| {"event": "hub", "repo": "AdaptiveChunking/ali_hnet_s4", "keep_steps": [0, 1, 2, 4, 6, 8, 16, 32, 64, 128, 256, 512, 1000, 1024, 2048, 3000, 4096, 6000], "t": 1.5} |
| {"event": "start", "run": "ali_hnet_s4", "arch": "hnet", "objective": "ratio", "condition": "A", "params": 22516992, "langs": ["en", "tr", "fi"], "premium": {"en": 1.0, "tr": 1.12, "fi": 1.104}, "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": {}, "weights": {"en": 0.74, "tr": 0.1536, "fi": 0.1064}, "device": "cuda", "t": 1.5} |
| {"event": "probe_schedule", "n": 45, "steps": [0, 1, 2, 4, 6, 8, 16, 32, 64, 128, 200, 256, 400, 512, 600, 800, 1000, 1024, 1200, 1400, 1600, 1800, 2000, 2048, 2200, 2400, 2600, 2800, 3000, 3200, 3400, 3600, 3800, 4000, 4096, 4200, 4400, 4600, 4800, 5000, 5200, 5400, 5600, 5800, 6000], "t": 1.5} |
| {"event": "probe", "step": 0, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 5462.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.313854, "cps": 42.671875, "mean_p": 0.489807, "boundary_entropy": 0.68614, "bpb": 8.077753}, "tr": {"chunks": 13273.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.733193, "cps": 103.695312, "mean_p": 0.515287, "boundary_entropy": 0.686872, "bpb": 8.018145}, "fi": {"chunks": 9462.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.503004, "cps": 73.921875, "mean_p": 0.501981, "boundary_entropy": 0.6873, "bpb": 7.990695}}, "bpb": {"en": 8.07775, "tr": 8.01815, "fi": 7.9907}, "gini_cpc": 0.180355, "gini_cps": 0.184677, "bpic": 2.011987, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 2.3360975145821765, "fi": 1.6026677014027704}, "t": 2.7} |
| {"event": "train", "step": 0, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 5.57735, "bpb": 8.04642, "chunk_loss": 0.037766, "F": 0.27518, "G": 0.48384, "chunks": 281.78, "bytes_per_chunk": 3.634, "boundary_entropy": 0.69036, "router_noise": 0.0, "grad_norm": 13.5988, "lr": 1e-06, "router_trainable": true, "sec_per_step": 0.006, "alpha": 0.03, "t": 3.1} |
| {"event": "probe", "step": 1, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 5491.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.31552, "cps": 42.898438, "mean_p": 0.48996, "boundary_entropy": 0.68615, "bpb": 8.04278}, "tr": {"chunks": 13309.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.735182, "cps": 103.976562, "mean_p": 0.515387, "boundary_entropy": 0.686869, "bpb": 7.989188}, "fi": {"chunks": 9452.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.502472, "cps": 73.84375, "mean_p": 0.50196, "boundary_entropy": 0.687303, "bpb": 7.955177}}, "bpb": {"en": 8.04278, "tr": 7.98919, "fi": 7.95518}, "gini_cpc": 0.180131, "gini_cps": 0.184483, "bpic": 2.00807, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 2.3300623912180702, "fi": 1.5925185728066034}, "t": 3.6} |
| {"event": "probe", "step": 2, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 5553.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.319083, "cps": 43.382812, "mean_p": 0.490267, "boundary_entropy": 0.68617, "bpb": 7.972885}, "tr": {"chunks": 13369.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.738496, "cps": 104.445312, "mean_p": 0.515607, "boundary_entropy": 0.686864, "bpb": 7.932931}, "fi": {"chunks": 9442.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.50194, "cps": 73.765625, "mean_p": 0.501956, "boundary_entropy": 0.687308, "bpb": 7.888634}}, "bpb": {"en": 7.97288, "tr": 7.93293, "fi": 7.88863}, "gini_cpc": 0.179292, "gini_cps": 0.183707, "bpic": 2.000141, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 2.3144340955752534, "fi": 1.573071849720905}, "t": 4.5} |
| {"event": "probe", "step": 4, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 5791.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.332759, "cps": 45.242188, "mean_p": 0.491395, "boundary_entropy": 0.686245, "bpb": 7.744326}, "tr": {"chunks": 13548.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.748384, "cps": 105.84375, "mean_p": 0.516305, "boundary_entropy": 0.686849, "bpb": 7.746395}, "fi": {"chunks": 9419.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.500718, "cps": 73.585938, "mean_p": 0.501962, "boundary_entropy": 0.687329, "bpb": 7.67829}}, "bpb": {"en": 7.74433, "tr": 7.74639, "fi": 7.67829}, "gini_cpc": 0.175163, "gini_cps": 0.179822, "bpic": 1.972738, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 2.249029706096772, "fi": 1.5047469396311823}, "t": 5.6} |
| {"event": "probe", "step": 6, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 6192.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.355801, "cps": 48.375, "mean_p": 0.493366, "boundary_entropy": 0.686366, "bpb": 7.454989}, "tr": {"chunks": 13905.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.768105, "cps": 108.632812, "mean_p": 0.517449, "boundary_entropy": 0.68682, "bpb": 7.524562}, "fi": {"chunks": 9421.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.500824, "cps": 73.601562, "mean_p": 0.50209, "boundary_entropy": 0.687363, "bpb": 7.435345}}, "bpb": {"en": 7.45499, "tr": 7.52456, "fi": 7.43535}, "gini_cpc": 0.169179, "gini_cps": 0.174199, "bpic": 1.921946, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 2.1588059893708995, "fi": 1.4075968714083256}, "t": 6.6} |
| {"event": "probe", "step": 8, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 6815.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.391599, "cps": 53.242188, "mean_p": 0.495905, "boundary_entropy": 0.686506, "bpb": 7.195752}, "tr": {"chunks": 14297.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.789759, "cps": 111.695312, "mean_p": 0.518617, "boundary_entropy": 0.686792, "bpb": 7.330353}, "fi": {"chunks": 9341.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.496571, "cps": 72.976562, "mean_p": 0.501968, "boundary_entropy": 0.687403, "bpb": 7.225268}}, "bpb": {"en": 7.19575, "tr": 7.33035, "fi": 7.22527}, "gini_cpc": 0.158195, "gini_cps": 0.163793, "bpic": 1.862936, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 2.0167526012498223, "fi": 1.268059840574229}, "t": 7.5} |
| {"event": "probe", "step": 16, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 5969.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.342987, "cps": 46.632812, "mean_p": 0.49356, "boundary_entropy": 0.686674, "bpb": 6.583543}, "tr": {"chunks": 11624.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.642104, "cps": 90.8125, "mean_p": 0.508547, "boundary_entropy": 0.687471, "bpb": 6.864233}, "fi": {"chunks": 5795.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.308064, "cps": 45.273438, "mean_p": 0.491689, "boundary_entropy": 0.687305, "bpb": 6.714621}}, "bpb": {"en": 6.58354, "tr": 6.86423, "fi": 6.71462}, "gini_cpc": 0.172209, "gini_cps": 0.166154, "bpic": 2.425688, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.8720937404711577, "fi": 0.8981814846734533}, "t": 9.3} |
| {"event": "probe", "step": 32, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 5335.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.306556, "cps": 41.679688, "mean_p": 0.495475, "boundary_entropy": 0.687407, "bpb": 6.230488}, "tr": {"chunks": 6217.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.343424, "cps": 48.570312, "mean_p": 0.4951, "boundary_entropy": 0.687943, "bpb": 6.584795}, "fi": {"chunks": 3649.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.193982, "cps": 28.507812, "mean_p": 0.490598, "boundary_entropy": 0.687815, "bpb": 6.372852}}, "bpb": {"en": 6.23049, "tr": 6.5848, "fi": 6.37285}, "gini_cpc": 0.118048, "gini_cps": 0.112624, "bpic": 3.732123, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.1202630516691923, "fi": 0.6327784442062828}, "t": 11.6} |
| {"event": "train", "step": 50, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 4.14693, "bpb": 5.98275, "chunk_loss": 0.0343, "F": 0.2276, "G": 0.4894, "chunks": 233.06, "bytes_per_chunk": 4.394, "boundary_entropy": 0.69182, "router_noise": 0.0, "grad_norm": 2.4165, "lr": 5.1e-05, "router_trainable": true, "sec_per_step": 0.207, "alpha": 0.03, "t": 13.4} |
| {"event": "probe", "step": 64, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3903.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.224272, "cps": 30.492188, "mean_p": 0.479517, "boundary_entropy": 0.685159, "bpb": 5.523868}, "tr": {"chunks": 5106.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.282053, "cps": 39.890625, "mean_p": 0.488067, "boundary_entropy": 0.687012, "bpb": 5.923158}, "fi": {"chunks": 4234.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.225081, "cps": 33.078125, "mean_p": 0.483121, "boundary_entropy": 0.686487, "bpb": 5.685012}}, "bpb": {"en": 5.52387, "tr": 5.92316, "fi": 5.68501}, "gini_cpc": 0.052667, "gini_cps": 0.06056, "bpic": 4.283924, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.2576385116798772, "fi": 1.0036089813019686}, "t": 14.8} |
| {"event": "train", "step": 100, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 3.27229, "bpb": 4.72091, "chunk_loss": 0.031566, "F": 0.1904, "G": 0.46463, "chunks": 194.97, "bytes_per_chunk": 5.252, "boundary_entropy": 0.6777, "router_noise": 0.0, "grad_norm": 1.6311, "lr": 0.000101, "router_trainable": true, "sec_per_step": 0.093, "alpha": 0.03, "t": 18.1} |
| {"event": "probe", "step": 128, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3380.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.194219, "cps": 26.40625, "mean_p": 0.453546, "boundary_entropy": 0.634456, "bpb": 4.119808}, "tr": {"chunks": 4385.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.242225, "cps": 34.257812, "mean_p": 0.441152, "boundary_entropy": 0.657116, "bpb": 4.424718}, "fi": {"chunks": 3774.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.200627, "cps": 29.484375, "mean_p": 0.437454, "boundary_entropy": 0.656246, "bpb": 4.369061}}, "bpb": {"en": 4.11981, "tr": 4.42472, "fi": 4.36906}, "gini_cpc": 0.050236, "gini_cps": 0.058064, "bpic": 4.916544, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.2471723278400029, "fi": 1.0329931278406599}, "t": 20.5} |
| {"event": "train", "step": 150, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.74881, "bpb": 3.9657, "chunk_loss": 0.031074, "F": 0.18564, "G": 0.41996, "chunks": 190.09, "bytes_per_chunk": 5.387, "boundary_entropy": 0.61595, "router_noise": 0.0, "grad_norm": 2.3047, "lr": 0.000151, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 22.4} |
| {"event": "probe", "step": 200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3023.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.173706, "cps": 23.617188, "mean_p": 0.316132, "boundary_entropy": 0.459391, "bpb": 3.548814}, "tr": {"chunks": 3767.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.208087, "cps": 29.429688, "mean_p": 0.378089, "boundary_entropy": 0.59391, "bpb": 3.811135}, "fi": {"chunks": 3329.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.176971, "cps": 26.007812, "mean_p": 0.347562, "boundary_entropy": 0.564049, "bpb": 3.854529}}, "bpb": {"en": 3.54881, "tr": 3.81113, "fi": 3.85453}, "gini_cpc": 0.041021, "gini_cps": 0.049017, "bpic": 5.606483, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.1979288983871446, "fi": 1.0187975331963566}, "t": 26.6} |
| {"event": "eval", "step": 200, "phase": 1, "bpb": {"en": 3.55692, "tr": 3.82176, "fi": 3.85192}, "bytes_per_chunk": {"en": 5.557, "tr": 5.157, "fi": 5.724}, "bpb_mean": 3.74353, "frozen_inventory": false, "t": 27.0} |
| {"event": "train", "step": 200, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.55422, "bpb": 3.68496, "chunk_loss": 0.030494, "F": 0.17801, "G": 0.33488, "chunks": 182.28, "bytes_per_chunk": 5.618, "boundary_entropy": 0.50099, "router_noise": 0.0, "grad_norm": 1.1454, "lr": 0.000201, "router_trainable": true, "sec_per_step": 0.098, "alpha": 0.03, "t": 27.3} |
| {"event": "train", "step": 250, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.35807, "bpb": 3.40198, "chunk_loss": 0.029889, "F": 0.1644, "G": 0.33492, "chunks": 168.34, "bytes_per_chunk": 6.083, "boundary_entropy": 0.4757, "router_noise": 0.0, "grad_norm": 1.2654, "lr": 0.000251, "router_trainable": true, "sec_per_step": 0.078, "alpha": 0.03, "t": 31.1} |
| {"event": "probe", "step": 256, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3004.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.172614, "cps": 23.46875, "mean_p": 0.311106, "boundary_entropy": 0.424643, "bpb": 3.286165}, "tr": {"chunks": 3245.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.179252, "cps": 25.351562, "mean_p": 0.364237, "boundary_entropy": 0.532966, "bpb": 3.522638}, "fi": {"chunks": 2988.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.158843, "cps": 23.34375, "mean_p": 0.332974, "boundary_entropy": 0.520079, "bpb": 3.679127}}, "bpb": {"en": 3.28616, "tr": 3.52264, "fi": 3.67913}, "gini_cpc": 0.026641, "gini_cps": 0.018549, "bpic": 6.141821, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0384565777733024, "fi": 0.9202226138897539}, "t": 31.9} |
| {"event": "train", "step": 300, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.24333, "bpb": 3.23644, "chunk_loss": 0.030008, "F": 0.16763, "G": 0.30721, "chunks": 171.66, "bytes_per_chunk": 5.965, "boundary_entropy": 0.40661, "router_noise": 0.0, "grad_norm": 1.3043, "lr": 0.0003, "router_trainable": true, "sec_per_step": 0.085, "alpha": 0.03, "t": 35.4} |
| {"event": "train", "step": 350, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.14927, "bpb": 3.10074, "chunk_loss": 0.030374, "F": 0.18015, "G": 0.29908, "chunks": 184.47, "bytes_per_chunk": 5.551, "boundary_entropy": 0.38107, "router_noise": 0.0, "grad_norm": 2.0478, "lr": 0.0002999, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 38.9} |
| {"event": "probe", "step": 400, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2897.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.166466, "cps": 22.632812, "mean_p": 0.26354, "boundary_entropy": 0.337933, "bpb": 2.880454}, "tr": {"chunks": 3100.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.171242, "cps": 24.21875, "mean_p": 0.347175, "boundary_entropy": 0.491579, "bpb": 3.148684}, "fi": {"chunks": 2971.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.15794, "cps": 23.210938, "mean_p": 0.305378, "boundary_entropy": 0.456253, "bpb": 3.368991}}, "bpb": {"en": 2.88045, "tr": 3.14868, "fi": 3.36899}, "gini_cpc": 0.017893, "gini_cps": 0.015091, "bpic": 6.326048, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.028695327971298, "fi": 0.9487819051078138}, "t": 42.9} |
| {"event": "eval", "step": 400, "phase": 1, "bpb": {"en": 2.89535, "tr": 3.16008, "fi": 3.36352}, "bytes_per_chunk": {"en": 5.492, "tr": 5.949, "fi": 6.146}, "bpb_mean": 3.13965, "frozen_inventory": false, "t": 43.3} |
| {"event": "train", "step": 400, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.05872, "bpb": 2.97011, "chunk_loss": 0.030303, "F": 0.1799, "G": 0.28035, "chunks": 184.22, "bytes_per_chunk": 5.559, "boundary_entropy": 0.34915, "router_noise": 0.0, "grad_norm": 1.7219, "lr": 0.0002998, "router_trainable": true, "sec_per_step": 0.097, "alpha": 0.03, "t": 43.7} |
| {"event": "train", "step": 450, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.05959, "bpb": 2.97136, "chunk_loss": 0.030396, "F": 0.18274, "G": 0.29463, "chunks": 187.12, "bytes_per_chunk": 5.472, "boundary_entropy": 0.36922, "router_noise": 0.0, "grad_norm": 1.2113, "lr": 0.0002995, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 47.4} |
| {"event": "train", "step": 500, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.89956, "bpb": 2.74049, "chunk_loss": 0.030227, "F": 0.17859, "G": 0.25924, "chunks": 182.88, "bytes_per_chunk": 5.599, "boundary_entropy": 0.30594, "router_noise": 0.0, "grad_norm": 1.2578, "lr": 0.0002992, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 51.0} |
| {"event": "probe", "step": 512, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2907.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.16704, "cps": 22.710938, "mean_p": 0.231453, "boundary_entropy": 0.269515, "bpb": 2.683024}, "tr": {"chunks": 3009.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.166216, "cps": 23.507812, "mean_p": 0.311873, "boundary_entropy": 0.444482, "bpb": 2.990877}, "fi": {"chunks": 3052.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.162245, "cps": 23.84375, "mean_p": 0.286796, "boundary_entropy": 0.417461, "bpb": 3.203354}}, "bpb": {"en": 2.68302, "tr": 2.99088, "fi": 3.20335}, "gini_cpc": 0.006451, "gini_cps": 0.010779, "bpic": 6.326048, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9950633364054228, "fi": 0.9712962998013096}, "t": 52.2} |
| {"event": "train", "step": 550, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.9276, "bpb": 2.78093, "chunk_loss": 0.030327, "F": 0.18399, "G": 0.26661, "chunks": 188.41, "bytes_per_chunk": 5.435, "boundary_entropy": 0.31252, "router_noise": 0.0, "grad_norm": 1.2691, "lr": 0.0002987, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 55.5} |
| {"event": "probe", "step": 600, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2909.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.167155, "cps": 22.726562, "mean_p": 0.220702, "boundary_entropy": 0.246474, "bpb": 2.587491}, "tr": {"chunks": 3111.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.17185, "cps": 24.304688, "mean_p": 0.316814, "boundary_entropy": 0.447479, "bpb": 2.898778}, "fi": {"chunks": 3056.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.162458, "cps": 23.875, "mean_p": 0.292659, "boundary_entropy": 0.422265, "bpb": 3.123128}}, "bpb": {"en": 2.58749, "tr": 2.89878, "fi": 3.12313}, "gini_cpc": 0.012486, "gini_cps": 0.014838, "bpic": 6.250771, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0280869787786846, "fi": 0.9719006339755092}, "t": 59.4} |
| {"event": "eval", "step": 600, "phase": 1, "bpb": {"en": 2.60007, "tr": 2.90813, "fi": 3.1216}, "bytes_per_chunk": {"en": 5.415, "tr": 5.899, "fi": 5.906}, "bpb_mean": 2.8766, "frozen_inventory": false, "t": 59.9} |
| {"event": "train", "step": 600, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.88976, "bpb": 2.72634, "chunk_loss": 0.030198, "F": 0.18027, "G": 0.2455, "chunks": 184.59, "bytes_per_chunk": 5.547, "boundary_entropy": 0.27599, "router_noise": 0.0, "grad_norm": 1.1086, "lr": 0.0002982, "router_trainable": true, "sec_per_step": 0.095, "alpha": 0.03, "t": 60.2} |
| {"event": "train", "step": 650, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.80281, "bpb": 2.60091, "chunk_loss": 0.030152, "F": 0.17877, "G": 0.24, "chunks": 183.06, "bytes_per_chunk": 5.594, "boundary_entropy": 0.26189, "router_noise": 0.0, "grad_norm": 1.45, "lr": 0.0002975, "router_trainable": true, "sec_per_step": 0.073, "alpha": 0.03, "t": 63.9} |
| {"event": "train", "step": 700, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.85565, "bpb": 2.67713, "chunk_loss": 0.030177, "F": 0.17551, "G": 0.26123, "chunks": 179.72, "bytes_per_chunk": 5.698, "boundary_entropy": 0.29905, "router_noise": 0.0, "grad_norm": 1.0717, "lr": 0.0002967, "router_trainable": true, "sec_per_step": 0.078, "alpha": 0.03, "t": 67.8} |
| {"event": "train", "step": 750, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.78133, "bpb": 2.56991, "chunk_loss": 0.03029, "F": 0.18481, "G": 0.25215, "chunks": 189.25, "bytes_per_chunk": 5.411, "boundary_entropy": 0.27114, "router_noise": 0.0, "grad_norm": 1.0954, "lr": 0.0002959, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 71.5} |
| {"event": "probe", "step": 800, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2909.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.167155, "cps": 22.726562, "mean_p": 0.209635, "boundary_entropy": 0.208782, "bpb": 2.438821}, "tr": {"chunks": 3001.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.165774, "cps": 23.445312, "mean_p": 0.307818, "boundary_entropy": 0.428854, "bpb": 2.73673}, "fi": {"chunks": 2906.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.154484, "cps": 22.703125, "mean_p": 0.263465, "boundary_entropy": 0.373364, "bpb": 2.973448}}, "bpb": {"en": 2.43882, "tr": 2.73673, "fi": 2.97345}, "gini_cpc": 0.017331, "gini_cps": 0.007184, "bpic": 6.435118, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9917354623319935, "fi": 0.924196087150795}, "t": 75.3} |
| {"event": "eval", "step": 800, "phase": 1, "bpb": {"en": 2.45549, "tr": 2.74539, "fi": 2.9766}, "bytes_per_chunk": {"en": 5.418, "tr": 6.139, "fi": 6.225}, "bpb_mean": 2.72583, "frozen_inventory": false, "t": 75.8} |
| {"event": "train", "step": 800, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.76884, "bpb": 2.5519, "chunk_loss": 0.030279, "F": 0.18414, "G": 0.23648, "chunks": 188.56, "bytes_per_chunk": 5.431, "boundary_entropy": 0.23383, "router_noise": 0.0, "grad_norm": 1.2586, "lr": 0.0002949, "router_trainable": true, "sec_per_step": 0.095, "alpha": 0.03, "t": 76.2} |
| {"event": "train", "step": 850, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.70441, "bpb": 2.45895, "chunk_loss": 0.030265, "F": 0.18143, "G": 0.23863, "chunks": 185.78, "bytes_per_chunk": 5.512, "boundary_entropy": 0.24162, "router_noise": 0.0, "grad_norm": 1.0977, "lr": 0.0002938, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 79.9} |
| {"event": "train", "step": 900, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.697, "bpb": 2.44826, "chunk_loss": 0.030185, "F": 0.18137, "G": 0.23501, "chunks": 185.72, "bytes_per_chunk": 5.514, "boundary_entropy": 0.2292, "router_noise": 0.0, "grad_norm": 0.9494, "lr": 0.0002927, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 83.5} |
| {"event": "train", "step": 950, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.75207, "bpb": 2.5277, "chunk_loss": 0.030315, "F": 0.18576, "G": 0.25315, "chunks": 190.22, "bytes_per_chunk": 5.383, "boundary_entropy": 0.25955, "router_noise": 0.0, "grad_norm": 1.0712, "lr": 0.0002914, "router_trainable": true, "sec_per_step": 0.069, "alpha": 0.03, "t": 87.0} |
| {"event": "probe", "step": 1000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2912.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.167327, "cps": 22.75, "mean_p": 0.203456, "boundary_entropy": 0.188817, "bpb": 2.333475}, "tr": {"chunks": 3106.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.171574, "cps": 24.265625, "mean_p": 0.304032, "boundary_entropy": 0.414733, "bpb": 2.623327}, "fi": {"chunks": 3102.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.164904, "cps": 24.234375, "mean_p": 0.277368, "boundary_entropy": 0.380497, "bpb": 2.861066}}, "bpb": {"en": 2.33348, "tr": 2.62333, "fi": 2.86107}, "gini_cpc": 0.008827, "gini_cps": 0.014181, "bpic": 6.220614, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0253771838557508, "fi": 0.9855136855861165}, "t": 90.9} |
| {"event": "eval", "step": 1000, "phase": 1, "bpb": {"en": 2.35096, "tr": 2.63076, "fi": 2.8664}, "bytes_per_chunk": {"en": 5.418, "tr": 5.931, "fi": 5.904}, "bpb_mean": 2.61604, "frozen_inventory": false, "t": 91.3} |
| {"event": "train", "step": 1000, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.7309, "bpb": 2.49716, "chunk_loss": 0.03016, "F": 0.17902, "G": 0.23722, "chunks": 183.31, "bytes_per_chunk": 5.586, "boundary_entropy": 0.23968, "router_noise": 0.0, "grad_norm": 1.0286, "lr": 0.0002901, "router_trainable": true, "sec_per_step": 0.092, "alpha": 0.03, "t": 91.6} |
| {"event": "probe", "step": 1024, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2916.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.167557, "cps": 22.78125, "mean_p": 0.20159, "boundary_entropy": 0.182941, "bpb": 2.327219}, "tr": {"chunks": 3102.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.171353, "cps": 24.234375, "mean_p": 0.291278, "boundary_entropy": 0.400796, "bpb": 2.616993}, "fi": {"chunks": 3010.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.160013, "cps": 23.515625, "mean_p": 0.25978, "boundary_entropy": 0.359724, "bpb": 2.851714}}, "bpb": {"en": 2.32722, "tr": 2.61699, "fi": 2.85171}, "gini_cpc": 0.015153, "gini_cps": 0.013735, "bpic": 6.284005, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0226519306874313, "fi": 0.954973263389143}, "t": 93.5} |
| {"event": "train", "step": 1050, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.63354, "bpb": 2.3567, "chunk_loss": 0.03023, "F": 0.18253, "G": 0.22792, "chunks": 186.91, "bytes_per_chunk": 5.479, "boundary_entropy": 0.20631, "router_noise": 0.0, "grad_norm": 0.9439, "lr": 0.0002886, "router_trainable": true, "sec_per_step": 0.083, "alpha": 0.03, "t": 95.7} |
| {"event": "train", "step": 1100, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.74655, "bpb": 2.51974, "chunk_loss": 0.030117, "F": 0.17523, "G": 0.23548, "chunks": 179.44, "bytes_per_chunk": 5.707, "boundary_entropy": 0.23988, "router_noise": 0.0, "grad_norm": 1.0025, "lr": 0.0002871, "router_trainable": true, "sec_per_step": 0.073, "alpha": 0.03, "t": 99.4} |
| {"event": "train", "step": 1150, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.6419, "bpb": 2.36876, "chunk_loss": 0.030156, "F": 0.17917, "G": 0.23742, "chunks": 183.47, "bytes_per_chunk": 5.581, "boundary_entropy": 0.23606, "router_noise": 0.0, "grad_norm": 0.8974, "lr": 0.0002855, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 102.9} |
| {"event": "probe", "step": 1200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2923.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.16796, "cps": 22.835938, "mean_p": 0.198499, "boundary_entropy": 0.175067, "bpb": 2.273376}, "tr": {"chunks": 3055.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.168757, "cps": 23.867188, "mean_p": 0.287797, "boundary_entropy": 0.395837, "bpb": 2.55597}, "fi": {"chunks": 2965.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.157621, "cps": 23.164062, "mean_p": 0.26489, "boundary_entropy": 0.367936, "bpb": 2.737202}}, "bpb": {"en": 2.27338, "tr": 2.55597, "fi": 2.7372}, "gini_cpc": 0.015018, "gini_cps": 0.00984, "bpic": 6.343733, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.004745264529467, "fi": 0.938443475206573}, "t": 106.8} |
| {"event": "eval", "step": 1200, "phase": 1, "bpb": {"en": 2.28694, "tr": 2.56629, "fi": 2.74498}, "bytes_per_chunk": {"en": 5.406, "tr": 6.022, "fi": 6.205}, "bpb_mean": 2.53274, "frozen_inventory": false, "t": 107.4} |
| {"event": "train", "step": 1200, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.72546, "bpb": 2.48931, "chunk_loss": 0.030234, "F": 0.18121, "G": 0.24448, "chunks": 185.56, "bytes_per_chunk": 5.518, "boundary_entropy": 0.25124, "router_noise": 0.0, "grad_norm": 0.9105, "lr": 0.0002837, "router_trainable": true, "sec_per_step": 0.097, "alpha": 0.03, "t": 107.7} |
| {"event": "train", "step": 1250, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.64386, "bpb": 2.37159, "chunk_loss": 0.030182, "F": 0.18234, "G": 0.22425, "chunks": 186.72, "bytes_per_chunk": 5.484, "boundary_entropy": 0.20119, "router_noise": 0.0, "grad_norm": 0.9748, "lr": 0.0002819, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 111.4} |
| {"event": "train", "step": 1300, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.69513, "bpb": 2.44555, "chunk_loss": 0.030258, "F": 0.18378, "G": 0.23477, "chunks": 188.19, "bytes_per_chunk": 5.441, "boundary_entropy": 0.22024, "router_noise": 0.0, "grad_norm": 0.9053, "lr": 0.00028, "router_trainable": true, "sec_per_step": 0.073, "alpha": 0.03, "t": 115.1} |
| {"event": "train", "step": 1350, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.59882, "bpb": 2.30661, "chunk_loss": 0.030205, "F": 0.18161, "G": 0.21966, "chunks": 185.97, "bytes_per_chunk": 5.506, "boundary_entropy": 0.18709, "router_noise": 0.0, "grad_norm": 0.8371, "lr": 0.000278, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 118.6} |
| {"event": "probe", "step": 1400, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2915.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.1675, "cps": 22.773438, "mean_p": 0.194624, "boundary_entropy": 0.154872, "bpb": 2.218486}, "tr": {"chunks": 3102.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.171353, "cps": 24.234375, "mean_p": 0.263302, "boundary_entropy": 0.349636, "bpb": 2.490113}, "fi": {"chunks": 2956.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.157142, "cps": 23.09375, "mean_p": 0.2553, "boundary_entropy": 0.348329, "bpb": 2.678489}}, "bpb": {"en": 2.21849, "tr": 2.49011, "fi": 2.67849}, "gini_cpc": 0.019101, "gini_cps": 0.013894, "bpic": 6.322523, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0230027546773754, "fi": 0.9381625819643318}, "t": 122.7} |
| {"event": "eval", "step": 1400, "phase": 1, "bpb": {"en": 2.23192, "tr": 2.49813, "fi": 2.68865}, "bytes_per_chunk": {"en": 5.407, "tr": 6.025, "fi": 6.163}, "bpb_mean": 2.4729, "frozen_inventory": false, "t": 123.2} |
| {"event": "train", "step": 1400, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.62258, "bpb": 2.34088, "chunk_loss": 0.030167, "F": 0.18079, "G": 0.21503, "chunks": 185.12, "bytes_per_chunk": 5.531, "boundary_entropy": 0.17675, "router_noise": 0.0, "grad_norm": 0.9054, "lr": 0.0002759, "router_trainable": true, "sec_per_step": 0.098, "alpha": 0.03, "t": 123.5} |
| {"event": "train", "step": 1450, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.5963, "bpb": 2.30298, "chunk_loss": 0.030215, "F": 0.1813, "G": 0.22608, "chunks": 185.66, "bytes_per_chunk": 5.516, "boundary_entropy": 0.20104, "router_noise": 0.0, "grad_norm": 0.8712, "lr": 0.0002738, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 127.2} |
| {"event": "train", "step": 1500, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.62643, "bpb": 2.34644, "chunk_loss": 0.030176, "F": 0.1814, "G": 0.23037, "chunks": 185.75, "bytes_per_chunk": 5.513, "boundary_entropy": 0.21136, "router_noise": 0.0, "grad_norm": 0.7764, "lr": 0.0002715, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 130.8} |
| {"event": "train", "step": 1550, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.5726, "bpb": 2.26879, "chunk_loss": 0.030172, "F": 0.17923, "G": 0.22398, "chunks": 183.53, "bytes_per_chunk": 5.579, "boundary_entropy": 0.19518, "router_noise": 0.0, "grad_norm": 0.8205, "lr": 0.0002692, "router_trainable": true, "sec_per_step": 0.073, "alpha": 0.03, "t": 134.5} |
| {"event": "probe", "step": 1600, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2924.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168017, "cps": 22.84375, "mean_p": 0.19387, "boundary_entropy": 0.148471, "bpb": 2.187484}, "tr": {"chunks": 3111.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.17185, "cps": 24.304688, "mean_p": 0.275123, "boundary_entropy": 0.363976, "bpb": 2.418431}, "fi": {"chunks": 2970.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.157886, "cps": 23.203125, "mean_p": 0.260181, "boundary_entropy": 0.353038, "bpb": 2.619628}}, "bpb": {"en": 2.18748, "tr": 2.41843, "fi": 2.61963}, "gini_cpc": 0.018702, "gini_cps": 0.013844, "bpic": 6.300056, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.022812934769902, "fi": 0.9397045242541892}, "t": 138.5} |
| {"event": "eval", "step": 1600, "phase": 1, "bpb": {"en": 2.20082, "tr": 2.42539, "fi": 2.62796}, "bytes_per_chunk": {"en": 5.4, "tr": 5.971, "fi": 6.137}, "bpb_mean": 2.41806, "frozen_inventory": false, "t": 139.0} |
| {"event": "train", "step": 1600, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.59049, "bpb": 2.29459, "chunk_loss": 0.030092, "F": 0.17676, "G": 0.2221, "chunks": 181.0, "bytes_per_chunk": 5.657, "boundary_entropy": 0.19235, "router_noise": 0.0, "grad_norm": 0.9911, "lr": 0.0002668, "router_trainable": true, "sec_per_step": 0.098, "alpha": 0.03, "t": 139.4} |
| {"event": "train", "step": 1650, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.52192, "bpb": 2.19566, "chunk_loss": 0.030157, "F": 0.18051, "G": 0.22408, "chunks": 184.84, "bytes_per_chunk": 5.54, "boundary_entropy": 0.19058, "router_noise": 0.0, "grad_norm": 0.8095, "lr": 0.0002643, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 143.1} |
| {"event": "train", "step": 1700, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.5972, "bpb": 2.30428, "chunk_loss": 0.03021, "F": 0.17957, "G": 0.23442, "chunks": 183.88, "bytes_per_chunk": 5.569, "boundary_entropy": 0.22191, "router_noise": 0.0, "grad_norm": 0.7929, "lr": 0.0002618, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 146.6} |
| {"event": "train", "step": 1750, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.55526, "bpb": 2.24377, "chunk_loss": 0.030162, "F": 0.18274, "G": 0.21344, "chunks": 187.12, "bytes_per_chunk": 5.472, "boundary_entropy": 0.17037, "router_noise": 0.0, "grad_norm": 0.831, "lr": 0.0002591, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 150.3} |
| {"event": "probe", "step": 1800, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2920.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.167787, "cps": 22.8125, "mean_p": 0.189043, "boundary_entropy": 0.126475, "bpb": 2.154704}, "tr": {"chunks": 3186.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.175993, "cps": 24.890625, "mean_p": 0.264037, "boundary_entropy": 0.339265, "bpb": 2.384325}, "fi": {"chunks": 2984.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.158631, "cps": 23.3125, "mean_p": 0.229023, "boundary_entropy": 0.296453, "bpb": 2.57779}}, "bpb": {"en": 2.1547, "tr": 2.38433, "fi": 2.57779}, "gini_cpc": 0.023039, "gini_cps": 0.019509, "bpic": 6.241144, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0489058046081818, "fi": 0.9454274422645449}, "t": 154.1} |
| {"event": "eval", "step": 1800, "phase": 1, "bpb": {"en": 2.16496, "tr": 2.38941, "fi": 2.58859}, "bytes_per_chunk": {"en": 5.402, "tr": 5.826, "fi": 6.16}, "bpb_mean": 2.38099, "frozen_inventory": false, "t": 154.7} |
| {"event": "train", "step": 1800, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.52263, "bpb": 2.19669, "chunk_loss": 0.030093, "F": 0.17682, "G": 0.21938, "chunks": 181.06, "bytes_per_chunk": 5.656, "boundary_entropy": 0.18698, "router_noise": 0.0, "grad_norm": 0.8825, "lr": 0.0002564, "router_trainable": true, "sec_per_step": 0.095, "alpha": 0.03, "t": 155.1} |
| {"event": "train", "step": 1850, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.62537, "bpb": 2.34491, "chunk_loss": 0.030099, "F": 0.17599, "G": 0.22153, "chunks": 180.22, "bytes_per_chunk": 5.682, "boundary_entropy": 0.1979, "router_noise": 0.0, "grad_norm": 0.8121, "lr": 0.0002537, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 158.6} |
| {"event": "train", "step": 1900, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.5482, "bpb": 2.23357, "chunk_loss": 0.030126, "F": 0.17615, "G": 0.20587, "chunks": 180.38, "bytes_per_chunk": 5.677, "boundary_entropy": 0.15301, "router_noise": 0.0, "grad_norm": 0.7866, "lr": 0.0002508, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 162.2} |
| {"event": "train", "step": 1950, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.47954, "bpb": 2.13452, "chunk_loss": 0.030097, "F": 0.17749, "G": 0.20405, "chunks": 181.75, "bytes_per_chunk": 5.634, "boundary_entropy": 0.1466, "router_noise": 0.0, "grad_norm": 0.8091, "lr": 0.0002479, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 165.9} |
| {"event": "probe", "step": 2000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2922.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.167902, "cps": 22.828125, "mean_p": 0.185468, "boundary_entropy": 0.11598, "bpb": 2.117606}, "tr": {"chunks": 3087.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.170524, "cps": 24.117188, "mean_p": 0.255146, "boundary_entropy": 0.328546, "bpb": 2.337182}, "fi": {"chunks": 3016.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.160332, "cps": 23.5625, "mean_p": 0.24494, "boundary_entropy": 0.323521, "bpb": 2.520102}}, "bpb": {"en": 2.11761, "tr": 2.33718, "fi": 2.5201}, "gini_cpc": 0.013624, "gini_cps": 0.012188, "bpic": 6.286094, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0156170582637953, "fi": 0.9549120250209667}, "t": 169.8} |
| {"event": "eval", "step": 2000, "phase": 1, "bpb": {"en": 2.13017, "tr": 2.34397, "fi": 2.52965}, "bytes_per_chunk": {"en": 5.404, "tr": 5.954, "fi": 6.081}, "bpb_mean": 2.33459, "frozen_inventory": false, "t": 170.3} |
| {"event": "train", "step": 2000, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.50665, "bpb": 2.17364, "chunk_loss": 0.030054, "F": 0.17468, "G": 0.20773, "chunks": 178.88, "bytes_per_chunk": 5.725, "boundary_entropy": 0.15899, "router_noise": 0.0, "grad_norm": 0.7474, "lr": 0.000245, "router_trainable": true, "sec_per_step": 0.095, "alpha": 0.03, "t": 170.6} |
| {"event": "probe", "step": 2048, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2920.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.167787, "cps": 22.8125, "mean_p": 0.185642, "boundary_entropy": 0.115878, "bpb": 2.105592}, "tr": {"chunks": 3081.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.170193, "cps": 24.070312, "mean_p": 0.236255, "boundary_entropy": 0.295271, "bpb": 2.337884}, "fi": {"chunks": 2886.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.153421, "cps": 22.546875, "mean_p": 0.224533, "boundary_entropy": 0.298201, "bpb": 2.516264}}, "bpb": {"en": 2.10559, "tr": 2.33788, "fi": 2.51626}, "gini_cpc": 0.022754, "gini_cps": 0.014628, "bpic": 6.383707, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.014337345887573, "fi": 0.9143778814931223}, "t": 174.3} |
| {"event": "train", "step": 2050, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.51495, "bpb": 2.18562, "chunk_loss": 0.030126, "F": 0.17728, "G": 0.21391, "chunks": 181.53, "bytes_per_chunk": 5.641, "boundary_entropy": 0.17456, "router_noise": 0.0, "grad_norm": 0.7627, "lr": 0.0002419, "router_trainable": true, "sec_per_step": 0.082, "alpha": 0.03, "t": 174.7} |
| {"event": "train", "step": 2100, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.55993, "bpb": 2.25051, "chunk_loss": 0.030144, "F": 0.17899, "G": 0.21785, "chunks": 183.28, "bytes_per_chunk": 5.587, "boundary_entropy": 0.17346, "router_noise": 0.0, "grad_norm": 0.7574, "lr": 0.0002388, "router_trainable": true, "sec_per_step": 0.073, "alpha": 0.03, "t": 178.4} |
| {"event": "train", "step": 2150, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.55611, "bpb": 2.24499, "chunk_loss": 0.030208, "F": 0.18414, "G": 0.22574, "chunks": 188.56, "bytes_per_chunk": 5.431, "boundary_entropy": 0.1889, "router_noise": 0.0, "grad_norm": 0.7812, "lr": 0.0002357, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 182.0} |
| {"event": "probe", "step": 2200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2924.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168017, "cps": 22.84375, "mean_p": 0.183201, "boundary_entropy": 0.106325, "bpb": 2.084853}, "tr": {"chunks": 3135.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.173176, "cps": 24.492188, "mean_p": 0.239042, "boundary_entropy": 0.293272, "bpb": 2.294915}, "fi": {"chunks": 2964.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.157567, "cps": 23.15625, "mean_p": 0.229146, "boundary_entropy": 0.297287, "bpb": 2.476184}}, "bpb": {"en": 2.08485, "tr": 2.29491, "fi": 2.47618}, "gini_cpc": 0.020863, "gini_cps": 0.01559, "bpic": 6.287488, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.030703487786449, "fi": 0.937806131275898}, "t": 185.9} |
| {"event": "eval", "step": 2200, "phase": 1, "bpb": {"en": 2.0984, "tr": 2.30162, "fi": 2.48654}, "bytes_per_chunk": {"en": 5.403, "tr": 5.909, "fi": 6.202}, "bpb_mean": 2.29552, "frozen_inventory": false, "t": 186.4} |
| {"event": "train", "step": 2200, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.49365, "bpb": 2.15488, "chunk_loss": 0.030164, "F": 0.18015, "G": 0.21268, "chunks": 184.47, "bytes_per_chunk": 5.551, "boundary_entropy": 0.1588, "router_noise": 0.0, "grad_norm": 0.7747, "lr": 0.0002325, "router_trainable": true, "sec_per_step": 0.097, "alpha": 0.03, "t": 186.8} |
| {"event": "train", "step": 2250, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.52635, "bpb": 2.20206, "chunk_loss": 0.030098, "F": 0.17572, "G": 0.20818, "chunks": 179.94, "bytes_per_chunk": 5.691, "boundary_entropy": 0.15812, "router_noise": 0.0, "grad_norm": 0.7771, "lr": 0.0002293, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 190.3} |
| {"event": "train", "step": 2300, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.48343, "bpb": 2.14014, "chunk_loss": 0.03008, "F": 0.17712, "G": 0.20724, "chunks": 181.38, "bytes_per_chunk": 5.646, "boundary_entropy": 0.15388, "router_noise": 0.0, "grad_norm": 0.8828, "lr": 0.000226, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 193.8} |
| {"event": "train", "step": 2350, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.50972, "bpb": 2.17807, "chunk_loss": 0.030164, "F": 0.18115, "G": 0.21617, "chunks": 185.5, "bytes_per_chunk": 5.52, "boundary_entropy": 0.16311, "router_noise": 0.0, "grad_norm": 0.7486, "lr": 0.0002226, "router_trainable": true, "sec_per_step": 0.065, "alpha": 0.03, "t": 197.1} |
| {"event": "probe", "step": 2400, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2918.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.167672, "cps": 22.796875, "mean_p": 0.183646, "boundary_entropy": 0.105576, "bpb": 2.057529}, "tr": {"chunks": 3158.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.174446, "cps": 24.671875, "mean_p": 0.243238, "boundary_entropy": 0.297905, "bpb": 2.273346}, "fi": {"chunks": 2891.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.153687, "cps": 22.585938, "mean_p": 0.21824, "boundary_entropy": 0.281175, "bpb": 2.450693}}, "bpb": {"en": 2.05753, "tr": 2.27335, "fi": 2.45069}, "gini_cpc": 0.027914, "gini_cps": 0.019851, "bpic": 6.326754, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0404001517930468, "fi": 0.9165898440199978}, "t": 201.0} |
| {"event": "eval", "step": 2400, "phase": 1, "bpb": {"en": 2.07088, "tr": 2.28148, "fi": 2.46039}, "bytes_per_chunk": {"en": 5.403, "tr": 5.9, "fi": 6.275}, "bpb_mean": 2.27092, "frozen_inventory": false, "t": 201.5} |
| {"event": "phase_transition", "step": 2400, "phase": 2, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 201.8} |
| {"event": "train", "step": 2400, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.49157, "bpb": 2.15188, "chunk_loss": 0.03012, "F": 0.17749, "G": 0.21028, "chunks": 181.75, "bytes_per_chunk": 5.634, "boundary_entropy": 0.15833, "router_noise": 0.0, "grad_norm": 0.723, "lr": 0.0002192, "router_trainable": true, "sec_per_step": 0.096, "alpha": 0.03, "t": 201.9} |
| {"event": "train", "step": 2450, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.5127, "bpb": 2.18237, "chunk_loss": 0.030215, "F": 0.1853, "G": 0.21992, "chunks": 189.75, "bytes_per_chunk": 5.397, "boundary_entropy": 0.15783, "router_noise": 0.0, "grad_norm": 0.7693, "lr": 0.0002158, "router_trainable": true, "sec_per_step": 0.073, "alpha": 0.03, "t": 205.5} |
| {"event": "train", "step": 2500, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.44067, "bpb": 2.07844, "chunk_loss": 0.030138, "F": 0.17981, "G": 0.21393, "chunks": 184.12, "bytes_per_chunk": 5.561, "boundary_entropy": 0.15728, "router_noise": 0.0, "grad_norm": 0.7382, "lr": 0.0002123, "router_trainable": true, "sec_per_step": 0.069, "alpha": 0.03, "t": 209.0} |
| {"event": "train", "step": 2550, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.43603, "bpb": 2.07176, "chunk_loss": 0.030223, "F": 0.1871, "G": 0.21259, "chunks": 191.59, "bytes_per_chunk": 5.345, "boundary_entropy": 0.13414, "router_noise": 0.0, "grad_norm": 0.7146, "lr": 0.0002088, "router_trainable": true, "sec_per_step": 0.073, "alpha": 0.03, "t": 212.6} |
| {"event": "probe", "step": 2600, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2914.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.167442, "cps": 22.765625, "mean_p": 0.183256, "boundary_entropy": 0.10231, "bpb": 2.038755}, "tr": {"chunks": 3175.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.175385, "cps": 24.804688, "mean_p": 0.251561, "boundary_entropy": 0.305958, "bpb": 2.243672}, "fi": {"chunks": 2983.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.158577, "cps": 23.304688, "mean_p": 0.233159, "boundary_entropy": 0.302326, "bpb": 2.421676}}, "bpb": {"en": 2.03876, "tr": 2.24367, "fi": 2.42168}, "gini_cpc": 0.022348, "gini_cps": 0.01918, "bpic": 6.253527, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0474366140430849, "fi": 0.9470566167585396}, "t": 216.5} |
| {"event": "eval", "step": 2600, "phase": 2, "bpb": {"en": 2.05208, "tr": 2.25274, "fi": 2.43189}, "bytes_per_chunk": {"en": 5.4, "tr": 5.877, "fi": 6.113}, "bpb_mean": 2.24557, "frozen_inventory": false, "t": 217.1} |
| {"event": "train", "step": 2600, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.52526, "bpb": 2.20049, "chunk_loss": 0.030108, "F": 0.17899, "G": 0.2104, "chunks": 183.28, "bytes_per_chunk": 5.587, "boundary_entropy": 0.15072, "router_noise": 0.0, "grad_norm": 0.763, "lr": 0.0002053, "router_trainable": true, "sec_per_step": 0.095, "alpha": 0.03, "t": 217.4} |
| {"event": "train", "step": 2650, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.53113, "bpb": 2.20896, "chunk_loss": 0.030057, "F": 0.17523, "G": 0.20508, "chunks": 179.44, "bytes_per_chunk": 5.707, "boundary_entropy": 0.14825, "router_noise": 0.0, "grad_norm": 0.823, "lr": 0.0002017, "router_trainable": true, "sec_per_step": 0.069, "alpha": 0.03, "t": 220.8} |
| {"event": "train", "step": 2700, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.44646, "bpb": 2.0868, "chunk_loss": 0.030082, "F": 0.17609, "G": 0.20574, "chunks": 180.31, "bytes_per_chunk": 5.679, "boundary_entropy": 0.14125, "router_noise": 0.0, "grad_norm": 0.7449, "lr": 0.0001981, "router_trainable": true, "sec_per_step": 0.069, "alpha": 0.03, "t": 224.3} |
| {"event": "train", "step": 2750, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.48423, "bpb": 2.1413, "chunk_loss": 0.030157, "F": 0.18106, "G": 0.21378, "chunks": 185.41, "bytes_per_chunk": 5.523, "boundary_entropy": 0.15649, "router_noise": 0.0, "grad_norm": 0.786, "lr": 0.0001945, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 227.9} |
| {"event": "probe", "step": 2800, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2918.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.167672, "cps": 22.796875, "mean_p": 0.183156, "boundary_entropy": 0.098373, "bpb": 2.017438}, "tr": {"chunks": 3193.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.17638, "cps": 24.945312, "mean_p": 0.229754, "boundary_entropy": 0.2656, "bpb": 2.224458}, "fi": {"chunks": 2908.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.15459, "cps": 22.71875, "mean_p": 0.207969, "boundary_entropy": 0.257753, "bpb": 2.390413}}, "bpb": {"en": 2.01744, "tr": 2.22446, "fi": 2.39041}, "gini_cpc": 0.029131, "gini_cps": 0.021067, "bpic": 6.290276, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0519308691181757, "fi": 0.9219796839882923}, "t": 231.6} |
| {"event": "eval", "step": 2800, "phase": 2, "bpb": {"en": 2.03436, "tr": 2.22967, "fi": 2.39748}, "bytes_per_chunk": {"en": 5.397, "tr": 5.824, "fi": 6.23}, "bpb_mean": 2.2205, "frozen_inventory": false, "t": 232.1} |
| {"event": "train", "step": 2800, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.4408, "bpb": 2.07863, "chunk_loss": 0.030103, "F": 0.17844, "G": 0.20679, "chunks": 182.72, "bytes_per_chunk": 5.604, "boundary_entropy": 0.14191, "router_noise": 0.0, "grad_norm": 0.7169, "lr": 0.0001909, "router_trainable": true, "sec_per_step": 0.092, "alpha": 0.03, "t": 232.5} |
| {"event": "train", "step": 2850, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.45845, "bpb": 2.1041, "chunk_loss": 0.030196, "F": 0.18219, "G": 0.2147, "chunks": 186.56, "bytes_per_chunk": 5.489, "boundary_entropy": 0.15353, "router_noise": 0.0, "grad_norm": 0.8502, "lr": 0.0001872, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 236.1} |
| {"event": "train", "step": 2900, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.36215, "bpb": 1.96517, "chunk_loss": 0.030172, "F": 0.18494, "G": 0.20212, "chunks": 189.38, "bytes_per_chunk": 5.407, "boundary_entropy": 0.10859, "router_noise": 0.0, "grad_norm": 0.6788, "lr": 0.0001835, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 239.7} |
| {"event": "train", "step": 2950, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.52119, "bpb": 2.19462, "chunk_loss": 0.030186, "F": 0.18356, "G": 0.20954, "chunks": 187.97, "bytes_per_chunk": 5.448, "boundary_entropy": 0.13839, "router_noise": 0.0, "grad_norm": 0.7437, "lr": 0.0001799, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 243.3} |
| {"event": "probe", "step": 3000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2925.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168074, "cps": 22.851562, "mean_p": 0.182529, "boundary_entropy": 0.095061, "bpb": 2.001481}, "tr": {"chunks": 3110.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.171795, "cps": 24.296875, "mean_p": 0.228943, "boundary_entropy": 0.267191, "bpb": 2.196194}, "fi": {"chunks": 2923.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.155388, "cps": 22.835938, "mean_p": 0.220522, "boundary_entropy": 0.281975, "bpb": 2.362998}}, "bpb": {"en": 2.00148, "tr": 2.19619, "fi": 2.363}, "gini_cpc": 0.022085, "gini_cps": 0.013917, "bpic": 6.33311, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0221345944927673, "fi": 0.9245175967689392}, "t": 247.2} |
| {"event": "eval", "step": 3000, "phase": 2, "bpb": {"en": 2.01618, "tr": 2.20344, "fi": 2.37383}, "bytes_per_chunk": {"en": 5.401, "tr": 5.918, "fi": 6.192}, "bpb_mean": 2.19782, "frozen_inventory": false, "t": 247.7} |
| {"event": "train", "step": 3000, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.43211, "bpb": 2.0661, "chunk_loss": 0.030199, "F": 0.18625, "G": 0.20806, "chunks": 190.72, "bytes_per_chunk": 5.369, "boundary_entropy": 0.1244, "router_noise": 0.0, "grad_norm": 0.7019, "lr": 0.0001761, "router_trainable": true, "sec_per_step": 0.096, "alpha": 0.03, "t": 248.1} |
| {"event": "train", "step": 3050, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.40861, "bpb": 2.03219, "chunk_loss": 0.030101, "F": 0.17624, "G": 0.2025, "chunks": 180.47, "bytes_per_chunk": 5.674, "boundary_entropy": 0.13255, "router_noise": 0.0, "grad_norm": 0.7044, "lr": 0.0001724, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 251.6} |
| {"event": "train", "step": 3100, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.41576, "bpb": 2.04252, "chunk_loss": 0.030162, "F": 0.18347, "G": 0.20379, "chunks": 187.88, "bytes_per_chunk": 5.45, "boundary_entropy": 0.11352, "router_noise": 0.0, "grad_norm": 0.757, "lr": 0.0001687, "router_trainable": true, "sec_per_step": 0.069, "alpha": 0.03, "t": 255.1} |
| {"event": "train", "step": 3150, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.34781, "bpb": 1.94448, "chunk_loss": 0.030125, "F": 0.17969, "G": 0.19861, "chunks": 184.0, "bytes_per_chunk": 5.565, "boundary_entropy": 0.10639, "router_noise": 0.0, "grad_norm": 0.6729, "lr": 0.000165, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 258.7} |
| {"event": "probe", "step": 3200, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2920.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.167787, "cps": 22.8125, "mean_p": 0.181817, "boundary_entropy": 0.08939, "bpb": 1.989049}, "tr": {"chunks": 3104.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.171463, "cps": 24.25, "mean_p": 0.233511, "boundary_entropy": 0.270815, "bpb": 2.181739}, "fi": {"chunks": 2907.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.154537, "cps": 22.710938, "mean_p": 0.21675, "boundary_entropy": 0.273681, "bpb": 2.335609}}, "bpb": {"en": 1.98905, "tr": 2.18174, "fi": 2.33561}, "gini_cpc": 0.022852, "gini_cps": 0.014705, "bpic": 6.352256, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0219094844644685, "fi": 0.9210313588012842}, "t": 262.4} |
| {"event": "eval", "step": 3200, "phase": 2, "bpb": {"en": 2.0032, "tr": 2.18762, "fi": 2.34701}, "bytes_per_chunk": {"en": 5.398, "tr": 5.926, "fi": 6.273}, "bpb_mean": 2.17928, "frozen_inventory": false, "t": 262.8} |
| {"event": "train", "step": 3200, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.3973, "bpb": 2.01587, "chunk_loss": 0.030131, "F": 0.18021, "G": 0.20712, "chunks": 184.53, "bytes_per_chunk": 5.549, "boundary_entropy": 0.13208, "router_noise": 0.0, "grad_norm": 0.7093, "lr": 0.0001613, "router_trainable": true, "sec_per_step": 0.091, "alpha": 0.03, "t": 263.2} |
| {"event": "train", "step": 3250, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.34207, "bpb": 1.93619, "chunk_loss": 0.030111, "F": 0.17752, "G": 0.20594, "chunks": 181.78, "bytes_per_chunk": 5.633, "boundary_entropy": 0.13791, "router_noise": 0.0, "grad_norm": 0.6911, "lr": 0.0001576, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 266.8} |
| {"event": "train", "step": 3300, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.41646, "bpb": 2.04352, "chunk_loss": 0.030103, "F": 0.17798, "G": 0.20523, "chunks": 182.25, "bytes_per_chunk": 5.619, "boundary_entropy": 0.13568, "router_noise": 0.0, "grad_norm": 0.7761, "lr": 0.0001539, "router_trainable": true, "sec_per_step": 0.069, "alpha": 0.03, "t": 270.3} |
| {"event": "train", "step": 3350, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.5935, "bpb": 2.29893, "chunk_loss": 0.03019, "F": 0.18134, "G": 0.21564, "chunks": 185.69, "bytes_per_chunk": 5.515, "boundary_entropy": 0.16063, "router_noise": 0.0, "grad_norm": 0.883, "lr": 0.0001501, "router_trainable": true, "sec_per_step": 0.073, "alpha": 0.03, "t": 273.9} |
| {"event": "probe", "step": 3400, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2925.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168074, "cps": 22.851562, "mean_p": 0.180974, "boundary_entropy": 0.083084, "bpb": 1.969966}, "tr": {"chunks": 3172.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.17522, "cps": 24.78125, "mean_p": 0.221027, "boundary_entropy": 0.242841, "bpb": 2.154641}, "fi": {"chunks": 2934.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.155973, "cps": 22.921875, "mean_p": 0.209414, "boundary_entropy": 0.255465, "bpb": 2.319616}}, "bpb": {"en": 1.96997, "tr": 2.15464, "fi": 2.31962}, "gini_cpc": 0.0257, "gini_cps": 0.018233, "bpic": 6.281918, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0425115542543595, "fi": 0.9279967940198658}, "t": 277.8} |
| {"event": "eval", "step": 3400, "phase": 2, "bpb": {"en": 1.98416, "tr": 2.16487, "fi": 2.32878}, "bytes_per_chunk": {"en": 5.398, "tr": 5.86, "fi": 6.118}, "bpb_mean": 2.15927, "frozen_inventory": false, "t": 278.3} |
| {"event": "train", "step": 3400, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.43249, "bpb": 2.06665, "chunk_loss": 0.030185, "F": 0.18466, "G": 0.20857, "chunks": 189.09, "bytes_per_chunk": 5.415, "boundary_entropy": 0.12548, "router_noise": 0.0, "grad_norm": 0.6832, "lr": 0.0001465, "router_trainable": true, "sec_per_step": 0.094, "alpha": 0.03, "t": 278.7} |
| {"event": "train", "step": 3450, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.41686, "bpb": 2.0441, "chunk_loss": 0.030163, "F": 0.18481, "G": 0.20695, "chunks": 189.25, "bytes_per_chunk": 5.411, "boundary_entropy": 0.11839, "router_noise": 0.0, "grad_norm": 0.7508, "lr": 0.0001428, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 282.1} |
| {"event": "train", "step": 3500, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.44695, "bpb": 2.08751, "chunk_loss": 0.030186, "F": 0.18213, "G": 0.20773, "chunks": 186.5, "bytes_per_chunk": 5.491, "boundary_entropy": 0.12825, "router_noise": 0.0, "grad_norm": 0.7631, "lr": 0.0001391, "router_trainable": true, "sec_per_step": 0.069, "alpha": 0.03, "t": 285.6} |
| {"event": "train", "step": 3550, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.47756, "bpb": 2.13167, "chunk_loss": 0.030143, "F": 0.17953, "G": 0.21209, "chunks": 183.84, "bytes_per_chunk": 5.57, "boundary_entropy": 0.153, "router_noise": 0.0, "grad_norm": 0.7148, "lr": 0.0001355, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 289.1} |
| {"event": "probe", "step": 3600, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2925.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168074, "cps": 22.851562, "mean_p": 0.180907, "boundary_entropy": 0.082355, "bpb": 1.952799}, "tr": {"chunks": 3169.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.175054, "cps": 24.757812, "mean_p": 0.228353, "boundary_entropy": 0.255247, "bpb": 2.14383}, "fi": {"chunks": 2925.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.155494, "cps": 22.851562, "mean_p": 0.214723, "boundary_entropy": 0.266503, "bpb": 2.302638}}, "bpb": {"en": 1.9528, "tr": 2.14383, "fi": 2.30264}, "gini_cpc": 0.026152, "gini_cps": 0.018036, "bpic": 6.290276, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0415255723304113, "fi": 0.9251501780872894}, "t": 292.8} |
| {"event": "eval", "step": 3600, "phase": 2, "bpb": {"en": 1.96934, "tr": 2.15077, "fi": 2.31242}, "bytes_per_chunk": {"en": 5.394, "tr": 5.863, "fi": 6.186}, "bpb_mean": 2.14418, "frozen_inventory": false, "t": 293.3} |
| {"event": "train", "step": 3600, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.4696, "bpb": 2.12019, "chunk_loss": 0.030177, "F": 0.18167, "G": 0.20772, "chunks": 186.03, "bytes_per_chunk": 5.504, "boundary_entropy": 0.12819, "router_noise": 0.0, "grad_norm": 0.773, "lr": 0.0001319, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 293.7} |
| {"event": "train", "step": 3650, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.39899, "bpb": 2.01832, "chunk_loss": 0.0301, "F": 0.17865, "G": 0.20171, "chunks": 182.94, "bytes_per_chunk": 5.598, "boundary_entropy": 0.12013, "router_noise": 0.0, "grad_norm": 0.7625, "lr": 0.0001283, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 297.2} |
| {"event": "train", "step": 3700, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.37692, "bpb": 1.98648, "chunk_loss": 0.030149, "F": 0.18362, "G": 0.20311, "chunks": 188.03, "bytes_per_chunk": 5.446, "boundary_entropy": 0.10585, "router_noise": 0.0, "grad_norm": 0.6852, "lr": 0.0001247, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 300.7} |
| {"event": "train", "step": 3750, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.3864, "bpb": 2.00015, "chunk_loss": 0.030122, "F": 0.17749, "G": 0.19991, "chunks": 181.75, "bytes_per_chunk": 5.634, "boundary_entropy": 0.1173, "router_noise": 0.0, "grad_norm": 0.6862, "lr": 0.0001212, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 304.2} |
| {"event": "probe", "step": 3800, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2926.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168132, "cps": 22.859375, "mean_p": 0.179457, "boundary_entropy": 0.076567, "bpb": 1.94206}, "tr": {"chunks": 3164.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.174778, "cps": 24.71875, "mean_p": 0.218894, "boundary_entropy": 0.236721, "bpb": 2.126526}, "fi": {"chunks": 3010.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.160013, "cps": 23.515625, "mean_p": 0.215523, "boundary_entropy": 0.260852, "bpb": 2.278649}}, "bpb": {"en": 1.94206, "tr": 2.12653, "fi": 2.27865}, "gini_cpc": 0.019572, "gini_cps": 0.017436, "bpic": 6.234286, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.039526875320303, "fi": 0.9517095133433837}, "t": 308.1} |
| {"event": "eval", "step": 3800, "phase": 2, "bpb": {"en": 1.958, "tr": 2.13356, "fi": 2.28862}, "bytes_per_chunk": {"en": 5.398, "tr": 5.878, "fi": 6.072}, "bpb_mean": 2.12673, "frozen_inventory": false, "t": 308.5} |
| {"event": "train", "step": 3800, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.35052, "bpb": 1.94839, "chunk_loss": 0.030104, "F": 0.17926, "G": 0.20136, "chunks": 183.56, "bytes_per_chunk": 5.578, "boundary_entropy": 0.11699, "router_noise": 0.0, "grad_norm": 0.7998, "lr": 0.0001177, "router_trainable": true, "sec_per_step": 0.094, "alpha": 0.03, "t": 308.9} |
| {"event": "train", "step": 3850, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.37503, "bpb": 1.98375, "chunk_loss": 0.030208, "F": 0.18655, "G": 0.20233, "chunks": 191.03, "bytes_per_chunk": 5.36, "boundary_entropy": 0.09307, "router_noise": 0.0, "grad_norm": 0.7133, "lr": 0.0001142, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 312.4} |
| {"event": "train", "step": 3900, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.40558, "bpb": 2.02783, "chunk_loss": 0.030199, "F": 0.18539, "G": 0.20705, "chunks": 189.84, "bytes_per_chunk": 5.394, "boundary_entropy": 0.11586, "router_noise": 0.0, "grad_norm": 0.7274, "lr": 0.0001108, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 316.0} |
| {"event": "train", "step": 3950, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.39079, "bpb": 2.00649, "chunk_loss": 0.030158, "F": 0.18405, "G": 0.20269, "chunks": 188.47, "bytes_per_chunk": 5.433, "boundary_entropy": 0.10419, "router_noise": 0.0, "grad_norm": 0.7504, "lr": 0.0001074, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 319.6} |
| {"event": "probe", "step": 4000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2926.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168132, "cps": 22.859375, "mean_p": 0.179704, "boundary_entropy": 0.075731, "bpb": 1.930938}, "tr": {"chunks": 3119.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.172292, "cps": 24.367188, "mean_p": 0.222062, "boundary_entropy": 0.241287, "bpb": 2.106862}, "fi": {"chunks": 2959.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.157302, "cps": 23.117188, "mean_p": 0.212098, "boundary_entropy": 0.255234, "bpb": 2.255977}}, "bpb": {"en": 1.93094, "tr": 2.10686, "fi": 2.25598}, "gini_cpc": 0.020078, "gini_cps": 0.01429, "bpic": 6.300755, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0247422010505767, "fi": 0.9355842026521834}, "t": 323.6} |
| {"event": "eval", "step": 4000, "phase": 2, "bpb": {"en": 1.94565, "tr": 2.11466, "fi": 2.26779}, "bytes_per_chunk": {"en": 5.399, "tr": 5.925, "fi": 6.115}, "bpb_mean": 2.10937, "frozen_inventory": false, "t": 324.1} |
| {"event": "train", "step": 4000, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.42516, "bpb": 2.05608, "chunk_loss": 0.030112, "F": 0.17679, "G": 0.2009, "chunks": 181.03, "bytes_per_chunk": 5.656, "boundary_entropy": 0.12022, "router_noise": 0.0, "grad_norm": 0.7011, "lr": 0.000104, "router_trainable": true, "sec_per_step": 0.097, "alpha": 0.03, "t": 324.4} |
| {"event": "train", "step": 4050, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.44117, "bpb": 2.07917, "chunk_loss": 0.030127, "F": 0.181, "G": 0.20473, "chunks": 185.34, "bytes_per_chunk": 5.525, "boundary_entropy": 0.11652, "router_noise": 0.0, "grad_norm": 0.8795, "lr": 0.0001007, "router_trainable": true, "sec_per_step": 0.076, "alpha": 0.03, "t": 328.2} |
| {"event": "probe", "step": 4096, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2925.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168074, "cps": 22.851562, "mean_p": 0.178249, "boundary_entropy": 0.071797, "bpb": 1.92345}, "tr": {"chunks": 3160.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.174557, "cps": 24.6875, "mean_p": 0.218852, "boundary_entropy": 0.231254, "bpb": 2.100688}, "fi": {"chunks": 2922.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.155335, "cps": 22.828125, "mean_p": 0.202915, "boundary_entropy": 0.240923, "bpb": 2.25701}}, "bpb": {"en": 1.92345, "tr": 2.10069, "fi": 2.25701}, "gini_cpc": 0.025734, "gini_cps": 0.017616, "bpic": 6.298657, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0385676265585673, "fi": 0.9242013061097639}, "t": 332.0} |
| {"event": "train", "step": 4100, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.47379, "bpb": 2.12623, "chunk_loss": 0.030119, "F": 0.1792, "G": 0.20607, "chunks": 183.5, "bytes_per_chunk": 5.58, "boundary_entropy": 0.13071, "router_noise": 0.0, "grad_norm": 0.7767, "lr": 9.75e-05, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 332.7} |
| {"event": "train", "step": 4150, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.38714, "bpb": 2.00121, "chunk_loss": 0.030158, "F": 0.1803, "G": 0.20305, "chunks": 184.62, "bytes_per_chunk": 5.546, "boundary_entropy": 0.11688, "router_noise": 0.0, "grad_norm": 0.715, "lr": 9.43e-05, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 336.3} |
| {"event": "probe", "step": 4200, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2926.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168132, "cps": 22.859375, "mean_p": 0.17908, "boundary_entropy": 0.073772, "bpb": 1.918259}, "tr": {"chunks": 3141.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.173507, "cps": 24.539062, "mean_p": 0.213942, "boundary_entropy": 0.226141, "bpb": 2.099158}, "fi": {"chunks": 2956.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.157142, "cps": 23.09375, "mean_p": 0.206338, "boundary_entropy": 0.243312, "bpb": 2.2432}}, "bpb": {"en": 1.91826, "tr": 2.09916, "fi": 2.2432}, "gini_cpc": 0.021873, "gini_cps": 0.015885, "bpic": 6.287488, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0319702640268873, "fi": 0.9346356549644658}, "t": 340.2} |
| {"event": "eval", "step": 4200, "phase": 2, "bpb": {"en": 1.93425, "tr": 2.10592, "fi": 2.25412}, "bytes_per_chunk": {"en": 5.397, "tr": 5.877, "fi": 6.174}, "bpb_mean": 2.0981, "frozen_inventory": false, "t": 340.7} |
| {"event": "phase_transition", "step": 4200, "phase": 3, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 341.0} |
| {"event": "train", "step": 4200, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.4615, "bpb": 2.1085, "chunk_loss": 0.030175, "F": 0.18256, "G": 0.20851, "chunks": 186.94, "bytes_per_chunk": 5.478, "boundary_entropy": 0.12805, "router_noise": 0.0, "grad_norm": 0.7476, "lr": 9.12e-05, "router_trainable": true, "sec_per_step": 0.095, "alpha": 0.03, "t": 341.1} |
| {"event": "train", "step": 4250, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.37587, "bpb": 1.98496, "chunk_loss": 0.030156, "F": 0.18332, "G": 0.20274, "chunks": 187.72, "bytes_per_chunk": 5.455, "boundary_entropy": 0.10697, "router_noise": 0.0, "grad_norm": 0.7263, "lr": 8.81e-05, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 344.7} |
| {"event": "train", "step": 4300, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.42836, "bpb": 2.06068, "chunk_loss": 0.030181, "F": 0.18402, "G": 0.20255, "chunks": 188.44, "bytes_per_chunk": 5.434, "boundary_entropy": 0.1082, "router_noise": 0.0, "grad_norm": 0.7909, "lr": 8.5e-05, "router_trainable": true, "sec_per_step": 0.068, "alpha": 0.03, "t": 348.1} |
| {"event": "train", "step": 4350, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.44686, "bpb": 2.08738, "chunk_loss": 0.030127, "F": 0.17892, "G": 0.20232, "chunks": 183.22, "bytes_per_chunk": 5.589, "boundary_entropy": 0.11814, "router_noise": 0.0, "grad_norm": 0.7074, "lr": 8.21e-05, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 351.8} |
| {"event": "probe", "step": 4400, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2925.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168074, "cps": 22.851562, "mean_p": 0.178945, "boundary_entropy": 0.072113, "bpb": 1.909826}, "tr": {"chunks": 3154.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.174225, "cps": 24.640625, "mean_p": 0.217744, "boundary_entropy": 0.229596, "bpb": 2.085823}, "fi": {"chunks": 2916.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.155016, "cps": 22.78125, "mean_p": 0.207082, "boundary_entropy": 0.247365, "bpb": 2.227609}}, "bpb": {"en": 1.90983, "tr": 2.08582, "fi": 2.22761}, "gini_cpc": 0.025751, "gini_cps": 0.017639, "bpic": 6.307059, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0365956627106712, "fi": 0.9223035621547132}, "t": 355.7} |
| {"event": "eval", "step": 4400, "phase": 3, "bpb": {"en": 1.92597, "tr": 2.09042, "fi": 2.23841}, "bytes_per_chunk": {"en": 5.401, "tr": 5.9, "fi": 6.226}, "bpb_mean": 2.08493, "frozen_inventory": false, "t": 356.1} |
| {"event": "train", "step": 4400, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.41426, "bpb": 2.04035, "chunk_loss": 0.03022, "F": 0.1857, "G": 0.20644, "chunks": 190.16, "bytes_per_chunk": 5.385, "boundary_entropy": 0.11208, "router_noise": 0.0, "grad_norm": 0.804, "lr": 7.92e-05, "router_trainable": true, "sec_per_step": 0.093, "alpha": 0.03, "t": 356.5} |
| {"event": "train", "step": 4450, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.38889, "bpb": 2.00374, "chunk_loss": 0.030178, "F": 0.18433, "G": 0.20894, "chunks": 188.75, "bytes_per_chunk": 5.425, "boundary_entropy": 0.12394, "router_noise": 0.0, "grad_norm": 0.7287, "lr": 7.63e-05, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 360.2} |
| {"event": "train", "step": 4500, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.47951, "bpb": 2.13448, "chunk_loss": 0.030307, "F": 0.18802, "G": 0.20645, "chunks": 192.53, "bytes_per_chunk": 5.319, "boundary_entropy": 0.11037, "router_noise": 0.0, "grad_norm": 0.7276, "lr": 7.36e-05, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 363.7} |
| {"event": "train", "step": 4550, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.39236, "bpb": 2.00875, "chunk_loss": 0.030223, "F": 0.18631, "G": 0.20701, "chunks": 190.78, "bytes_per_chunk": 5.367, "boundary_entropy": 0.1132, "router_noise": 0.0, "grad_norm": 0.7995, "lr": 7.09e-05, "router_trainable": true, "sec_per_step": 0.069, "alpha": 0.03, "t": 367.1} |
| {"event": "probe", "step": 4600, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2925.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168074, "cps": 22.851562, "mean_p": 0.178228, "boundary_entropy": 0.070047, "bpb": 1.899562}, "tr": {"chunks": 3128.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.172789, "cps": 24.4375, "mean_p": 0.21341, "boundary_entropy": 0.225141, "bpb": 2.073155}, "fi": {"chunks": 2923.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.155388, "cps": 22.835938, "mean_p": 0.207398, "boundary_entropy": 0.247262, "bpb": 2.219728}}, "bpb": {"en": 1.89956, "tr": 2.07315, "fi": 2.21973}, "gini_cpc": 0.023377, "gini_cps": 0.015226, "bpic": 6.32041, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0280504860364552, "fi": 0.9245175967689392}, "t": 370.9} |
| {"event": "eval", "step": 4600, "phase": 3, "bpb": {"en": 1.91573, "tr": 2.07839, "fi": 2.2301}, "bytes_per_chunk": {"en": 5.401, "tr": 5.909, "fi": 6.224}, "bpb_mean": 2.07474, "frozen_inventory": false, "t": 371.4} |
| {"event": "train", "step": 4600, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.33579, "bpb": 1.92713, "chunk_loss": 0.030086, "F": 0.17816, "G": 0.19747, "chunks": 182.44, "bytes_per_chunk": 5.613, "boundary_entropy": 0.10519, "router_noise": 0.0, "grad_norm": 0.6947, "lr": 6.82e-05, "router_trainable": true, "sec_per_step": 0.095, "alpha": 0.03, "t": 371.8} |
| {"event": "train", "step": 4650, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.33439, "bpb": 1.92512, "chunk_loss": 0.030129, "F": 0.18085, "G": 0.19912, "chunks": 185.19, "bytes_per_chunk": 5.53, "boundary_entropy": 0.10019, "router_noise": 0.0, "grad_norm": 0.7525, "lr": 6.57e-05, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 375.4} |
| {"event": "train", "step": 4700, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.39037, "bpb": 2.00589, "chunk_loss": 0.03013, "F": 0.18176, "G": 0.20206, "chunks": 186.12, "bytes_per_chunk": 5.502, "boundary_entropy": 0.10657, "router_noise": 0.0, "grad_norm": 0.8296, "lr": 6.32e-05, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 379.0} |
| {"event": "train", "step": 4750, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.40364, "bpb": 2.02502, "chunk_loss": 0.030185, "F": 0.18497, "G": 0.20321, "chunks": 189.41, "bytes_per_chunk": 5.406, "boundary_entropy": 0.1017, "router_noise": 0.0, "grad_norm": 0.717, "lr": 6.08e-05, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 382.6} |
| {"event": "probe", "step": 4800, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2925.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168074, "cps": 22.851562, "mean_p": 0.178662, "boundary_entropy": 0.069885, "bpb": 1.892474}, "tr": {"chunks": 3181.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.175717, "cps": 24.851562, "mean_p": 0.216526, "boundary_entropy": 0.223106, "bpb": 2.059791}, "fi": {"chunks": 2966.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.157674, "cps": 23.171875, "mean_p": 0.20654, "boundary_entropy": 0.242398, "bpb": 2.203475}}, "bpb": {"en": 1.89247, "tr": 2.05979, "fi": 2.20348}, "gini_cpc": 0.023987, "gini_cps": 0.018812, "bpic": 6.253527, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0454695000262033, "fi": 0.9381180951134702}, "t": 386.6} |
| {"event": "eval", "step": 4800, "phase": 3, "bpb": {"en": 1.90798, "tr": 2.06541, "fi": 2.21555}, "bytes_per_chunk": {"en": 5.395, "tr": 5.853, "fi": 6.178}, "bpb_mean": 2.06298, "frozen_inventory": false, "t": 387.1} |
| {"event": "train", "step": 4800, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.43344, "bpb": 2.06802, "chunk_loss": 0.030146, "F": 0.18222, "G": 0.20308, "chunks": 186.59, "bytes_per_chunk": 5.488, "boundary_entropy": 0.11407, "router_noise": 0.0, "grad_norm": 0.7579, "lr": 5.85e-05, "router_trainable": true, "sec_per_step": 0.098, "alpha": 0.03, "t": 387.5} |
| {"event": "train", "step": 4850, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.34015, "bpb": 1.93343, "chunk_loss": 0.030148, "F": 0.18304, "G": 0.19797, "chunks": 187.44, "bytes_per_chunk": 5.463, "boundary_entropy": 0.08753, "router_noise": 0.0, "grad_norm": 0.6683, "lr": 5.62e-05, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 391.0} |
| {"event": "train", "step": 4900, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.28722, "bpb": 1.85707, "chunk_loss": 0.030156, "F": 0.18341, "G": 0.2034, "chunks": 187.81, "bytes_per_chunk": 5.452, "boundary_entropy": 0.10613, "router_noise": 0.0, "grad_norm": 0.7676, "lr": 5.41e-05, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 394.6} |
| {"event": "train", "step": 4950, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.36337, "bpb": 1.96692, "chunk_loss": 0.030146, "F": 0.18271, "G": 0.20067, "chunks": 187.09, "bytes_per_chunk": 5.473, "boundary_entropy": 0.09932, "router_noise": 0.0, "grad_norm": 0.7292, "lr": 5.2e-05, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 398.1} |
| {"event": "probe", "step": 5000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2926.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168132, "cps": 22.859375, "mean_p": 0.178284, "boundary_entropy": 0.06789, "bpb": 1.884368}, "tr": {"chunks": 3153.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.17417, "cps": 24.632812, "mean_p": 0.21157, "boundary_entropy": 0.216709, "bpb": 2.052576}, "fi": {"chunks": 2927.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.1556, "cps": 22.867188, "mean_p": 0.206504, "boundary_entropy": 0.244332, "bpb": 2.196848}}, "bpb": {"en": 1.88437, "tr": 2.05258, "fi": 2.19685}, "gini_cpc": 0.024864, "gini_cps": 0.016804, "bpic": 6.299356, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0359128438321477, "fi": 0.9254663606498619}, "t": 401.8} |
| {"event": "eval", "step": 5000, "phase": 3, "bpb": {"en": 1.90035, "tr": 2.05836, "fi": 2.20789}, "bytes_per_chunk": {"en": 5.397, "tr": 5.894, "fi": 6.197}, "bpb_mean": 2.05553, "frozen_inventory": false, "t": 402.3} |
| {"event": "train", "step": 5000, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.31696, "bpb": 1.89997, "chunk_loss": 0.030143, "F": 0.18082, "G": 0.19935, "chunks": 185.16, "bytes_per_chunk": 5.53, "boundary_entropy": 0.10181, "router_noise": 0.0, "grad_norm": 0.7976, "lr": 5e-05, "router_trainable": true, "sec_per_step": 0.092, "alpha": 0.03, "t": 402.7} |
| {"event": "train", "step": 5050, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.40061, "bpb": 2.02066, "chunk_loss": 0.030096, "F": 0.17587, "G": 0.19849, "chunks": 180.09, "bytes_per_chunk": 5.686, "boundary_entropy": 0.11525, "router_noise": 0.0, "grad_norm": 0.7239, "lr": 4.81e-05, "router_trainable": true, "sec_per_step": 0.069, "alpha": 0.03, "t": 406.1} |
| {"event": "train", "step": 5100, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.3455, "bpb": 1.94115, "chunk_loss": 0.030123, "F": 0.18073, "G": 0.20095, "chunks": 185.06, "bytes_per_chunk": 5.533, "boundary_entropy": 0.10695, "router_noise": 0.0, "grad_norm": 0.7442, "lr": 4.63e-05, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 409.7} |
| {"event": "train", "step": 5150, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.37955, "bpb": 1.99027, "chunk_loss": 0.030097, "F": 0.17816, "G": 0.20109, "chunks": 182.44, "bytes_per_chunk": 5.613, "boundary_entropy": 0.11514, "router_noise": 0.0, "grad_norm": 0.7835, "lr": 4.45e-05, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 413.3} |
| {"event": "probe", "step": 5200, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2926.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168132, "cps": 22.859375, "mean_p": 0.178432, "boundary_entropy": 0.068305, "bpb": 1.877456}, "tr": {"chunks": 3204.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.176987, "cps": 25.03125, "mean_p": 0.213817, "boundary_entropy": 0.214889, "bpb": 2.045705}, "fi": {"chunks": 2990.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.15895, "cps": 23.359375, "mean_p": 0.211921, "boundary_entropy": 0.250371, "bpb": 2.188269}}, "bpb": {"en": 1.87746, "tr": 2.0457, "fi": 2.18827}, "gini_cpc": 0.023856, "gini_cps": 0.020322, "bpic": 6.220614, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.052668808004504, "fi": 0.9453858620919328}, "t": 417.1} |
| {"event": "eval", "step": 5200, "phase": 3, "bpb": {"en": 1.89349, "tr": 2.05075, "fi": 2.19971}, "bytes_per_chunk": {"en": 5.397, "tr": 5.847, "fi": 6.15}, "bpb_mean": 2.04798, "frozen_inventory": false, "t": 417.6} |
| {"event": "train", "step": 5200, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.34783, "bpb": 1.9445, "chunk_loss": 0.030064, "F": 0.17532, "G": 0.20037, "chunks": 179.53, "bytes_per_chunk": 5.704, "boundary_entropy": 0.11938, "router_noise": 0.0, "grad_norm": 0.7591, "lr": 4.29e-05, "router_trainable": true, "sec_per_step": 0.094, "alpha": 0.03, "t": 418.0} |
| {"event": "train", "step": 5250, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.40065, "bpb": 2.02071, "chunk_loss": 0.030292, "F": 0.18576, "G": 0.2022, "chunks": 190.22, "bytes_per_chunk": 5.383, "boundary_entropy": 0.09907, "router_noise": 0.0, "grad_norm": 0.8692, "lr": 4.14e-05, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 421.5} |
| {"event": "train", "step": 5300, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.33842, "bpb": 1.93093, "chunk_loss": 0.030096, "F": 0.17761, "G": 0.19736, "chunks": 181.88, "bytes_per_chunk": 5.63, "boundary_entropy": 0.10258, "router_noise": 0.0, "grad_norm": 0.7531, "lr": 3.99e-05, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 425.1} |
| {"event": "train", "step": 5350, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.43299, "bpb": 2.06736, "chunk_loss": 0.030159, "F": 0.18185, "G": 0.20294, "chunks": 186.22, "bytes_per_chunk": 5.499, "boundary_entropy": 0.10826, "router_noise": 0.0, "grad_norm": 0.7962, "lr": 3.86e-05, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 428.6} |
| {"event": "probe", "step": 5400, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2926.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168132, "cps": 22.859375, "mean_p": 0.177456, "boundary_entropy": 0.064395, "bpb": 1.872738}, "tr": {"chunks": 3160.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.174557, "cps": 24.6875, "mean_p": 0.211084, "boundary_entropy": 0.211979, "bpb": 2.039483}, "fi": {"chunks": 2898.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.154059, "cps": 22.640625, "mean_p": 0.20358, "boundary_entropy": 0.23788, "bpb": 2.184076}}, "bpb": {"en": 1.87274, "tr": 2.03948, "fi": 2.18408}, "gini_cpc": 0.027509, "gini_cps": 0.019442, "bpic": 6.314782, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0382126820518829, "fi": 0.9162970663352579}, "t": 432.5} |
| {"event": "eval", "step": 5400, "phase": 3, "bpb": {"en": 1.88822, "tr": 2.04611, "fi": 2.19371}, "bytes_per_chunk": {"en": 5.4, "tr": 5.873, "fi": 6.286}, "bpb_mean": 2.04268, "frozen_inventory": false, "t": 433.0} |
| {"event": "train", "step": 5400, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.32427, "bpb": 1.91052, "chunk_loss": 0.030178, "F": 0.18649, "G": 0.20172, "chunks": 190.97, "bytes_per_chunk": 5.362, "boundary_entropy": 0.09063, "router_noise": 0.0, "grad_norm": 0.7836, "lr": 3.73e-05, "router_trainable": true, "sec_per_step": 0.094, "alpha": 0.03, "t": 433.3} |
| {"event": "train", "step": 5450, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.42714, "bpb": 2.05893, "chunk_loss": 0.030197, "F": 0.18665, "G": 0.20806, "chunks": 191.12, "bytes_per_chunk": 5.358, "boundary_entropy": 0.116, "router_noise": 0.0, "grad_norm": 0.8, "lr": 3.62e-05, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 436.9} |
| {"event": "train", "step": 5500, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.35713, "bpb": 1.95793, "chunk_loss": 0.030188, "F": 0.18481, "G": 0.20306, "chunks": 189.25, "bytes_per_chunk": 5.411, "boundary_entropy": 0.10137, "router_noise": 0.0, "grad_norm": 0.8053, "lr": 3.51e-05, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 440.4} |
| {"event": "train", "step": 5550, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.3315, "bpb": 1.92095, "chunk_loss": 0.030124, "F": 0.18011, "G": 0.20191, "chunks": 184.44, "bytes_per_chunk": 5.552, "boundary_entropy": 0.11163, "router_noise": 0.0, "grad_norm": 0.7241, "lr": 3.41e-05, "router_trainable": true, "sec_per_step": 0.068, "alpha": 0.03, "t": 443.7} |
| {"event": "probe", "step": 5600, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2926.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168132, "cps": 22.859375, "mean_p": 0.177714, "boundary_entropy": 0.065269, "bpb": 1.868709}, "tr": {"chunks": 3154.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.174225, "cps": 24.640625, "mean_p": 0.211843, "boundary_entropy": 0.21425, "bpb": 2.030804}, "fi": {"chunks": 2919.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.155175, "cps": 22.804688, "mean_p": 0.206156, "boundary_entropy": 0.24199, "bpb": 2.177977}}, "bpb": {"en": 1.86871, "tr": 2.0308, "fi": 2.17798}, "gini_cpc": 0.025526, "gini_cps": 0.017409, "bpic": 6.304256, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0362413921492526, "fi": 0.9229369001492813}, "t": 447.7} |
| {"event": "eval", "step": 5600, "phase": 3, "bpb": {"en": 1.88408, "tr": 2.03775, "fi": 2.18936}, "bytes_per_chunk": {"en": 5.401, "tr": 5.884, "fi": 6.273}, "bpb_mean": 2.03706, "frozen_inventory": false, "t": 448.1} |
| {"event": "train", "step": 5600, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.38275, "bpb": 1.99489, "chunk_loss": 0.030084, "F": 0.17838, "G": 0.19475, "chunks": 182.66, "bytes_per_chunk": 5.606, "boundary_entropy": 0.09074, "router_noise": 0.0, "grad_norm": 0.7978, "lr": 3.33e-05, "router_trainable": true, "sec_per_step": 0.095, "alpha": 0.03, "t": 448.5} |
| {"event": "train", "step": 5650, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.29063, "bpb": 1.86198, "chunk_loss": 0.030122, "F": 0.1799, "G": 0.19621, "chunks": 184.22, "bytes_per_chunk": 5.559, "boundary_entropy": 0.09375, "router_noise": 0.0, "grad_norm": 0.7361, "lr": 3.25e-05, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 452.2} |
| {"event": "train", "step": 5700, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.34807, "bpb": 1.94485, "chunk_loss": 0.030112, "F": 0.17926, "G": 0.19694, "chunks": 183.56, "bytes_per_chunk": 5.578, "boundary_entropy": 0.09312, "router_noise": 0.0, "grad_norm": 0.7492, "lr": 3.18e-05, "router_trainable": true, "sec_per_step": 0.069, "alpha": 0.03, "t": 455.7} |
| {"event": "train", "step": 5750, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.38082, "bpb": 1.99211, "chunk_loss": 0.030136, "F": 0.18158, "G": 0.20197, "chunks": 185.94, "bytes_per_chunk": 5.507, "boundary_entropy": 0.10622, "router_noise": 0.0, "grad_norm": 0.7936, "lr": 3.13e-05, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 459.2} |
| {"event": "probe", "step": 5800, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2926.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168132, "cps": 22.859375, "mean_p": 0.177691, "boundary_entropy": 0.064278, "bpb": 1.863498}, "tr": {"chunks": 3161.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.174612, "cps": 24.695312, "mean_p": 0.212843, "boundary_entropy": 0.215039, "bpb": 2.026775}, "fi": {"chunks": 2949.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.15677, "cps": 23.039062, "mean_p": 0.204319, "boundary_entropy": 0.235528, "bpb": 2.170452}}, "bpb": {"en": 1.8635, "tr": 2.02677, "fi": 2.17045}, "gini_cpc": 0.023812, "gini_cps": 0.017338, "bpic": 6.278442, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0385412303689878, "fi": 0.932422377026458}, "t": 462.9} |
| {"event": "eval", "step": 5800, "phase": 3, "bpb": {"en": 1.87914, "tr": 2.03325, "fi": 2.18252}, "bytes_per_chunk": {"en": 5.401, "tr": 5.876, "fi": 6.187}, "bpb_mean": 2.03164, "frozen_inventory": false, "t": 463.4} |
| {"event": "train", "step": 5800, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.3379, "bpb": 1.93018, "chunk_loss": 0.030291, "F": 0.18878, "G": 0.2048, "chunks": 193.31, "bytes_per_chunk": 5.297, "boundary_entropy": 0.09195, "router_noise": 0.0, "grad_norm": 0.7948, "lr": 3.08e-05, "router_trainable": true, "sec_per_step": 0.092, "alpha": 0.03, "t": 463.8} |
| {"event": "train", "step": 5850, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.32215, "bpb": 1.90746, "chunk_loss": 0.030061, "F": 0.17471, "G": 0.19387, "chunks": 178.91, "bytes_per_chunk": 5.724, "boundary_entropy": 0.10051, "router_noise": 0.0, "grad_norm": 0.7531, "lr": 3.05e-05, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 467.3} |
| {"event": "train", "step": 5900, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.32335, "bpb": 1.90919, "chunk_loss": 0.030072, "F": 0.17538, "G": 0.1964, "chunks": 179.59, "bytes_per_chunk": 5.702, "boundary_entropy": 0.11052, "router_noise": 0.0, "grad_norm": 0.7578, "lr": 3.02e-05, "router_trainable": true, "sec_per_step": 0.069, "alpha": 0.03, "t": 470.8} |
| {"event": "train", "step": 5950, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.46017, "bpb": 2.10658, "chunk_loss": 0.030181, "F": 0.18231, "G": 0.20636, "chunks": 186.69, "bytes_per_chunk": 5.485, "boundary_entropy": 0.1222, "router_noise": 0.0, "grad_norm": 0.8476, "lr": 3.01e-05, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 474.3} |
| {"event": "train", "step": 5999, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.32697, "bpb": 1.91442, "chunk_loss": 0.030084, "F": 0.17853, "G": 0.19523, "chunks": 182.81, "bytes_per_chunk": 5.601, "boundary_entropy": 0.09225, "router_noise": 0.0, "grad_norm": 0.7273, "lr": 3e-05, "router_trainable": true, "sec_per_step": 0.073, "alpha": 0.03, "t": 477.9} |
| {"event": "probe", "step": 6000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2926.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.168132, "cps": 22.859375, "mean_p": 0.177446, "boundary_entropy": 0.063363, "bpb": 1.860025}, "tr": {"chunks": 3192.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.176324, "cps": 24.9375, "mean_p": 0.210628, "boundary_entropy": 0.208001, "bpb": 2.022279}, "fi": {"chunks": 2959.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.157302, "cps": 23.117188, "mean_p": 0.206199, "boundary_entropy": 0.23812, "bpb": 2.164351}}, "bpb": {"en": 1.86002, "tr": 2.02228, "fi": 2.16435}, "gini_cpc": 0.025275, "gini_cps": 0.019537, "bpic": 6.250083, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0487262281992435, "fi": 0.9355842026521834}, "t": 478.3} |
| {"event": "eval", "step": 6000, "phase": 3, "bpb": {"en": 1.87618, "tr": 2.02923, "fi": 2.17686}, "bytes_per_chunk": {"en": 5.399, "tr": 5.859, "fi": 6.139}, "bpb_mean": 2.02742, "frozen_inventory": false, "t": 478.9} |
| {"event": "reps_dumped", "dirs": ["/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s4/probe/reps/th0.30", "/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s4/probe/reps/th0.40", "/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s4/probe/reps/th0.50", "/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s4/probe/reps/th0.60", "/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s4/probe/reps/th0.70"], "t": 479.8} |
| {"event": "final", "run": "ali_hnet_s4", "arch": "hnet", "objective": "ratio", "condition": "A", "step": 6000, "params": 22516992, "seed": 4, "per_lang": {"en": {"cpc": 0.16813193127621676, "cps": 22.859375, "bpb": 1.8820559051223926, "bytes_per_chunk": 5.398686080276788}, "tr": {"cpc": 0.17632436612716124, "cps": 24.9375, "bpb": 2.022296008336176, "bytes_per_chunk": 5.869640178231567}, "fi": {"cpc": 0.15730157886343096, "cps": 23.1171875, "bpb": 2.195087751579969, "bytes_per_chunk": 6.104606213031531}}, "bpb_mean": 2.033146555012846, "gini_cpc": 0.02527485621187675, "gini_cps": 0.019536557599794335, "bpic": 6.25008262641842, "mean_boundary_entropy": 0.09225419163703918, "dechunk": "ema", "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}], "probe_overhead_sec_total": 0.0, "probe_overhead_seq_frac": 0.0, "wall_sec": 481.22227001190186, "t": 481.2} |
| {"event": "hub_done", "repo": "AdaptiveChunking/ali_hnet_s4", "pushed": [0, 1, 2, 4, 6, 8, 16, 32, 64, 128, 256, 512, 1000, 1024, 2048, 3000, 4096, 6000], "failed": [], "ok": true, "t": 496.0} |
|
|
