| {"event": "start", "run": "emactl_genreg_s1", "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": 2.0} |
| {"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": 2.0} |
| {"event": "probe", "step": 0, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2674.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.153652, "cps": 20.890625, "mean_p": 0.478126, "boundary_entropy": 0.685438, "bpb": 7.975542}, "de": {"chunks": 2056.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.100131, "cps": 16.0625, "mean_p": 0.469682, "boundary_entropy": 0.685273, "bpb": 7.980804}, "vi": {"chunks": 14185.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.76946, "cps": 110.820312, "mean_p": 0.507219, "boundary_entropy": 0.688283, "bpb": 7.810944}, "tr": {"chunks": 5645.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.311827, "cps": 44.101562, "mean_p": 0.486063, "boundary_entropy": 0.686375, "bpb": 7.940182}, "fi": {"chunks": 4732.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.251555, "cps": 36.96875, "mean_p": 0.48451, "boundary_entropy": 0.686595, "bpb": 7.978067}, "ru": {"chunks": 17548.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.926016, "cps": 137.09375, "mean_p": 0.502342, "boundary_entropy": 0.690066, "bpb": 8.056575}, "el": {"chunks": 31473.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.533548, "cps": 245.882812, "mean_p": 0.517965, "boundary_entropy": 0.689704, "bpb": 7.856999}, "hi": {"chunks": 1659.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.098603, "cps": 12.960938, "mean_p": 0.476241, "boundary_entropy": 0.689492, "bpb": 8.113177}, "ta": {"chunks": 39693.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.123188, "cps": 310.101562, "mean_p": 0.514597, "boundary_entropy": 0.690542, "bpb": 8.099115}, "th": {"chunks": 44424.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.761828, "cps": 347.0625, "mean_p": 0.52078, "boundary_entropy": 0.690224, "bpb": 8.255005}, "zh": {"chunks": 12933.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.189807, "cps": 101.039062, "mean_p": 0.528309, "boundary_entropy": 0.684752, "bpb": 8.116607}, "my": {"chunks": 7048.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.371965, "cps": 55.0625, "mean_p": 0.491145, "boundary_entropy": 0.691021, "bpb": 8.11144}}, "bpb": {"en": 7.97554, "de": 7.9808, "vi": 7.81094, "tr": 7.94018, "fi": 7.97807, "ru": 8.05657, "el": 7.857, "hi": 8.11318, "ta": 8.09911, "th": 8.25501, "zh": 8.11661, "my": 8.11144}, "gini_cpc": 0.507116, "gini_cps": 0.500863, "bpic": 2.084973, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 0.6516785410034256, "vi": 5.0078223656304, "tr": 2.029439514461233, "fi": 1.6371767549398106, "ru": 6.026721581613623, "el": 9.980677296913099, "hi": 0.6417324941848254, "ta": 13.818190381639045, "th": 17.9746032262692, "zh": 14.25176170627287, "my": 2.4208352616119475}, "t": 3.9} |
| {"event": "train", "step": 0, "phase": 1, "objective": "global", "lm_loss": 5.54764, "bpb": 8.00355, "chunk_loss": 0.043969, "F": 0.34537, "G": 0.49096, "chunks": 353.66, "bytes_per_chunk": 2.895, "boundary_entropy": 0.6904, "router_noise": 0.0, "grad_norm": 6.9677, "lr": 1e-06, "router_trainable": true, "sec_per_step": 0.005, "alpha": 0.03, "t": 4.2} |
| {"event": "probe", "step": 1, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2671.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.153479, "cps": 20.867188, "mean_p": 0.478155, "boundary_entropy": 0.685445, "bpb": 7.949514}, "de": {"chunks": 2047.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.099693, "cps": 15.992188, "mean_p": 0.469689, "boundary_entropy": 0.685277, "bpb": 7.958521}, "vi": {"chunks": 14156.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.767887, "cps": 110.59375, "mean_p": 0.50713, "boundary_entropy": 0.688286, "bpb": 7.793604}, "tr": {"chunks": 5645.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.311827, "cps": 44.101562, "mean_p": 0.486067, "boundary_entropy": 0.686378, "bpb": 7.922259}, "fi": {"chunks": 4720.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.250917, "cps": 36.875, "mean_p": 0.484468, "boundary_entropy": 0.686593, "bpb": 7.955702}, "ru": {"chunks": 17506.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.923799, "cps": 136.765625, "mean_p": 0.502295, "boundary_entropy": 0.690066, "bpb": 8.047038}, "el": {"chunks": 31463.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.53306, "cps": 245.804688, "mean_p": 0.517948, "boundary_entropy": 0.689705, "bpb": 7.846411}, "hi": {"chunks": 1643.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.097652, "cps": 12.835938, "mean_p": 0.476155, "boundary_entropy": 0.689484, "bpb": 8.093398}, "ta": {"chunks": 39591.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.117732, "cps": 309.304688, "mean_p": 0.514537, "boundary_entropy": 0.690545, "bpb": 8.08481}, "th": {"chunks": 44418.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.761455, "cps": 347.015625, "mean_p": 0.52079, "boundary_entropy": 0.690224, "bpb": 8.23927}, "zh": {"chunks": 12935.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.190146, "cps": 101.054688, "mean_p": 0.528311, "boundary_entropy": 0.684752, "bpb": 8.112235}, "my": {"chunks": 7051.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.372124, "cps": 55.085938, "mean_p": 0.491149, "boundary_entropy": 0.691021, "bpb": 8.097941}}, "bpb": {"en": 7.94951, "de": 7.95852, "vi": 7.7936, "tr": 7.92226, "fi": 7.9557, "ru": 8.04704, "el": 7.84641, "hi": 8.0934, "ta": 8.08481, "th": 8.23927, "zh": 8.11224, "my": 8.09794}, "gini_cpc": 0.507408, "gini_cps": 0.501014, "bpic": 2.087513, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 0.6495546074501287, "vi": 5.003197468881173, "tr": 2.031718929864971, "fi": 1.634859169064772, "ru": 6.019049875087547, "el": 9.988712593606797, "hi": 0.6362572187209158, "ta": 13.798161826294107, "th": 17.992361433497422, "zh": 14.26997533934549, "my": 2.424585875216864}, "t": 5.5} |
| {"event": "probe", "step": 2, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2681.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.154054, "cps": 20.945312, "mean_p": 0.478179, "boundary_entropy": 0.685452, "bpb": 7.904603}, "de": {"chunks": 2040.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.099352, "cps": 15.9375, "mean_p": 0.46969, "boundary_entropy": 0.685281, "bpb": 7.921359}, "vi": {"chunks": 14085.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.764036, "cps": 110.039062, "mean_p": 0.506995, "boundary_entropy": 0.688292, "bpb": 7.765662}, "tr": {"chunks": 5634.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.311219, "cps": 44.015625, "mean_p": 0.486066, "boundary_entropy": 0.686383, "bpb": 7.892285}, "fi": {"chunks": 4697.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.249694, "cps": 36.695312, "mean_p": 0.484376, "boundary_entropy": 0.686588, "bpb": 7.919815}, "ru": {"chunks": 17211.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.908232, "cps": 134.460938, "mean_p": 0.502015, "boundary_entropy": 0.690066, "bpb": 8.010401}, "el": {"chunks": 31500.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.534863, "cps": 246.09375, "mean_p": 0.517996, "boundary_entropy": 0.689703, "bpb": 7.826459}, "hi": {"chunks": 1621.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.096345, "cps": 12.664062, "mean_p": 0.476101, "boundary_entropy": 0.689478, "bpb": 8.058642}, "ta": {"chunks": 39507.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.113239, "cps": 308.648438, "mean_p": 0.514486, "boundary_entropy": 0.690547, "bpb": 8.052256}, "th": {"chunks": 44427.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.762014, "cps": 347.085938, "mean_p": 0.520844, "boundary_entropy": 0.690219, "bpb": 8.185998}, "zh": {"chunks": 12928.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.18896, "cps": 101.0, "mean_p": 0.5283, "boundary_entropy": 0.684752, "bpb": 8.103236}, "my": {"chunks": 7059.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.372546, "cps": 55.148438, "mean_p": 0.491144, "boundary_entropy": 0.69102, "bpb": 8.08433}}, "bpb": {"en": 7.9046, "de": 7.92136, "vi": 7.76566, "tr": 7.89229, "fi": 7.91982, "ru": 8.0104, "el": 7.82646, "hi": 8.05864, "ta": 8.05226, "th": 8.186, "zh": 8.10324, "my": 8.08433}, "gini_cpc": 0.508108, "gini_cps": 0.501479, "bpic": 2.092704, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 0.6449188434450902, "vi": 4.9595356407640905, "tr": 2.0201964213110757, "fi": 1.6208244634375224, "ru": 5.895548229060357, "el": 9.9631579369208, "hi": 0.6253962145148874, "ta": 13.71752906623821, "th": 17.928882822129776, "zh": 14.209055382517073, "my": 2.4182829396175416}, "t": 7.0} |
| {"event": "probe", "step": 4, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2691.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.154629, "cps": 21.023438, "mean_p": 0.478427, "boundary_entropy": 0.685496, "bpb": 7.75258}, "de": {"chunks": 2061.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.100375, "cps": 16.101562, "mean_p": 0.469876, "boundary_entropy": 0.685324, "bpb": 7.790456}, "vi": {"chunks": 13908.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.754434, "cps": 108.65625, "mean_p": 0.506487, "boundary_entropy": 0.688312, "bpb": 7.673564}, "tr": {"chunks": 5608.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.309783, "cps": 43.8125, "mean_p": 0.48619, "boundary_entropy": 0.686408, "bpb": 7.789616}, "fi": {"chunks": 4644.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.246877, "cps": 36.28125, "mean_p": 0.484202, "boundary_entropy": 0.686582, "bpb": 7.797578}, "ru": {"chunks": 15931.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.840686, "cps": 124.460938, "mean_p": 0.500759, "boundary_entropy": 0.690065, "bpb": 7.864194}, "el": {"chunks": 31545.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.537056, "cps": 246.445312, "mean_p": 0.518036, "boundary_entropy": 0.689701, "bpb": 7.754696}, "hi": {"chunks": 1561.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.092779, "cps": 12.195312, "mean_p": 0.475793, "boundary_entropy": 0.689447, "bpb": 7.934245}, "ta": {"chunks": 39118.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 2.092431, "cps": 305.609375, "mean_p": 0.51422, "boundary_entropy": 0.690559, "bpb": 7.939716}, "th": {"chunks": 44462.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.76419, "cps": 347.359375, "mean_p": 0.520977, "boundary_entropy": 0.690208, "bpb": 8.037471}, "zh": {"chunks": 12928.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.18896, "cps": 101.0, "mean_p": 0.528344, "boundary_entropy": 0.684745, "bpb": 8.079738}, "my": {"chunks": 7066.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.372915, "cps": 55.203125, "mean_p": 0.491108, "boundary_entropy": 0.691017, "bpb": 8.002622}}, "bpb": {"en": 7.75258, "de": 7.79046, "vi": 7.67356, "tr": 7.78962, "fi": 7.79758, "ru": 7.86419, "el": 7.7547, "hi": 7.93424, "ta": 7.93972, "th": 8.03747, "zh": 8.07974, "my": 8.00262}, "gini_cpc": 0.510489, "gini_cps": 0.502507, "bpic": 2.114228, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 0.6491364663490552, "vi": 4.879012856343313, "tr": 2.0034009527175445, "fi": 1.5965802404449543, "ru": 5.436811123563446, "el": 9.940314129982562, "hi": 0.6000096739834448, "ta": 13.531987728850408, "th": 17.876329518940107, "zh": 14.15625324434347, "my": 2.4116855401240125}, "t": 8.6} |
| {"event": "probe", "step": 8, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2574.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.147906, "cps": 20.109375, "mean_p": 0.478198, "boundary_entropy": 0.685533, "bpb": 7.305576}, "de": {"chunks": 1963.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.095602, "cps": 15.335938, "mean_p": 0.469421, "boundary_entropy": 0.685327, "bpb": 7.395705}, "vi": {"chunks": 12940.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.701926, "cps": 101.09375, "mean_p": 0.504308, "boundary_entropy": 0.688379, "bpb": 7.427738}, "tr": {"chunks": 5393.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.297906, "cps": 42.132812, "mean_p": 0.485772, "boundary_entropy": 0.686434, "bpb": 7.488602}, "fi": {"chunks": 4218.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.224231, "cps": 32.953125, "mean_p": 0.482623, "boundary_entropy": 0.686482, "bpb": 7.435824}, "ru": {"chunks": 11189.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.590449, "cps": 87.414062, "mean_p": 0.495917, "boundary_entropy": 0.690002, "bpb": 7.346449}, "el": {"chunks": 31649.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.542123, "cps": 247.257812, "mean_p": 0.51816, "boundary_entropy": 0.689695, "bpb": 7.597125}, "hi": {"chunks": 1360.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.080832, "cps": 10.625, "mean_p": 0.474725, "boundary_entropy": 0.689335, "bpb": 7.571439}, "ta": {"chunks": 36843.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 1.970741, "cps": 287.835938, "mean_p": 0.512713, "boundary_entropy": 0.690623, "bpb": 7.569584}, "th": {"chunks": 44514.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.767423, "cps": 347.765625, "mean_p": 0.521092, "boundary_entropy": 0.690198, "bpb": 7.640152}, "zh": {"chunks": 12946.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.192008, "cps": 101.140625, "mean_p": 0.52841, "boundary_entropy": 0.684736, "bpb": 7.985577}, "my": {"chunks": 7209.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.380462, "cps": 56.320312, "mean_p": 0.491101, "boundary_entropy": 0.691013, "bpb": 7.752999}}, "bpb": {"en": 7.30558, "de": 7.3957, "vi": 7.42774, "tr": 7.4886, "fi": 7.43582, "ru": 7.34645, "el": 7.59713, "hi": 7.57144, "ta": 7.56958, "th": 7.64015, "zh": 7.98558, "my": 7.753}, "gini_cpc": 0.524578, "gini_cps": 0.512352, "bpic": 2.220981, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 0.6463733915397092, "vi": 4.745770277096558, "tr": 2.0141668620650006, "fi": 1.5160386368190313, "ru": 3.9920653049676798, "el": 10.426408224464064, "hi": 0.5465116359618589, "ta": 13.324321225711973, "th": 18.7107473192024, "zh": 14.820325345723246, "my": 2.572333224090665}, "t": 10.4} |
| {"event": "probe", "step": 16, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 1446.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.083089, "cps": 11.296875, "mean_p": 0.470275, "boundary_entropy": 0.684695, "bpb": 6.687578}, "de": {"chunks": 1190.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.057955, "cps": 9.296875, "mean_p": 0.462617, "boundary_entropy": 0.684467, "bpb": 6.815}, "vi": {"chunks": 8392.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.455221, "cps": 65.5625, "mean_p": 0.493574, "boundary_entropy": 0.68834, "bpb": 7.081828}, "tr": {"chunks": 3858.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.213114, "cps": 30.140625, "mean_p": 0.479447, "boundary_entropy": 0.686013, "bpb": 7.045782}, "fi": {"chunks": 2240.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.119079, "cps": 17.5, "mean_p": 0.47244, "boundary_entropy": 0.685496, "bpb": 6.902342}, "ru": {"chunks": 3236.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.170765, "cps": 25.28125, "mean_p": 0.480218, "boundary_entropy": 0.689132, "bpb": 6.55387}, "el": {"chunks": 32754.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.595966, "cps": 255.890625, "mean_p": 0.519839, "boundary_entropy": 0.689606, "bpb": 6.95174}, "hi": {"chunks": 1540.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.09153, "cps": 12.03125, "mean_p": 0.47625, "boundary_entropy": 0.689506, "bpb": 6.768819}, "ta": {"chunks": 34044.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 1.821022, "cps": 265.96875, "mean_p": 0.511621, "boundary_entropy": 0.690646, "bpb": 6.706725}, "th": {"chunks": 45595.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.834629, "cps": 356.210938, "mean_p": 0.529294, "boundary_entropy": 0.689396, "bpb": 6.581897}, "zh": {"chunks": 13107.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 2.219269, "cps": 102.398438, "mean_p": 0.529118, "boundary_entropy": 0.684694, "bpb": 7.723783}, "my": {"chunks": 5399.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.284938, "cps": 42.179688, "mean_p": 0.488584, "boundary_entropy": 0.690886, "bpb": 6.989975}}, "bpb": {"en": 6.68758, "de": 6.815, "vi": 7.08183, "tr": 7.04578, "fi": 6.90234, "ru": 6.55387, "el": 6.95174, "hi": 6.76882, "ta": 6.70673, "th": 6.5819, "zh": 7.72378, "my": 6.98997}, "gini_cpc": 0.588017, "gini_cps": 0.584463, "bpic": 2.511639, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 0.6975099086522596, "vi": 5.478708077162442, "tr": 2.5648826459381917, "fi": 1.4331510365944178, "ru": 2.055204896046596, "el": 19.207875265136188, "hi": 1.1015937802494558, "ta": 21.91648686770266, "th": 34.11551895596139, "zh": 26.70949544242932, "my": 3.429302362350946}, "t": 12.4} |
| {"event": "probe", "step": 32, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4228.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.242947, "cps": 33.03125, "mean_p": 0.490928, "boundary_entropy": 0.687023, "bpb": 6.233538}, "de": {"chunks": 3988.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.194224, "cps": 31.15625, "mean_p": 0.486766, "boundary_entropy": 0.68763, "bpb": 6.334783}, "vi": {"chunks": 7835.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.425007, "cps": 61.210938, "mean_p": 0.493615, "boundary_entropy": 0.688533, "bpb": 6.757279}, "tr": {"chunks": 6981.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.385627, "cps": 54.539062, "mean_p": 0.494555, "boundary_entropy": 0.687434, "bpb": 6.652858}, "fi": {"chunks": 4154.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.220828, "cps": 32.453125, "mean_p": 0.486652, "boundary_entropy": 0.687211, "bpb": 6.418787}, "ru": {"chunks": 886.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.046755, "cps": 6.921875, "mean_p": 0.47555, "boundary_entropy": 0.68887, "bpb": 5.835347}, "el": {"chunks": 27738.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 1.351557, "cps": 216.703125, "mean_p": 0.511654, "boundary_entropy": 0.690113, "bpb": 6.129604}, "hi": {"chunks": 463.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.027519, "cps": 3.617188, "mean_p": 0.470214, "boundary_entropy": 0.688916, "bpb": 6.04779}, "ta": {"chunks": 17840.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.954266, "cps": 139.375, "mean_p": 0.494893, "boundary_entropy": 0.69097, "bpb": 6.133825}, "th": {"chunks": 43592.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 2.710103, "cps": 340.5625, "mean_p": 0.515733, "boundary_entropy": 0.690584, "bpb": 5.571662}, "zh": {"chunks": 9855.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.668642, "cps": 76.992188, "mean_p": 0.511193, "boundary_entropy": 0.686211, "bpb": 7.355967}, "my": {"chunks": 1198.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.063226, "cps": 9.359375, "mean_p": 0.475947, "boundary_entropy": 0.689886, "bpb": 6.049491}}, "bpb": {"en": 6.23354, "de": 6.33478, "vi": 6.75728, "tr": 6.65286, "fi": 6.41879, "ru": 5.83535, "el": 6.1296, "hi": 6.04779, "ta": 6.13383, "th": 5.57166, "zh": 7.35597, "my": 6.04949}, "gini_cpc": 0.579758, "gini_cps": 0.566769, "bpic": 2.980638, "worst_lang": "th", "cpc_parity_vs_en": {"en": 1.0, "de": 0.7994510624865315, "vi": 1.7493833969822632, "tr": 1.587289809792812, "fi": 0.9089578618199148, "ru": 0.1924481094749665, "el": 5.563184204204794, "hi": 0.11327004151197492, "ta": 3.9278828113396265, "th": 11.155136045529177, "zh": 6.868348569403787, "my": 0.2602451134000358}, "t": 15.3} |
| {"event": "train", "step": 50, "phase": 1, "objective": "global", "lm_loss": 4.18797, "bpb": 6.04197, "chunk_loss": 0.029696, "F": 0.16016, "G": 0.48524, "chunks": 164.0, "bytes_per_chunk": 6.244, "boundary_entropy": 0.69149, "router_noise": 0.0, "grad_norm": 1.636, "lr": 5.1e-05, "router_trainable": true, "sec_per_step": 0.253, "alpha": 0.03, "t": 16.8} |
| {"event": "probe", "step": 64, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 5124.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.294432, "cps": 40.03125, "mean_p": 0.491045, "boundary_entropy": 0.686724, "bpb": 5.669774}, "de": {"chunks": 5240.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.255199, "cps": 40.9375, "mean_p": 0.489323, "boundary_entropy": 0.687647, "bpb": 5.795925}, "vi": {"chunks": 4283.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.23233, "cps": 33.460938, "mean_p": 0.480118, "boundary_entropy": 0.687185, "bpb": 6.167252}, "tr": {"chunks": 6134.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.338839, "cps": 47.921875, "mean_p": 0.490371, "boundary_entropy": 0.687224, "bpb": 6.14411}, "fi": {"chunks": 4852.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.257934, "cps": 37.90625, "mean_p": 0.488067, "boundary_entropy": 0.687194, "bpb": 5.871953}, "ru": {"chunks": 11387.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.600897, "cps": 88.960938, "mean_p": 0.496012, "boundary_entropy": 0.690225, "bpb": 5.330609}, "el": {"chunks": 16521.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.804999, "cps": 129.070312, "mean_p": 0.500093, "boundary_entropy": 0.690535, "bpb": 5.604154}, "hi": {"chunks": 4533.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.269421, "cps": 35.414062, "mean_p": 0.489962, "boundary_entropy": 0.690614, "bpb": 5.170812}, "ta": {"chunks": 28995.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 1.550949, "cps": 226.523438, "mean_p": 0.504127, "boundary_entropy": 0.691002, "bpb": 5.184416}, "th": {"chunks": 10551.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.655953, "cps": 82.429688, "mean_p": 0.493747, "boundary_entropy": 0.690901, "bpb": 4.963355}, "zh": {"chunks": 5435.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.920251, "cps": 42.460938, "mean_p": 0.496112, "boundary_entropy": 0.686721, "bpb": 6.965478}, "my": {"chunks": 1910.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.100802, "cps": 14.921875, "mean_p": 0.483752, "boundary_entropy": 0.690638, "bpb": 5.076518}}, "bpb": {"en": 5.66977, "de": 5.79592, "vi": 6.16725, "tr": 6.14411, "fi": 5.87195, "ru": 5.33061, "el": 5.60415, "hi": 5.17081, "ta": 5.18442, "th": 4.96335, "zh": 6.96548, "my": 5.07652}, "gini_cpc": 0.384968, "gini_cps": 0.387252, "bpic": 3.656276, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.86675005711375, "vi": 0.7890779935071575, "tr": 1.1508221689031317, "fi": 0.8760399422481514, "ru": 2.040868889534273, "el": 2.7340753864887524, "hi": 0.9150517275577563, "ta": 5.267598225677443, "th": 2.227858260310601, "zh": 3.1255115268004356, "my": 0.34236155503117366}, "t": 18.8} |
| {"event": "train", "step": 100, "phase": 1, "objective": "global", "lm_loss": 3.25007, "bpb": 4.68886, "chunk_loss": 0.036097, "F": 0.24942, "G": 0.46483, "chunks": 255.41, "bytes_per_chunk": 4.009, "boundary_entropy": 0.67932, "router_noise": 0.0, "grad_norm": 1.9374, "lr": 0.000101, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 21.2} |
| {"event": "probe", "step": 128, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4637.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.266448, "cps": 36.226562, "mean_p": 0.442711, "boundary_entropy": 0.659536, "bpb": 4.381846}, "de": {"chunks": 6666.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.324648, "cps": 52.078125, "mean_p": 0.450894, "boundary_entropy": 0.664686, "bpb": 4.580108}, "vi": {"chunks": 7234.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.392406, "cps": 56.515625, "mean_p": 0.446264, "boundary_entropy": 0.659118, "bpb": 4.674396}, "tr": {"chunks": 4018.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.221952, "cps": 31.390625, "mean_p": 0.447012, "boundary_entropy": 0.667808, "bpb": 4.855002}, "fi": {"chunks": 3357.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.178459, "cps": 26.226562, "mean_p": 0.430587, "boundary_entropy": 0.662823, "bpb": 4.662274}, "ru": {"chunks": 10093.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.532612, "cps": 78.851562, "mean_p": 0.493226, "boundary_entropy": 0.689635, "bpb": 3.575777}, "el": {"chunks": 13649.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.665059, "cps": 106.632812, "mean_p": 0.496879, "boundary_entropy": 0.690013, "bpb": 4.048881}, "hi": {"chunks": 17892.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 1.063418, "cps": 139.78125, "mean_p": 0.495112, "boundary_entropy": 0.690507, "bpb": 3.311596}, "ta": {"chunks": 23384.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 1.250816, "cps": 182.6875, "mean_p": 0.510571, "boundary_entropy": 0.689262, "bpb": 3.391365}, "th": {"chunks": 16883.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 1.049611, "cps": 131.898438, "mean_p": 0.514111, "boundary_entropy": 0.685871, "bpb": 3.131293}, "zh": {"chunks": 2820.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.477481, "cps": 22.03125, "mean_p": 0.480911, "boundary_entropy": 0.68533, "bpb": 6.114234}, "my": {"chunks": 3334.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.175955, "cps": 26.046875, "mean_p": 0.463157, "boundary_entropy": 0.68565, "bpb": 3.574655}}, "bpb": {"en": 4.38185, "de": 4.58011, "vi": 4.6744, "tr": 4.855, "fi": 4.66227, "ru": 3.57578, "el": 4.04888, "hi": 3.3116, "ta": 3.39136, "th": 3.13129, "zh": 6.11423, "my": 3.57465}, "gini_cpc": 0.357179, "gini_cps": 0.378033, "bpic": 3.367475, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 1.2184281563992658, "vi": 1.4727274673431043, "tr": 0.8330026927242861, "fi": 0.6697712201507505, "ru": 1.9989322889253185, "el": 2.4960139767970313, "hi": 3.99108374688892, "ta": 4.6944028643174605, "th": 3.9392684660133592, "zh": 1.792019330082587, "my": 0.6603729016667024}, "t": 24.0} |
| {"event": "train", "step": 150, "phase": 1, "objective": "global", "lm_loss": 2.83863, "bpb": 4.09527, "chunk_loss": 0.032802, "F": 0.20947, "G": 0.44803, "chunks": 214.5, "bytes_per_chunk": 4.774, "boundary_entropy": 0.67019, "router_noise": 0.0, "grad_norm": 1.9646, "lr": 0.000151, "router_trainable": true, "sec_per_step": 0.085, "alpha": 0.03, "t": 25.5} |
| {"event": "probe", "step": 200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2442.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.140321, "cps": 19.078125, "mean_p": 0.419742, "boundary_entropy": 0.652188, "bpb": 3.821097}, "de": {"chunks": 3128.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.15234, "cps": 24.4375, "mean_p": 0.422795, "boundary_entropy": 0.658899, "bpb": 4.037268}, "vi": {"chunks": 5942.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.322322, "cps": 46.421875, "mean_p": 0.42568, "boundary_entropy": 0.649147, "bpb": 3.548487}, "tr": {"chunks": 3676.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.20306, "cps": 28.71875, "mean_p": 0.438673, "boundary_entropy": 0.665133, "bpb": 4.099851}, "fi": {"chunks": 2684.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.142682, "cps": 20.96875, "mean_p": 0.415792, "boundary_entropy": 0.656648, "bpb": 4.089151}, "ru": {"chunks": 9192.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.485066, "cps": 71.8125, "mean_p": 0.495118, "boundary_entropy": 0.689999, "bpb": 2.797961}, "el": {"chunks": 11597.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.565073, "cps": 90.601562, "mean_p": 0.494273, "boundary_entropy": 0.689883, "bpb": 2.94346}, "hi": {"chunks": 14068.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.836137, "cps": 109.90625, "mean_p": 0.487011, "boundary_entropy": 0.68936, "bpb": 2.418574}, "ta": {"chunks": 22488.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 1.202888, "cps": 175.6875, "mean_p": 0.490667, "boundary_entropy": 0.687459, "bpb": 2.366663}, "th": {"chunks": 12192.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.757973, "cps": 95.25, "mean_p": 0.483201, "boundary_entropy": 0.689009, "bpb": 2.29487}, "zh": {"chunks": 8224.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.392482, "cps": 64.25, "mean_p": 0.41419, "boundary_entropy": 0.618575, "bpb": 4.849761}, "my": {"chunks": 12273.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.64772, "cps": 95.882812, "mean_p": 0.389852, "boundary_entropy": 0.629795, "bpb": 2.350426}}, "bpb": {"en": 3.8211, "de": 4.03727, "vi": 3.54849, "tr": 4.09985, "fi": 4.08915, "ru": 2.79796, "el": 2.94346, "hi": 2.41857, "ta": 2.36666, "th": 2.29487, "zh": 4.84976, "my": 2.35043}, "gini_cpc": 0.388589, "gini_cps": 0.345586, "bpic": 3.556623, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.0856574022209826, "vi": 2.2970368696975694, "tr": 1.4471162217944518, "fi": 1.0168317272671108, "ru": 3.456839866971793, "el": 4.027015234660316, "hi": 5.958757990734216, "ta": 8.572427556540955, "th": 5.401723491372232, "zh": 9.923574160621232, "my": 4.61600019761831}, "t": 29.3} |
| {"event": "eval", "step": 200, "phase": 1, "bpb": {"en": 3.82666, "de": 4.03921, "vi": 3.55552, "tr": 4.1103, "fi": 4.09162, "ru": 2.78215, "el": 2.93459, "hi": 2.41198, "ta": 2.38487, "th": 2.3286, "zh": 4.85631, "my": 2.36772}, "bytes_per_chunk": {"en": 6.48, "de": 6.134, "vi": 4.121, "tr": 5.188, "fi": 7.356, "ru": 3.888, "el": 3.688, "hi": 3.241, "ta": 2.27, "th": 4.328, "zh": 1.713, "my": 4.44}, "bpb_mean": 3.30746, "t": 30.8} |
| {"event": "train", "step": 200, "phase": 1, "objective": "global", "lm_loss": 2.48477, "bpb": 3.58477, "chunk_loss": 0.034068, "F": 0.23111, "G": 0.44183, "chunks": 236.66, "bytes_per_chunk": 4.327, "boundary_entropy": 0.66526, "router_noise": 0.0, "grad_norm": 1.3285, "lr": 0.000201, "router_trainable": true, "sec_per_step": 0.108, "alpha": 0.03, "t": 30.9} |
| {"event": "train", "step": 250, "phase": 1, "objective": "global", "lm_loss": 2.17112, "bpb": 3.13226, "chunk_loss": 0.03118, "F": 0.17899, "G": 0.43338, "chunks": 183.28, "bytes_per_chunk": 5.587, "boundary_entropy": 0.66246, "router_noise": 0.0, "grad_norm": 1.9203, "lr": 0.000251, "router_trainable": true, "sec_per_step": 0.053, "alpha": 0.03, "t": 33.5} |
| {"event": "probe", "step": 256, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2821.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.162098, "cps": 22.039062, "mean_p": 0.430643, "boundary_entropy": 0.658602, "bpb": 3.604339}, "de": {"chunks": 3789.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.184532, "cps": 29.601562, "mean_p": 0.434398, "boundary_entropy": 0.664798, "bpb": 3.844632}, "vi": {"chunks": 5602.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.303878, "cps": 43.765625, "mean_p": 0.424985, "boundary_entropy": 0.651707, "bpb": 3.178753}, "tr": {"chunks": 4284.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.236646, "cps": 33.46875, "mean_p": 0.449731, "boundary_entropy": 0.670818, "bpb": 3.859596}, "fi": {"chunks": 3053.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.162299, "cps": 23.851562, "mean_p": 0.427812, "boundary_entropy": 0.665583, "bpb": 3.881141}, "ru": {"chunks": 5791.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.305594, "cps": 45.242188, "mean_p": 0.481948, "boundary_entropy": 0.688601, "bpb": 2.666258}, "el": {"chunks": 8202.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.399649, "cps": 64.078125, "mean_p": 0.490087, "boundary_entropy": 0.689796, "bpb": 2.765946}, "hi": {"chunks": 14366.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.853848, "cps": 112.234375, "mean_p": 0.477425, "boundary_entropy": 0.685797, "bpb": 2.153823}, "ta": {"chunks": 22801.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 1.219631, "cps": 178.132812, "mean_p": 0.481839, "boundary_entropy": 0.682142, "bpb": 1.938195}, "th": {"chunks": 14739.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.91632, "cps": 115.148438, "mean_p": 0.474096, "boundary_entropy": 0.6861, "bpb": 2.097571}, "zh": {"chunks": 8165.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 1.382492, "cps": 63.789062, "mean_p": 0.402905, "boundary_entropy": 0.600256, "bpb": 4.494328}, "my": {"chunks": 25161.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 1.327897, "cps": 196.570312, "mean_p": 0.425153, "boundary_entropy": 0.651439, "bpb": 2.189391}}, "bpb": {"en": 3.60434, "de": 3.84463, "vi": 3.17875, "tr": 3.8596, "fi": 3.88114, "ru": 2.66626, "el": 2.76595, "hi": 2.15382, "ta": 1.9382, "th": 2.09757, "zh": 4.49433, "my": 2.18939}, "gini_cpc": 0.405805, "gini_cps": 0.393398, "bpic": 3.231187, "worst_lang": "zh", "cpc_parity_vs_en": {"en": 1.0, "de": 1.138395662023245, "vi": 1.8746534548944058, "tr": 1.4598893648133393, "fi": 1.0012348435662866, "ru": 1.8852345254541183, "el": 2.465471314007889, "hi": 5.267466996886543, "ta": 7.524011646499115, "th": 5.652856848647356, "zh": 8.52871850411258, "my": 8.191917232086592}, "t": 34.9} |
| {"event": "train", "step": 300, "phase": 1, "objective": "global", "lm_loss": 2.17713, "bpb": 3.14093, "chunk_loss": 0.038293, "F": 0.30249, "G": 0.44986, "chunks": 309.75, "bytes_per_chunk": 3.306, "boundary_entropy": 0.66915, "router_noise": 0.0, "grad_norm": 1.3437, "lr": 0.0003, "router_trainable": true, "sec_per_step": 0.075, "alpha": 0.03, "t": 37.2} |
| {"event": "train", "step": 350, "phase": 1, "objective": "global", "lm_loss": 2.38702, "bpb": 3.44374, "chunk_loss": 0.033587, "F": 0.23401, "G": 0.42938, "chunks": 239.62, "bytes_per_chunk": 4.273, "boundary_entropy": 0.65672, "router_noise": 0.0, "grad_norm": 1.3079, "lr": 0.0002999, "router_trainable": true, "sec_per_step": 0.05, "alpha": 0.03, "t": 39.7} |
| {"event": "probe", "step": 400, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3181.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.182785, "cps": 24.851562, "mean_p": 0.437314, "boundary_entropy": 0.66525, "bpb": 3.171701}, "de": {"chunks": 3000.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.146106, "cps": 23.4375, "mean_p": 0.436444, "boundary_entropy": 0.671161, "bpb": 3.467661}, "vi": {"chunks": 7489.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.406238, "cps": 58.507812, "mean_p": 0.451719, "boundary_entropy": 0.663192, "bpb": 2.798031}, "tr": {"chunks": 3336.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.184279, "cps": 26.0625, "mean_p": 0.447979, "boundary_entropy": 0.669504, "bpb": 3.435488}, "fi": {"chunks": 2836.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.150763, "cps": 22.15625, "mean_p": 0.426802, "boundary_entropy": 0.666811, "bpb": 3.616152}, "ru": {"chunks": 5926.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.312718, "cps": 46.296875, "mean_p": 0.417352, "boundary_entropy": 0.626431, "bpb": 2.468132}, "el": {"chunks": 4750.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.231448, "cps": 37.109375, "mean_p": 0.45747, "boundary_entropy": 0.682422, "bpb": 2.570336}, "hi": {"chunks": 12715.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.755721, "cps": 99.335938, "mean_p": 0.463764, "boundary_entropy": 0.679345, "bpb": 1.960592}, "ta": {"chunks": 17757.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.949826, "cps": 138.726562, "mean_p": 0.480598, "boundary_entropy": 0.675207, "bpb": 1.736891}, "th": {"chunks": 12983.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.80715, "cps": 101.429688, "mean_p": 0.445191, "boundary_entropy": 0.674209, "bpb": 1.942811}, "zh": {"chunks": 3555.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.60193, "cps": 27.773438, "mean_p": 0.401509, "boundary_entropy": 0.624294, "bpb": 4.163197}, "my": {"chunks": 19682.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 1.038738, "cps": 153.765625, "mean_p": 0.438275, "boundary_entropy": 0.643164, "bpb": 1.950392}}, "bpb": {"en": 3.1717, "de": 3.46766, "vi": 2.79803, "tr": 3.43549, "fi": 3.61615, "ru": 2.46813, "el": 2.57034, "hi": 1.96059, "ta": 1.73689, "th": 1.94281, "zh": 4.1632, "my": 1.95039}, "gini_cpc": 0.367611, "gini_cps": 0.382853, "bpic": 3.947958, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 0.7993358633412944, "vi": 2.222496776399948, "tr": 1.0081750411296093, "fi": 0.8248116645883535, "ru": 1.710853741584059, "el": 1.2662317982984501, "hi": 4.134488065982409, "ta": 5.196423956463217, "th": 4.415851325044754, "zh": 3.2931128495012456, "my": 5.682851434031105}, "t": 43.3} |
| {"event": "eval", "step": 400, "phase": 1, "bpb": {"en": 3.18042, "de": 3.46434, "vi": 2.80509, "tr": 3.44392, "fi": 3.60881, "ru": 2.45113, "el": 2.56471, "hi": 1.9566, "ta": 1.7516, "th": 1.97388, "zh": 4.17109, "my": 1.96607}, "bytes_per_chunk": {"en": 5.003, "de": 6.593, "vi": 3.306, "tr": 6.147, "fi": 6.534, "ru": 6.704, "el": 7.171, "hi": 2.732, "ta": 2.984, "th": 3.204, "zh": 3.955, "my": 2.753}, "bpb_mean": 2.77814, "t": 44.7} |
| {"event": "train", "step": 400, "phase": 1, "objective": "global", "lm_loss": 2.0107, "bpb": 2.90082, "chunk_loss": 0.034399, "F": 0.23749, "G": 0.44055, "chunks": 243.19, "bytes_per_chunk": 4.211, "boundary_entropy": 0.66457, "router_noise": 0.0, "grad_norm": 1.078, "lr": 0.0002998, "router_trainable": true, "sec_per_step": 0.1, "alpha": 0.03, "t": 44.7} |
| {"event": "train", "step": 450, "phase": 1, "objective": "global", "lm_loss": 2.05868, "bpb": 2.97004, "chunk_loss": 0.032083, "F": 0.20029, "G": 0.43098, "chunks": 205.09, "bytes_per_chunk": 4.993, "boundary_entropy": 0.66106, "router_noise": 0.0, "grad_norm": 0.8984, "lr": 0.0002995, "router_trainable": true, "sec_per_step": 0.047, "alpha": 0.03, "t": 47.1} |
| {"event": "train", "step": 500, "phase": 1, "objective": "global", "lm_loss": 2.06099, "bpb": 2.97338, "chunk_loss": 0.031066, "F": 0.18448, "G": 0.42578, "chunks": 188.91, "bytes_per_chunk": 5.421, "boundary_entropy": 0.66005, "router_noise": 0.0, "grad_norm": 0.8853, "lr": 0.0002992, "router_trainable": true, "sec_per_step": 0.046, "alpha": 0.03, "t": 49.4} |
| {"event": "train", "step": 550, "phase": 1, "objective": "global", "lm_loss": 1.9705, "bpb": 2.84284, "chunk_loss": 0.031308, "F": 0.19049, "G": 0.42413, "chunks": 195.06, "bytes_per_chunk": 5.25, "boundary_entropy": 0.66052, "router_noise": 0.0, "grad_norm": 0.8843, "lr": 0.0002987, "router_trainable": true, "sec_per_step": 0.046, "alpha": 0.03, "t": 51.7} |
| {"event": "probe", "step": 600, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3252.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.186864, "cps": 25.40625, "mean_p": 0.438201, "boundary_entropy": 0.665958, "bpb": 2.899856}, "de": {"chunks": 2888.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.140652, "cps": 22.5625, "mean_p": 0.434215, "boundary_entropy": 0.669937, "bpb": 3.197829}, "vi": {"chunks": 5483.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.297423, "cps": 42.835938, "mean_p": 0.445659, "boundary_entropy": 0.660262, "bpb": 2.542356}, "tr": {"chunks": 4235.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.233939, "cps": 33.085938, "mean_p": 0.452669, "boundary_entropy": 0.668575, "bpb": 3.126755}, "fi": {"chunks": 3378.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179576, "cps": 26.390625, "mean_p": 0.436521, "boundary_entropy": 0.668568, "bpb": 3.366432}, "ru": {"chunks": 8353.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.440792, "cps": 65.257812, "mean_p": 0.414236, "boundary_entropy": 0.62962, "bpb": 2.200213}, "el": {"chunks": 4417.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.215222, "cps": 34.507812, "mean_p": 0.410404, "boundary_entropy": 0.654363, "bpb": 2.348004}, "hi": {"chunks": 6767.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.402199, "cps": 52.867188, "mean_p": 0.42087, "boundary_entropy": 0.656404, "bpb": 1.774394}, "ta": {"chunks": 15273.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.816956, "cps": 119.320312, "mean_p": 0.493391, "boundary_entropy": 0.658694, "bpb": 1.485577}, "th": {"chunks": 11085.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.689151, "cps": 86.601562, "mean_p": 0.410598, "boundary_entropy": 0.658088, "bpb": 1.87491}, "zh": {"chunks": 3729.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.631392, "cps": 29.132812, "mean_p": 0.424852, "boundary_entropy": 0.651814, "bpb": 3.946796}, "my": {"chunks": 18428.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.972556, "cps": 143.96875, "mean_p": 0.434055, "boundary_entropy": 0.659378, "bpb": 1.845076}}, "bpb": {"en": 2.89986, "de": 3.19783, "vi": 2.54236, "tr": 3.12675, "fi": 3.36643, "ru": 2.20021, "el": 2.348, "hi": 1.77439, "ta": 1.48558, "th": 1.87491, "zh": 3.9468, "my": 1.84508}, "gini_cpc": 0.339458, "gini_cps": 0.34833, "bpic": 4.396721, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 0.7526938455475238, "vi": 1.5916540813089173, "tr": 1.251919622300681, "fi": 0.9609954801903025, "ru": 2.3588854433399216, "el": 1.1517550824536165, "hi": 2.152358882132655, "ta": 4.371922608902473, "th": 3.6879770981421345, "zh": 3.3788781000746004, "my": 5.204612624023644}, "t": 54.8} |
| {"event": "eval", "step": 600, "phase": 1, "bpb": {"en": 2.9078, "de": 3.19416, "vi": 2.54931, "tr": 3.13783, "fi": 3.36386, "ru": 2.18507, "el": 2.34217, "hi": 1.76882, "ta": 1.49904, "th": 1.9024, "zh": 3.96021, "my": 1.85855}, "bytes_per_chunk": {"en": 4.992, "de": 6.774, "vi": 4.783, "tr": 4.815, "fi": 5.371, "ru": 4.286, "el": 8.575, "hi": 6.902, "ta": 3.343, "th": 3.834, "zh": 4.275, "my": 2.951}, "bpb_mean": 2.55577, "t": 56.2} |
| {"event": "train", "step": 600, "phase": 1, "objective": "global", "lm_loss": 1.88773, "bpb": 2.72342, "chunk_loss": 0.032903, "F": 0.21765, "G": 0.42437, "chunks": 222.88, "bytes_per_chunk": 4.595, "boundary_entropy": 0.65649, "router_noise": 0.0, "grad_norm": 0.7167, "lr": 0.0002982, "router_trainable": true, "sec_per_step": 0.092, "alpha": 0.03, "t": 56.2} |
| {"event": "train", "step": 650, "phase": 1, "objective": "global", "lm_loss": 1.90337, "bpb": 2.74598, "chunk_loss": 0.031213, "F": 0.18683, "G": 0.42924, "chunks": 191.31, "bytes_per_chunk": 5.352, "boundary_entropy": 0.66218, "router_noise": 0.0, "grad_norm": 0.9137, "lr": 0.0002975, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 58.5} |
| {"event": "train", "step": 700, "phase": 1, "objective": "global", "lm_loss": 1.84538, "bpb": 2.66232, "chunk_loss": 0.033079, "F": 0.21817, "G": 0.43666, "chunks": 223.41, "bytes_per_chunk": 4.584, "boundary_entropy": 0.66433, "router_noise": 0.0, "grad_norm": 0.8447, "lr": 0.0002967, "router_trainable": true, "sec_per_step": 0.045, "alpha": 0.03, "t": 60.7} |
| {"event": "train", "step": 750, "phase": 1, "objective": "global", "lm_loss": 1.91698, "bpb": 2.76561, "chunk_loss": 0.032703, "F": 0.21136, "G": 0.43518, "chunks": 216.44, "bytes_per_chunk": 4.731, "boundary_entropy": 0.66515, "router_noise": 0.0, "grad_norm": 0.9065, "lr": 0.0002959, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 62.9} |
| {"event": "probe", "step": 800, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3681.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.211515, "cps": 28.757812, "mean_p": 0.446435, "boundary_entropy": 0.668812, "bpb": 2.700405}, "de": {"chunks": 3873.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.188623, "cps": 30.257812, "mean_p": 0.445085, "boundary_entropy": 0.672554, "bpb": 3.023645}, "vi": {"chunks": 5288.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.286846, "cps": 41.3125, "mean_p": 0.450414, "boundary_entropy": 0.667357, "bpb": 2.412511}, "tr": {"chunks": 3551.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.196155, "cps": 27.742188, "mean_p": 0.451632, "boundary_entropy": 0.670612, "bpb": 2.975929}, "fi": {"chunks": 3665.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.194833, "cps": 28.632812, "mean_p": 0.44244, "boundary_entropy": 0.671824, "bpb": 3.221271}, "ru": {"chunks": 9593.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.506227, "cps": 74.945312, "mean_p": 0.422763, "boundary_entropy": 0.647833, "bpb": 2.040176}, "el": {"chunks": 4346.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.211762, "cps": 33.953125, "mean_p": 0.418976, "boundary_entropy": 0.65915, "bpb": 2.165339}, "hi": {"chunks": 7861.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.467221, "cps": 61.414062, "mean_p": 0.438216, "boundary_entropy": 0.667085, "bpb": 1.664161}, "ta": {"chunks": 19033.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 1.01808, "cps": 148.695312, "mean_p": 0.510163, "boundary_entropy": 0.660185, "bpb": 1.372801}, "th": {"chunks": 14807.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.920547, "cps": 115.679688, "mean_p": 0.432866, "boundary_entropy": 0.666641, "bpb": 1.693609}, "zh": {"chunks": 3500.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.592618, "cps": 27.34375, "mean_p": 0.42346, "boundary_entropy": 0.652322, "bpb": 3.809418}, "my": {"chunks": 712.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.037577, "cps": 5.5625, "mean_p": 0.420185, "boundary_entropy": 0.670437, "bpb": 1.585948}}, "bpb": {"en": 2.7004, "de": 3.02365, "vi": 2.41251, "tr": 2.97593, "fi": 3.22127, "ru": 2.04018, "el": 2.16534, "hi": 1.66416, "ta": 1.3728, "th": 1.69361, "zh": 3.80942, "my": 1.58595}, "gini_cpc": 0.388898, "gini_cps": 0.387793, "bpic": 4.802665, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8917710973223735, "vi": 1.3561464994410857, "tr": 0.9273814905299886, "fi": 0.9211288787530333, "ru": 2.3933352256721565, "el": 1.0011685063145606, "hi": 2.208925283159617, "ta": 4.813268412663019, "th": 4.3521545963920785, "zh": 2.801772733432303, "my": 0.17765397134593386}, "t": 66.1} |
| {"event": "eval", "step": 800, "phase": 1, "bpb": {"en": 2.7107, "de": 3.02099, "vi": 2.41991, "tr": 2.98826, "fi": 3.21679, "ru": 2.02105, "el": 2.15379, "hi": 1.65829, "ta": 1.38455, "th": 1.7225, "zh": 3.82301, "my": 1.59656}, "bytes_per_chunk": {"en": 4.436, "de": 5.03, "vi": 4.908, "tr": 5.665, "fi": 5.211, "ru": 3.663, "el": 9.516, "hi": 5.615, "ta": 2.682, "th": 3.067, "zh": 4.491, "my": 49.284}, "bpb_mean": 2.39303, "t": 67.5} |
| {"event": "train", "step": 800, "phase": 1, "objective": "global", "lm_loss": 1.64614, "bpb": 2.37488, "chunk_loss": 0.033324, "F": 0.21985, "G": 0.44019, "chunks": 225.12, "bytes_per_chunk": 4.549, "boundary_entropy": 0.66632, "router_noise": 0.0, "grad_norm": 0.8498, "lr": 0.0002949, "router_trainable": true, "sec_per_step": 0.093, "alpha": 0.03, "t": 67.5} |
| {"event": "train", "step": 850, "phase": 1, "objective": "global", "lm_loss": 1.72637, "bpb": 2.49063, "chunk_loss": 0.03316, "F": 0.22046, "G": 0.43538, "chunks": 225.75, "bytes_per_chunk": 4.536, "boundary_entropy": 0.66528, "router_noise": 0.0, "grad_norm": 0.7736, "lr": 0.0002938, "router_trainable": true, "sec_per_step": 0.045, "alpha": 0.03, "t": 69.8} |
| {"event": "train", "step": 900, "phase": 1, "objective": "global", "lm_loss": 1.63613, "bpb": 2.36043, "chunk_loss": 0.033006, "F": 0.21567, "G": 0.43855, "chunks": 220.84, "bytes_per_chunk": 4.637, "boundary_entropy": 0.66545, "router_noise": 0.0, "grad_norm": 0.7928, "lr": 0.0002927, "router_trainable": true, "sec_per_step": 0.046, "alpha": 0.03, "t": 72.1} |
| {"event": "train", "step": 950, "phase": 1, "objective": "global", "lm_loss": 1.74102, "bpb": 2.51176, "chunk_loss": 0.032587, "F": 0.21045, "G": 0.43563, "chunks": 215.5, "bytes_per_chunk": 4.752, "boundary_entropy": 0.66566, "router_noise": 0.0, "grad_norm": 0.7885, "lr": 0.0002914, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 74.3} |
| {"event": "probe", "step": 1000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3644.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.209389, "cps": 28.46875, "mean_p": 0.445228, "boundary_entropy": 0.668028, "bpb": 2.57653}, "de": {"chunks": 3413.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.16622, "cps": 26.664062, "mean_p": 0.438263, "boundary_entropy": 0.670592, "bpb": 2.904971}, "vi": {"chunks": 5034.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.273068, "cps": 39.328125, "mean_p": 0.44482, "boundary_entropy": 0.664818, "bpb": 2.28719}, "tr": {"chunks": 4664.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.257637, "cps": 36.4375, "mean_p": 0.457482, "boundary_entropy": 0.67157, "bpb": 2.87994}, "fi": {"chunks": 2709.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.144011, "cps": 21.164062, "mean_p": 0.43464, "boundary_entropy": 0.669296, "bpb": 3.16544}, "ru": {"chunks": 9144.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.482533, "cps": 71.4375, "mean_p": 0.428376, "boundary_entropy": 0.653908, "bpb": 1.959488}, "el": {"chunks": 3313.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.161429, "cps": 25.882812, "mean_p": 0.413797, "boundary_entropy": 0.663049, "bpb": 2.079721}, "hi": {"chunks": 7034.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.418068, "cps": 54.953125, "mean_p": 0.433223, "boundary_entropy": 0.667711, "bpb": 1.577217}, "ta": {"chunks": 12239.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.654667, "cps": 95.617188, "mean_p": 0.485142, "boundary_entropy": 0.664579, "bpb": 1.396644}, "th": {"chunks": 7374.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.45844, "cps": 57.609375, "mean_p": 0.451096, "boundary_entropy": 0.677538, "bpb": 1.605977}, "zh": {"chunks": 3055.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.517271, "cps": 23.867188, "mean_p": 0.422374, "boundary_entropy": 0.654896, "bpb": 3.706516}, "my": {"chunks": 1641.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.086605, "cps": 12.820312, "mean_p": 0.427754, "boundary_entropy": 0.67115, "bpb": 1.462954}}, "bpb": {"en": 2.57653, "de": 2.90497, "vi": 2.28719, "tr": 2.87994, "fi": 3.16544, "ru": 1.95949, "el": 2.07972, "hi": 1.57722, "ta": 1.39664, "th": 1.60598, "zh": 3.70652, "my": 1.46295}, "gini_cpc": 0.303758, "gini_cps": 0.301928, "bpic": 6.066341, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.7938338851388755, "vi": 1.3041147926214185, "tr": 1.2304210211199689, "fi": 0.6877694381005672, "ru": 2.3044790004315474, "el": 0.7709502305062891, "hi": 1.996609085259736, "ta": 3.126556038950992, "th": 2.189413642603253, "zh": 2.4703786415273608, "my": 0.41360993007148567}, "t": 77.5} |
| {"event": "eval", "step": 1000, "phase": 1, "bpb": {"en": 2.58947, "de": 2.91039, "vi": 2.29479, "tr": 2.89056, "fi": 3.16326, "ru": 1.94042, "el": 2.06959, "hi": 1.57234, "ta": 1.40542, "th": 1.62873, "zh": 3.71912, "my": 1.47287}, "bytes_per_chunk": {"en": 4.583, "de": 5.908, "vi": 5.257, "tr": 4.046, "fi": 7.157, "ru": 3.857, "el": 12.385, "hi": 5.906, "ta": 4.195, "th": 5.83, "zh": 5.436, "my": 22.478}, "bpb_mean": 2.30475, "t": 78.8} |
| {"event": "train", "step": 1000, "phase": 1, "objective": "global", "lm_loss": 1.7539, "bpb": 2.53035, "chunk_loss": 0.031298, "F": 0.18793, "G": 0.43362, "chunks": 192.44, "bytes_per_chunk": 5.321, "boundary_entropy": 0.66744, "router_noise": 0.0, "grad_norm": 0.8121, "lr": 0.0002901, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 78.8} |
| {"event": "train", "step": 1050, "phase": 1, "objective": "global", "lm_loss": 1.52086, "bpb": 2.19413, "chunk_loss": 0.034097, "F": 0.23486, "G": 0.44227, "chunks": 240.5, "bytes_per_chunk": 4.258, "boundary_entropy": 0.66983, "router_noise": 0.0, "grad_norm": 0.8024, "lr": 0.0002886, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 81.0} |
| {"event": "train", "step": 1100, "phase": 1, "objective": "global", "lm_loss": 1.76989, "bpb": 2.5534, "chunk_loss": 0.032483, "F": 0.20749, "G": 0.44051, "chunks": 212.47, "bytes_per_chunk": 4.82, "boundary_entropy": 0.66964, "router_noise": 0.0, "grad_norm": 0.829, "lr": 0.0002871, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 83.1} |
| {"event": "train", "step": 1150, "phase": 1, "objective": "global", "lm_loss": 1.68369, "bpb": 2.42905, "chunk_loss": 0.031745, "F": 0.19611, "G": 0.43358, "chunks": 200.81, "bytes_per_chunk": 5.099, "boundary_entropy": 0.6657, "router_noise": 0.0, "grad_norm": 0.7665, "lr": 0.0002855, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 85.3} |
| {"event": "probe", "step": 1200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3495.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.200827, "cps": 27.304688, "mean_p": 0.445521, "boundary_entropy": 0.668478, "bpb": 2.481139}, "de": {"chunks": 3524.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.171626, "cps": 27.53125, "mean_p": 0.443884, "boundary_entropy": 0.672638, "bpb": 2.782676}, "vi": {"chunks": 5450.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.295633, "cps": 42.578125, "mean_p": 0.447862, "boundary_entropy": 0.665033, "bpb": 2.224903}, "tr": {"chunks": 4444.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.245484, "cps": 34.71875, "mean_p": 0.455403, "boundary_entropy": 0.670152, "bpb": 2.789218}, "fi": {"chunks": 3100.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.164797, "cps": 24.21875, "mean_p": 0.436382, "boundary_entropy": 0.669335, "bpb": 3.042167}, "ru": {"chunks": 9805.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.517414, "cps": 76.601562, "mean_p": 0.433294, "boundary_entropy": 0.659668, "bpb": 1.889764}, "el": {"chunks": 5155.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.251182, "cps": 40.273438, "mean_p": 0.428652, "boundary_entropy": 0.66935, "bpb": 1.973475}, "hi": {"chunks": 10215.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.607132, "cps": 79.804688, "mean_p": 0.453867, "boundary_entropy": 0.674479, "bpb": 1.568878}, "ta": {"chunks": 15386.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.823001, "cps": 120.203125, "mean_p": 0.499803, "boundary_entropy": 0.667631, "bpb": 1.265967}, "th": {"chunks": 9791.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.608704, "cps": 76.492188, "mean_p": 0.444598, "boundary_entropy": 0.675164, "bpb": 1.501062}, "zh": {"chunks": 3004.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.508635, "cps": 23.46875, "mean_p": 0.416226, "boundary_entropy": 0.652067, "bpb": 3.614952}, "my": {"chunks": 3934.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.207621, "cps": 30.734375, "mean_p": 0.443973, "boundary_entropy": 0.677532, "bpb": 1.441098}}, "bpb": {"en": 2.48114, "de": 2.78268, "vi": 2.2249, "tr": 2.78922, "fi": 3.04217, "ru": 1.88976, "el": 1.97347, "hi": 1.56888, "ta": 1.26597, "th": 1.50106, "zh": 3.61495, "my": 1.4411}, "gini_cpc": 0.296986, "gini_cps": 0.302922, "bpic": 4.964633, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8545951675404128, "vi": 1.4720762311846167, "tr": 1.2223636847709067, "fi": 0.8205910020230606, "ru": 2.5764120624639038, "el": 1.2507334490623834, "hi": 3.0231537730614946, "ta": 4.098049488800911, "th": 3.0309789863666863, "zh": 2.532698107256896, "my": 1.0338271175414802}, "t": 88.4} |
| {"event": "eval", "step": 1200, "phase": 1, "bpb": {"en": 2.49124, "de": 2.78491, "vi": 2.23136, "tr": 2.80218, "fi": 3.04475, "ru": 1.87374, "el": 1.96194, "hi": 1.56468, "ta": 1.27639, "th": 1.52409, "zh": 3.626, "my": 1.45085}, "bytes_per_chunk": {"en": 4.976, "de": 5.648, "vi": 4.588, "tr": 4.352, "fi": 6.313, "ru": 3.602, "el": 6.85, "hi": 4.447, "ta": 3.535, "th": 4.547, "zh": 5.55, "my": 9.538}, "bpb_mean": 2.21934, "t": 89.7} |
| {"event": "train", "step": 1200, "phase": 1, "objective": "global", "lm_loss": 1.62563, "bpb": 2.34529, "chunk_loss": 0.033656, "F": 0.22675, "G": 0.44344, "chunks": 232.19, "bytes_per_chunk": 4.41, "boundary_entropy": 0.66932, "router_noise": 0.0, "grad_norm": 0.7552, "lr": 0.0002837, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 89.7} |
| {"event": "train", "step": 1250, "phase": 1, "objective": "global", "lm_loss": 1.66224, "bpb": 2.3981, "chunk_loss": 0.032061, "F": 0.20074, "G": 0.4367, "chunks": 205.56, "bytes_per_chunk": 4.981, "boundary_entropy": 0.66864, "router_noise": 0.0, "grad_norm": 0.9662, "lr": 0.0002819, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 91.8} |
| {"event": "train", "step": 1300, "phase": 1, "objective": "global", "lm_loss": 1.59279, "bpb": 2.29791, "chunk_loss": 0.033492, "F": 0.22385, "G": 0.44254, "chunks": 229.22, "bytes_per_chunk": 4.467, "boundary_entropy": 0.66982, "router_noise": 0.0, "grad_norm": 0.7652, "lr": 0.00028, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 94.0} |
| {"event": "train", "step": 1350, "phase": 1, "objective": "global", "lm_loss": 1.71578, "bpb": 2.47534, "chunk_loss": 0.032595, "F": 0.21075, "G": 0.43899, "chunks": 215.81, "bytes_per_chunk": 4.745, "boundary_entropy": 0.67089, "router_noise": 0.0, "grad_norm": 0.9116, "lr": 0.000278, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 96.2} |
| {"event": "probe", "step": 1400, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3779.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.217146, "cps": 29.523438, "mean_p": 0.444829, "boundary_entropy": 0.66781, "bpb": 2.430971}, "de": {"chunks": 3875.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.188721, "cps": 30.273438, "mean_p": 0.443741, "boundary_entropy": 0.671519, "bpb": 2.729554}, "vi": {"chunks": 5269.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.285815, "cps": 41.164062, "mean_p": 0.447236, "boundary_entropy": 0.663777, "bpb": 2.188014}, "tr": {"chunks": 3944.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.217864, "cps": 30.8125, "mean_p": 0.449631, "boundary_entropy": 0.668831, "bpb": 2.726346}, "fi": {"chunks": 2695.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.143267, "cps": 21.054688, "mean_p": 0.43272, "boundary_entropy": 0.668171, "bpb": 2.956527}, "ru": {"chunks": 9423.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.497256, "cps": 73.617188, "mean_p": 0.434261, "boundary_entropy": 0.661575, "bpb": 1.832511}, "el": {"chunks": 6227.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.303416, "cps": 48.648438, "mean_p": 0.431937, "boundary_entropy": 0.670104, "bpb": 1.921032}, "hi": {"chunks": 7968.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.473581, "cps": 62.25, "mean_p": 0.44271, "boundary_entropy": 0.671087, "bpb": 1.481012}, "ta": {"chunks": 14266.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.763092, "cps": 111.453125, "mean_p": 0.487416, "boundary_entropy": 0.668099, "bpb": 1.207934}, "th": {"chunks": 10967.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.681815, "cps": 85.679688, "mean_p": 0.447299, "boundary_entropy": 0.676459, "bpb": 1.424507}, "zh": {"chunks": 3481.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.589401, "cps": 27.195312, "mean_p": 0.417382, "boundary_entropy": 0.651736, "bpb": 3.575205}, "my": {"chunks": 14909.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.786838, "cps": 116.476562, "mean_p": 0.447687, "boundary_entropy": 0.676008, "bpb": 1.54895}}, "bpb": {"en": 2.43097, "de": 2.72955, "vi": 2.18801, "tr": 2.72635, "fi": 2.95653, "ru": 1.83251, "el": 1.92103, "hi": 1.48101, "ta": 1.20793, "th": 1.42451, "zh": 3.5752, "my": 1.54895}, "gini_cpc": 0.294009, "gini_cps": 0.309614, "bpic": 4.421287, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8690935523060375, "vi": 1.3162315145301073, "tr": 1.0033064012568578, "fi": 0.6597723551059129, "ru": 2.2899563551106272, "el": 1.3972858106634811, "hi": 2.1809287660320367, "ta": 3.5141798034930862, "th": 3.1398869116445507, "zh": 2.7142997639565687, "my": 3.623534219054302}, "t": 99.2} |
| {"event": "eval", "step": 1400, "phase": 1, "bpb": {"en": 2.44028, "de": 2.73284, "vi": 2.19602, "tr": 2.73713, "fi": 2.95908, "ru": 1.81334, "el": 1.91103, "hi": 1.4783, "ta": 1.21679, "th": 1.4477, "zh": 3.58203, "my": 1.55841}, "bytes_per_chunk": {"en": 4.7, "de": 5.384, "vi": 4.97, "tr": 4.99, "fi": 6.905, "ru": 3.732, "el": 5.941, "hi": 5.501, "ta": 3.68, "th": 4.035, "zh": 4.733, "my": 3.856}, "bpb_mean": 2.17275, "t": 100.6} |
| {"event": "train", "step": 1400, "phase": 1, "objective": "global", "lm_loss": 1.58787, "bpb": 2.29082, "chunk_loss": 0.032762, "F": 0.21359, "G": 0.43674, "chunks": 218.72, "bytes_per_chunk": 4.682, "boundary_entropy": 0.66946, "router_noise": 0.0, "grad_norm": 1.0044, "lr": 0.0002759, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 100.6} |
| {"event": "train", "step": 1450, "phase": 1, "objective": "global", "lm_loss": 1.73349, "bpb": 2.50089, "chunk_loss": 0.031764, "F": 0.19659, "G": 0.43445, "chunks": 201.31, "bytes_per_chunk": 5.087, "boundary_entropy": 0.66876, "router_noise": 0.0, "grad_norm": 0.7808, "lr": 0.0002738, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 102.8} |
| {"event": "train", "step": 1500, "phase": 1, "objective": "global", "lm_loss": 1.57246, "bpb": 2.26858, "chunk_loss": 0.032819, "F": 0.21411, "G": 0.4386, "chunks": 219.25, "bytes_per_chunk": 4.67, "boundary_entropy": 0.66846, "router_noise": 0.0, "grad_norm": 0.8184, "lr": 0.0002715, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 104.9} |
| {"event": "train", "step": 1550, "phase": 1, "objective": "global", "lm_loss": 1.62578, "bpb": 2.3455, "chunk_loss": 0.033233, "F": 0.22086, "G": 0.43818, "chunks": 226.16, "bytes_per_chunk": 4.528, "boundary_entropy": 0.66909, "router_noise": 0.0, "grad_norm": 0.8147, "lr": 0.0002692, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 107.1} |
| {"event": "probe", "step": 1600, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3620.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.20801, "cps": 28.28125, "mean_p": 0.443997, "boundary_entropy": 0.668257, "bpb": 2.355843}, "de": {"chunks": 3832.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.186626, "cps": 29.9375, "mean_p": 0.442608, "boundary_entropy": 0.671798, "bpb": 2.655587}, "vi": {"chunks": 5682.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.308218, "cps": 44.390625, "mean_p": 0.447671, "boundary_entropy": 0.665278, "bpb": 2.126098}, "tr": {"chunks": 4025.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.222339, "cps": 31.445312, "mean_p": 0.452418, "boundary_entropy": 0.670462, "bpb": 2.645621}, "fi": {"chunks": 3353.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.178247, "cps": 26.195312, "mean_p": 0.4384, "boundary_entropy": 0.66962, "bpb": 2.906973}, "ru": {"chunks": 8727.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.460528, "cps": 68.179688, "mean_p": 0.431754, "boundary_entropy": 0.66269, "bpb": 1.767966}, "el": {"chunks": 4430.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.215855, "cps": 34.609375, "mean_p": 0.430713, "boundary_entropy": 0.671125, "bpb": 1.879174}, "hi": {"chunks": 18238.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 1.083982, "cps": 142.484375, "mean_p": 0.454268, "boundary_entropy": 0.670054, "bpb": 1.446802}, "ta": {"chunks": 12862.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.687991, "cps": 100.484375, "mean_p": 0.48088, "boundary_entropy": 0.666784, "bpb": 1.226772}, "th": {"chunks": 8747.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.543799, "cps": 68.335938, "mean_p": 0.441502, "boundary_entropy": 0.674929, "bpb": 1.375184}, "zh": {"chunks": 3846.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.651202, "cps": 30.046875, "mean_p": 0.417223, "boundary_entropy": 0.650823, "bpb": 3.532932}, "my": {"chunks": 20370.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 1.075047, "cps": 159.140625, "mean_p": 0.454693, "boundary_entropy": 0.677052, "bpb": 1.318159}}, "bpb": {"en": 2.35584, "de": 2.65559, "vi": 2.1261, "tr": 2.64562, "fi": 2.90697, "ru": 1.76797, "el": 1.87917, "hi": 1.4468, "ta": 1.22677, "th": 1.37518, "zh": 3.53293, "my": 1.31816}, "gini_cpc": 0.352296, "gini_cps": 0.359676, "bpic": 3.926871, "worst_lang": "hi", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8971987150182446, "vi": 1.4817455686471956, "tr": 1.0688847549153204, "fi": 0.8569139632946634, "ru": 2.2139678566742957, "el": 1.0377158035601077, "hi": 5.211199362957977, "ta": 3.3074903473762447, "th": 2.614289092648345, "zh": 3.130627435719458, "my": 5.168246302471318}, "t": 110.3} |
| {"event": "eval", "step": 1600, "phase": 1, "bpb": {"en": 2.36806, "de": 2.66047, "vi": 2.13104, "tr": 2.65406, "fi": 2.91009, "ru": 1.74996, "el": 1.86744, "hi": 1.44554, "ta": 1.23403, "th": 1.39586, "zh": 3.53904, "my": 1.32643}, "bytes_per_chunk": {"en": 4.572, "de": 5.33, "vi": 4.445, "tr": 4.921, "fi": 5.546, "ru": 3.949, "el": 8.401, "hi": 2.367, "ta": 4.064, "th": 5.074, "zh": 4.459, "my": 2.619}, "bpb_mean": 2.10683, "t": 111.6} |
| {"event": "train", "step": 1600, "phase": 1, "objective": "global", "lm_loss": 1.60682, "bpb": 2.31815, "chunk_loss": 0.033036, "F": 0.21774, "G": 0.43685, "chunks": 222.97, "bytes_per_chunk": 4.593, "boundary_entropy": 0.66734, "router_noise": 0.0, "grad_norm": 0.8029, "lr": 0.0002668, "router_trainable": true, "sec_per_step": 0.093, "alpha": 0.03, "t": 111.7} |
| {"event": "train", "step": 1650, "phase": 1, "objective": "global", "lm_loss": 1.53402, "bpb": 2.21313, "chunk_loss": 0.033609, "F": 0.22574, "G": 0.44254, "chunks": 231.16, "bytes_per_chunk": 4.43, "boundary_entropy": 0.66959, "router_noise": 0.0, "grad_norm": 0.7971, "lr": 0.0002643, "router_trainable": true, "sec_per_step": 0.044, "alpha": 0.03, "t": 113.9} |
| {"event": "train", "step": 1700, "phase": 1, "objective": "global", "lm_loss": 1.45651, "bpb": 2.1013, "chunk_loss": 0.03282, "F": 0.21408, "G": 0.43722, "chunks": 219.22, "bytes_per_chunk": 4.671, "boundary_entropy": 0.66851, "router_noise": 0.0, "grad_norm": 0.7132, "lr": 0.0002618, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 116.0} |
| {"event": "train", "step": 1750, "phase": 1, "objective": "global", "lm_loss": 1.63551, "bpb": 2.35954, "chunk_loss": 0.031339, "F": 0.18991, "G": 0.43432, "chunks": 194.47, "bytes_per_chunk": 5.266, "boundary_entropy": 0.66954, "router_noise": 0.0, "grad_norm": 0.8568, "lr": 0.0002591, "router_trainable": true, "sec_per_step": 0.043, "alpha": 0.03, "t": 118.2} |
| {"event": "probe", "step": 1800, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3788.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.217664, "cps": 29.59375, "mean_p": 0.446024, "boundary_entropy": 0.669022, "bpb": 2.324433}, "de": {"chunks": 3518.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.171334, "cps": 27.484375, "mean_p": 0.440179, "boundary_entropy": 0.6713, "bpb": 2.586937}, "vi": {"chunks": 4886.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.265039, "cps": 38.171875, "mean_p": 0.439438, "boundary_entropy": 0.662668, "bpb": 2.11551}, "tr": {"chunks": 4273.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.236038, "cps": 33.382812, "mean_p": 0.454132, "boundary_entropy": 0.670666, "bpb": 2.621236}, "fi": {"chunks": 3664.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.19478, "cps": 28.625, "mean_p": 0.442519, "boundary_entropy": 0.670875, "bpb": 2.84472}, "ru": {"chunks": 8938.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.471662, "cps": 69.828125, "mean_p": 0.432746, "boundary_entropy": 0.66241, "bpb": 1.724828}, "el": {"chunks": 4097.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.19963, "cps": 32.007812, "mean_p": 0.427939, "boundary_entropy": 0.670676, "bpb": 1.835781}, "hi": {"chunks": 8658.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.514591, "cps": 67.640625, "mean_p": 0.44315, "boundary_entropy": 0.671264, "bpb": 1.424736}, "ta": {"chunks": 16154.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.864081, "cps": 126.203125, "mean_p": 0.487445, "boundary_entropy": 0.664838, "bpb": 1.146274}, "th": {"chunks": 3338.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.207523, "cps": 26.078125, "mean_p": 0.432665, "boundary_entropy": 0.674284, "bpb": 1.354827}, "zh": {"chunks": 3804.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.644091, "cps": 29.71875, "mean_p": 0.413841, "boundary_entropy": 0.649285, "bpb": 3.494047}, "my": {"chunks": 5222.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.275596, "cps": 40.796875, "mean_p": 0.445704, "boundary_entropy": 0.677698, "bpb": 1.315965}}, "bpb": {"en": 2.32443, "de": 2.58694, "vi": 2.11551, "tr": 2.62124, "fi": 2.84472, "ru": 1.72483, "el": 1.83578, "hi": 1.42474, "ta": 1.14627, "th": 1.35483, "zh": 3.49405, "my": 1.31597}, "gini_cpc": 0.300957, "gini_cps": 0.278772, "bpic": 5.456085, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.7871501414443415, "vi": 1.2176556003584613, "tr": 1.0844174345311741, "fi": 0.8948654309693316, "ru": 2.166932014164993, "el": 0.9171476733472492, "hi": 2.364158875058643, "ta": 3.9698012021465003, "th": 0.9534093725289259, "zh": 2.9591107212574617, "my": 1.266157235976682}, "t": 121.3} |
| {"event": "eval", "step": 1800, "phase": 1, "bpb": {"en": 2.33331, "de": 2.5922, "vi": 2.11984, "tr": 2.63313, "fi": 2.84357, "ru": 1.70517, "el": 1.82619, "hi": 1.42434, "ta": 1.15367, "th": 1.37388, "zh": 3.49933, "my": 1.32514}, "bytes_per_chunk": {"en": 4.402, "de": 5.899, "vi": 5.356, "tr": 4.574, "fi": 4.859, "ru": 3.906, "el": 8.712, "hi": 5.155, "ta": 3.149, "th": 12.216, "zh": 4.434, "my": 9.765}, "bpb_mean": 2.06915, "t": 122.6} |
| {"event": "train", "step": 1800, "phase": 1, "objective": "global", "lm_loss": 1.58778, "bpb": 2.29069, "chunk_loss": 0.032358, "F": 0.20538, "G": 0.43971, "chunks": 210.31, "bytes_per_chunk": 4.869, "boundary_entropy": 0.671, "router_noise": 0.0, "grad_norm": 1.0566, "lr": 0.0002564, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 122.6} |
| {"event": "train", "step": 1850, "phase": 1, "objective": "global", "lm_loss": 1.56345, "bpb": 2.25558, "chunk_loss": 0.03203, "F": 0.2009, "G": 0.43737, "chunks": 205.72, "bytes_per_chunk": 4.978, "boundary_entropy": 0.67009, "router_noise": 0.0, "grad_norm": 0.9142, "lr": 0.0002537, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 124.7} |
| {"event": "train", "step": 1900, "phase": 1, "objective": "global", "lm_loss": 1.6136, "bpb": 2.32793, "chunk_loss": 0.032374, "F": 0.20779, "G": 0.43265, "chunks": 212.78, "bytes_per_chunk": 4.812, "boundary_entropy": 0.66754, "router_noise": 0.0, "grad_norm": 0.7849, "lr": 0.0002508, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 126.8} |
| {"event": "train", "step": 1950, "phase": 1, "objective": "global", "lm_loss": 1.61608, "bpb": 2.33151, "chunk_loss": 0.031567, "F": 0.19376, "G": 0.43399, "chunks": 198.41, "bytes_per_chunk": 5.161, "boundary_entropy": 0.6682, "router_noise": 0.0, "grad_norm": 0.9026, "lr": 0.0002479, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 128.9} |
| {"event": "probe", "step": 2000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3893.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.223697, "cps": 30.414062, "mean_p": 0.445614, "boundary_entropy": 0.668042, "bpb": 2.290107}, "de": {"chunks": 3915.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.190669, "cps": 30.585938, "mean_p": 0.443569, "boundary_entropy": 0.671766, "bpb": 2.572561}, "vi": {"chunks": 5308.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.287931, "cps": 41.46875, "mean_p": 0.444646, "boundary_entropy": 0.664653, "bpb": 2.087619}, "tr": {"chunks": 4371.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.241452, "cps": 34.148438, "mean_p": 0.45134, "boundary_entropy": 0.669687, "bpb": 2.570845}, "fi": {"chunks": 3108.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.165222, "cps": 24.28125, "mean_p": 0.438413, "boundary_entropy": 0.669801, "bpb": 2.786095}, "ru": {"chunks": 8825.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.465699, "cps": 68.945312, "mean_p": 0.433583, "boundary_entropy": 0.662815, "bpb": 1.697669}, "el": {"chunks": 4975.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.242411, "cps": 38.867188, "mean_p": 0.432154, "boundary_entropy": 0.671789, "bpb": 1.799228}, "hi": {"chunks": 10695.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.635661, "cps": 83.554688, "mean_p": 0.446396, "boundary_entropy": 0.670685, "bpb": 1.426613}, "ta": {"chunks": 12437.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.665258, "cps": 97.164062, "mean_p": 0.473445, "boundary_entropy": 0.666328, "bpb": 1.136411}, "th": {"chunks": 7195.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.447311, "cps": 56.210938, "mean_p": 0.438295, "boundary_entropy": 0.673425, "bpb": 1.305859}, "zh": {"chunks": 2823.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.477988, "cps": 22.054688, "mean_p": 0.410867, "boundary_entropy": 0.649562, "bpb": 3.475666}, "my": {"chunks": 8885.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.468915, "cps": 69.414062, "mean_p": 0.448146, "boundary_entropy": 0.677103, "bpb": 1.275413}}, "bpb": {"en": 2.29011, "de": 2.57256, "vi": 2.08762, "tr": 2.57085, "fi": 2.7861, "ru": 1.69767, "el": 1.79923, "hi": 1.42661, "ta": 1.13641, "th": 1.30586, "zh": 3.47567, "my": 1.27541}, "gini_cpc": 0.244742, "gini_cps": 0.264224, "bpic": 5.02134, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8523521789318514, "vi": 1.2871450436691672, "tr": 1.0793690765058617, "fi": 0.7385992174403533, "ru": 2.081829777687711, "el": 1.0836572886923719, "hi": 2.8416162815950754, "ta": 2.9739241066542856, "th": 1.9996291056191418, "zh": 2.1367669218649326, "my": 2.096205096528463}, "t": 132.0} |
| {"event": "eval", "step": 2000, "phase": 1, "bpb": {"en": 2.302, "de": 2.57937, "vi": 2.09122, "tr": 2.57917, "fi": 2.78756, "ru": 1.67857, "el": 1.78887, "hi": 1.4247, "ta": 1.14406, "th": 1.32552, "zh": 3.47843, "my": 1.28412}, "bytes_per_chunk": {"en": 4.531, "de": 5.313, "vi": 4.779, "tr": 4.444, "fi": 6.2, "ru": 4.011, "el": 7.436, "hi": 4.094, "ta": 4.09, "th": 6.36, "zh": 5.854, "my": 5.625}, "bpb_mean": 2.03863, "t": 133.3} |
| {"event": "train", "step": 2000, "phase": 1, "objective": "global", "lm_loss": 1.54315, "bpb": 2.22629, "chunk_loss": 0.032554, "F": 0.20938, "G": 0.43724, "chunks": 214.41, "bytes_per_chunk": 4.776, "boundary_entropy": 0.66934, "router_noise": 0.0, "grad_norm": 0.8407, "lr": 0.000245, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 133.3} |
| {"event": "train", "step": 2050, "phase": 1, "objective": "global", "lm_loss": 1.4215, "bpb": 2.0508, "chunk_loss": 0.033085, "F": 0.21866, "G": 0.43826, "chunks": 223.91, "bytes_per_chunk": 4.573, "boundary_entropy": 0.66906, "router_noise": 0.0, "grad_norm": 0.8268, "lr": 0.0002419, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 135.4} |
| {"event": "train", "step": 2100, "phase": 1, "objective": "global", "lm_loss": 1.57854, "bpb": 2.27736, "chunk_loss": 0.032222, "F": 0.20404, "G": 0.43824, "chunks": 208.94, "bytes_per_chunk": 4.901, "boundary_entropy": 0.66999, "router_noise": 0.0, "grad_norm": 0.8655, "lr": 0.0002388, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 137.5} |
| {"event": "train", "step": 2150, "phase": 1, "objective": "global", "lm_loss": 1.49758, "bpb": 2.16055, "chunk_loss": 0.032555, "F": 0.20938, "G": 0.438, "chunks": 214.41, "bytes_per_chunk": 4.776, "boundary_entropy": 0.66895, "router_noise": 0.0, "grad_norm": 0.7727, "lr": 0.0002357, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 139.6} |
| {"event": "probe", "step": 2200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3782.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.217319, "cps": 29.546875, "mean_p": 0.442825, "boundary_entropy": 0.667025, "bpb": 2.256389}, "de": {"chunks": 3912.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.190523, "cps": 30.5625, "mean_p": 0.442763, "boundary_entropy": 0.671109, "bpb": 2.520129}, "vi": {"chunks": 5046.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.273718, "cps": 39.421875, "mean_p": 0.438069, "boundary_entropy": 0.66101, "bpb": 2.062391}, "tr": {"chunks": 4407.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.24344, "cps": 34.429688, "mean_p": 0.449925, "boundary_entropy": 0.667575, "bpb": 2.514412}, "fi": {"chunks": 3610.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.191909, "cps": 28.203125, "mean_p": 0.442828, "boundary_entropy": 0.670751, "bpb": 2.741499}, "ru": {"chunks": 7862.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.414881, "cps": 61.421875, "mean_p": 0.432664, "boundary_entropy": 0.66508, "bpb": 1.669875}, "el": {"chunks": 3948.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.19237, "cps": 30.84375, "mean_p": 0.430238, "boundary_entropy": 0.671948, "bpb": 1.764423}, "hi": {"chunks": 8710.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.517682, "cps": 68.046875, "mean_p": 0.443969, "boundary_entropy": 0.671959, "bpb": 1.386151}, "ta": {"chunks": 14291.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.764429, "cps": 111.648438, "mean_p": 0.477074, "boundary_entropy": 0.664704, "bpb": 1.113131}, "th": {"chunks": 7442.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.462667, "cps": 58.140625, "mean_p": 0.442242, "boundary_entropy": 0.675528, "bpb": 1.286872}, "zh": {"chunks": 3590.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.607856, "cps": 28.046875, "mean_p": 0.416413, "boundary_entropy": 0.650951, "bpb": 3.416164}, "my": {"chunks": 20625.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 1.088505, "cps": 161.132812, "mean_p": 0.462022, "boundary_entropy": 0.679066, "bpb": 1.253389}}, "bpb": {"en": 2.25639, "de": 2.52013, "vi": 2.06239, "tr": 2.51441, "fi": 2.7415, "ru": 1.66988, "el": 1.76442, "hi": 1.38615, "ta": 1.11313, "th": 1.28687, "zh": 3.41616, "my": 1.25339}, "gini_cpc": 0.326531, "gini_cps": 0.333315, "bpic": 4.399897, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8766960193549469, "vi": 1.2595247325883947, "tr": 1.1201987785213894, "fi": 0.8830756591473069, "ru": 1.9090900795184522, "el": 0.8851948825366327, "hi": 2.3821312025446257, "ta": 3.5175456749634995, "th": 2.1289781102409577, "zh": 2.797071715648097, "my": 5.008794072720051}, "t": 142.6} |
| {"event": "eval", "step": 2200, "phase": 1, "bpb": {"en": 2.26717, "de": 2.52447, "vi": 2.06572, "tr": 2.52387, "fi": 2.74327, "ru": 1.65092, "el": 1.75349, "hi": 1.38381, "ta": 1.11959, "th": 1.30697, "zh": 3.422, "my": 1.26102}, "bytes_per_chunk": {"en": 4.585, "de": 5.381, "vi": 5.194, "tr": 4.518, "fi": 5.168, "ru": 4.405, "el": 9.825, "hi": 4.895, "ta": 3.588, "th": 5.743, "zh": 4.571, "my": 2.589}, "bpb_mean": 2.00186, "t": 144.0} |
| {"event": "train", "step": 2200, "phase": 1, "objective": "global", "lm_loss": 1.51576, "bpb": 2.18678, "chunk_loss": 0.031941, "F": 0.20013, "G": 0.43375, "chunks": 204.94, "bytes_per_chunk": 4.997, "boundary_entropy": 0.66868, "router_noise": 0.0, "grad_norm": 0.793, "lr": 0.0002325, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 144.0} |
| {"event": "train", "step": 2250, "phase": 1, "objective": "global", "lm_loss": 1.72734, "bpb": 2.49202, "chunk_loss": 0.032229, "F": 0.20456, "G": 0.43555, "chunks": 209.47, "bytes_per_chunk": 4.889, "boundary_entropy": 0.66815, "router_noise": 0.0, "grad_norm": 0.8439, "lr": 0.0002293, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 146.1} |
| {"event": "train", "step": 2300, "phase": 1, "objective": "global", "lm_loss": 1.50866, "bpb": 2.17654, "chunk_loss": 0.032509, "F": 0.20959, "G": 0.43625, "chunks": 214.62, "bytes_per_chunk": 4.771, "boundary_entropy": 0.66887, "router_noise": 0.0, "grad_norm": 0.7175, "lr": 0.000226, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 148.2} |
| {"event": "train", "step": 2350, "phase": 1, "objective": "global", "lm_loss": 1.52048, "bpb": 2.19359, "chunk_loss": 0.031942, "F": 0.1998, "G": 0.43473, "chunks": 204.59, "bytes_per_chunk": 5.005, "boundary_entropy": 0.66928, "router_noise": 0.0, "grad_norm": 0.7615, "lr": 0.0002226, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 150.3} |
| {"event": "probe", "step": 2400, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3549.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.20393, "cps": 27.726562, "mean_p": 0.44041, "boundary_entropy": 0.666427, "bpb": 2.226793}, "de": {"chunks": 3640.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.177276, "cps": 28.4375, "mean_p": 0.442262, "boundary_entropy": 0.67184, "bpb": 2.48585}, "vi": {"chunks": 5476.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.297044, "cps": 42.78125, "mean_p": 0.441098, "boundary_entropy": 0.660528, "bpb": 2.056292}, "tr": {"chunks": 4089.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.225874, "cps": 31.945312, "mean_p": 0.447596, "boundary_entropy": 0.667425, "bpb": 2.476007}, "fi": {"chunks": 2853.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.151667, "cps": 22.289062, "mean_p": 0.436107, "boundary_entropy": 0.669466, "bpb": 2.714949}, "ru": {"chunks": 7155.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.377573, "cps": 55.898438, "mean_p": 0.427177, "boundary_entropy": 0.662503, "bpb": 1.649065}, "el": {"chunks": 6228.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.303464, "cps": 48.65625, "mean_p": 0.437646, "boundary_entropy": 0.673304, "bpb": 1.730321}, "hi": {"chunks": 7377.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.438455, "cps": 57.632812, "mean_p": 0.440095, "boundary_entropy": 0.669796, "bpb": 1.358621}, "ta": {"chunks": 12103.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.647392, "cps": 94.554688, "mean_p": 0.465469, "boundary_entropy": 0.667348, "bpb": 1.092926}, "th": {"chunks": 6159.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.382903, "cps": 48.117188, "mean_p": 0.438402, "boundary_entropy": 0.674981, "bpb": 1.264583}, "zh": {"chunks": 3565.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.603623, "cps": 27.851562, "mean_p": 0.410034, "boundary_entropy": 0.646027, "bpb": 3.408708}, "my": {"chunks": 16594.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.875765, "cps": 129.640625, "mean_p": 0.453257, "boundary_entropy": 0.677629, "bpb": 1.235886}}, "bpb": {"en": 2.22679, "de": 2.48585, "vi": 2.05629, "tr": 2.47601, "fi": 2.71495, "ru": 1.64906, "el": 1.73032, "hi": 1.35862, "ta": 1.09293, "th": 1.26458, "zh": 3.40871, "my": 1.23589}, "gini_cpc": 0.29151, "gini_cps": 0.29562, "bpic": 4.871059, "worst_lang": "my", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8692948312097973, "vi": 1.4565936702508022, "tr": 1.1076044195119232, "fi": 0.7437175142527576, "ru": 1.8514779517215076, "el": 1.4880786288924999, "hi": 2.1500216414057993, "ta": 3.1745756784002306, "th": 1.877618085030629, "zh": 2.9599488921820836, "my": 4.294432991052078}, "t": 153.3} |
| {"event": "eval", "step": 2400, "phase": 1, "bpb": {"en": 2.23397, "de": 2.492, "vi": 2.05725, "tr": 2.48284, "fi": 2.71392, "ru": 1.62678, "el": 1.72135, "hi": 1.35715, "ta": 1.09917, "th": 1.28235, "zh": 3.41269, "my": 1.24623}, "bytes_per_chunk": {"en": 5.035, "de": 5.709, "vi": 4.72, "tr": 4.911, "fi": 6.538, "ru": 4.844, "el": 6.177, "hi": 6.119, "ta": 4.149, "th": 7.111, "zh": 4.732, "my": 3.259}, "bpb_mean": 1.97714, "t": 154.6} |
| {"event": "phase_transition", "step": 2400, "phase": 2, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 154.6} |
| {"event": "train", "step": 2400, "phase": 2, "objective": "global", "lm_loss": 1.5385, "bpb": 2.21958, "chunk_loss": 0.031633, "F": 0.19412, "G": 0.43571, "chunks": 198.78, "bytes_per_chunk": 5.151, "boundary_entropy": 0.66818, "router_noise": 0.0, "grad_norm": 0.8299, "lr": 0.0002192, "router_trainable": true, "sec_per_step": 0.087, "alpha": 0.03, "t": 154.6} |
| {"event": "train", "step": 2450, "phase": 2, "objective": "global", "lm_loss": 1.33269, "bpb": 1.92267, "chunk_loss": 0.032444, "F": 0.20776, "G": 0.43739, "chunks": 212.75, "bytes_per_chunk": 4.813, "boundary_entropy": 0.66891, "router_noise": 0.0, "grad_norm": 0.6818, "lr": 0.0002158, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 156.7} |
| {"event": "train", "step": 2500, "phase": 2, "objective": "global", "lm_loss": 1.43355, "bpb": 2.06817, "chunk_loss": 0.032249, "F": 0.20551, "G": 0.43471, "chunks": 210.44, "bytes_per_chunk": 4.866, "boundary_entropy": 0.66845, "router_noise": 0.0, "grad_norm": 0.7733, "lr": 0.0002123, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 158.8} |
| {"event": "train", "step": 2550, "phase": 2, "objective": "global", "lm_loss": 1.5952, "bpb": 2.30139, "chunk_loss": 0.032279, "F": 0.20596, "G": 0.43411, "chunks": 210.91, "bytes_per_chunk": 4.855, "boundary_entropy": 0.6674, "router_noise": 0.0, "grad_norm": 0.7692, "lr": 0.0002088, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 160.9} |
| {"event": "probe", "step": 2600, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3857.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.221628, "cps": 30.132812, "mean_p": 0.44037, "boundary_entropy": 0.665685, "bpb": 2.202613}, "de": {"chunks": 3906.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.19023, "cps": 30.515625, "mean_p": 0.441102, "boundary_entropy": 0.670546, "bpb": 2.455797}, "vi": {"chunks": 5205.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.282343, "cps": 40.664062, "mean_p": 0.435542, "boundary_entropy": 0.658935, "bpb": 2.036525}, "tr": {"chunks": 4589.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.253494, "cps": 35.851562, "mean_p": 0.450867, "boundary_entropy": 0.668535, "bpb": 2.453737}, "fi": {"chunks": 3422.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.181915, "cps": 26.734375, "mean_p": 0.440012, "boundary_entropy": 0.66964, "bpb": 2.671344}, "ru": {"chunks": 8285.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.437203, "cps": 64.726562, "mean_p": 0.433715, "boundary_entropy": 0.665407, "bpb": 1.596425}, "el": {"chunks": 3864.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.188277, "cps": 30.1875, "mean_p": 0.431179, "boundary_entropy": 0.672503, "bpb": 1.715475}, "hi": {"chunks": 6955.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.413373, "cps": 54.335938, "mean_p": 0.443064, "boundary_entropy": 0.672843, "bpb": 1.354307}, "ta": {"chunks": 13503.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.722279, "cps": 105.492188, "mean_p": 0.467657, "boundary_entropy": 0.665422, "bpb": 1.079371}, "th": {"chunks": 4940.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.307118, "cps": 38.59375, "mean_p": 0.436181, "boundary_entropy": 0.674387, "bpb": 1.238968}, "zh": {"chunks": 2916.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.493735, "cps": 22.78125, "mean_p": 0.409907, "boundary_entropy": 0.648494, "bpb": 3.371191}, "my": {"chunks": 8143.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.429755, "cps": 63.617188, "mean_p": 0.446047, "boundary_entropy": 0.676899, "bpb": 1.218873}}, "bpb": {"en": 2.20261, "de": 2.4558, "vi": 2.03652, "tr": 2.45374, "fi": 2.67134, "ru": 1.59642, "el": 1.71547, "hi": 1.35431, "ta": 1.07937, "th": 1.23897, "zh": 3.37119, "my": 1.21887}, "gini_cpc": 0.241971, "gini_cps": 0.251406, "bpic": 5.515283, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8583300415965339, "vi": 1.2739490909029538, "tr": 1.143778655113422, "fi": 0.820809932438347, "ru": 1.972685170299965, "el": 0.8495144175809831, "hi": 1.8651619313180217, "ta": 3.2589618719447584, "th": 1.3857355704225245, "zh": 2.2277608029204576, "my": 1.9390791653414508}, "t": 163.9} |
| {"event": "eval", "step": 2600, "phase": 2, "bpb": {"en": 2.21287, "de": 2.46162, "vi": 2.03774, "tr": 2.459, "fi": 2.67277, "ru": 1.57422, "el": 1.70771, "hi": 1.35111, "ta": 1.0873, "th": 1.26332, "zh": 3.37623, "my": 1.22751}, "bytes_per_chunk": {"en": 4.478, "de": 5.459, "vi": 4.993, "tr": 4.282, "fi": 5.437, "ru": 4.153, "el": 9.783, "hi": 6.343, "ta": 3.753, "th": 8.764, "zh": 5.63, "my": 6.868}, "bpb_mean": 1.95262, "t": 165.1} |
| {"event": "train", "step": 2600, "phase": 2, "objective": "global", "lm_loss": 1.69357, "bpb": 2.44331, "chunk_loss": 0.031877, "F": 0.19861, "G": 0.42916, "chunks": 203.38, "bytes_per_chunk": 5.035, "boundary_entropy": 0.66544, "router_noise": 0.0, "grad_norm": 0.7723, "lr": 0.0002053, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 165.2} |
| {"event": "train", "step": 2650, "phase": 2, "objective": "global", "lm_loss": 1.44529, "bpb": 2.08511, "chunk_loss": 0.032853, "F": 0.21481, "G": 0.4357, "chunks": 219.97, "bytes_per_chunk": 4.655, "boundary_entropy": 0.66874, "router_noise": 0.0, "grad_norm": 0.7316, "lr": 0.0002017, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 167.2} |
| {"event": "train", "step": 2700, "phase": 2, "objective": "global", "lm_loss": 1.38808, "bpb": 2.00257, "chunk_loss": 0.032206, "F": 0.20486, "G": 0.43416, "chunks": 209.78, "bytes_per_chunk": 4.881, "boundary_entropy": 0.66778, "router_noise": 0.0, "grad_norm": 0.7298, "lr": 0.0001981, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 169.3} |
| {"event": "train", "step": 2750, "phase": 2, "objective": "global", "lm_loss": 1.36666, "bpb": 1.97167, "chunk_loss": 0.031482, "F": 0.19193, "G": 0.4347, "chunks": 196.53, "bytes_per_chunk": 5.21, "boundary_entropy": 0.66994, "router_noise": 0.0, "grad_norm": 0.8102, "lr": 0.0001945, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 171.4} |
| {"event": "probe", "step": 2800, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3727.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.214158, "cps": 29.117188, "mean_p": 0.441, "boundary_entropy": 0.665845, "bpb": 2.181097}, "de": {"chunks": 3819.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.185993, "cps": 29.835938, "mean_p": 0.440734, "boundary_entropy": 0.669965, "bpb": 2.426989}, "vi": {"chunks": 4952.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.268619, "cps": 38.6875, "mean_p": 0.432204, "boundary_entropy": 0.657418, "bpb": 2.020123}, "tr": {"chunks": 4397.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.242888, "cps": 34.351562, "mean_p": 0.447847, "boundary_entropy": 0.66705, "bpb": 2.413578}, "fi": {"chunks": 3137.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.166764, "cps": 24.507812, "mean_p": 0.437579, "boundary_entropy": 0.669135, "bpb": 2.637979}, "ru": {"chunks": 8022.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.423325, "cps": 62.671875, "mean_p": 0.431625, "boundary_entropy": 0.664177, "bpb": 1.57166}, "el": {"chunks": 4756.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.23174, "cps": 37.15625, "mean_p": 0.433375, "boundary_entropy": 0.672694, "bpb": 1.672207}, "hi": {"chunks": 7384.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.438871, "cps": 57.6875, "mean_p": 0.440014, "boundary_entropy": 0.670256, "bpb": 1.344587}, "ta": {"chunks": 11829.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.632736, "cps": 92.414062, "mean_p": 0.460108, "boundary_entropy": 0.663684, "bpb": 1.055905}, "th": {"chunks": 6637.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.41262, "cps": 51.851562, "mean_p": 0.443245, "boundary_entropy": 0.676438, "bpb": 1.227346}, "zh": {"chunks": 3645.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.617169, "cps": 28.476562, "mean_p": 0.416552, "boundary_entropy": 0.650115, "bpb": 3.349067}, "my": {"chunks": 9956.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.525438, "cps": 77.78125, "mean_p": 0.452014, "boundary_entropy": 0.678136, "bpb": 1.196265}}, "bpb": {"en": 2.1811, "de": 2.42699, "vi": 2.02012, "tr": 2.41358, "fi": 2.63798, "ru": 1.57166, "el": 1.67221, "hi": 1.34459, "ta": 1.0559, "th": 1.22735, "zh": 3.34907, "my": 1.19627}, "gini_cpc": 0.24666, "gini_cps": 0.238968, "bpic": 5.311039, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8684843134908669, "vi": 1.2543023082936888, "tr": 1.1341503774106256, "fi": 0.7786949580520061, "ru": 1.9766882048065426, "el": 1.0820958419749105, "hi": 2.0492801934081286, "ta": 2.954522418789239, "th": 1.9267061331257784, "zh": 2.8818330483177506, "my": 2.4535009999264226}, "t": 174.4} |
| {"event": "eval", "step": 2800, "phase": 2, "bpb": {"en": 2.19183, "de": 2.43347, "vi": 2.02204, "tr": 2.41925, "fi": 2.64189, "ru": 1.54997, "el": 1.66279, "hi": 1.34285, "ta": 1.0638, "th": 1.2454, "zh": 3.3523, "my": 1.20559}, "bytes_per_chunk": {"en": 4.638, "de": 5.507, "vi": 5.292, "tr": 4.492, "fi": 5.947, "ru": 4.298, "el": 8.185, "hi": 6.02, "ta": 4.203, "th": 6.901, "zh": 4.503, "my": 5.226}, "bpb_mean": 1.9276, "t": 175.6} |
| {"event": "train", "step": 2800, "phase": 2, "objective": "global", "lm_loss": 1.47953, "bpb": 2.13451, "chunk_loss": 0.031905, "F": 0.19977, "G": 0.43286, "chunks": 204.56, "bytes_per_chunk": 5.006, "boundary_entropy": 0.6676, "router_noise": 0.0, "grad_norm": 0.7641, "lr": 0.0001909, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 175.7} |
| {"event": "train", "step": 2850, "phase": 2, "objective": "global", "lm_loss": 1.42427, "bpb": 2.05478, "chunk_loss": 0.033365, "F": 0.22412, "G": 0.43691, "chunks": 229.5, "bytes_per_chunk": 4.462, "boundary_entropy": 0.6686, "router_noise": 0.0, "grad_norm": 0.8111, "lr": 0.0001872, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 177.7} |
| {"event": "train", "step": 2900, "phase": 2, "objective": "global", "lm_loss": 1.49494, "bpb": 2.15675, "chunk_loss": 0.031839, "F": 0.19876, "G": 0.43198, "chunks": 203.53, "bytes_per_chunk": 5.031, "boundary_entropy": 0.66652, "router_noise": 0.0, "grad_norm": 0.8, "lr": 0.0001835, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 179.8} |
| {"event": "train", "step": 2950, "phase": 2, "objective": "global", "lm_loss": 1.39039, "bpb": 2.00591, "chunk_loss": 0.033161, "F": 0.22015, "G": 0.43637, "chunks": 225.44, "bytes_per_chunk": 4.542, "boundary_entropy": 0.66849, "router_noise": 0.0, "grad_norm": 0.7614, "lr": 0.0001799, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 181.8} |
| {"event": "probe", "step": 3000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3449.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.198184, "cps": 26.945312, "mean_p": 0.437051, "boundary_entropy": 0.664002, "bpb": 2.16845}, "de": {"chunks": 3484.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.169678, "cps": 27.21875, "mean_p": 0.43929, "boundary_entropy": 0.669709, "bpb": 2.41248}, "vi": {"chunks": 4864.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.263846, "cps": 38.0, "mean_p": 0.432077, "boundary_entropy": 0.657892, "bpb": 2.007273}, "tr": {"chunks": 4048.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.223609, "cps": 31.625, "mean_p": 0.444388, "boundary_entropy": 0.666116, "bpb": 2.388059}, "fi": {"chunks": 2936.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.156079, "cps": 22.9375, "mean_p": 0.435094, "boundary_entropy": 0.668168, "bpb": 2.597072}, "ru": {"chunks": 8329.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.439525, "cps": 65.070312, "mean_p": 0.435223, "boundary_entropy": 0.666449, "bpb": 1.555076}, "el": {"chunks": 4259.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.207523, "cps": 33.273438, "mean_p": 0.433478, "boundary_entropy": 0.67349, "bpb": 1.652309}, "hi": {"chunks": 7285.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.432987, "cps": 56.914062, "mean_p": 0.443343, "boundary_entropy": 0.673406, "bpb": 1.307941}, "ta": {"chunks": 13364.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.714844, "cps": 104.40625, "mean_p": 0.458793, "boundary_entropy": 0.662059, "bpb": 1.044258}, "th": {"chunks": 7107.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.44184, "cps": 55.523438, "mean_p": 0.444016, "boundary_entropy": 0.676931, "bpb": 1.210178}, "zh": {"chunks": 3334.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.564511, "cps": 26.046875, "mean_p": 0.412231, "boundary_entropy": 0.648091, "bpb": 3.319724}, "my": {"chunks": 5713.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.301509, "cps": 44.632812, "mean_p": 0.446479, "boundary_entropy": 0.677334, "bpb": 1.18159}}, "bpb": {"en": 2.16845, "de": 2.41248, "vi": 2.00727, "tr": 2.38806, "fi": 2.59707, "ru": 1.55508, "el": 1.65231, "hi": 1.30794, "ta": 1.04426, "th": 1.21018, "zh": 3.31972, "my": 1.18159}, "gini_cpc": 0.268005, "gini_cps": 0.256985, "bpic": 5.629599, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.85616341320283, "vi": 1.331316609028784, "tr": 1.1282903616810651, "fi": 0.7875444832891509, "ru": 2.217760140027586, "el": 1.0471230525583195, "hi": 2.1847684170573807, "ta": 3.6069649593965045, "th": 2.229441987527661, "zh": 2.848413783664184, "my": 1.5213592305210586}, "t": 184.9} |
| {"event": "eval", "step": 3000, "phase": 2, "bpb": {"en": 2.17864, "de": 2.41926, "vi": 2.00895, "tr": 2.3952, "fi": 2.60312, "ru": 1.53128, "el": 1.6441, "hi": 1.30526, "ta": 1.05241, "th": 1.2276, "zh": 3.32419, "my": 1.19083}, "bytes_per_chunk": {"en": 5.08, "de": 6.043, "vi": 5.495, "tr": 4.883, "fi": 6.43, "ru": 4.152, "el": 8.748, "hi": 6.162, "ta": 3.714, "th": 6.517, "zh": 4.995, "my": 8.758}, "bpb_mean": 1.90674, "t": 186.1} |
| {"event": "train", "step": 3000, "phase": 2, "objective": "global", "lm_loss": 1.49788, "bpb": 2.16099, "chunk_loss": 0.030787, "F": 0.1806, "G": 0.43215, "chunks": 184.94, "bytes_per_chunk": 5.537, "boundary_entropy": 0.66838, "router_noise": 0.0, "grad_norm": 0.8632, "lr": 0.0001761, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 186.2} |
| {"event": "train", "step": 3050, "phase": 2, "objective": "global", "lm_loss": 1.4055, "bpb": 2.02771, "chunk_loss": 0.031891, "F": 0.19925, "G": 0.43523, "chunks": 204.03, "bytes_per_chunk": 5.019, "boundary_entropy": 0.66896, "router_noise": 0.0, "grad_norm": 0.8361, "lr": 0.0001724, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 188.2} |
| {"event": "train", "step": 3100, "phase": 2, "objective": "global", "lm_loss": 1.43927, "bpb": 2.07643, "chunk_loss": 0.032105, "F": 0.20248, "G": 0.4351, "chunks": 207.34, "bytes_per_chunk": 4.939, "boundary_entropy": 0.66885, "router_noise": 0.0, "grad_norm": 0.7369, "lr": 0.0001687, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 190.3} |
| {"event": "train", "step": 3150, "phase": 2, "objective": "global", "lm_loss": 1.38338, "bpb": 1.9958, "chunk_loss": 0.032795, "F": 0.21332, "G": 0.43734, "chunks": 218.44, "bytes_per_chunk": 4.688, "boundary_entropy": 0.67065, "router_noise": 0.0, "grad_norm": 0.7753, "lr": 0.000165, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 192.3} |
| {"event": "probe", "step": 3200, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3635.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.208872, "cps": 28.398438, "mean_p": 0.438808, "boundary_entropy": 0.665011, "bpb": 2.157837}, "de": {"chunks": 3671.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.178785, "cps": 28.679688, "mean_p": 0.441146, "boundary_entropy": 0.670383, "bpb": 2.384723}, "vi": {"chunks": 4915.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.266612, "cps": 38.398438, "mean_p": 0.431563, "boundary_entropy": 0.657387, "bpb": 1.997574}, "tr": {"chunks": 3980.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.219853, "cps": 31.09375, "mean_p": 0.444455, "boundary_entropy": 0.666654, "bpb": 2.382422}, "fi": {"chunks": 3296.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.175217, "cps": 25.75, "mean_p": 0.438849, "boundary_entropy": 0.669574, "bpb": 2.576745}, "ru": {"chunks": 7873.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.415462, "cps": 61.507812, "mean_p": 0.430754, "boundary_entropy": 0.663985, "bpb": 1.527901}, "el": {"chunks": 4884.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.237977, "cps": 38.15625, "mean_p": 0.435054, "boundary_entropy": 0.673643, "bpb": 1.624099}, "hi": {"chunks": 7411.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.440475, "cps": 57.898438, "mean_p": 0.442934, "boundary_entropy": 0.672507, "bpb": 1.296673}, "ta": {"chunks": 13265.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.709548, "cps": 103.632812, "mean_p": 0.461022, "boundary_entropy": 0.665334, "bpb": 1.053168}, "th": {"chunks": 6943.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.431644, "cps": 54.242188, "mean_p": 0.444493, "boundary_entropy": 0.676938, "bpb": 1.193742}, "zh": {"chunks": 3022.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.511683, "cps": 23.609375, "mean_p": 0.410651, "boundary_entropy": 0.648002, "bpb": 3.310535}, "my": {"chunks": 12775.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.674214, "cps": 99.804688, "mean_p": 0.45182, "boundary_entropy": 0.677642, "bpb": 1.168463}}, "bpb": {"en": 2.15784, "de": 2.38472, "vi": 1.99757, "tr": 2.38242, "fi": 2.57675, "ru": 1.5279, "el": 1.6241, "hi": 1.29667, "ta": 1.05317, "th": 1.19374, "zh": 3.31054, "my": 1.16846}, "gini_cpc": 0.266369, "gini_cps": 0.281122, "bpic": 5.071772, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8559564765438478, "vi": 1.2764390592770092, "tr": 1.052573000242157, "fi": 0.8388706979300429, "ru": 1.9890730910635024, "el": 1.1393430700523806, "hi": 2.1088293890298013, "ta": 3.3970464853713027, "th": 2.066549464908526, "zh": 2.449744111203909, "my": 3.2278789354430963}, "t": 195.3} |
| {"event": "eval", "step": 3200, "phase": 2, "bpb": {"en": 2.1657, "de": 2.39202, "vi": 2.00043, "tr": 2.38844, "fi": 2.58118, "ru": 1.5051, "el": 1.61754, "hi": 1.29548, "ta": 1.06077, "th": 1.21145, "zh": 3.31029, "my": 1.17933}, "bytes_per_chunk": {"en": 4.889, "de": 5.902, "vi": 5.324, "tr": 5.102, "fi": 5.758, "ru": 4.399, "el": 7.878, "hi": 5.986, "ta": 3.761, "th": 6.578, "zh": 5.495, "my": 4.214}, "bpb_mean": 1.89231, "t": 196.6} |
| {"event": "train", "step": 3200, "phase": 2, "objective": "global", "lm_loss": 1.34159, "bpb": 1.93551, "chunk_loss": 0.031621, "F": 0.19427, "G": 0.43384, "chunks": 198.94, "bytes_per_chunk": 5.147, "boundary_entropy": 0.66729, "router_noise": 0.0, "grad_norm": 0.8683, "lr": 0.0001613, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 196.6} |
| {"event": "train", "step": 3250, "phase": 2, "objective": "global", "lm_loss": 1.56482, "bpb": 2.25756, "chunk_loss": 0.032319, "F": 0.20685, "G": 0.43311, "chunks": 211.81, "bytes_per_chunk": 4.834, "boundary_entropy": 0.66678, "router_noise": 0.0, "grad_norm": 0.7698, "lr": 0.0001576, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 198.7} |
| {"event": "train", "step": 3300, "phase": 2, "objective": "global", "lm_loss": 1.52729, "bpb": 2.20341, "chunk_loss": 0.031982, "F": 0.20059, "G": 0.43547, "chunks": 205.41, "bytes_per_chunk": 4.985, "boundary_entropy": 0.66887, "router_noise": 0.0, "grad_norm": 0.8328, "lr": 0.0001539, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 200.8} |
| {"event": "train", "step": 3350, "phase": 2, "objective": "global", "lm_loss": 1.36377, "bpb": 1.9675, "chunk_loss": 0.03159, "F": 0.194, "G": 0.43115, "chunks": 198.66, "bytes_per_chunk": 5.155, "boundary_entropy": 0.66708, "router_noise": 0.0, "grad_norm": 0.7212, "lr": 0.0001501, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 202.8} |
| {"event": "probe", "step": 3400, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3442.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.197782, "cps": 26.890625, "mean_p": 0.436084, "boundary_entropy": 0.663985, "bpb": 2.134598}, "de": {"chunks": 3374.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.164321, "cps": 26.359375, "mean_p": 0.437551, "boundary_entropy": 0.669576, "bpb": 2.36674}, "vi": {"chunks": 5134.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.278492, "cps": 40.109375, "mean_p": 0.433461, "boundary_entropy": 0.65664, "bpb": 1.975227}, "tr": {"chunks": 4077.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.225211, "cps": 31.851562, "mean_p": 0.445531, "boundary_entropy": 0.667024, "bpb": 2.354286}, "fi": {"chunks": 3324.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.176705, "cps": 25.96875, "mean_p": 0.440749, "boundary_entropy": 0.670258, "bpb": 2.555401}, "ru": {"chunks": 7959.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.42, "cps": 62.179688, "mean_p": 0.430541, "boundary_entropy": 0.664238, "bpb": 1.513718}, "el": {"chunks": 4983.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.242801, "cps": 38.929688, "mean_p": 0.435396, "boundary_entropy": 0.673633, "bpb": 1.613287}, "hi": {"chunks": 7481.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.444636, "cps": 58.445312, "mean_p": 0.444602, "boundary_entropy": 0.672977, "bpb": 1.284814}, "ta": {"chunks": 14061.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.752126, "cps": 109.851562, "mean_p": 0.453193, "boundary_entropy": 0.659353, "bpb": 1.030467}, "th": {"chunks": 8406.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.522599, "cps": 65.671875, "mean_p": 0.448607, "boundary_entropy": 0.677839, "bpb": 1.186005}, "zh": {"chunks": 2626.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.444633, "cps": 20.515625, "mean_p": 0.410123, "boundary_entropy": 0.649823, "bpb": 3.291085}, "my": {"chunks": 8388.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.442685, "cps": 65.53125, "mean_p": 0.447831, "boundary_entropy": 0.677357, "bpb": 1.159766}}, "bpb": {"en": 2.1346, "de": 2.36674, "vi": 1.97523, "tr": 2.35429, "fi": 2.5554, "ru": 1.51372, "el": 1.61329, "hi": 1.28481, "ta": 1.03047, "th": 1.18601, "zh": 3.29109, "my": 1.15977}, "gini_cpc": 0.254043, "gini_cps": 0.272529, "bpic": 5.238973, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8308180538899796, "vi": 1.4080756121788522, "tr": 1.1386845137529413, "fi": 0.8934338152127106, "ru": 2.123550261475886, "el": 1.2276181372506967, "hi": 2.2481114421709623, "ta": 3.80280444562621, "th": 2.642296652895525, "zh": 2.248094345796719, "my": 2.2382484887536482}, "t": 205.8} |
| {"event": "eval", "step": 3400, "phase": 2, "bpb": {"en": 2.14408, "de": 2.37211, "vi": 1.97649, "tr": 2.36341, "fi": 2.55893, "ru": 1.4898, "el": 1.60514, "hi": 1.28338, "ta": 1.0381, "th": 1.20368, "zh": 3.29168, "my": 1.16974}, "bytes_per_chunk": {"en": 5.155, "de": 6.486, "vi": 5.14, "tr": 4.915, "fi": 5.628, "ru": 4.383, "el": 7.988, "hi": 6.064, "ta": 3.572, "th": 5.534, "zh": 6.349, "my": 6.607}, "bpb_mean": 1.87471, "t": 207.1} |
| {"event": "train", "step": 3400, "phase": 2, "objective": "global", "lm_loss": 1.39032, "bpb": 2.00581, "chunk_loss": 0.030423, "F": 0.17404, "G": 0.43108, "chunks": 178.22, "bytes_per_chunk": 5.746, "boundary_entropy": 0.66857, "router_noise": 0.0, "grad_norm": 0.7722, "lr": 0.0001465, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 207.1} |
| {"event": "train", "step": 3450, "phase": 2, "objective": "global", "lm_loss": 1.38601, "bpb": 1.99959, "chunk_loss": 0.031656, "F": 0.1954, "G": 0.43437, "chunks": 200.09, "bytes_per_chunk": 5.118, "boundary_entropy": 0.66854, "router_noise": 0.0, "grad_norm": 0.7863, "lr": 0.0001428, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 209.1} |
| {"event": "train", "step": 3500, "phase": 2, "objective": "global", "lm_loss": 1.45969, "bpb": 2.10589, "chunk_loss": 0.032095, "F": 0.20291, "G": 0.43113, "chunks": 207.78, "bytes_per_chunk": 4.928, "boundary_entropy": 0.66576, "router_noise": 0.0, "grad_norm": 0.7399, "lr": 0.0001391, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 211.2} |
| {"event": "train", "step": 3550, "phase": 2, "objective": "global", "lm_loss": 1.40747, "bpb": 2.03055, "chunk_loss": 0.032818, "F": 0.21454, "G": 0.43635, "chunks": 219.69, "bytes_per_chunk": 4.661, "boundary_entropy": 0.6682, "router_noise": 0.0, "grad_norm": 0.7716, "lr": 0.0001355, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 213.2} |
| {"event": "probe", "step": 3600, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4015.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.230707, "cps": 31.367188, "mean_p": 0.440605, "boundary_entropy": 0.664853, "bpb": 2.121279}, "de": {"chunks": 4020.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.195782, "cps": 31.40625, "mean_p": 0.444046, "boundary_entropy": 0.671164, "bpb": 2.343934}, "vi": {"chunks": 5222.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.283266, "cps": 40.796875, "mean_p": 0.430458, "boundary_entropy": 0.654313, "bpb": 1.96737}, "tr": {"chunks": 4434.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.244932, "cps": 34.640625, "mean_p": 0.447332, "boundary_entropy": 0.667441, "bpb": 2.336478}, "fi": {"chunks": 3451.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.183456, "cps": 26.960938, "mean_p": 0.441581, "boundary_entropy": 0.670312, "bpb": 2.531818}, "ru": {"chunks": 8147.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.429921, "cps": 63.648438, "mean_p": 0.431973, "boundary_entropy": 0.664577, "bpb": 1.500477}, "el": {"chunks": 5087.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.247868, "cps": 39.742188, "mean_p": 0.43799, "boundary_entropy": 0.674784, "bpb": 1.592405}, "hi": {"chunks": 8129.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.48315, "cps": 63.507812, "mean_p": 0.447925, "boundary_entropy": 0.674499, "bpb": 1.270509}, "ta": {"chunks": 13900.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.743514, "cps": 108.59375, "mean_p": 0.448111, "boundary_entropy": 0.658023, "bpb": 1.00254}, "th": {"chunks": 9499.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.59055, "cps": 74.210938, "mean_p": 0.447799, "boundary_entropy": 0.677018, "bpb": 1.16544}, "zh": {"chunks": 3759.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.636471, "cps": 29.367188, "mean_p": 0.414811, "boundary_entropy": 0.648513, "bpb": 3.254323}, "my": {"chunks": 8004.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.422419, "cps": 62.53125, "mean_p": 0.448504, "boundary_entropy": 0.678692, "bpb": 1.159437}}, "bpb": {"en": 2.12128, "de": 2.34393, "vi": 1.96737, "tr": 2.33648, "fi": 2.53182, "ru": 1.50048, "el": 1.59241, "hi": 1.27051, "ta": 1.00254, "th": 1.16544, "zh": 3.25432, "my": 1.15944}, "gini_cpc": 0.257922, "gini_cps": 0.243468, "bpic": 4.941365, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8486179553989541, "vi": 1.2278131944184085, "tr": 1.061655729642796, "fi": 0.7951913485876054, "ru": 1.863490025859491, "el": 1.0743838292414263, "hi": 2.0942118911437233, "ta": 3.2227595300719383, "th": 2.5597372767880473, "zh": 2.7587824442627316, "my": 1.8309744096118339}, "t": 216.3} |
| {"event": "eval", "step": 3600, "phase": 2, "bpb": {"en": 2.13126, "de": 2.34969, "vi": 1.97045, "tr": 2.34489, "fi": 2.53683, "ru": 1.47741, "el": 1.58454, "hi": 1.26882, "ta": 1.01163, "th": 1.18515, "zh": 3.25705, "my": 1.16996}, "bytes_per_chunk": {"en": 4.38, "de": 5.147, "vi": 5.013, "tr": 4.482, "fi": 5.446, "ru": 4.23, "el": 7.559, "hi": 5.466, "ta": 3.569, "th": 4.893, "zh": 4.393, "my": 6.558}, "bpb_mean": 1.85731, "t": 217.6} |
| {"event": "train", "step": 3600, "phase": 2, "objective": "global", "lm_loss": 1.49585, "bpb": 2.15806, "chunk_loss": 0.032442, "F": 0.2085, "G": 0.43628, "chunks": 213.5, "bytes_per_chunk": 4.796, "boundary_entropy": 0.66963, "router_noise": 0.0, "grad_norm": 2.8154, "lr": 0.0001319, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 217.6} |
| {"event": "train", "step": 3650, "phase": 2, "objective": "global", "lm_loss": 1.34745, "bpb": 1.94397, "chunk_loss": 0.032678, "F": 0.21326, "G": 0.43218, "chunks": 218.38, "bytes_per_chunk": 4.689, "boundary_entropy": 0.66618, "router_noise": 0.0, "grad_norm": 0.7638, "lr": 0.0001283, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 219.7} |
| {"event": "train", "step": 3700, "phase": 2, "objective": "global", "lm_loss": 1.60288, "bpb": 2.31246, "chunk_loss": 0.031616, "F": 0.19446, "G": 0.4295, "chunks": 199.12, "bytes_per_chunk": 5.142, "boundary_entropy": 0.66602, "router_noise": 0.0, "grad_norm": 0.8365, "lr": 0.0001247, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 221.7} |
| {"event": "train", "step": 3750, "phase": 2, "objective": "global", "lm_loss": 1.32451, "bpb": 1.91087, "chunk_loss": 0.031964, "F": 0.20053, "G": 0.43618, "chunks": 205.34, "bytes_per_chunk": 4.987, "boundary_entropy": 0.66921, "router_noise": 0.0, "grad_norm": 0.7713, "lr": 0.0001212, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 223.8} |
| {"event": "probe", "step": 3800, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3657.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.210136, "cps": 28.570312, "mean_p": 0.437579, "boundary_entropy": 0.663785, "bpb": 2.109921}, "de": {"chunks": 3863.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.188136, "cps": 30.179688, "mean_p": 0.440266, "boundary_entropy": 0.669708, "bpb": 2.325741}, "vi": {"chunks": 4987.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.270518, "cps": 38.960938, "mean_p": 0.427426, "boundary_entropy": 0.653984, "bpb": 1.958276}, "tr": {"chunks": 4172.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.230459, "cps": 32.59375, "mean_p": 0.4462, "boundary_entropy": 0.667232, "bpb": 2.32466}, "fi": {"chunks": 3246.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.172559, "cps": 25.359375, "mean_p": 0.441066, "boundary_entropy": 0.670589, "bpb": 2.5168}, "ru": {"chunks": 8413.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.443958, "cps": 65.726562, "mean_p": 0.432404, "boundary_entropy": 0.664563, "bpb": 1.481594}, "el": {"chunks": 4884.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.237977, "cps": 38.15625, "mean_p": 0.437376, "boundary_entropy": 0.674936, "bpb": 1.576298}, "hi": {"chunks": 8159.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.484933, "cps": 63.742188, "mean_p": 0.446415, "boundary_entropy": 0.674543, "bpb": 1.262302}, "ta": {"chunks": 13208.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.706499, "cps": 103.1875, "mean_p": 0.440221, "boundary_entropy": 0.655532, "bpb": 0.997919}, "th": {"chunks": 8990.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.558906, "cps": 70.234375, "mean_p": 0.449057, "boundary_entropy": 0.678078, "bpb": 1.15384}, "zh": {"chunks": 3412.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.577718, "cps": 26.65625, "mean_p": 0.411305, "boundary_entropy": 0.647753, "bpb": 3.249697}, "my": {"chunks": 3325.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.17548, "cps": 25.976562, "mean_p": 0.433767, "boundary_entropy": 0.676049, "bpb": 1.15064}}, "bpb": {"en": 2.10992, "de": 2.32574, "vi": 1.95828, "tr": 2.32466, "fi": 2.5168, "ru": 1.48159, "el": 1.5763, "hi": 1.2623, "ta": 0.99792, "th": 1.15384, "zh": 3.2497, "my": 1.15064}, "gini_cpc": 0.2767, "gini_cps": 0.263297, "bpic": 5.457947, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8953059296252013, "vi": 1.287346291083698, "tr": 1.0967127897301907, "fi": 0.8211751375639434, "ru": 2.112714604513843, "el": 1.1324889416572064, "hi": 2.307708873982645, "ta": 3.362100959642954, "th": 2.659731435975258, "zh": 2.7492532031079198, "my": 0.8350787518640456}, "t": 226.8} |
| {"event": "eval", "step": 3800, "phase": 2, "bpb": {"en": 2.12045, "de": 2.33212, "vi": 1.95972, "tr": 2.33156, "fi": 2.52186, "ru": 1.45693, "el": 1.56798, "hi": 1.26031, "ta": 1.00667, "th": 1.17355, "zh": 3.24985, "my": 1.16075}, "bytes_per_chunk": {"en": 4.915, "de": 5.57, "vi": 5.284, "tr": 4.867, "fi": 5.772, "ru": 4.095, "el": 7.739, "hi": 5.576, "ta": 3.736, "th": 5.138, "zh": 4.977, "my": 16.627}, "bpb_mean": 1.84515, "t": 228.1} |
| {"event": "train", "step": 3800, "phase": 2, "objective": "global", "lm_loss": 1.3867, "bpb": 2.00059, "chunk_loss": 0.031714, "F": 0.19647, "G": 0.4323, "chunks": 201.19, "bytes_per_chunk": 5.09, "boundary_entropy": 0.66763, "router_noise": 0.0, "grad_norm": 0.8132, "lr": 0.0001177, "router_trainable": true, "sec_per_step": 0.087, "alpha": 0.03, "t": 228.1} |
| {"event": "train", "step": 3850, "phase": 2, "objective": "global", "lm_loss": 1.49199, "bpb": 2.15249, "chunk_loss": 0.032073, "F": 0.20291, "G": 0.43, "chunks": 207.78, "bytes_per_chunk": 4.928, "boundary_entropy": 0.66489, "router_noise": 0.0, "grad_norm": 0.7268, "lr": 0.0001142, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 230.1} |
| {"event": "train", "step": 3900, "phase": 2, "objective": "global", "lm_loss": 1.51182, "bpb": 2.18109, "chunk_loss": 0.031641, "F": 0.19543, "G": 0.42953, "chunks": 200.12, "bytes_per_chunk": 5.117, "boundary_entropy": 0.66577, "router_noise": 0.0, "grad_norm": 0.7734, "lr": 0.0001108, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 232.2} |
| {"event": "train", "step": 3950, "phase": 2, "objective": "global", "lm_loss": 1.43707, "bpb": 2.07326, "chunk_loss": 0.031426, "F": 0.19147, "G": 0.43148, "chunks": 196.06, "bytes_per_chunk": 5.223, "boundary_entropy": 0.66739, "router_noise": 0.0, "grad_norm": 0.7857, "lr": 0.0001074, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 234.3} |
| {"event": "probe", "step": 4000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3638.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.209044, "cps": 28.421875, "mean_p": 0.436856, "boundary_entropy": 0.663374, "bpb": 2.099455}, "de": {"chunks": 3947.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.192227, "cps": 30.835938, "mean_p": 0.441709, "boundary_entropy": 0.669892, "bpb": 2.31026}, "vi": {"chunks": 5045.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.273664, "cps": 39.414062, "mean_p": 0.430122, "boundary_entropy": 0.65546, "bpb": 1.943174}, "tr": {"chunks": 4427.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.244545, "cps": 34.585938, "mean_p": 0.447629, "boundary_entropy": 0.667123, "bpb": 2.290699}, "fi": {"chunks": 3162.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.168093, "cps": 24.703125, "mean_p": 0.437593, "boundary_entropy": 0.668751, "bpb": 2.494585}, "ru": {"chunks": 8106.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.427757, "cps": 63.328125, "mean_p": 0.432687, "boundary_entropy": 0.665757, "bpb": 1.466158}, "el": {"chunks": 5175.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.252156, "cps": 40.429688, "mean_p": 0.439764, "boundary_entropy": 0.675651, "bpb": 1.555005}, "hi": {"chunks": 7313.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.434651, "cps": 57.132812, "mean_p": 0.445179, "boundary_entropy": 0.674827, "bpb": 1.258438}, "ta": {"chunks": 12872.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.688526, "cps": 100.5625, "mean_p": 0.442053, "boundary_entropy": 0.657168, "bpb": 0.997462}, "th": {"chunks": 8409.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.522785, "cps": 65.695312, "mean_p": 0.448849, "boundary_entropy": 0.677817, "bpb": 1.147841}, "zh": {"chunks": 3508.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.593972, "cps": 27.40625, "mean_p": 0.416629, "boundary_entropy": 0.651053, "bpb": 3.222938}, "my": {"chunks": 10269.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.541957, "cps": 80.226562, "mean_p": 0.451762, "boundary_entropy": 0.679002, "bpb": 1.143461}}, "bpb": {"en": 2.09946, "de": 2.31026, "vi": 1.94317, "tr": 2.2907, "fi": 2.49458, "ru": 1.46616, "el": 1.555, "hi": 1.25844, "ta": 0.99746, "th": 1.14784, "zh": 3.22294, "my": 1.14346}, "gini_cpc": 0.252116, "gini_cps": 0.253324, "bpic": 5.058336, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9195516853533379, "vi": 1.3091199887633966, "tr": 1.16982364702718, "fi": 0.8041024912347468, "ru": 2.0462505566426503, "el": 1.2062322466511177, "hi": 2.0792270952068597, "ta": 3.293684432002924, "th": 2.500833122141974, "zh": 2.841368539592334, "my": 2.5925444023590267}, "t": 237.3} |
| {"event": "eval", "step": 4000, "phase": 2, "bpb": {"en": 2.10936, "de": 2.31658, "vi": 1.94409, "tr": 2.29761, "fi": 2.50031, "ru": 1.44164, "el": 1.54792, "hi": 1.25606, "ta": 1.00477, "th": 1.16632, "zh": 3.22064, "my": 1.15192}, "bytes_per_chunk": {"en": 4.902, "de": 5.405, "vi": 5.212, "tr": 4.51, "fi": 6.131, "ru": 4.207, "el": 7.36, "hi": 6.21, "ta": 3.837, "th": 5.475, "zh": 4.75, "my": 5.166}, "bpb_mean": 1.82977, "t": 238.6} |
| {"event": "train", "step": 4000, "phase": 2, "objective": "global", "lm_loss": 1.4262, "bpb": 2.05757, "chunk_loss": 0.032176, "F": 0.20441, "G": 0.43169, "chunks": 209.31, "bytes_per_chunk": 4.892, "boundary_entropy": 0.6667, "router_noise": 0.0, "grad_norm": 0.8088, "lr": 0.000104, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 238.7} |
| {"event": "train", "step": 4050, "phase": 2, "objective": "global", "lm_loss": 1.31796, "bpb": 1.90141, "chunk_loss": 0.032244, "F": 0.20541, "G": 0.43498, "chunks": 210.34, "bytes_per_chunk": 4.868, "boundary_entropy": 0.66853, "router_noise": 0.0, "grad_norm": 0.8239, "lr": 0.0001007, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 240.7} |
| {"event": "train", "step": 4100, "phase": 2, "objective": "global", "lm_loss": 1.4708, "bpb": 2.12192, "chunk_loss": 0.031591, "F": 0.19431, "G": 0.43329, "chunks": 198.97, "bytes_per_chunk": 5.147, "boundary_entropy": 0.66797, "router_noise": 0.0, "grad_norm": 0.7918, "lr": 9.75e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 242.8} |
| {"event": "train", "step": 4150, "phase": 2, "objective": "global", "lm_loss": 1.26198, "bpb": 1.82065, "chunk_loss": 0.031933, "F": 0.20044, "G": 0.43333, "chunks": 205.25, "bytes_per_chunk": 4.989, "boundary_entropy": 0.66822, "router_noise": 0.0, "grad_norm": 0.7993, "lr": 9.43e-05, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 244.9} |
| {"event": "probe", "step": 4200, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3666.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.210653, "cps": 28.640625, "mean_p": 0.436005, "boundary_entropy": 0.662729, "bpb": 2.082252}, "de": {"chunks": 3859.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.187941, "cps": 30.148438, "mean_p": 0.440299, "boundary_entropy": 0.669448, "bpb": 2.306552}, "vi": {"chunks": 4991.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.270735, "cps": 38.992188, "mean_p": 0.429077, "boundary_entropy": 0.654476, "bpb": 1.938779}, "tr": {"chunks": 4522.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.249793, "cps": 35.328125, "mean_p": 0.448135, "boundary_entropy": 0.667216, "bpb": 2.27939}, "fi": {"chunks": 3315.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.176227, "cps": 25.898438, "mean_p": 0.439277, "boundary_entropy": 0.669526, "bpb": 2.472783}, "ru": {"chunks": 8159.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.430554, "cps": 63.742188, "mean_p": 0.432018, "boundary_entropy": 0.664942, "bpb": 1.448319}, "el": {"chunks": 4983.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.242801, "cps": 38.929688, "mean_p": 0.437774, "boundary_entropy": 0.674864, "bpb": 1.554661}, "hi": {"chunks": 7130.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.423774, "cps": 55.703125, "mean_p": 0.44281, "boundary_entropy": 0.673757, "bpb": 1.241455}, "ta": {"chunks": 12588.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.673335, "cps": 98.34375, "mean_p": 0.43512, "boundary_entropy": 0.654536, "bpb": 0.980517}, "th": {"chunks": 8837.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.549394, "cps": 69.039062, "mean_p": 0.448568, "boundary_entropy": 0.677706, "bpb": 1.139499}, "zh": {"chunks": 3040.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.514731, "cps": 23.75, "mean_p": 0.411547, "boundary_entropy": 0.649138, "bpb": 3.208286}, "my": {"chunks": 6546.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.345472, "cps": 51.140625, "mean_p": 0.449503, "boundary_entropy": 0.678987, "bpb": 1.130513}}, "bpb": {"en": 2.08225, "de": 2.30655, "vi": 1.93878, "tr": 2.27939, "fi": 2.47278, "ru": 1.44832, "el": 1.55466, "hi": 1.24145, "ta": 0.98052, "th": 1.1395, "zh": 3.20829, "my": 1.13051}, "gini_cpc": 0.241405, "gini_cps": 0.242676, "bpic": 5.357376, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8921831791592042, "vi": 1.2852158932365985, "tr": 1.1858006010885338, "fi": 0.8365719695470171, "ru": 2.04389875156001, "el": 1.152608191057528, "hi": 2.011713435942831, "ta": 3.1964132650537596, "th": 2.6080472143944786, "zh": 2.4434969458782243, "my": 1.6400016480629043}, "t": 248.0} |
| {"event": "eval", "step": 4200, "phase": 2, "bpb": {"en": 2.09384, "de": 2.31444, "vi": 1.94067, "tr": 2.28663, "fi": 2.47901, "ru": 1.4222, "el": 1.54636, "hi": 1.23984, "ta": 0.9893, "th": 1.15658, "zh": 3.20771, "my": 1.14043}, "bytes_per_chunk": {"en": 4.852, "de": 5.482, "vi": 5.293, "tr": 4.46, "fi": 5.803, "ru": 4.266, "el": 7.68, "hi": 6.29, "ta": 3.911, "th": 5.172, "zh": 5.355, "my": 8.079}, "bpb_mean": 1.81808, "t": 249.3} |
| {"event": "phase_transition", "step": 4200, "phase": 3, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 249.3} |
| {"event": "train", "step": 4200, "phase": 3, "objective": "global", "lm_loss": 1.27197, "bpb": 1.83507, "chunk_loss": 0.031722, "F": 0.19675, "G": 0.4339, "chunks": 201.47, "bytes_per_chunk": 5.083, "boundary_entropy": 0.66839, "router_noise": 0.0, "grad_norm": 0.7722, "lr": 9.12e-05, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 249.3} |
| {"event": "train", "step": 4250, "phase": 3, "objective": "global", "lm_loss": 1.30036, "bpb": 1.87603, "chunk_loss": 0.032476, "F": 0.20966, "G": 0.43333, "chunks": 214.69, "bytes_per_chunk": 4.77, "boundary_entropy": 0.66588, "router_noise": 0.0, "grad_norm": 0.7867, "lr": 8.81e-05, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 251.4} |
| {"event": "train", "step": 4300, "phase": 3, "objective": "global", "lm_loss": 1.37788, "bpb": 1.98787, "chunk_loss": 0.031902, "F": 0.19992, "G": 0.43303, "chunks": 204.72, "bytes_per_chunk": 5.002, "boundary_entropy": 0.66772, "router_noise": 0.0, "grad_norm": 0.7732, "lr": 8.5e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 253.5} |
| {"event": "train", "step": 4350, "phase": 3, "objective": "global", "lm_loss": 1.34695, "bpb": 1.94324, "chunk_loss": 0.032038, "F": 0.202, "G": 0.43332, "chunks": 206.84, "bytes_per_chunk": 4.951, "boundary_entropy": 0.66669, "router_noise": 0.0, "grad_norm": 0.7742, "lr": 8.21e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 255.6} |
| {"event": "probe", "step": 4400, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3598.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.206746, "cps": 28.109375, "mean_p": 0.437387, "boundary_entropy": 0.663405, "bpb": 2.074854}, "de": {"chunks": 3687.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.179565, "cps": 28.804688, "mean_p": 0.440329, "boundary_entropy": 0.669791, "bpb": 2.292428}, "vi": {"chunks": 5089.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.276051, "cps": 39.757812, "mean_p": 0.431129, "boundary_entropy": 0.655364, "bpb": 1.928633}, "tr": {"chunks": 4506.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.248909, "cps": 35.203125, "mean_p": 0.448935, "boundary_entropy": 0.667857, "bpb": 2.262986}, "fi": {"chunks": 3352.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.178194, "cps": 26.1875, "mean_p": 0.440914, "boundary_entropy": 0.670029, "bpb": 2.458953}, "ru": {"chunks": 7860.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.414776, "cps": 61.40625, "mean_p": 0.429897, "boundary_entropy": 0.664634, "bpb": 1.441011}, "el": {"chunks": 5761.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.280709, "cps": 45.007812, "mean_p": 0.441808, "boundary_entropy": 0.676051, "bpb": 1.53787}, "hi": {"chunks": 6629.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.393997, "cps": 51.789062, "mean_p": 0.447298, "boundary_entropy": 0.675558, "bpb": 1.243206}, "ta": {"chunks": 12042.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.644129, "cps": 94.078125, "mean_p": 0.431905, "boundary_entropy": 0.653705, "bpb": 0.97576}, "th": {"chunks": 6793.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.422319, "cps": 53.070312, "mean_p": 0.443878, "boundary_entropy": 0.677482, "bpb": 1.132826}, "zh": {"chunks": 2960.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.501185, "cps": 23.125, "mean_p": 0.412908, "boundary_entropy": 0.650006, "bpb": 3.196507}, "my": {"chunks": 6265.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.330642, "cps": 48.945312, "mean_p": 0.449271, "boundary_entropy": 0.67922, "bpb": 1.126607}}, "bpb": {"en": 2.07485, "de": 2.29243, "vi": 1.92863, "tr": 2.26299, "fi": 2.45895, "ru": 1.44101, "el": 1.53787, "hi": 1.24321, "ta": 0.97576, "th": 1.13283, "zh": 3.19651, "my": 1.12661}, "gini_cpc": 0.219962, "gini_cps": 0.223097, "bpic": 5.599209, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8685277352984324, "vi": 1.3352182819222866, "tr": 1.2039365440736454, "fi": 0.8618964416199538, "ru": 2.006209547667203, "el": 1.357750562100678, "hi": 1.905706026214002, "ta": 3.1155599653368986, "th": 2.0426949279147415, "zh": 2.424159713969852, "my": 1.5992658383501355}, "t": 258.6} |
| {"event": "eval", "step": 4400, "phase": 3, "bpb": {"en": 2.08582, "de": 2.29766, "vi": 1.93001, "tr": 2.27111, "fi": 2.46419, "ru": 1.41576, "el": 1.53111, "hi": 1.24088, "ta": 0.98319, "th": 1.15052, "zh": 3.19501, "my": 1.13704}, "bytes_per_chunk": {"en": 4.877, "de": 5.755, "vi": 5.219, "tr": 4.442, "fi": 5.648, "ru": 4.391, "el": 6.63, "hi": 7.039, "ta": 4.079, "th": 6.792, "zh": 5.546, "my": 8.277}, "bpb_mean": 1.80852, "t": 259.8} |
| {"event": "train", "step": 4400, "phase": 3, "objective": "global", "lm_loss": 1.35646, "bpb": 1.95695, "chunk_loss": 0.031778, "F": 0.198, "G": 0.43165, "chunks": 202.75, "bytes_per_chunk": 5.051, "boundary_entropy": 0.66672, "router_noise": 0.0, "grad_norm": 0.7837, "lr": 7.92e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 259.9} |
| {"event": "train", "step": 4450, "phase": 3, "objective": "global", "lm_loss": 1.19842, "bpb": 1.72895, "chunk_loss": 0.031205, "F": 0.18793, "G": 0.43634, "chunks": 192.44, "bytes_per_chunk": 5.321, "boundary_entropy": 0.66976, "router_noise": 0.0, "grad_norm": 1.0646, "lr": 7.63e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 262.0} |
| {"event": "train", "step": 4500, "phase": 3, "objective": "global", "lm_loss": 1.31415, "bpb": 1.89592, "chunk_loss": 0.03181, "F": 0.19806, "G": 0.43422, "chunks": 202.81, "bytes_per_chunk": 5.049, "boundary_entropy": 0.66782, "router_noise": 0.0, "grad_norm": 0.7819, "lr": 7.36e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 264.0} |
| {"event": "train", "step": 4550, "phase": 3, "objective": "global", "lm_loss": 1.31731, "bpb": 1.90048, "chunk_loss": 0.03184, "F": 0.19891, "G": 0.43122, "chunks": 203.69, "bytes_per_chunk": 5.027, "boundary_entropy": 0.66682, "router_noise": 0.0, "grad_norm": 0.7535, "lr": 7.09e-05, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 266.1} |
| {"event": "probe", "step": 4600, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3714.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.213411, "cps": 29.015625, "mean_p": 0.438232, "boundary_entropy": 0.663653, "bpb": 2.065221}, "de": {"chunks": 3912.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.190523, "cps": 30.5625, "mean_p": 0.441969, "boundary_entropy": 0.669864, "bpb": 2.285455}, "vi": {"chunks": 5079.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.275509, "cps": 39.679688, "mean_p": 0.430919, "boundary_entropy": 0.653926, "bpb": 1.923948}, "tr": {"chunks": 4309.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.238027, "cps": 33.664062, "mean_p": 0.447151, "boundary_entropy": 0.667218, "bpb": 2.256021}, "fi": {"chunks": 3379.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179629, "cps": 26.398438, "mean_p": 0.440785, "boundary_entropy": 0.669933, "bpb": 2.447001}, "ru": {"chunks": 7725.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.407652, "cps": 60.351562, "mean_p": 0.428897, "boundary_entropy": 0.663384, "bpb": 1.428711}, "el": {"chunks": 6013.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.292988, "cps": 46.976562, "mean_p": 0.442003, "boundary_entropy": 0.675983, "bpb": 1.525275}, "hi": {"chunks": 6663.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.396018, "cps": 52.054688, "mean_p": 0.445615, "boundary_entropy": 0.675271, "bpb": 1.227}, "ta": {"chunks": 12343.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.66023, "cps": 96.429688, "mean_p": 0.430916, "boundary_entropy": 0.652508, "bpb": 0.966941}, "th": {"chunks": 7421.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.461362, "cps": 57.976562, "mean_p": 0.446606, "boundary_entropy": 0.677702, "bpb": 1.126171}, "zh": {"chunks": 3309.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.560278, "cps": 25.851562, "mean_p": 0.413372, "boundary_entropy": 0.648716, "bpb": 3.188792}, "my": {"chunks": 8129.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.429016, "cps": 63.507812, "mean_p": 0.454034, "boundary_entropy": 0.680048, "bpb": 1.120957}}, "bpb": {"en": 2.06522, "de": 2.28545, "vi": 1.92395, "tr": 2.25602, "fi": 2.447, "ru": 1.42871, "el": 1.52528, "hi": 1.227, "ta": 0.96694, "th": 1.12617, "zh": 3.18879, "my": 1.12096}, "gini_cpc": 0.228522, "gini_cps": 0.226566, "bpic": 5.330588, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8927475350566529, "vi": 1.290973393396788, "tr": 1.1153422745112176, "fi": 0.8417023294983712, "ru": 1.9101676889697825, "el": 1.3728800037343343, "hi": 1.8556538250113421, "ta": 3.0936948922007517, "th": 2.161840193680282, "zh": 2.6253399379727744, "my": 2.0102772983316637}, "t": 269.1} |
| {"event": "eval", "step": 4600, "phase": 3, "bpb": {"en": 2.07549, "de": 2.29125, "vi": 1.92391, "tr": 2.26152, "fi": 2.45415, "ru": 1.40415, "el": 1.51904, "hi": 1.22392, "ta": 0.9748, "th": 1.14301, "zh": 3.18713, "my": 1.13088}, "bytes_per_chunk": {"en": 4.755, "de": 5.408, "vi": 5.225, "tr": 4.577, "fi": 5.571, "ru": 4.458, "el": 6.313, "hi": 6.671, "ta": 4.015, "th": 6.22, "zh": 4.968, "my": 6.581}, "bpb_mean": 1.7991, "t": 270.4} |
| {"event": "train", "step": 4600, "phase": 3, "objective": "global", "lm_loss": 1.47023, "bpb": 2.12109, "chunk_loss": 0.03156, "F": 0.19437, "G": 0.43012, "chunks": 199.03, "bytes_per_chunk": 5.145, "boundary_entropy": 0.66582, "router_noise": 0.0, "grad_norm": 0.7343, "lr": 6.82e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 270.4} |
| {"event": "train", "step": 4650, "phase": 3, "objective": "global", "lm_loss": 1.35986, "bpb": 1.96187, "chunk_loss": 0.031997, "F": 0.20108, "G": 0.43495, "chunks": 205.91, "bytes_per_chunk": 4.973, "boundary_entropy": 0.66863, "router_noise": 0.0, "grad_norm": 0.7998, "lr": 6.57e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 272.5} |
| {"event": "train", "step": 4700, "phase": 3, "objective": "global", "lm_loss": 1.42152, "bpb": 2.05082, "chunk_loss": 0.031657, "F": 0.1955, "G": 0.43156, "chunks": 200.19, "bytes_per_chunk": 5.115, "boundary_entropy": 0.6662, "router_noise": 0.0, "grad_norm": 0.8246, "lr": 6.32e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 274.6} |
| {"event": "train", "step": 4750, "phase": 3, "objective": "global", "lm_loss": 1.39251, "bpb": 2.00896, "chunk_loss": 0.031896, "F": 0.19937, "G": 0.43342, "chunks": 204.16, "bytes_per_chunk": 5.016, "boundary_entropy": 0.66829, "router_noise": 0.0, "grad_norm": 0.7742, "lr": 6.08e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 276.7} |
| {"event": "probe", "step": 4800, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3623.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.208182, "cps": 28.304688, "mean_p": 0.437514, "boundary_entropy": 0.663535, "bpb": 2.057741}, "de": {"chunks": 3750.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.182633, "cps": 29.296875, "mean_p": 0.441329, "boundary_entropy": 0.670054, "bpb": 2.274104}, "vi": {"chunks": 5218.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.283049, "cps": 40.765625, "mean_p": 0.430812, "boundary_entropy": 0.654261, "bpb": 1.912578}, "tr": {"chunks": 4521.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.249738, "cps": 35.320312, "mean_p": 0.448868, "boundary_entropy": 0.667921, "bpb": 2.242806}, "fi": {"chunks": 3434.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.182553, "cps": 26.828125, "mean_p": 0.44173, "boundary_entropy": 0.670382, "bpb": 2.430457}, "ru": {"chunks": 7706.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.406649, "cps": 60.203125, "mean_p": 0.43052, "boundary_entropy": 0.664653, "bpb": 1.418753}, "el": {"chunks": 5887.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.286849, "cps": 45.992188, "mean_p": 0.441763, "boundary_entropy": 0.675876, "bpb": 1.522141}, "hi": {"chunks": 7645.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.454383, "cps": 59.726562, "mean_p": 0.44879, "boundary_entropy": 0.676, "bpb": 1.211859}, "ta": {"chunks": 13150.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.703397, "cps": 102.734375, "mean_p": 0.428935, "boundary_entropy": 0.651737, "bpb": 0.957834}, "th": {"chunks": 8646.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.537519, "cps": 67.546875, "mean_p": 0.44932, "boundary_entropy": 0.678354, "bpb": 1.116923}, "zh": {"chunks": 3046.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.515747, "cps": 23.796875, "mean_p": 0.412921, "boundary_entropy": 0.64949, "bpb": 3.169906}, "my": {"chunks": 8892.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.469284, "cps": 69.46875, "mean_p": 0.452988, "boundary_entropy": 0.679845, "bpb": 1.114962}}, "bpb": {"en": 2.05774, "de": 2.2741, "vi": 1.91258, "tr": 2.24281, "fi": 2.43046, "ru": 1.41875, "el": 1.52214, "hi": 1.21186, "ta": 0.95783, "th": 1.11692, "zh": 3.16991, "my": 1.11496}, "gini_cpc": 0.23824, "gini_cps": 0.248559, "bpic": 5.08198, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8772727647283527, "vi": 1.3596174158094465, "tr": 1.1996090729270537, "fi": 0.8768881345712812, "ru": 1.953329803651109, "el": 1.3778723235312667, "hi": 2.1826203646096283, "ta": 3.3787500838783924, "th": 2.581962629355452, "zh": 2.4773778056980467, "my": 2.254196982650951}, "t": 279.7} |
| {"event": "eval", "step": 4800, "phase": 3, "bpb": {"en": 2.06817, "de": 2.28029, "vi": 1.91253, "tr": 2.25015, "fi": 2.43774, "ru": 1.39302, "el": 1.51364, "hi": 1.2079, "ta": 0.96652, "th": 1.13469, "zh": 3.16942, "my": 1.12462}, "bytes_per_chunk": {"en": 4.9, "de": 5.615, "vi": 5.093, "tr": 4.476, "fi": 5.485, "ru": 4.456, "el": 6.548, "hi": 5.897, "ta": 3.752, "th": 5.25, "zh": 5.334, "my": 6.039}, "bpb_mean": 1.78822, "t": 281.0} |
| {"event": "train", "step": 4800, "phase": 3, "objective": "global", "lm_loss": 1.2784, "bpb": 1.84434, "chunk_loss": 0.031789, "F": 0.19772, "G": 0.43546, "chunks": 202.47, "bytes_per_chunk": 5.058, "boundary_entropy": 0.66925, "router_noise": 0.0, "grad_norm": 0.7773, "lr": 5.85e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 281.0} |
| {"event": "train", "step": 4850, "phase": 3, "objective": "global", "lm_loss": 1.36132, "bpb": 1.96397, "chunk_loss": 0.031488, "F": 0.19278, "G": 0.43241, "chunks": 197.41, "bytes_per_chunk": 5.187, "boundary_entropy": 0.66787, "router_noise": 0.0, "grad_norm": 0.8208, "lr": 5.62e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 283.1} |
| {"event": "train", "step": 4900, "phase": 3, "objective": "global", "lm_loss": 1.39866, "bpb": 2.01784, "chunk_loss": 0.031668, "F": 0.1955, "G": 0.43011, "chunks": 200.19, "bytes_per_chunk": 5.115, "boundary_entropy": 0.66624, "router_noise": 0.0, "grad_norm": 0.8164, "lr": 5.41e-05, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 285.2} |
| {"event": "train", "step": 4950, "phase": 3, "objective": "global", "lm_loss": 1.27911, "bpb": 1.84536, "chunk_loss": 0.031753, "F": 0.19724, "G": 0.43304, "chunks": 201.97, "bytes_per_chunk": 5.07, "boundary_entropy": 0.6674, "router_noise": 0.0, "grad_norm": 0.7828, "lr": 5.2e-05, "router_trainable": true, "sec_per_step": 0.042, "alpha": 0.03, "t": 287.3} |
| {"event": "probe", "step": 5000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3691.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.21209, "cps": 28.835938, "mean_p": 0.438048, "boundary_entropy": 0.663809, "bpb": 2.054043}, "de": {"chunks": 3915.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.190669, "cps": 30.585938, "mean_p": 0.442719, "boundary_entropy": 0.670418, "bpb": 2.259266}, "vi": {"chunks": 5274.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.286086, "cps": 41.203125, "mean_p": 0.429678, "boundary_entropy": 0.653392, "bpb": 1.908272}, "tr": {"chunks": 4757.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.262774, "cps": 37.164062, "mean_p": 0.450752, "boundary_entropy": 0.668288, "bpb": 2.233412}, "fi": {"chunks": 3367.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.178991, "cps": 26.304688, "mean_p": 0.441779, "boundary_entropy": 0.670308, "bpb": 2.41888}, "ru": {"chunks": 8025.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.423483, "cps": 62.695312, "mean_p": 0.43024, "boundary_entropy": 0.664112, "bpb": 1.4049}, "el": {"chunks": 5370.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.261658, "cps": 41.953125, "mean_p": 0.44068, "boundary_entropy": 0.675838, "bpb": 1.515579}, "hi": {"chunks": 6661.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.395899, "cps": 52.039062, "mean_p": 0.442168, "boundary_entropy": 0.674032, "bpb": 1.209924}, "ta": {"chunks": 11783.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.630275, "cps": 92.054688, "mean_p": 0.421832, "boundary_entropy": 0.64892, "bpb": 0.957862}, "th": {"chunks": 8086.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.502704, "cps": 63.171875, "mean_p": 0.447973, "boundary_entropy": 0.678006, "bpb": 1.112321}, "zh": {"chunks": 3151.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.533525, "cps": 24.617188, "mean_p": 0.414245, "boundary_entropy": 0.650424, "bpb": 3.164326}, "my": {"chunks": 7410.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.39107, "cps": 57.890625, "mean_p": 0.449985, "boundary_entropy": 0.67921, "bpb": 1.108735}}, "bpb": {"en": 2.05404, "de": 2.25927, "vi": 1.90827, "tr": 2.23341, "fi": 2.41888, "ru": 1.4049, "el": 1.51558, "hi": 1.20992, "ta": 0.95786, "th": 1.11232, "zh": 3.16433, "my": 1.10874}, "gini_cpc": 0.222476, "gini_cps": 0.222693, "bpic": 5.368317, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8989994669687612, "vi": 1.3488916258601014, "tr": 1.2389753757559339, "fi": 0.8439394878406674, "ru": 1.9967141781799698, "el": 1.233711223121971, "hi": 1.8666566238990225, "ta": 2.971737763926899, "th": 2.3702423128551944, "zh": 2.515562057165254, "my": 1.8438895665457407}, "t": 290.3} |
| {"event": "eval", "step": 5000, "phase": 3, "bpb": {"en": 2.06385, "de": 2.2666, "vi": 1.90869, "tr": 2.24034, "fi": 2.42595, "ru": 1.37859, "el": 1.5075, "hi": 1.20802, "ta": 0.9659, "th": 1.12871, "zh": 3.16205, "my": 1.11864}, "bytes_per_chunk": {"en": 4.858, "de": 5.376, "vi": 5.018, "tr": 4.218, "fi": 5.599, "ru": 4.277, "el": 7.134, "hi": 6.666, "ta": 4.163, "th": 5.593, "zh": 5.207, "my": 7.324}, "bpb_mean": 1.78124, "t": 291.6} |
| {"event": "train", "step": 5000, "phase": 3, "objective": "global", "lm_loss": 1.47631, "bpb": 2.12987, "chunk_loss": 0.031983, "F": 0.20074, "G": 0.43198, "chunks": 205.56, "bytes_per_chunk": 4.981, "boundary_entropy": 0.66613, "router_noise": 0.0, "grad_norm": 0.8428, "lr": 5e-05, "router_trainable": true, "sec_per_step": 0.087, "alpha": 0.03, "t": 291.6} |
| {"event": "train", "step": 5050, "phase": 3, "objective": "global", "lm_loss": 1.29854, "bpb": 1.87339, "chunk_loss": 0.031925, "F": 0.20074, "G": 0.43161, "chunks": 205.56, "bytes_per_chunk": 4.981, "boundary_entropy": 0.66732, "router_noise": 0.0, "grad_norm": 0.7481, "lr": 4.81e-05, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 293.6} |
| {"event": "train", "step": 5100, "phase": 3, "objective": "global", "lm_loss": 1.46011, "bpb": 2.1065, "chunk_loss": 0.031686, "F": 0.19635, "G": 0.42979, "chunks": 201.06, "bytes_per_chunk": 5.093, "boundary_entropy": 0.66574, "router_noise": 0.0, "grad_norm": 0.7843, "lr": 4.63e-05, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 295.7} |
| {"event": "train", "step": 5150, "phase": 3, "objective": "global", "lm_loss": 1.25901, "bpb": 1.81637, "chunk_loss": 0.032098, "F": 0.20343, "G": 0.4309, "chunks": 208.31, "bytes_per_chunk": 4.916, "boundary_entropy": 0.66586, "router_noise": 0.0, "grad_norm": 0.7344, "lr": 4.45e-05, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 297.7} |
| {"event": "probe", "step": 5200, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3777.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.217032, "cps": 29.507812, "mean_p": 0.43828, "boundary_entropy": 0.663504, "bpb": 2.045801}, "de": {"chunks": 3775.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.18385, "cps": 29.492188, "mean_p": 0.441469, "boundary_entropy": 0.670049, "bpb": 2.254047}, "vi": {"chunks": 5140.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.278817, "cps": 40.15625, "mean_p": 0.429014, "boundary_entropy": 0.653079, "bpb": 1.903839}, "tr": {"chunks": 4364.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.241065, "cps": 34.09375, "mean_p": 0.447409, "boundary_entropy": 0.667476, "bpb": 2.221415}, "fi": {"chunks": 3125.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.166126, "cps": 24.414062, "mean_p": 0.43967, "boundary_entropy": 0.669886, "bpb": 2.413518}, "ru": {"chunks": 7871.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.415356, "cps": 61.492188, "mean_p": 0.429922, "boundary_entropy": 0.664233, "bpb": 1.402145}, "el": {"chunks": 5070.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.24704, "cps": 39.609375, "mean_p": 0.440656, "boundary_entropy": 0.676154, "bpb": 1.502588}, "hi": {"chunks": 7601.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.451768, "cps": 59.382812, "mean_p": 0.448799, "boundary_entropy": 0.676048, "bpb": 1.200333}, "ta": {"chunks": 13229.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.707622, "cps": 103.351562, "mean_p": 0.42472, "boundary_entropy": 0.649298, "bpb": 0.944303}, "th": {"chunks": 7847.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.487846, "cps": 61.304688, "mean_p": 0.448623, "boundary_entropy": 0.678426, "bpb": 1.107135}, "zh": {"chunks": 3148.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.533017, "cps": 24.59375, "mean_p": 0.414309, "boundary_entropy": 0.650362, "bpb": 3.148727}, "my": {"chunks": 9619.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.507653, "cps": 75.148438, "mean_p": 0.453939, "boundary_entropy": 0.680376, "bpb": 1.109157}}, "bpb": {"en": 2.0458, "de": 2.25405, "vi": 1.90384, "tr": 2.22141, "fi": 2.41352, "ru": 1.40215, "el": 1.50259, "hi": 1.20033, "ta": 0.9443, "th": 1.10713, "zh": 3.14873, "my": 1.10916}, "gini_cpc": 0.246363, "gini_cps": 0.255016, "bpic": 5.146863, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.847113658491279, "vi": 1.2846863580323111, "tr": 1.1107372221886087, "fi": 0.765447261456918, "ru": 1.9138056546951658, "el": 1.138267273467607, "hi": 2.0815785069638277, "ta": 3.2604585417728384, "th": 2.247810640602459, "zh": 2.455943754241389, "my": 2.3390725052788595}, "t": 300.7} |
| {"event": "eval", "step": 5200, "phase": 3, "bpb": {"en": 2.05585, "de": 2.26063, "vi": 1.90383, "tr": 2.23035, "fi": 2.42019, "ru": 1.37518, "el": 1.49366, "hi": 1.19768, "ta": 0.95349, "th": 1.12446, "zh": 3.14705, "my": 1.11929}, "bytes_per_chunk": {"en": 4.712, "de": 5.595, "vi": 5.105, "tr": 4.596, "fi": 6.098, "ru": 4.37, "el": 7.501, "hi": 5.84, "ta": 3.781, "th": 5.786, "zh": 5.231, "my": 5.721}, "bpb_mean": 1.77347, "t": 302.0} |
| {"event": "train", "step": 5200, "phase": 3, "objective": "global", "lm_loss": 1.36975, "bpb": 1.97613, "chunk_loss": 0.031704, "F": 0.19626, "G": 0.43411, "chunks": 200.97, "bytes_per_chunk": 5.095, "boundary_entropy": 0.66819, "router_noise": 0.0, "grad_norm": 0.8129, "lr": 4.29e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 302.0} |
| {"event": "train", "step": 5250, "phase": 3, "objective": "global", "lm_loss": 1.33225, "bpb": 1.92203, "chunk_loss": 0.031798, "F": 0.198, "G": 0.43405, "chunks": 202.75, "bytes_per_chunk": 5.051, "boundary_entropy": 0.66806, "router_noise": 0.0, "grad_norm": 0.7992, "lr": 4.14e-05, "router_trainable": true, "sec_per_step": 0.041, "alpha": 0.03, "t": 304.0} |
| {"event": "train", "step": 5300, "phase": 3, "objective": "global", "lm_loss": 1.29437, "bpb": 1.86739, "chunk_loss": 0.031978, "F": 0.20123, "G": 0.43335, "chunks": 206.06, "bytes_per_chunk": 4.969, "boundary_entropy": 0.66845, "router_noise": 0.0, "grad_norm": 0.7842, "lr": 3.99e-05, "router_trainable": true, "sec_per_step": 0.04, "alpha": 0.03, "t": 306.0} |
| {"event": "train", "step": 5350, "phase": 3, "objective": "global", "lm_loss": 1.3761, "bpb": 1.98529, "chunk_loss": 0.031529, "F": 0.19348, "G": 0.4302, "chunks": 198.12, "bytes_per_chunk": 5.168, "boundary_entropy": 0.6648, "router_noise": 0.0, "grad_norm": 0.7891, "lr": 3.86e-05, "router_trainable": true, "sec_per_step": 0.04, "alpha": 0.03, "t": 308.1} |
| {"event": "probe", "step": 5400, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3698.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.212492, "cps": 28.890625, "mean_p": 0.438276, "boundary_entropy": 0.663688, "bpb": 2.040115}, "de": {"chunks": 3878.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.188867, "cps": 30.296875, "mean_p": 0.442558, "boundary_entropy": 0.670505, "bpb": 2.248057}, "vi": {"chunks": 5026.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.272634, "cps": 39.265625, "mean_p": 0.429293, "boundary_entropy": 0.653877, "bpb": 1.898139}, "tr": {"chunks": 4534.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.250456, "cps": 35.421875, "mean_p": 0.448596, "boundary_entropy": 0.667686, "bpb": 2.217292}, "fi": {"chunks": 3386.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.180001, "cps": 26.453125, "mean_p": 0.44152, "boundary_entropy": 0.670303, "bpb": 2.404816}, "ru": {"chunks": 7906.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.417203, "cps": 61.765625, "mean_p": 0.42934, "boundary_entropy": 0.664103, "bpb": 1.39374}, "el": {"chunks": 5556.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.270721, "cps": 43.40625, "mean_p": 0.442206, "boundary_entropy": 0.676511, "bpb": 1.497205}, "hi": {"chunks": 6801.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.40422, "cps": 53.132812, "mean_p": 0.446845, "boundary_entropy": 0.675954, "bpb": 1.198342}, "ta": {"chunks": 12659.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.677133, "cps": 98.898438, "mean_p": 0.421925, "boundary_entropy": 0.649595, "bpb": 0.944925}, "th": {"chunks": 8490.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.527821, "cps": 66.328125, "mean_p": 0.449786, "boundary_entropy": 0.678599, "bpb": 1.105331}, "zh": {"chunks": 3187.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.539621, "cps": 24.898438, "mean_p": 0.415573, "boundary_entropy": 0.650988, "bpb": 3.148949}, "my": {"chunks": 9461.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.499314, "cps": 73.914062, "mean_p": 0.454244, "boundary_entropy": 0.680257, "bpb": 1.09967}}, "bpb": {"en": 2.04011, "de": 2.24806, "vi": 1.89814, "tr": 2.21729, "fi": 2.40482, "ru": 1.39374, "el": 1.4972, "hi": 1.19834, "ta": 0.94493, "th": 1.10533, "zh": 3.14895, "my": 1.09967}, "gini_cpc": 0.237359, "gini_cps": 0.245683, "bpic": 5.145759, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8888175288117373, "vi": 1.2830292450210592, "tr": 1.1786590027314108, "fi": 0.8470953226077776, "ru": 1.9633820178061023, "el": 1.2740269217928457, "hi": 1.902282073680243, "ta": 3.1866263556441243, "th": 2.483955655402201, "zh": 2.539486065906888, "my": 2.349799893546004}, "t": 311.0} |
| {"event": "eval", "step": 5400, "phase": 3, "bpb": {"en": 2.05017, "de": 2.25466, "vi": 1.89822, "tr": 2.22357, "fi": 2.40999, "ru": 1.36832, "el": 1.49042, "hi": 1.19496, "ta": 0.9535, "th": 1.12265, "zh": 3.14571, "my": 1.10988}, "bytes_per_chunk": {"en": 4.802, "de": 5.492, "vi": 5.258, "tr": 4.44, "fi": 5.603, "ru": 4.361, "el": 6.901, "hi": 6.454, "ta": 3.924, "th": 5.4, "zh": 5.03, "my": 5.734}, "bpb_mean": 1.7685, "t": 312.3} |
| {"event": "train", "step": 5400, "phase": 3, "objective": "global", "lm_loss": 1.25785, "bpb": 1.8147, "chunk_loss": 0.031408, "F": 0.19168, "G": 0.43414, "chunks": 196.28, "bytes_per_chunk": 5.217, "boundary_entropy": 0.66923, "router_noise": 0.0, "grad_norm": 0.7822, "lr": 3.73e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 312.3} |
| {"event": "train", "step": 5450, "phase": 3, "objective": "global", "lm_loss": 1.3065, "bpb": 1.88488, "chunk_loss": 0.031037, "F": 0.18481, "G": 0.43285, "chunks": 189.25, "bytes_per_chunk": 5.411, "boundary_entropy": 0.66911, "router_noise": 0.0, "grad_norm": 0.7923, "lr": 3.62e-05, "router_trainable": true, "sec_per_step": 0.04, "alpha": 0.03, "t": 314.4} |
| {"event": "train", "step": 5500, "phase": 3, "objective": "global", "lm_loss": 1.46888, "bpb": 2.11915, "chunk_loss": 0.031816, "F": 0.19827, "G": 0.43208, "chunks": 203.03, "bytes_per_chunk": 5.044, "boundary_entropy": 0.66655, "router_noise": 0.0, "grad_norm": 0.8148, "lr": 3.51e-05, "router_trainable": true, "sec_per_step": 0.04, "alpha": 0.03, "t": 316.4} |
| {"event": "train", "step": 5550, "phase": 3, "objective": "global", "lm_loss": 1.31385, "bpb": 1.89548, "chunk_loss": 0.032368, "F": 0.20804, "G": 0.43358, "chunks": 213.03, "bytes_per_chunk": 4.807, "boundary_entropy": 0.66563, "router_noise": 0.0, "grad_norm": 0.7846, "lr": 3.41e-05, "router_trainable": true, "sec_per_step": 0.04, "alpha": 0.03, "t": 318.4} |
| {"event": "probe", "step": 5600, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3594.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.206516, "cps": 28.078125, "mean_p": 0.43742, "boundary_entropy": 0.663391, "bpb": 2.037183}, "de": {"chunks": 3829.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.18648, "cps": 29.914062, "mean_p": 0.442127, "boundary_entropy": 0.67026, "bpb": 2.242348}, "vi": {"chunks": 5068.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.274912, "cps": 39.59375, "mean_p": 0.430031, "boundary_entropy": 0.653998, "bpb": 1.893052}, "tr": {"chunks": 4339.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.239684, "cps": 33.898438, "mean_p": 0.448012, "boundary_entropy": 0.667619, "bpb": 2.209156}, "fi": {"chunks": 3196.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.169901, "cps": 24.96875, "mean_p": 0.440812, "boundary_entropy": 0.670307, "bpb": 2.397582}, "ru": {"chunks": 7546.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.398206, "cps": 58.953125, "mean_p": 0.427347, "boundary_entropy": 0.663398, "bpb": 1.390948}, "el": {"chunks": 5199.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.253326, "cps": 40.617188, "mean_p": 0.440541, "boundary_entropy": 0.67607, "bpb": 1.489197}, "hi": {"chunks": 7050.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.419019, "cps": 55.078125, "mean_p": 0.446418, "boundary_entropy": 0.675857, "bpb": 1.189746}, "ta": {"chunks": 13074.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.699331, "cps": 102.140625, "mean_p": 0.420794, "boundary_entropy": 0.647533, "bpb": 0.935399}, "th": {"chunks": 7363.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.457756, "cps": 57.523438, "mean_p": 0.448581, "boundary_entropy": 0.678831, "bpb": 1.100764}, "zh": {"chunks": 3076.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.520826, "cps": 24.03125, "mean_p": 0.412788, "boundary_entropy": 0.649646, "bpb": 3.131557}, "my": {"chunks": 5407.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.28536, "cps": 42.242188, "mean_p": 0.448393, "boundary_entropy": 0.679621, "bpb": 1.101589}}, "bpb": {"en": 2.03718, "de": 2.24235, "vi": 1.89305, "tr": 2.20916, "fi": 2.39758, "ru": 1.39095, "el": 1.4892, "hi": 1.18975, "ta": 0.9354, "th": 1.10076, "zh": 3.13156, "my": 1.10159}, "gini_cpc": 0.241879, "gini_cps": 0.236335, "bpic": 5.583, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.9029818199734992, "vi": 1.33118836067526, "tr": 1.16060689449364, "fi": 0.8226989341032211, "ru": 1.9282069059690514, "el": 1.2266623049330136, "hi": 2.0289908639212952, "ta": 3.38632828807574, "th": 2.21656148497418, "zh": 2.5219643078231186, "my": 1.3817804408374665}, "t": 321.3} |
| {"event": "eval", "step": 5600, "phase": 3, "bpb": {"en": 2.04772, "de": 2.24967, "vi": 1.89268, "tr": 2.21662, "fi": 2.40395, "ru": 1.36486, "el": 1.48199, "hi": 1.1867, "ta": 0.9445, "th": 1.11884, "zh": 3.12911, "my": 1.11171}, "bytes_per_chunk": {"en": 4.985, "de": 5.496, "vi": 5.25, "tr": 4.634, "fi": 5.911, "ru": 4.575, "el": 7.336, "hi": 6.338, "ta": 3.817, "th": 6.119, "zh": 5.351, "my": 9.598}, "bpb_mean": 1.76236, "t": 322.6} |
| {"event": "train", "step": 5600, "phase": 3, "objective": "global", "lm_loss": 1.33572, "bpb": 1.92704, "chunk_loss": 0.031558, "F": 0.19388, "G": 0.43213, "chunks": 198.53, "bytes_per_chunk": 5.158, "boundary_entropy": 0.6673, "router_noise": 0.0, "grad_norm": 0.7647, "lr": 3.33e-05, "router_trainable": true, "sec_per_step": 0.085, "alpha": 0.03, "t": 322.6} |
| {"event": "train", "step": 5650, "phase": 3, "objective": "global", "lm_loss": 1.24537, "bpb": 1.79669, "chunk_loss": 0.031567, "F": 0.19373, "G": 0.43496, "chunks": 198.38, "bytes_per_chunk": 5.162, "boundary_entropy": 0.6696, "router_noise": 0.0, "grad_norm": 0.7993, "lr": 3.25e-05, "router_trainable": true, "sec_per_step": 0.04, "alpha": 0.03, "t": 324.6} |
| {"event": "train", "step": 5700, "phase": 3, "objective": "global", "lm_loss": 1.30506, "bpb": 1.8828, "chunk_loss": 0.031458, "F": 0.19223, "G": 0.43456, "chunks": 196.84, "bytes_per_chunk": 5.202, "boundary_entropy": 0.6685, "router_noise": 0.0, "grad_norm": 0.7985, "lr": 3.18e-05, "router_trainable": true, "sec_per_step": 0.04, "alpha": 0.03, "t": 326.7} |
| {"event": "train", "step": 5750, "phase": 3, "objective": "global", "lm_loss": 1.32808, "bpb": 1.91601, "chunk_loss": 0.03131, "F": 0.19016, "G": 0.43243, "chunks": 194.72, "bytes_per_chunk": 5.259, "boundary_entropy": 0.66821, "router_noise": 0.0, "grad_norm": 0.8221, "lr": 3.13e-05, "router_trainable": true, "sec_per_step": 0.04, "alpha": 0.03, "t": 328.7} |
| {"event": "probe", "step": 5800, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3825.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.21979, "cps": 29.882812, "mean_p": 0.439145, "boundary_entropy": 0.663899, "bpb": 2.033848}, "de": {"chunks": 3974.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.193542, "cps": 31.046875, "mean_p": 0.44333, "boundary_entropy": 0.67059, "bpb": 2.236082}, "vi": {"chunks": 5081.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.275617, "cps": 39.695312, "mean_p": 0.430614, "boundary_entropy": 0.654155, "bpb": 1.890246}, "tr": {"chunks": 4500.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.248578, "cps": 35.15625, "mean_p": 0.449362, "boundary_entropy": 0.667963, "bpb": 2.206038}, "fi": {"chunks": 3371.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179204, "cps": 26.335938, "mean_p": 0.441784, "boundary_entropy": 0.670377, "bpb": 2.392867}, "ru": {"chunks": 7748.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.408865, "cps": 60.53125, "mean_p": 0.427863, "boundary_entropy": 0.663536, "bpb": 1.384422}, "el": {"chunks": 5375.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.261901, "cps": 41.992188, "mean_p": 0.441238, "boundary_entropy": 0.676183, "bpb": 1.485928}, "hi": {"chunks": 7494.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.445409, "cps": 58.546875, "mean_p": 0.447902, "boundary_entropy": 0.676342, "bpb": 1.187434}, "ta": {"chunks": 12372.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.661781, "cps": 96.65625, "mean_p": 0.416739, "boundary_entropy": 0.645899, "bpb": 0.937191}, "th": {"chunks": 7316.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.454834, "cps": 57.15625, "mean_p": 0.446481, "boundary_entropy": 0.678249, "bpb": 1.097618}, "zh": {"chunks": 3030.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.513038, "cps": 23.671875, "mean_p": 0.412731, "boundary_entropy": 0.649704, "bpb": 3.129722}, "my": {"chunks": 8412.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.443952, "cps": 65.71875, "mean_p": 0.453889, "boundary_entropy": 0.680504, "bpb": 1.094827}}, "bpb": {"en": 2.03385, "de": 2.23608, "vi": 1.89025, "tr": 2.20604, "fi": 2.39287, "ru": 1.38442, "el": 1.48593, "hi": 1.18743, "ta": 0.93719, "th": 1.09762, "zh": 3.12972, "my": 1.09483}, "gini_cpc": 0.22239, "gini_cps": 0.232114, "bpic": 5.293677, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8805786190188343, "vi": 1.2540034567664873, "tr": 1.1309792657050668, "fi": 0.8153415033548372, "ru": 1.8602575611775052, "el": 1.191599477582779, "hi": 2.026521877458264, "ta": 3.0109748123908564, "th": 2.069404133677095, "zh": 2.334220433855899, "my": 2.0198939512473766}, "t": 331.6} |
| {"event": "eval", "step": 5800, "phase": 3, "bpb": {"en": 2.04512, "de": 2.24336, "vi": 1.88908, "tr": 2.21206, "fi": 2.39893, "ru": 1.35861, "el": 1.47786, "hi": 1.18498, "ta": 0.94621, "th": 1.11463, "zh": 3.12618, "my": 1.10469}, "bytes_per_chunk": {"en": 4.646, "de": 5.283, "vi": 5.188, "tr": 4.5, "fi": 5.502, "ru": 4.472, "el": 7.141, "hi": 5.914, "ta": 4.01, "th": 6.243, "zh": 5.372, "my": 6.355}, "bpb_mean": 1.75848, "t": 332.9} |
| {"event": "train", "step": 5800, "phase": 3, "objective": "global", "lm_loss": 1.35374, "bpb": 1.95303, "chunk_loss": 0.031687, "F": 0.19553, "G": 0.43739, "chunks": 200.22, "bytes_per_chunk": 5.114, "boundary_entropy": 0.66996, "router_noise": 0.0, "grad_norm": 0.8083, "lr": 3.08e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 333.0} |
| {"event": "train", "step": 5850, "phase": 3, "objective": "global", "lm_loss": 1.32575, "bpb": 1.91265, "chunk_loss": 0.031498, "F": 0.19281, "G": 0.43477, "chunks": 197.44, "bytes_per_chunk": 5.186, "boundary_entropy": 0.66866, "router_noise": 0.0, "grad_norm": 0.7767, "lr": 3.05e-05, "router_trainable": true, "sec_per_step": 0.04, "alpha": 0.03, "t": 335.0} |
| {"event": "train", "step": 5900, "phase": 3, "objective": "global", "lm_loss": 1.23463, "bpb": 1.7812, "chunk_loss": 0.031674, "F": 0.19586, "G": 0.43581, "chunks": 200.56, "bytes_per_chunk": 5.106, "boundary_entropy": 0.66993, "router_noise": 0.0, "grad_norm": 0.8059, "lr": 3.02e-05, "router_trainable": true, "sec_per_step": 0.04, "alpha": 0.03, "t": 337.0} |
| {"event": "train", "step": 5950, "phase": 3, "objective": "global", "lm_loss": 1.25824, "bpb": 1.81526, "chunk_loss": 0.032181, "F": 0.2052, "G": 0.43271, "chunks": 210.12, "bytes_per_chunk": 4.873, "boundary_entropy": 0.66671, "router_noise": 0.0, "grad_norm": 0.8882, "lr": 3.01e-05, "router_trainable": true, "sec_per_step": 0.04, "alpha": 0.03, "t": 339.0} |
| {"event": "probe", "step": 6000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3603.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.207033, "cps": 28.148438, "mean_p": 0.437879, "boundary_entropy": 0.663669, "bpb": 2.031333}, "de": {"chunks": 3711.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.180733, "cps": 28.992188, "mean_p": 0.442422, "boundary_entropy": 0.670858, "bpb": 2.235349}, "vi": {"chunks": 5139.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.278763, "cps": 40.148438, "mean_p": 0.429787, "boundary_entropy": 0.65334, "bpb": 1.887135}, "tr": {"chunks": 4360.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.240844, "cps": 34.0625, "mean_p": 0.448613, "boundary_entropy": 0.668046, "bpb": 2.20246}, "fi": {"chunks": 3250.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.172771, "cps": 25.390625, "mean_p": 0.441461, "boundary_entropy": 0.670507, "bpb": 2.38859}, "ru": {"chunks": 7652.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.403799, "cps": 59.78125, "mean_p": 0.428383, "boundary_entropy": 0.663782, "bpb": 1.384052}, "el": {"chunks": 4939.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.240657, "cps": 38.585938, "mean_p": 0.440425, "boundary_entropy": 0.676055, "bpb": 1.482803}, "hi": {"chunks": 7540.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.448143, "cps": 58.90625, "mean_p": 0.447165, "boundary_entropy": 0.675733, "bpb": 1.181173}, "ta": {"chunks": 12327.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.659374, "cps": 96.304688, "mean_p": 0.416662, "boundary_entropy": 0.645867, "bpb": 0.934155}, "th": {"chunks": 7734.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.480821, "cps": 60.421875, "mean_p": 0.448366, "boundary_entropy": 0.67866, "bpb": 1.093868}, "zh": {"chunks": 3106.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.525906, "cps": 24.265625, "mean_p": 0.414972, "boundary_entropy": 0.651006, "bpb": 3.123397}, "my": {"chunks": 8436.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.445218, "cps": 65.90625, "mean_p": 0.453699, "boundary_entropy": 0.68048, "bpb": 1.094733}}, "bpb": {"en": 2.03133, "de": 2.23535, "vi": 1.88714, "tr": 2.20246, "fi": 2.38859, "ru": 1.38405, "el": 1.4828, "hi": 1.18117, "ta": 0.93415, "th": 1.09387, "zh": 3.1234, "my": 1.09473}, "gini_cpc": 0.236366, "gini_cps": 0.239642, "bpic": 5.345363, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8729681628237771, "vi": 1.3464658214709873, "tr": 1.1633108961151755, "fi": 0.8345095972200084, "ru": 1.9504086084814982, "el": 1.1624065248294972, "hi": 2.1645924087529833, "ta": 3.184870546720561, "th": 2.322431752422665, "zh": 2.5401997376043726, "my": 2.1504683399257214}, "t": 342.0} |
| {"event": "eval", "step": 6000, "phase": 3, "bpb": {"en": 2.04117, "de": 2.24239, "vi": 1.88713, "tr": 2.20932, "fi": 2.39543, "ru": 1.35878, "el": 1.47552, "hi": 1.17761, "ta": 0.9431, "th": 1.11093, "zh": 3.12008, "my": 1.10432}, "bytes_per_chunk": {"en": 4.902, "de": 5.717, "vi": 5.141, "tr": 4.552, "fi": 5.774, "ru": 4.504, "el": 7.641, "hi": 5.922, "ta": 4.055, "th": 5.836, "zh": 5.164, "my": 6.376}, "bpb_mean": 1.75548, "t": 343.2} |
| {"event": "probe", "step": 6000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3603.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.207033, "cps": 28.148438, "mean_p": 0.437879, "boundary_entropy": 0.663669, "bpb": 2.031333}, "de": {"chunks": 3711.0, "chars": 20533.0, "bytes": 20862.0, "cpc": 0.180733, "cps": 28.992188, "mean_p": 0.442422, "boundary_entropy": 0.670858, "bpb": 2.235349}, "vi": {"chunks": 5139.0, "chars": 18435.0, "bytes": 24274.0, "cpc": 0.278763, "cps": 40.148438, "mean_p": 0.429787, "boundary_entropy": 0.65334, "bpb": 1.887135}, "tr": {"chunks": 4360.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.240844, "cps": 34.0625, "mean_p": 0.448613, "boundary_entropy": 0.668046, "bpb": 2.20246}, "fi": {"chunks": 3250.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.172771, "cps": 25.390625, "mean_p": 0.441461, "boundary_entropy": 0.670507, "bpb": 2.38859}, "ru": {"chunks": 7652.0, "chars": 18950.0, "bytes": 34560.0, "cpc": 0.403799, "cps": 59.78125, "mean_p": 0.428383, "boundary_entropy": 0.663782, "bpb": 1.384052}, "el": {"chunks": 4939.0, "chars": 20523.0, "bytes": 37264.0, "cpc": 0.240657, "cps": 38.585938, "mean_p": 0.440425, "boundary_entropy": 0.676055, "bpb": 1.482803}, "hi": {"chunks": 7540.0, "chars": 16825.0, "bytes": 43173.0, "cpc": 0.448143, "cps": 58.90625, "mean_p": 0.447165, "boundary_entropy": 0.675733, "bpb": 1.181173}, "ta": {"chunks": 12327.0, "chars": 18695.0, "bytes": 50915.0, "cpc": 0.659374, "cps": 96.304688, "mean_p": 0.416662, "boundary_entropy": 0.645867, "bpb": 0.934155}, "th": {"chunks": 7734.0, "chars": 16085.0, "bytes": 46085.0, "cpc": 0.480821, "cps": 60.421875, "mean_p": 0.448366, "boundary_entropy": 0.67866, "bpb": 1.093868}, "zh": {"chunks": 3106.0, "chars": 5906.0, "bytes": 16173.0, "cpc": 0.525906, "cps": 24.265625, "mean_p": 0.414972, "boundary_entropy": 0.651006, "bpb": 3.123397}, "my": {"chunks": 8436.0, "chars": 18948.0, "bytes": 53743.0, "cpc": 0.445218, "cps": 65.90625, "mean_p": 0.453699, "boundary_entropy": 0.68048, "bpb": 1.094733}}, "bpb": {"en": 2.03133, "de": 2.23535, "vi": 1.88714, "tr": 2.20246, "fi": 2.38859, "ru": 1.38405, "el": 1.4828, "hi": 1.18117, "ta": 0.93415, "th": 1.09387, "zh": 3.1234, "my": 1.09473}, "gini_cpc": 0.236366, "gini_cps": 0.239642, "bpic": 5.345363, "worst_lang": "ta", "cpc_parity_vs_en": {"en": 1.0, "de": 0.8729681628237771, "vi": 1.3464658214709873, "tr": 1.1633108961151755, "fi": 0.8345095972200084, "ru": 1.9504086084814982, "el": 1.1624065248294972, "hi": 2.1645924087529833, "ta": 3.184870546720561, "th": 2.322431752422665, "zh": 2.5401997376043726, "my": 2.1504683399257214}, "t": 344.2} |
| {"event": "reps_dumped", "dirs": ["/local/scratch/sas245/beetle-hnet/scratch/reps/emactl_genreg_s1/th0.30", "/local/scratch/sas245/beetle-hnet/scratch/reps/emactl_genreg_s1/th0.40", "/local/scratch/sas245/beetle-hnet/scratch/reps/emactl_genreg_s1/th0.50", "/local/scratch/sas245/beetle-hnet/scratch/reps/emactl_genreg_s1/th0.60", "/local/scratch/sas245/beetle-hnet/scratch/reps/emactl_genreg_s1/th0.70"], "t": 349.1} |
| {"event": "final", "run": "emactl_genreg_s1", "condition": "A", "step": 6000, "params": 22516992, "per_lang": {"en": {"cpc": 0.20703327012584039, "cps": 28.1484375, "bpb": 2.050497110736466, "bytes_per_chunk": 4.915092482351948}, "de": {"cpc": 0.18073345346515365, "cps": 28.9921875, "bpb": 2.234483494945642, "bytes_per_chunk": 5.727357140516271}, "vi": {"cpc": 0.2787632221318145, "cps": 40.1484375, "bpb": 1.889786890741809, "bytes_per_chunk": 5.128263314911723}, "tr": {"cpc": 0.24084405899574657, "cps": 34.0625, "bpb": 2.200235880250896, "bytes_per_chunk": 4.5643048046871595}, "fi": {"cpc": 0.17277125086385625, "cps": 25.390625, "bpb": 2.410347579260111, "bytes_per_chunk": 5.786971014812689}, "ru": {"cpc": 0.4037994722955145, "cps": 59.78125, "bpb": 1.3596733775112122, "bytes_per_chunk": 4.511013215859031}, "el": {"cpc": 0.24065682405106467, "cps": 38.5859375, "bpb": 1.470116008487108, "bytes_per_chunk": 7.593644540358906}, "hi": {"cpc": 0.44814264487369987, "cps": 58.90625, "bpb": 1.1866793342998843, "bytes_per_chunk": 5.916870746763872}, "ta": {"cpc": 0.6593741642150308, "cps": 96.3046875, "bpb": 0.9440668483807373, "bytes_per_chunk": 4.056543773453519}, "th": {"cpc": 0.4808206403481505, "cps": 60.421875, "bpb": 1.1085081298711636, "bytes_per_chunk": 5.815535811342939}, "zh": {"cpc": 0.5259058584490349, "cps": 24.265625, "bpb": 3.1290566063685494, "bytes_per_chunk": 5.197248161145145}, "my": {"cpc": 0.44521849271690944, "cps": 65.90625, "bpb": 1.1147818581196989, "bytes_per_chunk": 6.384024353120243}}, "gini_cpc": 0.23636601785882339, "gini_cps": 0.23964209275225068, "bpic": 5.3453626196080615, "seed": 1, "mean_boundary_entropy": 0.6663709878921509, "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": 351.4} |
|
|
