": 1.18277}, "gini_cpc": 0.1984, "gini_cps": 0.19326, "bpic": 4.968039, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0186865298732974, "vi": 1.3245087599263319, "tr": 1.342546240177841, "fi": 1.1427236500417681, "ru": 1.2809306848210036, "el": 1.7347748788495332, "hi": 1.9245654300160482, "ta": 1.4330172475572243, "th": 2.647057363274243, "zh": 2.6127041945868474, "my": 2.6786340943167795}, "t": 175.2} |
| {"event": "eval", "step": 2800, "phase": 2, "bpb": {"en": 2.20424, "de": 2.45387, "vi": 2.0474, "tr": 2.43846, "fi": 2.65874, "ru": 1.54718, "el": 1.63203, "hi": 1.28924, "ta": 1.08347, "th": 1.22707, "zh": 3.36034, "my": 1.18985}, "bytes_per_chunk": {"en": 4.204, "de": 4.299, "vi": 4.083, "tr": 3.495, "fi": 4.058, "ru": 6.25, "el": 4.737, "hi": 6.073, "ta": 7.748, "th": 4.714, "zh": 4.527, "my": 4.654}, "bpb_mean": 1.92766, "t": 176.5} |
| {"event": "train", "step": 2800, "phase": 2, "objective": "global", "lm_loss": 1.48127, "bpb": 2.13702, "chunk_loss": 0.032994, "F": 0.21664, "G": 0.44157, "chunks": 221.84, "bytes_per_chunk": 4.616, "boundary_entropy": 0.67225, "router_noise": 0.0, "grad_norm": 0.9088, "lr": 0.0001909, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 176.6} |
| {"event": "train", "step": 2850, "phase": 2, "objective": "global", "lm_loss": 1.43929, "bpb": 2.07645, "chunk_loss": 0.033411, "F": 0.22311, "G": 0.43929, "chunks": 228.47, "bytes_per_chunk": 4.482, "boundary_entropy": 0.67022, "router_noise": 0.0, "grad_norm": 0.6897, "lr": 0.0001872, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 178.7} |
| {"event": "train", "step": 2900, "phase": 2, "objective": "global", "lm_loss": 1.45587, "bpb": 2.10038, "chunk_loss": 0.034442, "F": 0.24103, "G": 0.43903, "chunks": 246.81, "bytes_per_chunk": 4.149, "boundary_entropy": 0.66979, "router_noise": 0.0, "grad_norm": 0.7723, "lr": 0.0001835, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 180.8} |
| {"event": "train", "step": 2950, "phase": 2, "objective": "global", "lm_loss": 1.36207, "bpb": 1.96506, "chunk_loss": 0.033774, "F": 0.23035, "G": 0.43817, "chunks": 235.88, "bytes_per_chunk": 4.341, "boundary_entropy": 0.66976, "router_noise": 0.0, "grad_norm": 0.7227, "lr": 0.0001799, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 183.0} |
| {"event": "probe", "step": 3000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4178.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.240074, "cps": 32.640625, "mean_p": 0.44027, "boundary_entropy": 0.662855, "bpb": 2.166717}, "de": {"chunks": 5563.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.27093, "cps": 43.460938, "mean_p": 0.45019, "boundary_entropy": 0.670468, "bpb": 2.406462}, "vi": {"chunks": 5150.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.27936, "cps": 40.234375, "mean_p": 0.450831, "boundary_entropy": 0.672875, "bpb": 2.022914}, "tr": {"chunks": 6088.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.336298, "cps": 47.5625, "mean_p": 0.458447, "boundary_entropy": 0.667576, "bpb": 2.393601}, "fi": {"chunks": 5388.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.286428, "cps": 42.09375, "mean_p": 0.451774, "boundary_entropy": 0.66899, "bpb": 2.597073}, "ru": {"chunks": 7371.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.388971, "cps": 57.585938, "mean_p": 0.437609, "boundary_entropy": 0.672024, "bpb": 1.542015}, "el": {"chunks": 8134.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.396336, "cps": 63.546875, "mean_p": 0.460393, "boundary_entropy": 0.676704, "bpb": 1.605139}, "hi": {"chunks": 6984.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.415097, "cps": 54.5625, "mean_p": 0.420392, "boundary_entropy": 0.665603, "bpb": 1.282726}, "ta": {"chunks": 5737.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.306873, "cps": 44.820312, "mean_p": 0.414279, "boundary_entropy": 0.662976, "bpb": 1.059172}, "th": {"chunks": 10200.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.634131, "cps": 79.6875, "mean_p": 0.433958, "boundary_entropy": 0.669363, "bpb": 1.188975}, "zh": {"chunks": 3373.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.571114, "cps": 26.351562, "mean_p": 0.443742, "boundary_entropy": 0.670541, "bpb": 3.327529}, "my": {"chunks": 9711.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.512508, "cps": 75.867188, "mean_p": 0.447876, "boundary_entropy": 0.680028, "bpb": 1.169634}}, "bpb": {"en": 2.16672, "de": 2.40646, "vi": 2.02291, "tr": 2.3936, "fi": 2.59707, "ru": 1.54201, "el": 1.60514, "hi": 1.28273, "ta": 1.05917, "th": 1.18898, "zh": 3.32753, "my": 1.16963}, "gini_cpc": 0.173072, "gini_cps": 0.171543, "bpic": 4.92804, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1285279960195307, "vi": 1.1636430276612626, "tr": 1.40081175303269, "fi": 1.193085007069008, "ru": 1.620215868773611, "el": 1.650893309417796, "hi": 1.729039211265121, "ta": 1.2782478323842805, "th": 2.641403756045291, "zh": 2.378913128724681, "my": 2.1347954210525835}, "t": 186.1} |
| {"event": "eval", "step": 3000, "phase": 2, "bpb": {"en": 2.17607, "de": 2.41135, "vi": 2.02392, "tr": 2.4015, "fi": 2.60251, "ru": 1.52128, "el": 1.59794, "hi": 1.28011, "ta": 1.06477, "th": 1.20821, "zh": 3.32858, "my": 1.17995}, "bytes_per_chunk": {"en": 4.127, "de": 3.739, "vi": 4.564, "tr": 3.212, "fi": 3.662, "ru": 4.54, "el": 4.675, "hi": 6.716, "ta": 8.682, "th": 4.451, "zh": 4.805, "my": 5.592}, "bpb_mean": 1.89968, "t": 187.4} |
| {"event": "train", "step": 3000, "phase": 2, "objective": "global", "lm_loss": 1.43351, "bpb": 2.06812, "chunk_loss": 0.033781, "F": 0.23035, "G": 0.43692, "chunks": 235.88, "bytes_per_chunk": 4.341, "boundary_entropy": 0.66949, "router_noise": 0.0, "grad_norm": 0.7377, "lr": 0.0001761, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 187.5} |
| {"event": "train", "step": 3050, "phase": 2, "objective": "global", "lm_loss": 1.39956, "bpb": 2.01914, "chunk_loss": 0.034154, "F": 0.23599, "G": 0.4417, "chunks": 241.66, "bytes_per_chunk": 4.237, "boundary_entropy": 0.67156, "router_noise": 0.0, "grad_norm": 0.7272, "lr": 0.0001724, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 189.6} |
| {"event": "train", "step": 3100, "phase": 2, "objective": "global", "lm_loss": 1.40891, "bpb": 2.03263, "chunk_loss": 0.034023, "F": 0.23444, "G": 0.43937, "chunks": 240.06, "bytes_per_chunk": 4.266, "boundary_entropy": 0.66985, "router_noise": 0.0, "grad_norm": 0.8182, "lr": 0.0001687, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 191.7} |
| {"event": "train", "step": 3150, "phase": 2, "objective": "global", "lm_loss": 1.35874, "bpb": 1.96024, "chunk_loss": 0.033057, "F": 0.21838, "G": 0.43878, "chunks": 223.62, "bytes_per_chunk": 4.579, "boundary_entropy": 0.67083, "router_noise": 0.0, "grad_norm": 0.7548, "lr": 0.000165, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 193.9} |
| {"event": "probe", "step": 3200, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4274.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.24559, "cps": 33.390625, "mean_p": 0.439759, "boundary_entropy": 0.661848, "bpb": 2.147687}, "de": {"chunks": 5118.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.249257, "cps": 39.984375, "mean_p": 0.447895, "boundary_entropy": 0.669566, "bpb": 2.375533}, "vi": {"chunks": 4841.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.262598, "cps": 37.820312, "mean_p": 0.447127, "boundary_entropy": 0.67243, "bpb": 2.008239}, "tr": {"chunks": 6207.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.342871, "cps": 48.492188, "mean_p": 0.457513, "boundary_entropy": 0.66665, "bpb": 2.362578}, "fi": {"chunks": 5481.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.291372, "cps": 42.820312, "mean_p": 0.454108, "boundary_entropy": 0.669549, "bpb": 2.561475}, "ru": {"chunks": 5891.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.310871, "cps": 46.023438, "mean_p": 0.431145, "boundary_entropy": 0.670232, "bpb": 1.521602}, "el": {"chunks": 7574.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.369049, "cps": 59.171875, "mean_p": 0.456633, "boundary_entropy": 0.676168, "bpb": 1.588079}, "hi": {"chunks": 9325.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.554235, "cps": 72.851562, "mean_p": 0.417647, "boundary_entropy": 0.661519, "bpb": 1.271161}, "ta": {"chunks": 6906.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.369404, "cps": 53.953125, "mean_p": 0.413137, "boundary_entropy": 0.66171, "bpb": 1.035196}, "th": {"chunks": 10409.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.647125, "cps": 81.320312, "mean_p": 0.425193, "boundary_entropy": 0.664325, "bpb": 1.170132}, "zh": {"chunks": 3330.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.563833, "cps": 26.015625, "mean_p": 0.442864, "boundary_entropy": 0.670221, "bpb": 3.29098}, "my": {"chunks": 10235.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.540163, "cps": 79.960938, "mean_p": 0.440349, "boundary_entropy": 0.676987, "bpb": 1.152811}}, "bpb": {"en": 2.14769, "de": 2.37553, "vi": 2.00824, "tr": 2.36258, "fi": 2.56148, "ru": 1.5216, "el": 1.58808, "hi": 1.27116, "ta": 1.0352, "th": 1.17013, "zh": 3.29098, "my": 1.15281}, "gini_cpc": 0.188482, "gini_cps": 0.18823, "bpic": 4.821915, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.014933241103989, "vi": 1.0692556236300024, "tr": 1.3961137241218173, "fi": 1.1864174370838636, "ru": 1.2658125895918502, "el": 1.5027061298837006, "hi": 2.256749578118775, "ta": 1.504148472080261, "th": 2.6349813498102983, "zh": 2.295833524151022, "my": 2.199449897060833}, "t": 197.0} |
| {"event": "eval", "step": 3200, "phase": 2, "bpb": {"en": 2.15793, "de": 2.38624, "vi": 2.01192, "tr": 2.3692, "fi": 2.5706, "ru": 1.49989, "el": 1.58006, "hi": 1.26612, "ta": 1.04322, "th": 1.19095, "zh": 3.29405, "my": 1.16088}, "bytes_per_chunk": {"en": 3.951, "de": 4.024, "vi": 4.826, "tr": 3.14, "fi": 3.614, "ru": 5.647, "el": 5.099, "hi": 4.781, "ta": 7.322, "th": 4.45, "zh": 4.883, "my": 5.33}, "bpb_mean": 1.87759, "t": 198.3} |
| {"event": "train", "step": 3200, "phase": 2, "objective": "global", "lm_loss": 1.38181, "bpb": 1.99353, "chunk_loss": 0.032819, "F": 0.21365, "G": 0.43695, "chunks": 218.78, "bytes_per_chunk": 4.68, "boundary_entropy": 0.67022, "router_noise": 0.0, "grad_norm": 0.6687, "lr": 0.0001613, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 198.4} |
| {"event": "train", "step": 3250, "phase": 2, "objective": "global", "lm_loss": 1.45058, "bpb": 2.09275, "chunk_loss": 0.033032, "F": 0.21753, "G": 0.43678, "chunks": 222.75, "bytes_per_chunk": 4.597, "boundary_entropy": 0.66932, "router_noise": 0.0, "grad_norm": 0.7171, "lr": 0.0001576, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 200.5} |
| {"event": "train", "step": 3300, "phase": 2, "objective": "global", "lm_loss": 1.41432, "bpb": 2.04044, "chunk_loss": 0.034241, "F": 0.23792, "G": 0.44043, "chunks": 243.62, "bytes_per_chunk": 4.203, "boundary_entropy": 0.67007, "router_noise": 0.0, "grad_norm": 0.8032, "lr": 0.0001539, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 202.6} |
| {"event": "train", "step": 3350, "phase": 2, "objective": "global", "lm_loss": 1.47646, "bpb": 2.13008, "chunk_loss": 0.033959, "F": 0.2334, "G": 0.43982, "chunks": 239.0, "bytes_per_chunk": 4.285, "boundary_entropy": 0.67038, "router_noise": 0.0, "grad_norm": 0.7735, "lr": 0.0001501, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 204.7} |
| {"event": "probe", "step": 3400, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4206.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.241682, "cps": 32.859375, "mean_p": 0.440357, "boundary_entropy": 0.662739, "bpb": 2.131026}, "de": {"chunks": 5209.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.253689, "cps": 40.695312, "mean_p": 0.44645, "boundary_entropy": 0.668999, "bpb": 2.347969}, "vi": {"chunks": 5995.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.325197, "cps": 46.835938, "mean_p": 0.452691, "boundary_entropy": 0.673036, "bpb": 1.991159}, "tr": {"chunks": 5698.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.314754, "cps": 44.515625, "mean_p": 0.455756, "boundary_entropy": 0.666879, "bpb": 2.345267}, "fi": {"chunks": 5314.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.282494, "cps": 41.515625, "mean_p": 0.451373, "boundary_entropy": 0.668625, "bpb": 2.552937}, "ru": {"chunks": 7422.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.391662, "cps": 57.984375, "mean_p": 0.436978, "boundary_entropy": 0.671672, "bpb": 1.503274}, "el": {"chunks": 7856.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.38279, "cps": 61.375, "mean_p": 0.457888, "boundary_entropy": 0.675926, "bpb": 1.588041}, "hi": {"chunks": 6653.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.395423, "cps": 51.976562, "mean_p": 0.4091, "boundary_entropy": 0.659001, "bpb": 1.25397}, "ta": {"chunks": 7030.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.376036, "cps": 54.921875, "mean_p": 0.413566, "boundary_entropy": 0.661377, "bpb": 1.035718}, "th": {"chunks": 10494.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.652409, "cps": 81.984375, "mean_p": 0.423177, "boundary_entropy": 0.662233, "bpb": 1.14505}, "zh": {"chunks": 3333.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.564341, "cps": 26.039062, "mean_p": 0.444161, "boundary_entropy": 0.670983, "bpb": 3.270004}, "my": {"chunks": 10745.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.567078, "cps": 83.945312, "mean_p": 0.451984, "boundary_entropy": 0.68107, "bpb": 1.146842}}, "bpb": {"en": 2.13103, "de": 2.34797, "vi": 1.99116, "tr": 2.34527, "fi": 2.55294, "ru": 1.50327, "el": 1.58804, "hi": 1.25397, "ta": 1.03572, "th": 1.14505, "zh": 3.27, "my": 1.14684}, "gini_cpc": 0.174141, "gini_cps": 0.179108, "bpic": 4.799962, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0496797090769232, "vi": 1.3455532740820872, "tr": 1.3023470940535782, "fi": 1.1688654413589767, "ru": 1.6205654944500694, "el": 1.5838552243983872, "hi": 1.6361280955035578, "ta": 1.5559108429901272, "th": 2.6994472570602768, "zh": 2.335052894780955, "my": 2.346377554969122}, "t": 207.8} |
| {"event": "eval", "step": 3400, "phase": 2, "bpb": {"en": 2.13938, "de": 2.35828, "vi": 1.99348, "tr": 2.35145, "fi": 2.5618, "ru": 1.481, "el": 1.58104, "hi": 1.2513, "ta": 1.04258, "th": 1.16612, "zh": 3.27278, "my": 1.15533}, "bytes_per_chunk": {"en": 3.948, "de": 3.956, "vi": 3.954, "tr": 3.413, "fi": 3.76, "ru": 4.613, "el": 4.898, "hi": 6.844, "ta": 6.965, "th": 4.356, "zh": 4.785, "my": 4.971}, "bpb_mean": 1.86288, "t": 209.2} |
| {"event": "train", "step": 3400, "phase": 2, "objective": "global", "lm_loss": 1.4021, "bpb": 2.0228, "chunk_loss": 0.033591, "F": 0.22714, "G": 0.43846, "chunks": 232.59, "bytes_per_chunk": 4.403, "boundary_entropy": 0.67076, "router_noise": 0.0, "grad_norm": 0.7325, "lr": 0.0001465, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 209.2} |
| {"event": "train", "step": 3450, "phase": 2, "objective": "global", "lm_loss": 1.26353, "bpb": 1.8229, "chunk_loss": 0.033288, "F": 0.22208, "G": 0.43787, "chunks": 227.41, "bytes_per_chunk": 4.503, "boundary_entropy": 0.66947, "router_noise": 0.0, "grad_norm": 0.6924, "lr": 0.0001428, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 211.4} |
| {"event": "train", "step": 3500, "phase": 2, "objective": "global", "lm_loss": 1.45292, "bpb": 2.09612, "chunk_loss": 0.033406, "F": 0.2251, "G": 0.43429, "chunks": 230.5, "bytes_per_chunk": 4.443, "boundary_entropy": 0.66802, "router_noise": 0.0, "grad_norm": 0.8221, "lr": 0.0001391, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 213.5} |
| {"event": "train", "step": 3550, "phase": 2, "objective": "global", "lm_loss": 1.38251, "bpb": 1.99454, "chunk_loss": 0.034078, "F": 0.23483, "G": 0.44226, "chunks": 240.47, "bytes_per_chunk": 4.258, "boundary_entropy": 0.67149, "router_noise": 0.0, "grad_norm": 0.7416, "lr": 0.0001355, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 215.6} |
| {"event": "probe", "step": 3600, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4036.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.231914, "cps": 31.53125, "mean_p": 0.436238, "boundary_entropy": 0.660924, "bpb": 2.115811}, "de": {"chunks": 5233.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.254858, "cps": 40.882812, "mean_p": 0.447269, "boundary_entropy": 0.669045, "bpb": 2.340997}, "vi": {"chunks": 5193.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.281692, "cps": 40.570312, "mean_p": 0.447604, "boundary_entropy": 0.671788, "bpb": 2.002307}, "tr": {"chunks": 5755.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.317903, "cps": 44.960938, "mean_p": 0.454435, "boundary_entropy": 0.666035, "bpb": 2.323751}, "fi": {"chunks": 5256.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.279411, "cps": 41.0625, "mean_p": 0.450884, "boundary_entropy": 0.66876, "bpb": 2.517867}, "ru": {"chunks": 5618.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.296464, "cps": 43.890625, "mean_p": 0.429651, "boundary_entropy": 0.670193, "bpb": 1.496896}, "el": {"chunks": 7547.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.367734, "cps": 58.960938, "mean_p": 0.457221, "boundary_entropy": 0.676742, "bpb": 1.559715}, "hi": {"chunks": 8142.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.483923, "cps": 63.609375, "mean_p": 0.415165, "boundary_entropy": 0.660905, "bpb": 1.238734}, "ta": {"chunks": 6834.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.365552, "cps": 53.390625, "mean_p": 0.411346, "boundary_entropy": 0.660919, "bpb": 1.028932}, "th": {"chunks": 9924.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.616972, "cps": 77.53125, "mean_p": 0.422764, "boundary_entropy": 0.663438, "bpb": 1.138836}, "zh": {"chunks": 2962.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.501524, "cps": 23.140625, "mean_p": 0.442076, "boundary_entropy": 0.670901, "bpb": 3.28231}, "my": {"chunks": 9245.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.487914, "cps": 72.226562, "mean_p": 0.448699, "boundary_entropy": 0.680284, "bpb": 1.137486}}, "bpb": {"en": 2.11581, "de": 2.341, "vi": 2.00231, "tr": 2.32375, "fi": 2.51787, "ru": 1.4969, "el": 1.55971, "hi": 1.23873, "ta": 1.02893, "th": 1.13884, "zh": 3.28231, "my": 1.13749}, "gini_cpc": 0.17068, "gini_cps": 0.176631, "bpic": 5.06675, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.098933190145651, "vi": 1.2146415781159152, "tr": 1.3707799363262536, "fi": 1.2048040971325058, "ru": 1.278337364772457, "el": 1.5856468439504192, "hi": 2.0866470119592315, "ta": 1.5762404473671168, "th": 2.6603492410395346, "zh": 2.1625421158775224, "my": 2.1038583793985857}, "t": 218.7} |
| {"event": "eval", "step": 3600, "phase": 2, "bpb": {"en": 2.12855, "de": 2.35171, "vi": 2.00341, "tr": 2.33165, "fi": 2.52797, "ru": 1.47287, "el": 1.55065, "hi": 1.2353, "ta": 1.03552, "th": 1.15881, "zh": 3.28141, "my": 1.14693}, "bytes_per_chunk": {"en": 4.287, "de": 3.944, "vi": 4.667, "tr": 3.424, "fi": 3.76, "ru": 6.089, "el": 5.064, "hi": 5.447, "ta": 7.468, "th": 4.607, "zh": 5.7, "my": 5.84}, "bpb_mean": 1.85206, "t": 220.0} |
| {"event": "train", "step": 3600, "phase": 2, "objective": "global", "lm_loss": 1.64615, "bpb": 2.37489, "chunk_loss": 0.033676, "F": 0.22711, "G": 0.44029, "chunks": 232.56, "bytes_per_chunk": 4.403, "boundary_entropy": 0.67054, "router_noise": 0.0, "grad_norm": 0.8001, "lr": 0.0001319, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 220.1} |
| {"event": "train", "step": 3650, "phase": 2, "objective": "global", "lm_loss": 1.44699, "bpb": 2.08756, "chunk_loss": 0.033787, "F": 0.23056, "G": 0.43877, "chunks": 236.09, "bytes_per_chunk": 4.337, "boundary_entropy": 0.6701, "router_noise": 0.0, "grad_norm": 0.8211, "lr": 0.0001283, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 222.2} |
| {"event": "train", "step": 3700, "phase": 2, "objective": "global", "lm_loss": 1.35942, "bpb": 1.96122, "chunk_loss": 0.033361, "F": 0.22363, "G": 0.43794, "chunks": 229.0, "bytes_per_chunk": 4.472, "boundary_entropy": 0.66991, "router_noise": 0.0, "grad_norm": 0.7436, "lr": 0.0001247, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 224.3} |
| {"event": "train", "step": 3750, "phase": 2, "objective": "global", "lm_loss": 1.40326, "bpb": 2.02448, "chunk_loss": 0.032924, "F": 0.21631, "G": 0.43762, "chunks": 221.5, "bytes_per_chunk": 4.623, "boundary_entropy": 0.67025, "router_noise": 0.0, "grad_norm": 0.7592, "lr": 0.0001212, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 226.4} |
| {"event": "probe", "step": 3800, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4163.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.239212, "cps": 32.523438, "mean_p": 0.438517, "boundary_entropy": 0.661212, "bpb": 2.108277}, "de": {"chunks": 5261.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.256222, "cps": 41.101562, "mean_p": 0.448306, "boundary_entropy": 0.669397, "bpb": 2.311723}, "vi": {"chunks": 5147.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.279197, "cps": 40.210938, "mean_p": 0.448025, "boundary_entropy": 0.6721, "bpb": 1.983539}, "tr": {"chunks": 5650.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.312103, "cps": 44.140625, "mean_p": 0.455167, "boundary_entropy": 0.666839, "bpb": 2.307545}, "fi": {"chunks": 5321.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.282866, "cps": 41.570312, "mean_p": 0.453863, "boundary_entropy": 0.669692, "bpb": 2.503975}, "ru": {"chunks": 7466.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.393984, "cps": 58.328125, "mean_p": 0.438464, "boundary_entropy": 0.672284, "bpb": 1.475088}, "el": {"chunks": 7842.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.382108, "cps": 61.265625, "mean_p": 0.460918, "boundary_entropy": 0.677721, "bpb": 1.540894}, "hi": {"chunks": 8894.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.528618, "cps": 69.484375, "mean_p": 0.420554, "boundary_entropy": 0.663167, "bpb": 1.222701}, "ta": {"chunks": 8168.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.436908, "cps": 63.8125, "mean_p": 0.419417, "boundary_entropy": 0.664391, "bpb": 1.014379}, "th": {"chunks": 11882.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.738701, "cps": 92.828125, "mean_p": 0.426554, "boundary_entropy": 0.662422, "bpb": 1.118091}, "zh": {"chunks": 3240.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.548595, "cps": 25.3125, "mean_p": 0.444225, "boundary_entropy": 0.671336, "bpb": 3.246796}, "my": {"chunks": 11076.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.584547, "cps": 86.53125, "mean_p": 0.443357, "boundary_entropy": 0.677875, "bpb": 1.136662}}, "bpb": {"en": 2.10828, "de": 2.31172, "vi": 1.98354, "tr": 2.30755, "fi": 2.50397, "ru": 1.47509, "el": 1.54089, "hi": 1.2227, "ta": 1.01438, "th": 1.11809, "zh": 3.2468, "my": 1.13666}, "gini_cpc": 0.199846, "gini_cps": 0.205507, "bpic": 4.562846, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.071108840816689, "vi": 1.1671555395110595, "tr": 1.304714850720294, "fi": 1.1824943784776523, "ru": 1.647010927399753, "el": 1.5973632999349032, "hi": 2.2098345611002466, "ta": 1.8264507620945534, "th": 3.0880632861798705, "zh": 2.293344387558173, "my": 2.4436403084746243}, "t": 229.5} |
| {"event": "eval", "step": 3800, "phase": 2, "bpb": {"en": 2.11958, "de": 2.32064, "vi": 1.98444, "tr": 2.31261, "fi": 2.5111, "ru": 1.45369, "el": 1.532, "hi": 1.21698, "ta": 1.02195, "th": 1.13672, "zh": 3.2467, "my": 1.14496}, "bytes_per_chunk": {"en": 4.074, "de": 3.883, "vi": 4.6, "tr": 3.467, "fi": 3.723, "ru": 4.59, "el": 4.86, "hi": 4.951, "ta": 6.22, "th": 3.929, "zh": 4.973, "my": 4.851}, "bpb_mean": 1.83345, "t": 230.9} |
| {"event": "train", "step": 3800, "phase": 2, "objective": "global", "lm_loss": 1.47653, "bpb": 2.13019, "chunk_loss": 0.033466, "F": 0.22467, "G": 0.44079, "chunks": 230.06, "bytes_per_chunk": 4.451, "boundary_entropy": 0.67186, "router_noise": 0.0, "grad_norm": 0.777, "lr": 0.0001177, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 230.9} |
| {"event": "train", "step": 3850, "phase": 2, "objective": "global", "lm_loss": 1.48013, "bpb": 2.13538, "chunk_loss": 0.032678, "F": 0.21234, "G": 0.4339, "chunks": 217.44, "bytes_per_chunk": 4.709, "boundary_entropy": 0.66948, "router_noise": 0.0, "grad_norm": 0.7707, "lr": 0.0001142, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 233.0} |
| {"event": "train", "step": 3900, "phase": 2, "objective": "global", "lm_loss": 1.47472, "bpb": 2.12757, "chunk_loss": 0.034329, "F": 0.2388, "G": 0.44142, "chunks": 244.53, "bytes_per_chunk": 4.188, "boundary_entropy": 0.67005, "router_noise": 0.0, "grad_norm": 0.7991, "lr": 0.0001108, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 235.2} |
| {"event": "train", "step": 3950, "phase": 2, "objective": "global", "lm_loss": 1.55392, "bpb": 2.24183, "chunk_loss": 0.032672, "F": 0.21155, "G": 0.43701, "chunks": 216.62, "bytes_per_chunk": 4.727, "boundary_entropy": 0.67054, "router_noise": 0.0, "grad_norm": 0.7622, "lr": 0.0001074, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 237.3} |
| {"event": "probe", "step": 4000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4165.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.239327, "cps": 32.539062, "mean_p": 0.437772, "boundary_entropy": 0.661069, "bpb": 2.086971}, "de": {"chunks": 5455.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.26567, "cps": 42.617188, "mean_p": 0.447729, "boundary_entropy": 0.668711, "bpb": 2.310466}, "vi": {"chunks": 5429.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.294494, "cps": 42.414062, "mean_p": 0.450769, "boundary_entropy": 0.672625, "bpb": 1.97141}, "tr": {"chunks": 5827.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.32188, "cps": 45.523438, "mean_p": 0.455131, "boundary_entropy": 0.666621, "bpb": 2.285743}, "fi": {"chunks": 5444.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.289405, "cps": 42.53125, "mean_p": 0.452843, "boundary_entropy": 0.668821, "bpb": 2.477434}, "ru": {"chunks": 7123.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.375884, "cps": 55.648438, "mean_p": 0.437199, "boundary_entropy": 0.672023, "bpb": 1.459402}, "el": {"chunks": 8280.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.40345, "cps": 64.6875, "mean_p": 0.459834, "boundary_entropy": 0.676704, "bpb": 1.525919}, "hi": {"chunks": 8784.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.52208, "cps": 68.625, "mean_p": 0.419796, "boundary_entropy": 0.663336, "bpb": 1.216306}, "ta": {"chunks": 7861.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.420487, "cps": 61.414062, "mean_p": 0.416874, "boundary_entropy": 0.662893, "bpb": 1.006336}, "th": {"chunks": 11392.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.708237, "cps": 89.0, "mean_p": 0.421579, "boundary_entropy": 0.659783, "bpb": 1.104227}, "zh": {"chunks": 3179.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.538266, "cps": 24.835938, "mean_p": 0.444475, "boundary_entropy": 0.671545, "bpb": 3.215812}, "my": {"chunks": 8839.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.466487, "cps": 69.054688, "mean_p": 0.445039, "boundary_entropy": 0.679447, "bpb": 1.125739}}, "bpb": {"en": 2.08697, "de": 2.31047, "vi": 1.97141, "tr": 2.28574, "fi": 2.47743, "ru": 1.4594, "el": 1.52592, "hi": 1.21631, "ta": 1.00634, "th": 1.10423, "zh": 3.21581, "my": 1.12574}, "gini_cpc": 0.179257, "gini_cps": 0.181985, "bpic": 4.692961, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1100728023152961, "vi": 1.2305118890517963, "tr": 1.344942077808862, "fi": 1.209247915848068, "ru": 1.5705900597706102, "el": 1.6857711071565011, "hi": 2.181455552800615, "ta": 1.7569582483969786, "th": 2.9592933996725828, "zh": 2.249086712233149, "my": 1.9491662022630458}, "t": 240.4} |
| {"event": "eval", "step": 4000, "phase": 2, "bpb": {"en": 2.09911, "de": 2.31926, "vi": 1.97246, "tr": 2.29241, "fi": 2.48809, "ru": 1.43805, "el": 1.51845, "hi": 1.21225, "ta": 1.0135, "th": 1.12337, "zh": 3.21566, "my": 1.13366}, "bytes_per_chunk": {"en": 4.085, "de": 3.808, "vi": 4.3, "tr": 3.367, "fi": 3.629, "ru": 4.727, "el": 4.643, "hi": 4.979, "ta": 6.425, "th": 4.093, "zh": 5.288, "my": 6.158}, "bpb_mean": 1.81886, "t": 241.7} |
| {"event": "train", "step": 4000, "phase": 2, "objective": "global", "lm_loss": 1.34419, "bpb": 1.93925, "chunk_loss": 0.033553, "F": 0.22729, "G": 0.43671, "chunks": 232.75, "bytes_per_chunk": 4.4, "boundary_entropy": 0.66837, "router_noise": 0.0, "grad_norm": 0.7424, "lr": 0.000104, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 241.8} |
| {"event": "train", "step": 4050, "phase": 2, "objective": "global", "lm_loss": 1.54261, "bpb": 2.22552, "chunk_loss": 0.033577, "F": 0.22668, "G": 0.43742, "chunks": 232.12, "bytes_per_chunk": 4.411, "boundary_entropy": 0.6685, "router_noise": 0.0, "grad_norm": 0.7446, "lr": 0.0001007, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 243.9} |
| {"event": "train", "step": 4100, "phase": 2, "objective": "global", "lm_loss": 1.40932, "bpb": 2.03322, "chunk_loss": 0.033785, "F": 0.23041, "G": 0.44035, "chunks": 235.94, "bytes_per_chunk": 4.34, "boundary_entropy": 0.66972, "router_noise": 0.0, "grad_norm": 0.7293, "lr": 9.75e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 246.0} |
| {"event": "train", "step": 4150, "phase": 2, "objective": "global", "lm_loss": 1.38609, "bpb": 1.9997, "chunk_loss": 0.034181, "F": 0.23712, "G": 0.43925, "chunks": 242.81, "bytes_per_chunk": 4.217, "boundary_entropy": 0.6684, "router_noise": 0.0, "grad_norm": 0.8245, "lr": 9.43e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 248.2} |
| {"event": "probe", "step": 4200, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4195.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.24105, "cps": 32.773438, "mean_p": 0.43772, "boundary_entropy": 0.660647, "bpb": 2.074889}, "de": {"chunks": 5446.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.265232, "cps": 42.546875, "mean_p": 0.44757, "boundary_entropy": 0.668724, "bpb": 2.282801}, "vi": {"chunks": 4904.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.266016, "cps": 38.3125, "mean_p": 0.44774, "boundary_entropy": 0.672114, "bpb": 1.963391}, "tr": {"chunks": 5991.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.33094, "cps": 46.804688, "mean_p": 0.455618, "boundary_entropy": 0.6656, "bpb": 2.257664}, "fi": {"chunks": 5758.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.306097, "cps": 44.984375, "mean_p": 0.455254, "boundary_entropy": 0.668868, "bpb": 2.457951}, "ru": {"chunks": 6165.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.32533, "cps": 48.164062, "mean_p": 0.432143, "boundary_entropy": 0.67075, "bpb": 1.457058}, "el": {"chunks": 7616.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.371096, "cps": 59.5, "mean_p": 0.456804, "boundary_entropy": 0.676373, "bpb": 1.517779}, "hi": {"chunks": 8831.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.524874, "cps": 68.992188, "mean_p": 0.415627, "boundary_entropy": 0.660562, "bpb": 1.208805}, "ta": {"chunks": 10449.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.558919, "cps": 81.632812, "mean_p": 0.417338, "boundary_entropy": 0.662045, "bpb": 0.991676}, "th": {"chunks": 10410.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.647187, "cps": 81.328125, "mean_p": 0.416595, "boundary_entropy": 0.658662, "bpb": 1.100989}, "zh": {"chunks": 3440.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.582459, "cps": 26.875, "mean_p": 0.445341, "boundary_entropy": 0.671354, "bpb": 3.208836}, "my": {"chunks": 11030.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.582119, "cps": 86.171875, "mean_p": 0.447894, "boundary_entropy": 0.679307, "bpb": 1.117431}}, "bpb": {"en": 2.07489, "de": 2.2828, "vi": 1.96339, "tr": 2.25766, "fi": 2.45795, "ru": 1.45706, "el": 1.51778, "hi": 1.20881, "ta": 0.99168, "th": 1.10099, "zh": 3.20884, "my": 1.11743}, "gini_cpc": 0.190586, "gini_cps": 0.19884, "bpic": 4.556075, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1003158902263033, "vi": 1.1035689548660648, "tr": 1.3729063799077943, "fi": 1.269848563868084, "ru": 1.3496340347379248, "el": 1.5394948671501267, "hi": 2.1774438613859637, "ta": 2.3186831965740744, "th": 2.6848610795672423, "zh": 2.4163350577401137, "my": 2.414928580648424}, "t": 251.3} |
| {"event": "eval", "step": 4200, "phase": 2, "bpb": {"en": 2.08564, "de": 2.29307, "vi": 1.96463, "tr": 2.26428, "fi": 2.46634, "ru": 1.43621, "el": 1.50916, "hi": 1.20459, "ta": 0.99938, "th": 1.12025, "zh": 3.206, "my": 1.12581}, "bytes_per_chunk": {"en": 4.007, "de": 3.772, "vi": 4.78, "tr": 3.249, "fi": 3.398, "ru": 5.491, "el": 4.927, "hi": 4.946, "ta": 4.813, "th": 4.429, "zh": 4.824, "my": 4.864}, "bpb_mean": 1.80628, "t": 252.6} |
| {"event": "phase_transition", "step": 4200, "phase": 3, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 252.6} |
| {"event": "train", "step": 4200, "phase": 3, "objective": "global", "lm_loss": 1.20191, "bpb": 1.73398, "chunk_loss": 0.032992, "F": 0.21814, "G": 0.43389, "chunks": 223.38, "bytes_per_chunk": 4.584, "boundary_entropy": 0.66894, "router_noise": 0.0, "grad_norm": 0.6938, "lr": 9.12e-05, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 252.7} |
| {"event": "train", "step": 4250, "phase": 3, "objective": "global", "lm_loss": 1.25435, "bpb": 1.80964, "chunk_loss": 0.032774, "F": 0.21396, "G": 0.43242, "chunks": 219.09, "bytes_per_chunk": 4.674, "boundary_entropy": 0.66723, "router_noise": 0.0, "grad_norm": 0.7552, "lr": 8.81e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 254.8} |
| {"event": "train", "step": 4300, "phase": 3, "objective": "global", "lm_loss": 1.47913, "bpb": 2.13394, "chunk_loss": 0.034388, "F": 0.23999, "G": 0.43848, "chunks": 245.75, "bytes_per_chunk": 4.167, "boundary_entropy": 0.66756, "router_noise": 0.0, "grad_norm": 0.7685, "lr": 8.5e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 256.9} |
| {"event": "train", "step": 4350, "phase": 3, "objective": "global", "lm_loss": 1.47517, "bpb": 2.12821, "chunk_loss": 0.033641, "F": 0.22812, "G": 0.43896, "chunks": 233.59, "bytes_per_chunk": 4.384, "boundary_entropy": 0.67072, "router_noise": 0.0, "grad_norm": 0.857, "lr": 8.21e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 259.0} |
| {"event": "probe", "step": 4400, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3951.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.22703, "cps": 30.867188, "mean_p": 0.434426, "boundary_entropy": 0.659848, "bpb": 2.069637}, "de": {"chunks": 5100.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.248381, "cps": 39.84375, "mean_p": 0.445475, "boundary_entropy": 0.668186, "bpb": 2.278965}, "vi": {"chunks": 5018.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.2722, "cps": 39.203125, "mean_p": 0.446963, "boundary_entropy": 0.671559, "bpb": 1.957653}, "tr": {"chunks": 5697.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.314699, "cps": 44.507812, "mean_p": 0.454185, "boundary_entropy": 0.665763, "bpb": 2.242852}, "fi": {"chunks": 5442.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.289299, "cps": 42.515625, "mean_p": 0.452615, "boundary_entropy": 0.668485, "bpb": 2.441954}, "ru": {"chunks": 6440.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.339842, "cps": 50.3125, "mean_p": 0.432772, "boundary_entropy": 0.670701, "bpb": 1.432322}, "el": {"chunks": 7073.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.344638, "cps": 55.257812, "mean_p": 0.455207, "boundary_entropy": 0.676381, "bpb": 1.502588}, "hi": {"chunks": 7941.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.471976, "cps": 62.039062, "mean_p": 0.41421, "boundary_entropy": 0.660974, "bpb": 1.197784}, "ta": {"chunks": 8589.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.459428, "cps": 67.101562, "mean_p": 0.415358, "boundary_entropy": 0.662142, "bpb": 0.98532}, "th": {"chunks": 10477.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.651352, "cps": 81.851562, "mean_p": 0.418426, "boundary_entropy": 0.659779, "bpb": 1.087932}, "zh": {"chunks": 3156.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.534372, "cps": 24.65625, "mean_p": 0.443482, "boundary_entropy": 0.671201, "bpb": 3.187112}, "my": {"chunks": 9585.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.505858, "cps": 74.882812, "mean_p": 0.438604, "boundary_entropy": 0.676742, "bpb": 1.104345}}, "bpb": {"en": 2.06964, "de": 2.27897, "vi": 1.95765, "tr": 2.24285, "fi": 2.44195, "ru": 1.43232, "el": 1.50259, "hi": 1.19778, "ta": 0.98532, "th": 1.08793, "zh": 3.18711, "my": 1.10435}, "gini_cpc": 0.182966, "gini_cps": 0.18699, "bpic": 4.890861, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.094044178231009, "vi": 1.1989597549641364, "tr": 1.3861580726938751, "fi": 1.2742767069478687, "ru": 1.4969032922429042, "el": 1.5180284238388613, "hi": 2.0789173015803173, "ta": 2.02364451289462, "th": 2.8690159930079124, "zh": 2.3537516768013997, "my": 2.2281572201199973}, "t": 262.1} |
| {"event": "eval", "step": 4400, "phase": 3, "bpb": {"en": 2.08081, "de": 2.28755, "vi": 1.95852, "tr": 2.25009, "fi": 2.44886, "ru": 1.41146, "el": 1.49506, "hi": 1.19341, "ta": 0.99272, "th": 1.1078, "zh": 3.18742, "my": 1.11247}, "bytes_per_chunk": {"en": 4.36, "de": 4.023, "vi": 4.712, "tr": 3.433, "fi": 3.64, "ru": 5.271, "el": 5.328, "hi": 5.608, "ta": 5.89, "th": 4.38, "zh": 5.298, "my": 5.626}, "bpb_mean": 1.79385, "t": 263.4} |
| {"event": "train", "step": 4400, "phase": 3, "objective": "global", "lm_loss": 1.47979, "bpb": 2.13489, "chunk_loss": 0.033918, "F": 0.23273, "G": 0.43771, "chunks": 238.31, "bytes_per_chunk": 4.297, "boundary_entropy": 0.66856, "router_noise": 0.0, "grad_norm": 0.8115, "lr": 7.92e-05, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 263.5} |
| {"event": "train", "step": 4450, "phase": 3, "objective": "global", "lm_loss": 1.49683, "bpb": 2.15947, "chunk_loss": 0.033433, "F": 0.22528, "G": 0.43594, "chunks": 230.69, "bytes_per_chunk": 4.439, "boundary_entropy": 0.66948, "router_noise": 0.0, "grad_norm": 0.7615, "lr": 7.63e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 265.6} |
| {"event": "train", "step": 4500, "phase": 3, "objective": "global", "lm_loss": 1.42249, "bpb": 2.05221, "chunk_loss": 0.034542, "F": 0.24237, "G": 0.44266, "chunks": 248.19, "bytes_per_chunk": 4.126, "boundary_entropy": 0.67049, "router_noise": 0.0, "grad_norm": 0.8102, "lr": 7.36e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 267.7} |
| {"event": "train", "step": 4550, "phase": 3, "objective": "global", "lm_loss": 1.23271, "bpb": 1.77843, "chunk_loss": 0.033407, "F": 0.22467, "G": 0.43461, "chunks": 230.06, "bytes_per_chunk": 4.451, "boundary_entropy": 0.66805, "router_noise": 0.0, "grad_norm": 0.7386, "lr": 7.09e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 269.8} |
| {"event": "probe", "step": 4600, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3885.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.223237, "cps": 30.351562, "mean_p": 0.433853, "boundary_entropy": 0.660151, "bpb": 2.059297}, "de": {"chunks": 5172.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.251887, "cps": 40.40625, "mean_p": 0.445068, "boundary_entropy": 0.668265, "bpb": 2.265301}, "vi": {"chunks": 4707.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.25533, "cps": 36.773438, "mean_p": 0.444663, "boundary_entropy": 0.671486, "bpb": 1.951068}, "tr": {"chunks": 5762.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.31829, "cps": 45.015625, "mean_p": 0.454494, "boundary_entropy": 0.666073, "bpb": 2.231994}, "fi": {"chunks": 5426.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.288448, "cps": 42.390625, "mean_p": 0.452159, "boundary_entropy": 0.668499, "bpb": 2.42229}, "ru": {"chunks": 6318.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.333404, "cps": 49.359375, "mean_p": 0.434131, "boundary_entropy": 0.671399, "bpb": 1.428729}, "el": {"chunks": 7505.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.365687, "cps": 58.632812, "mean_p": 0.456104, "boundary_entropy": 0.676316, "bpb": 1.496909}, "hi": {"chunks": 7727.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.459257, "cps": 60.367188, "mean_p": 0.41561, "boundary_entropy": 0.661866, "bpb": 1.194812}, "ta": {"chunks": 8222.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.439797, "cps": 64.234375, "mean_p": 0.416457, "boundary_entropy": 0.663209, "bpb": 0.98516}, "th": {"chunks": 10377.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.645135, "cps": 81.070312, "mean_p": 0.414083, "boundary_entropy": 0.657025, "bpb": 1.078556}, "zh": {"chunks": 3170.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.536742, "cps": 24.765625, "mean_p": 0.444651, "boundary_entropy": 0.671674, "bpb": 3.180004}, "my": {"chunks": 7912.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.417564, "cps": 61.8125, "mean_p": 0.434456, "boundary_entropy": 0.676041, "bpb": 1.104805}}, "bpb": {"en": 2.0593, "de": 2.2653, "vi": 1.95107, "tr": 2.23199, "fi": 2.42229, "ru": 1.42873, "el": 1.49691, "hi": 1.19481, "ta": 0.98516, "th": 1.07856, "zh": 3.18, "my": 1.10481}, "gini_cpc": 0.1786, "gini_cps": 0.17536, "bpic": 5.03762, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1283379782089178, "vi": 1.1437580228141662, "tr": 1.425790772109062, "fi": 1.2921145087005488, "ru": 1.4934940760587199, "el": 1.6381095736453142, "hi": 2.057258836420798, "ta": 1.9700856153597528, "th": 2.8899068774356307, "zh": 2.404356802257243, "my": 1.8704926223608935}, "t": 272.8} |
| {"event": "eval", "step": 4600, "phase": 3, "bpb": {"en": 2.07013, "de": 2.275, "vi": 1.95064, "tr": 2.23991, "fi": 2.43061, "ru": 1.40787, "el": 1.48862, "hi": 1.19039, "ta": 0.99222, "th": 1.09738, "zh": 3.17642, "my": 1.11304}, "bytes_per_chunk": {"en": 4.483, "de": 4.006, "vi": 5.08, "tr": 3.385, "fi": 3.665, "ru": 5.429, "el": 5.084, "hi": 5.86, "ta": 6.348, "th": 4.447, "zh": 5.198, "my": 6.702}, "bpb_mean": 1.78602, "t": 274.2} |
| {"event": "train", "step": 4600, "phase": 3, "objective": "global", "lm_loss": 1.31192, "bpb": 1.8927, "chunk_loss": 0.032926, "F": 0.21567, "G": 0.43782, "chunks": 220.84, "bytes_per_chunk": 4.637, "boundary_entropy": 0.67044, "router_noise": 0.0, "grad_norm": 0.7557, "lr": 6.82e-05, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 274.2} |
| {"event": "train", "step": 4650, "phase": 3, "objective": "global", "lm_loss": 1.41235, "bpb": 2.03759, "chunk_loss": 0.03278, "F": 0.21344, "G": 0.43666, "chunks": 218.56, "bytes_per_chunk": 4.685, "boundary_entropy": 0.6697, "router_noise": 0.0, "grad_norm": 0.7471, "lr": 6.57e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 276.4} |
| {"event": "train", "step": 4700, "phase": 3, "objective": "global", "lm_loss": 1.3248, "bpb": 1.91129, "chunk_loss": 0.034409, "F": 0.2403, "G": 0.44053, "chunks": 246.06, "bytes_per_chunk": 4.162, "boundary_entropy": 0.66939, "router_noise": 0.0, "grad_norm": 0.8049, "lr": 6.32e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 278.5} |
| {"event": "train", "step": 4750, "phase": 3, "objective": "global", "lm_loss": 1.35066, "bpb": 1.94859, "chunk_loss": 0.033756, "F": 0.22931, "G": 0.43947, "chunks": 234.81, "bytes_per_chunk": 4.361, "boundary_entropy": 0.66976, "router_noise": 0.0, "grad_norm": 0.8072, "lr": 6.08e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 280.7} |
| {"event": "probe", "step": 4800, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4014.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.23065, "cps": 31.359375, "mean_p": 0.435094, "boundary_entropy": 0.659994, "bpb": 2.049668}, "de": {"chunks": 5349.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.260507, "cps": 41.789062, "mean_p": 0.446907, "boundary_entropy": 0.668615, "bpb": 2.250814}, "vi": {"chunks": 4983.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.270301, "cps": 38.929688, "mean_p": 0.446676, "boundary_entropy": 0.67179, "bpb": 1.943901}, "tr": {"chunks": 5596.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.30912, "cps": 43.71875, "mean_p": 0.454236, "boundary_entropy": 0.665845, "bpb": 2.223372}, "fi": {"chunks": 5427.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.288501, "cps": 42.398438, "mean_p": 0.452536, "boundary_entropy": 0.668598, "bpb": 2.41185}, "ru": {"chunks": 6798.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.358734, "cps": 53.109375, "mean_p": 0.433989, "boundary_entropy": 0.671002, "bpb": 1.413351}, "el": {"chunks": 7644.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.37246, "cps": 59.71875, "mean_p": 0.456745, "boundary_entropy": 0.676339, "bpb": 1.478518}, "hi": {"chunks": 7315.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.43477, "cps": 57.148438, "mean_p": 0.413093, "boundary_entropy": 0.661037, "bpb": 1.189112}, "ta": {"chunks": 8606.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.460337, "cps": 67.234375, "mean_p": 0.416468, "boundary_entropy": 0.662836, "bpb": 0.977487}, "th": {"chunks": 10874.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.676034, "cps": 84.953125, "mean_p": 0.412865, "boundary_entropy": 0.655102, "bpb": 1.073034}, "zh": {"chunks": 3201.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.541991, "cps": 25.007812, "mean_p": 0.443595, "boundary_entropy": 0.671111, "bpb": 3.172315}, "my": {"chunks": 9383.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.495197, "cps": 73.304688, "mean_p": 0.435764, "boundary_entropy": 0.675807, "bpb": 1.093577}}, "bpb": {"en": 2.04967, "de": 2.25081, "vi": 1.9439, "tr": 2.22337, "fi": 2.41185, "ru": 1.41335, "el": 1.47852, "hi": 1.18911, "ta": 0.97749, "th": 1.07303, "zh": 3.17231, "my": 1.09358}, "gini_cpc": 0.181911, "gini_cps": 0.185657, "bpic": 4.846332, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1294498258190522, "vi": 1.1719106398557044, "tr": 1.3402132474471153, "fi": 1.2508196353960437, "ru": 1.5553162085734231, "el": 1.6148291679312436, "hi": 1.8849768011069727, "ta": 1.9958257625457196, "th": 2.930994580825984, "zh": 2.349843740271731, "my": 2.146965631373158}, "t": 283.7} |
| {"event": "eval", "step": 4800, "phase": 3, "bpb": {"en": 2.06088, "de": 2.26026, "vi": 1.94272, "tr": 2.23017, "fi": 2.41953, "ru": 1.39235, "el": 1.46962, "hi": 1.18526, "ta": 0.98489, "th": 1.09074, "zh": 3.16668, "my": 1.1011}, "bytes_per_chunk": {"en": 4.261, "de": 3.866, "vi": 4.758, "tr": 3.477, "fi": 3.69, "ru": 5.049, "el": 4.981, "hi": 6.16, "ta": 6.016, "th": 4.269, "zh": 5.187, "my": 5.768}, "bpb_mean": 1.77535, "t": 285.1} |
| {"event": "train", "step": 4800, "phase": 3, "objective": "global", "lm_loss": 1.25875, "bpb": 1.816, "chunk_loss": 0.033169, "F": 0.22089, "G": 0.43359, "chunks": 226.19, "bytes_per_chunk": 4.527, "boundary_entropy": 0.66733, "router_noise": 0.0, "grad_norm": 0.7442, "lr": 5.85e-05, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 285.1} |
| {"event": "train", "step": 4850, "phase": 3, "objective": "global", "lm_loss": 1.49465, "bpb": 2.15632, "chunk_loss": 0.033993, "F": 0.23428, "G": 0.43892, "chunks": 239.91, "bytes_per_chunk": 4.268, "boundary_entropy": 0.66872, "router_noise": 0.0, "grad_norm": 0.8122, "lr": 5.62e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 287.3} |
| {"event": "train", "step": 4900, "phase": 3, "objective": "global", "lm_loss": 1.40874, "bpb": 2.03238, "chunk_loss": 0.033116, "F": 0.21942, "G": 0.43674, "chunks": 224.69, "bytes_per_chunk": 4.557, "boundary_entropy": 0.67026, "router_noise": 0.0, "grad_norm": 0.7514, "lr": 5.41e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 289.4} |
| {"event": "train", "step": 4950, "phase": 3, "objective": "global", "lm_loss": 1.52416, "bpb": 2.19889, "chunk_loss": 0.033442, "F": 0.22501, "G": 0.43731, "chunks": 230.41, "bytes_per_chunk": 4.444, "boundary_entropy": 0.66892, "router_noise": 0.0, "grad_norm": 0.8115, "lr": 5.2e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 291.6} |
| {"event": "probe", "step": 5000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4134.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.237545, "cps": 32.296875, "mean_p": 0.436298, "boundary_entropy": 0.660138, "bpb": 2.040441}, "de": {"chunks": 5518.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.268738, "cps": 43.109375, "mean_p": 0.447049, "boundary_entropy": 0.66805, "bpb": 2.241297}, "vi": {"chunks": 5038.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.273285, "cps": 39.359375, "mean_p": 0.445977, "boundary_entropy": 0.671568, "bpb": 1.939052}, "tr": {"chunks": 5819.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.321438, "cps": 45.460938, "mean_p": 0.455614, "boundary_entropy": 0.665542, "bpb": 2.210824}, "fi": {"chunks": 5390.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.286534, "cps": 42.109375, "mean_p": 0.451115, "boundary_entropy": 0.667992, "bpb": 2.400161}, "ru": {"chunks": 6881.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.363113, "cps": 53.757812, "mean_p": 0.435199, "boundary_entropy": 0.671429, "bpb": 1.407383}, "el": {"chunks": 7636.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.37207, "cps": 59.65625, "mean_p": 0.458316, "boundary_entropy": 0.67672, "bpb": 1.473743}, "hi": {"chunks": 7992.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.475007, "cps": 62.4375, "mean_p": 0.419358, "boundary_entropy": 0.664291, "bpb": 1.182704}, "ta": {"chunks": 8640.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.462156, "cps": 67.5, "mean_p": 0.418367, "boundary_entropy": 0.664002, "bpb": 0.970443}, "th": {"chunks": 10485.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.65185, "cps": 81.914062, "mean_p": 0.410506, "boundary_entropy": 0.654305, "bpb": 1.067824}, "zh": {"chunks": 3078.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.521165, "cps": 24.046875, "mean_p": 0.443247, "boundary_entropy": 0.671287, "bpb": 3.168789}, "my": {"chunks": 9334.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.492611, "cps": 72.921875, "mean_p": 0.434938, "boundary_entropy": 0.675332, "bpb": 1.086038}}, "bpb": {"en": 2.04044, "de": 2.2413, "vi": 1.93905, "tr": 2.21082, "fi": 2.40016, "ru": 1.40738, "el": 1.47374, "hi": 1.1827, "ta": 0.97044, "th": 1.06782, "zh": 3.16879, "my": 1.08604}, "gini_cpc": 0.171726, "gini_cps": 0.181288, "bpic": 4.800563, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1313134147678447, "vi": 1.1504524389029356, "tr": 1.353167173456003, "fi": 1.2062311223731228, "ru": 1.5286075188315442, "el": 1.5663136130959163, "hi": 1.9996502888739847, "ta": 1.9455478692990489, "th": 2.7441068471062433, "zh": 2.1939605832449054, "my": 2.073757971170353}, "t": 294.7} |
| {"event": "eval", "step": 5000, "phase": 3, "bpb": {"en": 2.05101, "de": 2.24935, "vi": 1.93916, "tr": 2.21936, "fi": 2.40881, "ru": 1.38503, "el": 1.46542, "hi": 1.178, "ta": 0.97811, "th": 1.08673, "zh": 3.16512, "my": 1.09404}, "bytes_per_chunk": {"en": 4.157, "de": 3.766, "vi": 4.678, "tr": 3.336, "fi": 3.741, "ru": 4.972, "el": 5.02, "hi": 5.631, "ta": 5.945, "th": 4.406, "zh": 5.461, "my": 5.721}, "bpb_mean": 1.76834, "t": 296.0} |
| {"event": "train", "step": 5000, "phase": 3, "objective": "global", "lm_loss": 1.29238, "bpb": 1.86451, "chunk_loss": 0.033435, "F": 0.2254, "G": 0.43822, "chunks": 230.81, "bytes_per_chunk": 4.437, "boundary_entropy": 0.6692, "router_noise": 0.0, "grad_norm": 0.7981, "lr": 5e-05, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 296.1} |
| {"event": "train", "step": 5050, "phase": 3, "objective": "global", "lm_loss": 1.2645, "bpb": 1.82428, "chunk_loss": 0.033522, "F": 0.22656, "G": 0.43444, "chunks": 232.0, "bytes_per_chunk": 4.414, "boundary_entropy": 0.66704, "router_noise": 0.0, "grad_norm": 0.7978, "lr": 4.81e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 298.2} |
| {"event": "train", "step": 5100, "phase": 3, "objective": "global", "lm_loss": 1.30584, "bpb": 1.88393, "chunk_loss": 0.033821, "F": 0.23178, "G": 0.4352, "chunks": 237.34, "bytes_per_chunk": 4.314, "boundary_entropy": 0.66668, "router_noise": 0.0, "grad_norm": 0.7767, "lr": 4.63e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 300.3} |
| {"event": "train", "step": 5150, "phase": 3, "objective": "global", "lm_loss": 1.35201, "bpb": 1.95053, "chunk_loss": 0.033109, "F": 0.21991, "G": 0.43443, "chunks": 225.19, "bytes_per_chunk": 4.547, "boundary_entropy": 0.66843, "router_noise": 0.0, "grad_norm": 0.7892, "lr": 4.45e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 302.5} |
| {"event": "probe", "step": 5200, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4017.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.230822, "cps": 31.382812, "mean_p": 0.433507, "boundary_entropy": 0.659308, "bpb": 2.038919}, "de": {"chunks": 5252.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.255783, "cps": 41.03125, "mean_p": 0.444782, "boundary_entropy": 0.66759, "bpb": 2.233667}, "vi": {"chunks": 4795.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.260103, "cps": 37.460938, "mean_p": 0.445445, "boundary_entropy": 0.67129, "bpb": 1.939912}, "tr": {"chunks": 5850.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.323151, "cps": 45.703125, "mean_p": 0.455104, "boundary_entropy": 0.665358, "bpb": 2.203731}, "fi": {"chunks": 5399.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.287013, "cps": 42.179688, "mean_p": 0.450872, "boundary_entropy": 0.667658, "bpb": 2.387824}, "ru": {"chunks": 6937.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.366069, "cps": 54.195312, "mean_p": 0.435068, "boundary_entropy": 0.671291, "bpb": 1.395929}, "el": {"chunks": 7338.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.35755, "cps": 57.328125, "mean_p": 0.456488, "boundary_entropy": 0.676424, "bpb": 1.468351}, "hi": {"chunks": 7886.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.468707, "cps": 61.609375, "mean_p": 0.416774, "boundary_entropy": 0.662222, "bpb": 1.175706}, "ta": {"chunks": 8317.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.444878, "cps": 64.976562, "mean_p": 0.417745, "boundary_entropy": 0.663494, "bpb": 0.970252}, "th": {"chunks": 10749.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.668262, "cps": 83.976562, "mean_p": 0.412162, "boundary_entropy": 0.654829, "bpb": 1.055727}, "zh": {"chunks": 3288.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.556722, "cps": 25.6875, "mean_p": 0.444747, "boundary_entropy": 0.671525, "bpb": 3.138152}, "my": {"chunks": 9994.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.527444, "cps": 78.078125, "mean_p": 0.434494, "boundary_entropy": 0.674671, "bpb": 1.081006}}, "bpb": {"en": 2.03892, "de": 2.23367, "vi": 1.93991, "tr": 2.20373, "fi": 2.38782, "ru": 1.39593, "el": 1.46835, "hi": 1.17571, "ta": 0.97025, "th": 1.05573, "zh": 3.13815, "my": 1.08101}, "gini_cpc": 0.187398, "gini_cps": 0.188536, "bpic": 4.80796, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1081399159014214, "vi": 1.1268542785918187, "tr": 1.3999986056057865, "fi": 1.2434368462766432, "ru": 1.585932754132667, "el": 1.5490275814704624, "hi": 2.03059815990066, "ta": 1.927363013158239, "th": 2.895138109413606, "zh": 2.4119075372268606, "my": 2.2850634171459077}, "t": 305.6} |
| {"event": "eval", "step": 5200, "phase": 3, "bpb": {"en": 2.04978, "de": 2.2424, "vi": 1.94005, "tr": 2.20945, "fi": 2.39728, "ru": 1.37388, "el": 1.46155, "hi": 1.17116, "ta": 0.97814, "th": 1.07493, "zh": 3.13595, "my": 1.08868}, "bytes_per_chunk": {"en": 4.327, "de": 3.911, "vi": 4.916, "tr": 3.33, "fi": 3.716, "ru": 4.948, "el": 5.206, "hi": 5.84, "ta": 6.224, "th": 4.307, "zh": 5.098, "my": 5.399}, "bpb_mean": 1.76027, "t": 307.0} |
| {"event": "train", "step": 5200, "phase": 3, "objective": "global", "lm_loss": 1.29538, "bpb": 1.86883, "chunk_loss": 0.032781, "F": 0.21417, "G": 0.43614, "chunks": 219.31, "bytes_per_chunk": 4.669, "boundary_entropy": 0.66978, "router_noise": 0.0, "grad_norm": 0.8479, "lr": 4.29e-05, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 307.0} |
| {"event": "train", "step": 5250, "phase": 3, "objective": "global", "lm_loss": 1.38001, "bpb": 1.99093, "chunk_loss": 0.033574, "F": 0.22702, "G": 0.4356, "chunks": 232.47, "bytes_per_chunk": 4.405, "boundary_entropy": 0.66795, "router_noise": 0.0, "grad_norm": 0.7479, "lr": 4.14e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 309.2} |
| {"event": "train", "step": 5300, "phase": 3, "objective": "global", "lm_loss": 1.3799, "bpb": 1.99077, "chunk_loss": 0.033148, "F": 0.22037, "G": 0.43608, "chunks": 225.66, "bytes_per_chunk": 4.538, "boundary_entropy": 0.6688, "router_noise": 0.0, "grad_norm": 0.7975, "lr": 3.99e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 311.3} |
| {"event": "train", "step": 5350, "phase": 3, "objective": "global", "lm_loss": 1.43634, "bpb": 2.07221, "chunk_loss": 0.033712, "F": 0.22958, "G": 0.43846, "chunks": 235.09, "bytes_per_chunk": 4.356, "boundary_entropy": 0.66953, "router_noise": 0.0, "grad_norm": 0.8308, "lr": 3.86e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 313.4} |
| {"event": "probe", "step": 5400, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4002.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.22996, "cps": 31.265625, "mean_p": 0.434003, "boundary_entropy": 0.659469, "bpb": 2.029419}, "de": {"chunks": 5446.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.265232, "cps": 42.546875, "mean_p": 0.446068, "boundary_entropy": 0.667912, "bpb": 2.223226}, "vi": {"chunks": 4832.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.26211, "cps": 37.75, "mean_p": 0.445715, "boundary_entropy": 0.671277, "bpb": 1.929851}, "tr": {"chunks": 5777.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.319118, "cps": 45.132812, "mean_p": 0.45508, "boundary_entropy": 0.665737, "bpb": 2.195411}, "fi": {"chunks": 5451.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.289777, "cps": 42.585938, "mean_p": 0.452413, "boundary_entropy": 0.668355, "bpb": 2.376171}, "ru": {"chunks": 6511.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.343588, "cps": 50.867188, "mean_p": 0.434219, "boundary_entropy": 0.671269, "bpb": 1.392768}, "el": {"chunks": 7575.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.369098, "cps": 59.179688, "mean_p": 0.457902, "boundary_entropy": 0.676763, "bpb": 1.458284}, "hi": {"chunks": 8438.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.501516, "cps": 65.921875, "mean_p": 0.417487, "boundary_entropy": 0.662358, "bpb": 1.165657}, "ta": {"chunks": 10022.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.536079, "cps": 78.296875, "mean_p": 0.418702, "boundary_entropy": 0.663103, "bpb": 0.961573}, "th": {"chunks": 10401.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.646627, "cps": 81.257812, "mean_p": 0.409365, "boundary_entropy": 0.653109, "bpb": 1.050114}, "zh": {"chunks": 3257.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.551473, "cps": 25.445312, "mean_p": 0.444252, "boundary_entropy": 0.671345, "bpb": 3.138964}, "my": {"chunks": 9960.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.525649, "cps": 77.8125, "mean_p": 0.435142, "boundary_entropy": 0.675012, "bpb": 1.077148}}, "bpb": {"en": 2.02942, "de": 2.22323, "vi": 1.92985, "tr": 2.19541, "fi": 2.37617, "ru": 1.39277, "el": 1.45828, "hi": 1.16566, "ta": 0.96157, "th": 1.05011, "zh": 3.13896, "my": 1.07715}, "gini_cpc": 0.186555, "gini_cps": 0.194606, "bpic": 4.699052, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1533796000747982, "vi": 1.1398056870669628, "tr": 1.3877104286060582, "fi": 1.260118346015446, "ru": 1.4941201299086606, "el": 1.605050968157202, "hi": 2.180878565175065, "ta": 2.331180839107058, "th": 2.8119077390120606, "zh": 2.398122433872803, "my": 2.285825105180152}, "t": 316.5} |
| {"event": "eval", "step": 5400, "phase": 3, "bpb": {"en": 2.04033, "de": 2.23208, "vi": 1.92915, "tr": 2.20152, "fi": 2.38607, "ru": 1.37095, "el": 1.44989, "hi": 1.16096, "ta": 0.9691, "th": 1.06988, "zh": 3.13469, "my": 1.08468}, "bytes_per_chunk": {"en": 4.249, "de": 3.78, "vi": 4.882, "tr": 3.371, "fi": 3.654, "ru": 5.197, "el": 5.05, "hi": 5.395, "ta": 5.145, "th": 4.429, "zh": 5.145, "my": 5.422}, "bpb_mean": 1.75244, "t": 317.9} |
| {"event": "train", "step": 5400, "phase": 3, "objective": "global", "lm_loss": 1.1349, "bpb": 1.63732, "chunk_loss": 0.032585, "F": 0.21133, "G": 0.43276, "chunks": 216.41, "bytes_per_chunk": 4.732, "boundary_entropy": 0.66883, "router_noise": 0.0, "grad_norm": 0.7385, "lr": 3.73e-05, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 317.9} |
| {"event": "train", "step": 5450, "phase": 3, "objective": "global", "lm_loss": 1.19715, "bpb": 1.72712, "chunk_loss": 0.033152, "F": 0.22061, "G": 0.43618, "chunks": 225.91, "bytes_per_chunk": 4.533, "boundary_entropy": 0.66924, "router_noise": 0.0, "grad_norm": 0.79, "lr": 3.62e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 320.1} |
| {"event": "train", "step": 5500, "phase": 3, "objective": "global", "lm_loss": 1.39881, "bpb": 2.01806, "chunk_loss": 0.033981, "F": 0.23392, "G": 0.43594, "chunks": 239.53, "bytes_per_chunk": 4.275, "boundary_entropy": 0.66712, "router_noise": 0.0, "grad_norm": 0.822, "lr": 3.51e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 322.2} |
| {"event": "train", "step": 5550, "phase": 3, "objective": "global", "lm_loss": 1.41096, "bpb": 2.03558, "chunk_loss": 0.033877, "F": 0.23199, "G": 0.4376, "chunks": 237.56, "bytes_per_chunk": 4.31, "boundary_entropy": 0.66849, "router_noise": 0.0, "grad_norm": 0.8388, "lr": 3.41e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 324.3} |
| {"event": "probe", "step": 5600, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3886.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.223295, "cps": 30.359375, "mean_p": 0.432518, "boundary_entropy": 0.659209, "bpb": 2.027779}, "de": {"chunks": 5335.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.259826, "cps": 41.679688, "mean_p": 0.44522, "boundary_entropy": 0.667845, "bpb": 2.218721}, "vi": {"chunks": 4847.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.262924, "cps": 37.867188, "mean_p": 0.446091, "boundary_entropy": 0.671319, "bpb": 1.927715}, "tr": {"chunks": 5697.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.314699, "cps": 44.507812, "mean_p": 0.454118, "boundary_entropy": 0.66545, "bpb": 2.187097}, "fi": {"chunks": 5362.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.285046, "cps": 41.890625, "mean_p": 0.451396, "boundary_entropy": 0.668293, "bpb": 2.370607}, "ru": {"chunks": 7127.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.376095, "cps": 55.679688, "mean_p": 0.434432, "boundary_entropy": 0.670872, "bpb": 1.387303}, "el": {"chunks": 7696.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.374994, "cps": 60.125, "mean_p": 0.457887, "boundary_entropy": 0.676535, "bpb": 1.451291}, "hi": {"chunks": 8245.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.490045, "cps": 64.414062, "mean_p": 0.417101, "boundary_entropy": 0.662375, "bpb": 1.161951}, "ta": {"chunks": 8939.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.478149, "cps": 69.835938, "mean_p": 0.418172, "boundary_entropy": 0.663598, "bpb": 0.958783}, "th": {"chunks": 10154.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.631271, "cps": 79.328125, "mean_p": 0.409569, "boundary_entropy": 0.654213, "bpb": 1.048605}, "zh": {"chunks": 3313.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.560955, "cps": 25.882812, "mean_p": 0.444441, "boundary_entropy": 0.671345, "bpb": 3.128732}, "my": {"chunks": 10349.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.546179, "cps": 80.851562, "mean_p": 0.43442, "boundary_entropy": 0.674312, "bpb": 1.072781}}, "bpb": {"en": 2.02778, "de": 2.21872, "vi": 1.92771, "tr": 2.1871, "fi": 2.37061, "ru": 1.3873, "el": 1.45129, "hi": 1.16195, "ta": 0.95878, "th": 1.0486, "zh": 3.12873, "my": 1.07278}, "gini_cpc": 0.18558, "gini_cps": 0.190894, "bpic": 4.740964, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1635990031887962, "vi": 1.1774736625226607, "tr": 1.4093439385521103, "fi": 1.2765453563829245, "ru": 1.6842977497192408, "el": 1.6793667017678142, "hi": 2.1946077625399667, "ta": 2.1413358684531065, "th": 2.827075570450407, "zh": 2.5121716899812623, "my": 2.4459993360488945}, "t": 327.4} |
| {"event": "eval", "step": 5600, "phase": 3, "bpb": {"en": 2.0394, "de": 2.22891, "vi": 1.92644, "tr": 2.19271, "fi": 2.37961, "ru": 1.36427, "el": 1.44266, "hi": 1.15778, "ta": 0.96645, "th": 1.06758, "zh": 3.12579, "my": 1.08063}, "bytes_per_chunk": {"en": 4.45, "de": 3.897, "vi": 4.869, "tr": 3.448, "fi": 3.749, "ru": 4.821, "el": 4.996, "hi": 5.503, "ta": 5.741, "th": 4.519, "zh": 5.066, "my": 5.189}, "bpb_mean": 1.74769, "t": 328.8} |
| {"event": "train", "step": 5600, "phase": 3, "objective": "global", "lm_loss": 1.47089, "bpb": 2.12204, "chunk_loss": 0.033383, "F": 0.22412, "G": 0.43672, "chunks": 229.5, "bytes_per_chunk": 4.462, "boundary_entropy": 0.66947, "router_noise": 0.0, "grad_norm": 0.8303, "lr": 3.33e-05, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 328.8} |
| {"event": "train", "step": 5650, "phase": 3, "objective": "global", "lm_loss": 1.27845, "bpb": 1.84442, "chunk_loss": 0.033102, "F": 0.21957, "G": 0.43386, "chunks": 224.84, "bytes_per_chunk": 4.554, "boundary_entropy": 0.6684, "router_noise": 0.0, "grad_norm": 0.769, "lr": 3.25e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 331.0} |
| {"event": "train", "step": 5700, "phase": 3, "objective": "global", "lm_loss": 1.32829, "bpb": 1.91632, "chunk_loss": 0.033792, "F": 0.23093, "G": 0.43645, "chunks": 236.47, "bytes_per_chunk": 4.33, "boundary_entropy": 0.66796, "router_noise": 0.0, "grad_norm": 0.8081, "lr": 3.18e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 333.1} |
| {"event": "train", "step": 5750, "phase": 3, "objective": "global", "lm_loss": 1.20669, "bpb": 1.74088, "chunk_loss": 0.033528, "F": 0.22668, "G": 0.4332, "chunks": 232.12, "bytes_per_chunk": 4.411, "boundary_entropy": 0.66708, "router_noise": 0.0, "grad_norm": 0.7641, "lr": 3.13e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 335.3} |
| {"event": "probe", "step": 5800, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3995.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.229558, "cps": 31.210938, "mean_p": 0.433111, "boundary_entropy": 0.659017, "bpb": 2.020256}, "de": {"chunks": 5366.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.261335, "cps": 41.921875, "mean_p": 0.444815, "boundary_entropy": 0.667494, "bpb": 2.215515}, "vi": {"chunks": 4929.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.267372, "cps": 38.507812, "mean_p": 0.445681, "boundary_entropy": 0.671087, "bpb": 1.921169}, "tr": {"chunks": 5690.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.314313, "cps": 44.453125, "mean_p": 0.453711, "boundary_entropy": 0.665182, "bpb": 2.178621}, "fi": {"chunks": 5473.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.290947, "cps": 42.757812, "mean_p": 0.451171, "boundary_entropy": 0.667553, "bpb": 2.361736}, "ru": {"chunks": 6940.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.366227, "cps": 54.21875, "mean_p": 0.433989, "boundary_entropy": 0.670674, "bpb": 1.384357}, "el": {"chunks": 7214.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.351508, "cps": 56.359375, "mean_p": 0.456668, "boundary_entropy": 0.676562, "bpb": 1.448985}, "hi": {"chunks": 7796.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.463358, "cps": 60.90625, "mean_p": 0.416365, "boundary_entropy": 0.662553, "bpb": 1.161712}, "ta": {"chunks": 8849.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.473335, "cps": 69.132812, "mean_p": 0.418188, "boundary_entropy": 0.663819, "bpb": 0.954448}, "th": {"chunks": 10281.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.639167, "cps": 80.320312, "mean_p": 0.410032, "boundary_entropy": 0.653849, "bpb": 1.045145}, "zh": {"chunks": 3352.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.567558, "cps": 26.1875, "mean_p": 0.444954, "boundary_entropy": 0.671407, "bpb": 3.125353}, "my": {"chunks": 9288.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.490184, "cps": 72.5625, "mean_p": 0.427516, "boundary_entropy": 0.672226, "bpb": 1.071898}}, "bpb": {"en": 2.02026, "de": 2.21552, "vi": 1.92117, "tr": 2.17862, "fi": 2.36174, "ru": 1.38436, "el": 1.44898, "hi": 1.16171, "ta": 0.95445, "th": 1.04515, "zh": 3.12535, "my": 1.0719}, "gini_cpc": 0.180196, "gini_cps": 0.178848, "bpic": 4.847372, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1384280757609457, "vi": 1.1647239684267663, "tr": 1.3692068131620914, "fi": 1.2674210074272176, "ru": 1.5953559363452336, "el": 1.531237757180396, "hi": 2.0184783430997513, "ta": 2.0619401826500834, "th": 2.7843359219229193, "zh": 2.4723952688913973, "my": 2.135335730778102}, "t": 338.3} |
| {"event": "eval", "step": 5800, "phase": 3, "bpb": {"en": 2.03165, "de": 2.22605, "vi": 1.92006, "tr": 2.18509, "fi": 2.37187, "ru": 1.36144, "el": 1.44123, "hi": 1.15676, "ta": 0.96263, "th": 1.06388, "zh": 3.12175, "my": 1.08059}, "bytes_per_chunk": {"en": 4.307, "de": 3.866, "vi": 4.83, "tr": 3.431, "fi": 3.654, "ru": 4.925, "el": 5.25, "hi": 5.807, "ta": 5.841, "th": 4.479, "zh": 5.059, "my": 5.661}, "bpb_mean": 1.74358, "t": 339.7} |
| {"event": "train", "step": 5800, "phase": 3, "objective": "global", "lm_loss": 1.26402, "bpb": 1.82359, "chunk_loss": 0.032952, "F": 0.21762, "G": 0.43059, "chunks": 222.84, "bytes_per_chunk": 4.595, "boundary_entropy": 0.66641, "router_noise": 0.0, "grad_norm": 0.7679, "lr": 3.08e-05, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 339.7} |
| {"event": "train", "step": 5850, "phase": 3, "objective": "global", "lm_loss": 1.37359, "bpb": 1.98168, "chunk_loss": 0.033574, "F": 0.22708, "G": 0.43677, "chunks": 232.53, "bytes_per_chunk": 4.404, "boundary_entropy": 0.66888, "router_noise": 0.0, "grad_norm": 0.7586, "lr": 3.05e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 341.9} |
| {"event": "train", "step": 5900, "phase": 3, "objective": "global", "lm_loss": 1.32497, "bpb": 1.91153, "chunk_loss": 0.033651, "F": 0.22913, "G": 0.43649, "chunks": 234.62, "bytes_per_chunk": 4.364, "boundary_entropy": 0.66766, "router_noise": 0.0, "grad_norm": 0.7634, "lr": 3.02e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 344.0} |
| {"event": "train", "step": 5950, "phase": 3, "objective": "global", "lm_loss": 1.28555, "bpb": 1.85466, "chunk_loss": 0.033033, "F": 0.21863, "G": 0.4352, "chunks": 223.88, "bytes_per_chunk": 4.574, "boundary_entropy": 0.66868, "router_noise": 0.0, "grad_norm": 0.7743, "lr": 3.01e-05, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 346.1} |
| {"event": "probe", "step": 6000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4020.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.230995, "cps": 31.40625, "mean_p": 0.433259, "boundary_entropy": 0.6587, "bpb": 2.017392}, "de": {"chunks": 5386.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.262309, "cps": 42.078125, "mean_p": 0.444977, "boundary_entropy": 0.667423, "bpb": 2.210801}, "vi": {"chunks": 4987.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.270518, "cps": 38.960938, "mean_p": 0.446588, "boundary_entropy": 0.671409, "bpb": 1.920715}, "tr": {"chunks": 5663.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.312821, "cps": 44.242188, "mean_p": 0.454014, "boundary_entropy": 0.665219, "bpb": 2.179768}, "fi": {"chunks": 5378.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.285897, "cps": 42.015625, "mean_p": 0.451754, "boundary_entropy": 0.667832, "bpb": 2.364606}, "ru": {"chunks": 7137.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.376623, "cps": 55.757812, "mean_p": 0.435144, "boundary_entropy": 0.670826, "bpb": 1.37747}, "el": {"chunks": 7470.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.363982, "cps": 58.359375, "mean_p": 0.457781, "boundary_entropy": 0.676748, "bpb": 1.442782}, "hi": {"chunks": 8241.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.489807, "cps": 64.382812, "mean_p": 0.416997, "boundary_entropy": 0.662156, "bpb": 1.159408}, "ta": {"chunks": 8376.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.448034, "cps": 65.4375, "mean_p": 0.41599, "boundary_entropy": 0.662784, "bpb": 0.954679}, "th": {"chunks": 10465.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.650606, "cps": 81.757812, "mean_p": 0.409212, "boundary_entropy": 0.652972, "bpb": 1.041289}, "zh": {"chunks": 3406.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.576702, "cps": 26.609375, "mean_p": 0.444676, "boundary_entropy": 0.67115, "bpb": 3.117714}, "my": {"chunks": 10110.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.533566, "cps": 78.984375, "mean_p": 0.431741, "boundary_entropy": 0.673122, "bpb": 1.066457}}, "bpb": {"en": 2.01739, "de": 2.2108, "vi": 1.92072, "tr": 2.17977, "fi": 2.36461, "ru": 1.37747, "el": 1.44278, "hi": 1.15941, "ta": 0.95468, "th": 1.04129, "zh": 3.11771, "my": 1.06646}, "gini_cpc": 0.186116, "gini_cps": 0.184178, "bpic": 4.759248, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1355650278292462, "vi": 1.1711008424112148, "tr": 1.3542351354314985, "fi": 1.2376760342670254, "ru": 1.630438979246249, "el": 1.5757155604814694, "hi": 2.1204249628528977, "ta": 1.9395870074606907, "th": 2.816542017917871, "zh": 2.4966017356495547, "my": 2.3098610021457}, "t": 349.2} |
| {"event": "eval", "step": 6000, "phase": 3, "bpb": {"en": 2.02911, "de": 2.221, "vi": 1.9193, "tr": 2.18511, "fi": 2.3739, "ru": 1.35583, "el": 1.43478, "hi": 1.15503, "ta": 0.96281, "th": 1.0602, "zh": 3.11373, "my": 1.07465}, "bytes_per_chunk": {"en": 4.31, "de": 3.853, "vi": 4.746, "tr": 3.45, "fi": 3.702, "ru": 4.767, "el": 5.085, "hi": 5.353, "ta": 6.27, "th": 4.438, "zh": 4.975, "my": 5.303}, "bpb_mean": 1.74045, "t": 350.6} |
| {"event": "probe", "step": 6000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4020.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.230995, "cps": 31.40625, "mean_p": 0.433259, "boundary_entropy": 0.6587, "bpb": 2.017392}, "de": {"chunks": 5386.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.262309, "cps": 42.078125, "mean_p": 0.444977, "boundary_entropy": 0.667423, "bpb": 2.210801}, "vi": {"chunks": 4987.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.270518, "cps": 38.960938, "mean_p": 0.446588, "boundary_entropy": 0.671409, "bpb": 1.920715}, "tr": {"chunks": 5663.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.312821, "cps": 44.242188, "mean_p": 0.454014, "boundary_entropy": 0.665219, "bpb": 2.179768}, "fi": {"chunks": 5378.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.285897, "cps": 42.015625, "mean_p": 0.451754, "boundary_entropy": 0.667832, "bpb": 2.364606}, "ru": {"chunks": 7137.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.376623, "cps": 55.757812, "mean_p": 0.435144, "boundary_entropy": 0.670826, "bpb": 1.37747}, "el": {"chunks": 7470.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.363982, "cps": 58.359375, "mean_p": 0.457781, "boundary_entropy": 0.676748, "bpb": 1.442782}, "hi": {"chunks": 8241.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.489807, "cps": 64.382812, "mean_p": 0.416997, "boundary_entropy": 0.662156, "bpb": 1.159408}, "ta": {"chunks": 8376.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.448034, "cps": 65.4375, "mean_p": 0.41599, "boundary_entropy": 0.662784, "bpb": 0.954679}, "th": {"chunks": 10465.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.650606, "cps": 81.757812, "mean_p": 0.409212, "boundary_entropy": 0.652972, "bpb": 1.041289}, "zh": {"chunks": 3406.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.576702, "cps": 26.609375, "mean_p": 0.444676, "boundary_entropy": 0.67115, "bpb": 3.117714}, "my": {"chunks": 10110.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.533566, "cps": 78.984375, "mean_p": 0.431741, "boundary_entropy": 0.673122, "bpb": 1.066457}}, "bpb": {"en": 2.01739, "de": 2.2108, "vi": 1.92072, "tr": 2.17977, "fi": 2.36461, "ru": 1.37747, "el": 1.44278, "hi": 1.15941, "ta": 0.95468, "th": 1.04129, "zh": 3.11771, "my": 1.06646}, "gini_cpc": 0.186116, "gini_cps": 0.184178, "bpic": 4.759248, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 1.1355650278292462, "vi": 1.1711008424112148, "tr": 1.3542351354314985, "fi": 1.2376760342670254, "ru": 1.630438979246249, "el": 1.5757155604814694, "hi": 2.1204249628528977, "ta": 1.9395870074606907, "th": 2.816542017917871, "zh": 2.4966017356495547, "my": 2.3098610021457}, "t": 351.6} |
| {"event": "reps_dumped", "dirs": ["/local/scratch/sas245/beetle-hnet/scratch/reps/emactl_noema_s0/th0.30", "/local/scratch/sas245/beetle-hnet/scratch/reps/emactl_noema_s0/th0.40", "/local/scratch/sas245/beetle-hnet/scratch/reps/emactl_noema_s0/th0.50", "/local/scratch/sas245/beetle-hnet/scratch/reps/emactl_noema_s0/th0.60", "/local/scratch/sas245/beetle-hnet/scratch/reps/emactl_noema_s0/th0.70"], "t": 356.3} |
| {"event": "final", "run": "emactl_noema_s0", "condition": "A", "step": 6000, "params": 22516992, "per_lang": {"en": {"cpc": 0.23099465609377692, "cps": 31.40625, "bpb": 2.0382534508432286, "bytes_per_chunk": 4.302768180288718}, "de": {"cpc": 0.26230945307553694, "cps": 42.078125, "bpb": 2.2124327286163514, "bytes_per_chunk": 3.8386598428771204}, "vi": {"cpc": 0.27051803634391103, "cps": 38.9609375, "bpb": 1.9233381799299585, "bytes_per_chunk": 4.745675570480733}, "tr": {"cpc": 0.3128210793791084, "cps": 44.2421875, "bpb": 2.1762234583080327, "bytes_per_chunk": 3.449739766678291}, "fi": {"cpc": 0.2858965498910212, "cps": 42.015625, "bpb": 2.3890020062928516, "bytes_per_chunk": 3.6961775446611114}, "ru": {"cpc": 0.376622691292876, "cps": 55.7578125, "bpb": 1.3563091872564799, "bytes_per_chunk": 4.759763958238765}, "el": {"cpc": 0.36398187399502996, "cps": 58.359375, "bpb": 1.431533232486521, "bytes_per_chunk": 5.101963760923298}, "hi": {"cpc": 0.4898068350668648, "cps": 64.3828125, "bpb": 1.1639179249083127, "bytes_per_chunk": 5.340885244231651}, "ta": {"cpc": 0.4480342337523402, "cps": 65.4375, "bpb": 0.9624026067198248, "bytes_per_chunk": 6.288613546676902}, "th": {"cpc": 0.6506061548026111, "cps": 81.7578125, "bpb": 1.0589655444660309, "bytes_per_chunk": 4.42780893182893}, "zh": {"cpc": 0.5767016593294955, "cps": 26.609375, "bpb": 3.1217361863324053, "bytes_per_chunk": 4.9474200732268905}, "my": {"cpc": 0.5335655478150728, "cps": 78.984375, "bpb": 1.0843777332042663, "bytes_per_chunk": 5.2943944580770905}}, "gini_cpc": 0.18611578592940914, "gini_cps": 0.18417783785347885, "bpic": 4.7592480065477005, "seed": 0, "mean_boundary_entropy": 0.6694673299789429, "dechunk": "ste", "router_noise": 0.0, "transitions": [{"step": 0, "phase": 1, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0}, {"step": 2400, "phase": 2, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0}, {"step": 4200, "phase": 3, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0}], "t": 358.8} |
|
|
