| {"event": "hub", "repo": "AdaptiveChunking/ali_hnet_s3", "keep_steps": [0, 1, 2, 4, 6, 8, 16, 32, 64, 128, 256, 512, 1000, 1024, 2048, 3000, 4096, 6000], "t": 1.7} |
| {"event": "start", "run": "ali_hnet_s3", "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": 1.7} |
| {"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": 1.7} |
| {"event": "probe", "step": 0, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 15894.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.913291, "cps": 124.171875, "mean_p": 0.532199, "boundary_entropy": 0.685519, "bpb": 8.014977}, "tr": {"chunks": 15976.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.882506, "cps": 124.8125, "mean_p": 0.523031, "boundary_entropy": 0.686614, "bpb": 8.046717}, "fi": {"chunks": 12092.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.642815, "cps": 94.46875, "mean_p": 0.5099, "boundary_entropy": 0.687389, "bpb": 8.01587}}, "bpb": {"en": 8.01498, "tr": 8.04672, "fi": 8.01587}, "gini_cpc": 0.073942, "gini_cps": 0.058899, "bpic": 1.290478, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9662920621975919, "fi": 0.7038452216831197}, "t": 3.2} |
| {"event": "train", "step": 0, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 5.5638, "bpb": 8.02686, "chunk_loss": 0.092026, "F": 0.95703, "G": 0.52969, "chunks": 980.0, "bytes_per_chunk": 1.045, "boundary_entropy": 0.6902, "router_noise": 0.0, "grad_norm": 13.6074, "lr": 1e-06, "router_trainable": true, "sec_per_step": 0.009, "alpha": 0.03, "t": 3.7} |
| {"event": "probe", "step": 1, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 15873.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.912084, "cps": 124.007812, "mean_p": 0.532017, "boundary_entropy": 0.685541, "bpb": 7.976205}, "tr": {"chunks": 15942.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.880628, "cps": 124.546875, "mean_p": 0.522881, "boundary_entropy": 0.686629, "bpb": 8.015666}, "fi": {"chunks": 12054.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.640795, "cps": 94.171875, "mean_p": 0.509762, "boundary_entropy": 0.687394, "bpb": 7.982759}}, "bpb": {"en": 7.97621, "tr": 8.01567, "fi": 7.98276}, "gini_cpc": 0.07432, "gini_cps": 0.059085, "bpic": 1.293214, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9655112918660398, "fi": 0.7025615981014418}, "t": 4.5} |
| {"event": "probe", "step": 2, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 15836.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.909958, "cps": 123.71875, "mean_p": 0.53166, "boundary_entropy": 0.685584, "bpb": 7.902382}, "tr": {"chunks": 15887.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.877589, "cps": 124.117188, "mean_p": 0.522604, "boundary_entropy": 0.686657, "bpb": 7.954572}, "fi": {"chunks": 11990.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.637393, "cps": 93.671875, "mean_p": 0.50952, "boundary_entropy": 0.687404, "bpb": 7.918459}}, "bpb": {"en": 7.90238, "tr": 7.95457, "fi": 7.91846}, "gini_cpc": 0.074934, "gini_cps": 0.059433, "bpic": 1.297829, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9644283566000211, "fi": 0.7004641724720004}, "t": 6.0} |
| {"event": "probe", "step": 4, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 15730.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.903867, "cps": 122.890625, "mean_p": 0.530351, "boundary_entropy": 0.685738, "bpb": 7.664597}, "tr": {"chunks": 15699.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.867204, "cps": 122.648438, "mean_p": 0.521526, "boundary_entropy": 0.686761, "bpb": 7.757562}, "fi": {"chunks": 11715.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.622774, "cps": 91.523438, "mean_p": 0.50861, "boundary_entropy": 0.687438, "bpb": 7.708648}}, "bpb": {"en": 7.6646, "tr": 7.75756, "fi": 7.70865}, "gini_cpc": 0.078282, "gini_cps": 0.06204, "bpic": 1.314945, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9594378238053571, "fi": 0.6890104473167575}, "t": 7.7} |
| {"event": "probe", "step": 6, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 15484.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.889732, "cps": 120.96875, "mean_p": 0.528073, "boundary_entropy": 0.685989, "bpb": 7.380967}, "tr": {"chunks": 15324.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.84649, "cps": 119.71875, "mean_p": 0.51968, "boundary_entropy": 0.686928, "bpb": 7.522808}, "fi": {"chunks": 11165.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.593536, "cps": 87.226562, "mean_p": 0.506968, "boundary_entropy": 0.687493, "bpb": 7.4576}}, "bpb": {"en": 7.38097, "tr": 7.52281, "fi": 7.4576}, "gini_cpc": 0.084757, "gini_cps": 0.0686, "bpic": 1.351631, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9513986907462398, "fi": 0.6670951781416032}, "t": 9.1} |
| {"event": "probe", "step": 8, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 15043.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.864391, "cps": 117.523438, "mean_p": 0.524655, "boundary_entropy": 0.686327, "bpb": 7.132232}, "tr": {"chunks": 14720.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.813125, "cps": 115.0, "mean_p": 0.516948, "boundary_entropy": 0.687145, "bpb": 7.319916}, "fi": {"chunks": 10307.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.547924, "cps": 80.523438, "mean_p": 0.504437, "boundary_entropy": 0.687558, "bpb": 7.240654}}, "bpb": {"en": 7.13223, "tr": 7.31992, "fi": 7.24065}, "gini_cpc": 0.094803, "gini_cps": 0.078795, "bpic": 1.415822, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9406908577080233, "fi": 0.6338843904504216}, "t": 10.6} |
| {"event": "probe", "step": 16, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 5093.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.292651, "cps": 39.789062, "mean_p": 0.494261, "boundary_entropy": 0.687125, "bpb": 6.587553}, "tr": {"chunks": 6034.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.333315, "cps": 47.140625, "mean_p": 0.492628, "boundary_entropy": 0.687482, "bpb": 6.902386}, "fi": {"chunks": 2868.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.152464, "cps": 22.40625, "mean_p": 0.481677, "boundary_entropy": 0.686779, "bpb": 6.781008}}, "bpb": {"en": 6.58755, "tr": 6.90239, "fi": 6.78101}, "gini_cpc": 0.154885, "gini_cps": 0.150816, "bpic": 4.053733, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.1389514148585738, "fi": 0.5209759887599344}, "t": 13.2} |
| {"event": "probe", "step": 32, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 8033.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.461587, "cps": 62.757812, "mean_p": 0.502933, "boundary_entropy": 0.687621, "bpb": 6.271697}, "tr": {"chunks": 10422.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.575706, "cps": 81.421875, "mean_p": 0.504102, "boundary_entropy": 0.688067, "bpb": 6.617762}, "fi": {"chunks": 7523.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.399926, "cps": 58.773438, "mean_p": 0.499565, "boundary_entropy": 0.688094, "bpb": 6.45839}}, "bpb": {"en": 6.2717, "tr": 6.61776, "fi": 6.45839}, "gini_cpc": 0.081537, "gini_cps": 0.074396, "bpic": 2.183848, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.2472309250717635, "fi": 0.8664141403897271}, "t": 16.5} |
| {"event": "train", "step": 50, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 4.2279, "bpb": 6.09957, "chunk_loss": 0.031698, "F": 0.19028, "G": 0.49071, "chunks": 194.84, "bytes_per_chunk": 5.255, "boundary_entropy": 0.69201, "router_noise": 0.0, "grad_norm": 2.5055, "lr": 5.1e-05, "router_trainable": true, "sec_per_step": 0.305, "alpha": 0.03, "t": 19.0} |
| {"event": "probe", "step": 64, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3300.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.189622, "cps": 25.78125, "mean_p": 0.485352, "boundary_entropy": 0.686548, "bpb": 5.61239}, "tr": {"chunks": 4352.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.240402, "cps": 34.0, "mean_p": 0.487794, "boundary_entropy": 0.687312, "bpb": 6.000183}, "fi": {"chunks": 4822.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.256339, "cps": 37.671875, "mean_p": 0.488608, "boundary_entropy": 0.687248, "bpb": 5.801785}}, "bpb": {"en": 5.61239, "tr": 6.00018, "fi": 5.80179}, "gini_cpc": 0.064802, "gini_cps": 0.081343, "bpic": 4.54802, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.267793484756419, "fi": 1.3518406541626997}, "t": 20.8} |
| {"event": "train", "step": 100, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 3.33417, "bpb": 4.81019, "chunk_loss": 0.031361, "F": 0.18698, "G": 0.46884, "chunks": 191.47, "bytes_per_chunk": 5.348, "boundary_entropy": 0.68812, "router_noise": 0.0, "grad_norm": 2.8452, "lr": 0.000101, "router_trainable": true, "sec_per_step": 0.12, "alpha": 0.03, "t": 24.9} |
| {"event": "probe", "step": 128, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3278.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.188358, "cps": 25.609375, "mean_p": 0.430094, "boundary_entropy": 0.665394, "bpb": 4.153348}, "tr": {"chunks": 4525.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.249959, "cps": 35.351562, "mean_p": 0.435409, "boundary_entropy": 0.667279, "bpb": 4.469938}, "fi": {"chunks": 4559.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.242358, "cps": 35.617188, "mean_p": 0.43353, "boundary_entropy": 0.668596, "bpb": 4.396325}}, "bpb": {"en": 4.15335, "tr": 4.46994, "fi": 4.39632}, "gini_cpc": 0.060332, "gini_cps": 0.069083, "bpic": 4.589225, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.3270375230988751, "fi": 1.2866869011287225}, "t": 28.0} |
| {"event": "train", "step": 150, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.74237, "bpb": 3.95641, "chunk_loss": 0.031528, "F": 0.20074, "G": 0.37107, "chunks": 205.56, "bytes_per_chunk": 4.981, "boundary_entropy": 0.62368, "router_noise": 0.0, "grad_norm": 1.0933, "lr": 0.000151, "router_trainable": true, "sec_per_step": 0.11, "alpha": 0.03, "t": 30.5} |
| {"event": "probe", "step": 200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3442.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.197782, "cps": 26.890625, "mean_p": 0.362441, "boundary_entropy": 0.550377, "bpb": 3.659763}, "tr": {"chunks": 4360.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.240844, "cps": 34.0625, "mean_p": 0.341342, "boundary_entropy": 0.539372, "bpb": 3.808937}, "fi": {"chunks": 4598.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.244431, "cps": 35.921875, "mean_p": 0.353746, "boundary_entropy": 0.553317, "bpb": 3.881982}}, "bpb": {"en": 3.65976, "tr": 3.80894, "fi": 3.88198}, "gini_cpc": 0.04553, "gini_cps": 0.062151, "bpic": 4.575161, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.2177249153698366, "fi": 1.2358630211636712}, "t": 35.9} |
| {"event": "eval", "step": 200, "phase": 1, "bpb": {"en": 3.66332, "tr": 3.81914, "fi": 3.8805}, "bytes_per_chunk": {"en": 4.946, "tr": 4.767, "fi": 4.578}, "bpb_mean": 3.78765, "frozen_inventory": false, "t": 36.6} |
| {"event": "train", "step": 200, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.59889, "bpb": 3.7494, "chunk_loss": 0.031298, "F": 0.19937, "G": 0.3507, "chunks": 204.16, "bytes_per_chunk": 5.016, "boundary_entropy": 0.54505, "router_noise": 0.0, "grad_norm": 1.522, "lr": 0.000201, "router_trainable": true, "sec_per_step": 0.131, "alpha": 0.03, "t": 37.0} |
| {"event": "train", "step": 250, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.37964, "bpb": 3.43309, "chunk_loss": 0.031435, "F": 0.21967, "G": 0.2881, "chunks": 224.94, "bytes_per_chunk": 4.552, "boundary_entropy": 0.41085, "router_noise": 0.0, "grad_norm": 1.5144, "lr": 0.000251, "router_trainable": true, "sec_per_step": 0.104, "alpha": 0.03, "t": 42.2} |
| {"event": "probe", "step": 256, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4101.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.235649, "cps": 32.039062, "mean_p": 0.294904, "boundary_entropy": 0.376147, "bpb": 3.327818}, "tr": {"chunks": 4252.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.234878, "cps": 33.21875, "mean_p": 0.303587, "boundary_entropy": 0.433601, "bpb": 3.522469}, "fi": {"chunks": 4487.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.238531, "cps": 35.054688, "mean_p": 0.313872, "boundary_entropy": 0.468096, "bpb": 3.671843}}, "bpb": {"en": 3.32782, "tr": 3.52247, "fi": 3.67184}, "gini_cpc": 0.003434, "gini_cps": 0.020042, "bpic": 4.41838, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9967289105367599, "fi": 1.0122284440569782}, "t": 43.2} |
| {"event": "train", "step": 300, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.30237, "bpb": 3.32162, "chunk_loss": 0.031437, "F": 0.22375, "G": 0.27628, "chunks": 229.12, "bytes_per_chunk": 4.469, "boundary_entropy": 0.35535, "router_noise": 0.0, "grad_norm": 1.5157, "lr": 0.0003, "router_trainable": true, "sec_per_step": 0.116, "alpha": 0.03, "t": 48.0} |
| {"event": "train", "step": 350, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.2265, "bpb": 3.21216, "chunk_loss": 0.030922, "F": 0.21234, "G": 0.2573, "chunks": 217.44, "bytes_per_chunk": 4.709, "boundary_entropy": 0.31044, "router_noise": 0.0, "grad_norm": 1.1723, "lr": 0.0002999, "router_trainable": true, "sec_per_step": 0.099, "alpha": 0.03, "t": 53.0} |
| {"event": "probe", "step": 400, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4025.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.231282, "cps": 31.445312, "mean_p": 0.260439, "boundary_entropy": 0.283631, "bpb": 2.887007}, "tr": {"chunks": 4257.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.235154, "cps": 33.257812, "mean_p": 0.260172, "boundary_entropy": 0.343751, "bpb": 3.200381}, "fi": {"chunks": 4545.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.241614, "cps": 35.507812, "mean_p": 0.278317, "boundary_entropy": 0.377051, "bpb": 3.370379}}, "bpb": {"en": 2.88701, "tr": 3.20038, "fi": 3.37038}, "gini_cpc": 0.009728, "gini_cps": 0.027026, "bpic": 4.422858, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.016743335152919, "fi": 1.0446726855668895}, "t": 57.8} |
| {"event": "eval", "step": 400, "phase": 1, "bpb": {"en": 2.90046, "tr": 3.2072, "fi": 3.36285}, "bytes_per_chunk": {"en": 4.345, "tr": 5.022, "fi": 4.449}, "bpb_mean": 3.15684, "frozen_inventory": false, "t": 58.5} |
| {"event": "train", "step": 400, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.15556, "bpb": 3.10982, "chunk_loss": 0.031406, "F": 0.22922, "G": 0.2625, "chunks": 234.72, "bytes_per_chunk": 4.363, "boundary_entropy": 0.30808, "router_noise": 0.0, "grad_norm": 1.0848, "lr": 0.0002998, "router_trainable": true, "sec_per_step": 0.116, "alpha": 0.03, "t": 58.8} |
| {"event": "train", "step": 450, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.04737, "bpb": 2.95373, "chunk_loss": 0.031301, "F": 0.22806, "G": 0.26032, "chunks": 233.53, "bytes_per_chunk": 4.385, "boundary_entropy": 0.28132, "router_noise": 0.0, "grad_norm": 1.3169, "lr": 0.0002995, "router_trainable": true, "sec_per_step": 0.099, "alpha": 0.03, "t": 63.7} |
| {"event": "train", "step": 500, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.97799, "bpb": 2.85364, "chunk_loss": 0.031378, "F": 0.23184, "G": 0.26062, "chunks": 237.41, "bytes_per_chunk": 4.313, "boundary_entropy": 0.26748, "router_noise": 0.0, "grad_norm": 1.238, "lr": 0.0002992, "router_trainable": true, "sec_per_step": 0.093, "alpha": 0.03, "t": 68.4} |
| {"event": "probe", "step": 512, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4000.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.229845, "cps": 31.25, "mean_p": 0.254839, "boundary_entropy": 0.254421, "bpb": 2.704102}, "tr": {"chunks": 4417.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.243993, "cps": 34.507812, "mean_p": 0.257103, "boundary_entropy": 0.310124, "bpb": 3.05394}, "fi": {"chunks": 4851.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.257881, "cps": 37.898438, "mean_p": 0.267498, "boundary_entropy": 0.32965, "bpb": 3.238355}}, "bpb": {"en": 2.7041, "tr": 3.05394, "fi": 3.23836}, "gini_cpc": 0.025543, "gini_cps": 0.04276, "bpic": 4.275852, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.061551276031597, "fi": 1.1219758784753602}, "t": 70.0} |
| {"event": "train", "step": 550, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.92184, "bpb": 2.77263, "chunk_loss": 0.031262, "F": 0.23135, "G": 0.25216, "chunks": 236.91, "bytes_per_chunk": 4.322, "boundary_entropy": 0.2584, "router_noise": 0.0, "grad_norm": 1.3465, "lr": 0.0002987, "router_trainable": true, "sec_per_step": 0.127, "alpha": 0.03, "t": 74.8} |
| {"event": "probe", "step": 600, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 4017.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.230822, "cps": 31.382812, "mean_p": 0.254149, "boundary_entropy": 0.234321, "bpb": 2.588702}, "tr": {"chunks": 4356.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.240623, "cps": 34.03125, "mean_p": 0.261055, "boundary_entropy": 0.300182, "bpb": 2.94115}, "fi": {"chunks": 5019.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.266812, "cps": 39.210938, "mean_p": 0.27299, "boundary_entropy": 0.300369, "bpb": 3.138731}}, "bpb": {"en": 2.5887, "tr": 2.94115, "fi": 3.13873}, "gini_cpc": 0.0325, "gini_cps": 0.049881, "bpic": 4.23626, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.042460500174155, "fi": 1.1559195279611914}, "t": 80.1} |
| {"event": "eval", "step": 600, "phase": 1, "bpb": {"en": 2.60094, "tr": 2.9495, "fi": 3.13942}, "bytes_per_chunk": {"en": 4.319, "tr": 5.033, "fi": 4.052}, "bpb_mean": 2.89662, "frozen_inventory": false, "t": 80.7} |
| {"event": "train", "step": 600, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.84474, "bpb": 2.66139, "chunk_loss": 0.031218, "F": 0.22787, "G": 0.25477, "chunks": 233.34, "bytes_per_chunk": 4.388, "boundary_entropy": 0.24032, "router_noise": 0.0, "grad_norm": 1.2428, "lr": 0.0002982, "router_trainable": true, "sec_per_step": 0.125, "alpha": 0.03, "t": 81.1} |
| {"event": "train", "step": 650, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.83797, "bpb": 2.65163, "chunk_loss": 0.031082, "F": 0.22342, "G": 0.2487, "chunks": 228.78, "bytes_per_chunk": 4.476, "boundary_entropy": 0.23348, "router_noise": 0.0, "grad_norm": 1.1958, "lr": 0.0002975, "router_trainable": true, "sec_per_step": 0.095, "alpha": 0.03, "t": 85.8} |
| {"event": "train", "step": 700, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.7282, "bpb": 2.49327, "chunk_loss": 0.031114, "F": 0.22549, "G": 0.25089, "chunks": 230.91, "bytes_per_chunk": 4.435, "boundary_entropy": 0.23327, "router_noise": 0.0, "grad_norm": 1.2336, "lr": 0.0002967, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 90.2} |
| {"event": "train", "step": 750, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.77318, "bpb": 2.55816, "chunk_loss": 0.031275, "F": 0.23111, "G": 0.25362, "chunks": 236.66, "bytes_per_chunk": 4.327, "boundary_entropy": 0.2222, "router_noise": 0.0, "grad_norm": 1.2184, "lr": 0.0002959, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 94.6} |
| {"event": "probe", "step": 800, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3858.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.221686, "cps": 30.140625, "mean_p": 0.246163, "boundary_entropy": 0.216674, "bpb": 2.421023}, "tr": {"chunks": 4175.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.230625, "cps": 32.617188, "mean_p": 0.254643, "boundary_entropy": 0.266071, "bpb": 2.771132}, "fi": {"chunks": 4890.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.259954, "cps": 38.203125, "mean_p": 0.264176, "boundary_entropy": 0.269666, "bpb": 2.979987}}, "bpb": {"en": 2.42102, "tr": 2.77113, "fi": 2.97999}, "gini_cpc": 0.035818, "gini_cps": 0.053238, "bpic": 4.390002, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.040322101910538, "fi": 1.172624253718726}, "t": 99.7} |
| {"event": "eval", "step": 800, "phase": 1, "bpb": {"en": 2.43315, "tr": 2.78092, "fi": 2.98349}, "bytes_per_chunk": {"en": 4.491, "tr": 5.111, "fi": 4.242}, "bpb_mean": 2.73252, "frozen_inventory": false, "t": 100.4} |
| {"event": "train", "step": 800, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.74816, "bpb": 2.52206, "chunk_loss": 0.030896, "F": 0.21503, "G": 0.2415, "chunks": 220.19, "bytes_per_chunk": 4.651, "boundary_entropy": 0.22312, "router_noise": 0.0, "grad_norm": 1.0451, "lr": 0.0002949, "router_trainable": true, "sec_per_step": 0.123, "alpha": 0.03, "t": 100.8} |
| {"event": "train", "step": 850, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.77524, "bpb": 2.56113, "chunk_loss": 0.031496, "F": 0.22827, "G": 0.25136, "chunks": 233.75, "bytes_per_chunk": 4.381, "boundary_entropy": 0.22348, "router_noise": 0.0, "grad_norm": 1.0213, "lr": 0.0002938, "router_trainable": true, "sec_per_step": 0.087, "alpha": 0.03, "t": 105.1} |
| {"event": "train", "step": 900, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.69697, "bpb": 2.44821, "chunk_loss": 0.031029, "F": 0.22433, "G": 0.24487, "chunks": 229.72, "bytes_per_chunk": 4.458, "boundary_entropy": 0.20983, "router_noise": 0.0, "grad_norm": 1.0383, "lr": 0.0002927, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 109.4} |
| {"event": "train", "step": 950, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.7074, "bpb": 2.46326, "chunk_loss": 0.03095, "F": 0.21973, "G": 0.24501, "chunks": 225.0, "bytes_per_chunk": 4.551, "boundary_entropy": 0.21625, "router_noise": 0.0, "grad_norm": 1.193, "lr": 0.0002914, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 113.8} |
| {"event": "probe", "step": 1000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3824.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.219732, "cps": 29.875, "mean_p": 0.241125, "boundary_entropy": 0.204623, "bpb": 2.342676}, "tr": {"chunks": 4214.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.232779, "cps": 32.921875, "mean_p": 0.250326, "boundary_entropy": 0.238801, "bpb": 2.667449}, "fi": {"chunks": 4947.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.262984, "cps": 38.648438, "mean_p": 0.267501, "boundary_entropy": 0.250619, "bpb": 2.843853}}, "bpb": {"en": 2.34268, "tr": 2.66745, "fi": 2.84385}, "gini_cpc": 0.0403, "gini_cps": 0.057656, "bpic": 4.369041, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.059376211252348, "fi": 1.1968404631270453}, "t": 118.4} |
| {"event": "eval", "step": 1000, "phase": 1, "bpb": {"en": 2.35457, "tr": 2.67598, "fi": 2.84736}, "bytes_per_chunk": {"en": 4.499, "tr": 4.934, "fi": 4.144}, "bpb_mean": 2.62597, "frozen_inventory": false, "t": 119.1} |
| {"event": "train", "step": 1000, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.72628, "bpb": 2.4905, "chunk_loss": 0.031018, "F": 0.22183, "G": 0.24557, "chunks": 227.16, "bytes_per_chunk": 4.508, "boundary_entropy": 0.21242, "router_noise": 0.0, "grad_norm": 1.1338, "lr": 0.0002901, "router_trainable": true, "sec_per_step": 0.114, "alpha": 0.03, "t": 119.5} |
| {"event": "probe", "step": 1024, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3891.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.223582, "cps": 30.398438, "mean_p": 0.242912, "boundary_entropy": 0.196071, "bpb": 2.320705}, "tr": {"chunks": 4211.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.232613, "cps": 32.898438, "mean_p": 0.250532, "boundary_entropy": 0.23667, "bpb": 2.621527}, "fi": {"chunks": 5070.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.269523, "cps": 39.609375, "mean_p": 0.271997, "boundary_entropy": 0.246723, "bpb": 2.8163}}, "bpb": {"en": 2.3207, "tr": 2.62153, "fi": 2.8163}, "gini_cpc": 0.042203, "gini_cps": 0.059672, "bpic": 4.307015, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0403933782321848, "fi": 1.2054771017482797}, "t": 121.8} |
| {"event": "train", "step": 1050, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.6726, "bpb": 2.41306, "chunk_loss": 0.031032, "F": 0.22195, "G": 0.24642, "chunks": 227.28, "bytes_per_chunk": 4.505, "boundary_entropy": 0.21038, "router_noise": 0.0, "grad_norm": 0.9962, "lr": 0.0002886, "router_trainable": true, "sec_per_step": 0.107, "alpha": 0.03, "t": 124.8} |
| {"event": "train", "step": 1100, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.65665, "bpb": 2.39004, "chunk_loss": 0.030943, "F": 0.21814, "G": 0.24322, "chunks": 223.38, "bytes_per_chunk": 4.584, "boundary_entropy": 0.20537, "router_noise": 0.0, "grad_norm": 0.8961, "lr": 0.0002871, "router_trainable": true, "sec_per_step": 0.098, "alpha": 0.03, "t": 129.7} |
| {"event": "train", "step": 1150, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.68787, "bpb": 2.43508, "chunk_loss": 0.030891, "F": 0.21643, "G": 0.24188, "chunks": 221.62, "bytes_per_chunk": 4.62, "boundary_entropy": 0.20698, "router_noise": 0.0, "grad_norm": 0.9244, "lr": 0.0002855, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 134.2} |
| {"event": "probe", "step": 1200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3798.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.218238, "cps": 29.671875, "mean_p": 0.239548, "boundary_entropy": 0.197696, "bpb": 2.26834}, "tr": {"chunks": 4093.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.226095, "cps": 31.976562, "mean_p": 0.245328, "boundary_entropy": 0.236872, "bpb": 2.566188}, "fi": {"chunks": 4921.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.261602, "cps": 38.445312, "mean_p": 0.270808, "boundary_entropy": 0.234052, "bpb": 2.724401}}, "bpb": {"en": 2.26834, "tr": 2.56619, "fi": 2.7244}, "gini_cpc": 0.040952, "gini_cps": 0.058435, "bpic": 4.428036, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0360014255795715, "fi": 1.1987003755575436}, "t": 138.7} |
| {"event": "eval", "step": 1200, "phase": 1, "bpb": {"en": 2.28018, "tr": 2.5747, "fi": 2.73101}, "bytes_per_chunk": {"en": 4.569, "tr": 5.118, "fi": 4.23}, "bpb_mean": 2.52863, "frozen_inventory": false, "t": 139.2} |
| {"event": "train", "step": 1200, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.64783, "bpb": 2.37732, "chunk_loss": 0.031119, "F": 0.22458, "G": 0.24588, "chunks": 229.97, "bytes_per_chunk": 4.453, "boundary_entropy": 0.20494, "router_noise": 0.0, "grad_norm": 0.8559, "lr": 0.0002837, "router_trainable": true, "sec_per_step": 0.106, "alpha": 0.03, "t": 139.5} |
| {"event": "train", "step": 1250, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.64821, "bpb": 2.37786, "chunk_loss": 0.030871, "F": 0.21759, "G": 0.24113, "chunks": 222.81, "bytes_per_chunk": 4.596, "boundary_entropy": 0.19919, "router_noise": 0.0, "grad_norm": 0.8892, "lr": 0.0002819, "router_trainable": true, "sec_per_step": 0.085, "alpha": 0.03, "t": 143.7} |
| {"event": "train", "step": 1300, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.59942, "bpb": 2.30748, "chunk_loss": 0.030827, "F": 0.21674, "G": 0.23819, "chunks": 221.94, "bytes_per_chunk": 4.614, "boundary_entropy": 0.19298, "router_noise": 0.0, "grad_norm": 0.7649, "lr": 0.00028, "router_trainable": true, "sec_per_step": 0.091, "alpha": 0.03, "t": 148.3} |
| {"event": "train", "step": 1350, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.63563, "bpb": 2.35971, "chunk_loss": 0.031141, "F": 0.22745, "G": 0.24687, "chunks": 232.91, "bytes_per_chunk": 4.397, "boundary_entropy": 0.19147, "router_noise": 0.0, "grad_norm": 1.0073, "lr": 0.000278, "router_trainable": true, "sec_per_step": 0.091, "alpha": 0.03, "t": 152.8} |
| {"event": "probe", "step": 1400, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3775.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.216917, "cps": 29.492188, "mean_p": 0.234921, "boundary_entropy": 0.182118, "bpb": 2.208375}, "tr": {"chunks": 4149.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.229189, "cps": 32.414062, "mean_p": 0.243641, "boundary_entropy": 0.219082, "bpb": 2.488515}, "fi": {"chunks": 4770.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.253575, "cps": 37.265625, "mean_p": 0.266302, "boundary_entropy": 0.222549, "bpb": 2.672495}}, "bpb": {"en": 2.20837, "tr": 2.48852, "fi": 2.67249}, "gini_cpc": 0.034929, "gini_cps": 0.052256, "bpic": 4.469198, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.056574311893715, "fi": 1.1689977084705618}, "t": 157.4} |
| {"event": "eval", "step": 1400, "phase": 1, "bpb": {"en": 2.22239, "tr": 2.49539, "fi": 2.67655}, "bytes_per_chunk": {"en": 4.583, "tr": 5.007, "fi": 4.357}, "bpb_mean": 2.46478, "frozen_inventory": false, "t": 158.0} |
| {"event": "train", "step": 1400, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.61011, "bpb": 2.3229, "chunk_loss": 0.030803, "F": 0.21606, "G": 0.23589, "chunks": 221.25, "bytes_per_chunk": 4.628, "boundary_entropy": 0.18807, "router_noise": 0.0, "grad_norm": 0.7964, "lr": 0.0002759, "router_trainable": true, "sec_per_step": 0.111, "alpha": 0.03, "t": 158.4} |
| {"event": "train", "step": 1450, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.61056, "bpb": 2.32354, "chunk_loss": 0.031047, "F": 0.22571, "G": 0.24593, "chunks": 231.12, "bytes_per_chunk": 4.431, "boundary_entropy": 0.18937, "router_noise": 0.0, "grad_norm": 0.7683, "lr": 0.0002738, "router_trainable": true, "sec_per_step": 0.084, "alpha": 0.03, "t": 162.6} |
| {"event": "train", "step": 1500, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.60463, "bpb": 2.31499, "chunk_loss": 0.031017, "F": 0.22177, "G": 0.24122, "chunks": 227.09, "bytes_per_chunk": 4.509, "boundary_entropy": 0.18712, "router_noise": 0.0, "grad_norm": 0.8059, "lr": 0.0002715, "router_trainable": true, "sec_per_step": 0.077, "alpha": 0.03, "t": 166.5} |
| {"event": "train", "step": 1550, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.46947, "bpb": 2.12, "chunk_loss": 0.030799, "F": 0.21759, "G": 0.23588, "chunks": 222.81, "bytes_per_chunk": 4.596, "boundary_entropy": 0.18251, "router_noise": 0.0, "grad_norm": 0.7836, "lr": 0.0002692, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 170.8} |
| {"event": "probe", "step": 1600, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3831.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.220134, "cps": 29.929688, "mean_p": 0.234208, "boundary_entropy": 0.176561, "bpb": 2.173448}, "tr": {"chunks": 4193.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.231619, "cps": 32.757812, "mean_p": 0.241649, "boundary_entropy": 0.208636, "bpb": 2.43443}, "fi": {"chunks": 4731.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.251502, "cps": 36.960938, "mean_p": 0.264988, "boundary_entropy": 0.206807, "bpb": 2.610368}}, "bpb": {"en": 2.17345, "tr": 2.43443, "fi": 2.61037}, "gini_cpc": 0.029735, "gini_cps": 0.04704, "bpic": 4.447824, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0521708827815566, "fi": 1.1424916451399023}, "t": 175.5} |
| {"event": "eval", "step": 1600, "phase": 1, "bpb": {"en": 2.18284, "tr": 2.44148, "fi": 2.61515}, "bytes_per_chunk": {"en": 4.511, "tr": 4.945, "fi": 4.437}, "bpb_mean": 2.41316, "frozen_inventory": false, "t": 175.8} |
| {"event": "train", "step": 1600, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.5798, "bpb": 2.27916, "chunk_loss": 0.030921, "F": 0.22351, "G": 0.23952, "chunks": 228.88, "bytes_per_chunk": 4.474, "boundary_entropy": 0.17795, "router_noise": 0.0, "grad_norm": 0.861, "lr": 0.0002668, "router_trainable": true, "sec_per_step": 0.106, "alpha": 0.03, "t": 176.1} |
| {"event": "train", "step": 1650, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.51049, "bpb": 2.17917, "chunk_loss": 0.030593, "F": 0.20895, "G": 0.2262, "chunks": 213.97, "bytes_per_chunk": 4.786, "boundary_entropy": 0.17852, "router_noise": 0.0, "grad_norm": 0.8497, "lr": 0.0002643, "router_trainable": true, "sec_per_step": 0.099, "alpha": 0.03, "t": 181.1} |
| {"event": "train", "step": 1700, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.55938, "bpb": 2.24971, "chunk_loss": 0.030738, "F": 0.21609, "G": 0.23197, "chunks": 221.28, "bytes_per_chunk": 4.628, "boundary_entropy": 0.17258, "router_noise": 0.0, "grad_norm": 0.8465, "lr": 0.0002618, "router_trainable": true, "sec_per_step": 0.092, "alpha": 0.03, "t": 185.7} |
| {"event": "train", "step": 1750, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.55362, "bpb": 2.24139, "chunk_loss": 0.030725, "F": 0.21442, "G": 0.23053, "chunks": 219.56, "bytes_per_chunk": 4.664, "boundary_entropy": 0.17234, "router_noise": 0.0, "grad_norm": 0.7396, "lr": 0.0002591, "router_trainable": true, "sec_per_step": 0.084, "alpha": 0.03, "t": 189.9} |
| {"event": "probe", "step": 1800, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3771.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.216687, "cps": 29.460938, "mean_p": 0.229303, "boundary_entropy": 0.163595, "bpb": 2.13354}, "tr": {"chunks": 4102.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.226592, "cps": 32.046875, "mean_p": 0.2336, "boundary_entropy": 0.201771, "bpb": 2.365447}, "fi": {"chunks": 4696.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.249641, "cps": 36.6875, "mean_p": 0.26062, "boundary_entropy": 0.196063, "bpb": 2.531174}}, "bpb": {"en": 2.13354, "tr": 2.36545, "fi": 2.53117}, "gini_cpc": 0.031706, "gini_cps": 0.049063, "bpic": 4.513645, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0457134462301396, "fi": 1.1520830645181412}, "t": 194.5} |
| {"event": "eval", "step": 1800, "phase": 1, "bpb": {"en": 2.14387, "tr": 2.37653, "fi": 2.54296}, "bytes_per_chunk": {"en": 4.563, "tr": 5.048, "fi": 4.457}, "bpb_mean": 2.35445, "frozen_inventory": false, "t": 195.0} |
| {"event": "train", "step": 1800, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.51699, "bpb": 2.18856, "chunk_loss": 0.030758, "F": 0.21579, "G": 0.23056, "chunks": 220.97, "bytes_per_chunk": 4.634, "boundary_entropy": 0.16975, "router_noise": 0.0, "grad_norm": 0.725, "lr": 0.0002564, "router_trainable": true, "sec_per_step": 0.108, "alpha": 0.03, "t": 195.3} |
| {"event": "train", "step": 1850, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.53825, "bpb": 2.21922, "chunk_loss": 0.030597, "F": 0.21082, "G": 0.2236, "chunks": 215.88, "bytes_per_chunk": 4.743, "boundary_entropy": 0.16817, "router_noise": 0.0, "grad_norm": 0.7272, "lr": 0.0002537, "router_trainable": true, "sec_per_step": 0.083, "alpha": 0.03, "t": 199.4} |
| {"event": "train", "step": 1900, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.46336, "bpb": 2.11118, "chunk_loss": 0.030881, "F": 0.22159, "G": 0.23689, "chunks": 226.91, "bytes_per_chunk": 4.513, "boundary_entropy": 0.16352, "router_noise": 0.0, "grad_norm": 0.7699, "lr": 0.0002508, "router_trainable": true, "sec_per_step": 0.092, "alpha": 0.03, "t": 204.0} |
| {"event": "train", "step": 1950, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.46245, "bpb": 2.10987, "chunk_loss": 0.030685, "F": 0.21344, "G": 0.22695, "chunks": 218.56, "bytes_per_chunk": 4.685, "boundary_entropy": 0.16159, "router_noise": 0.0, "grad_norm": 0.8027, "lr": 0.0002479, "router_trainable": true, "sec_per_step": 0.092, "alpha": 0.03, "t": 208.6} |
| {"event": "probe", "step": 2000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3748.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.215365, "cps": 29.28125, "mean_p": 0.22873, "boundary_entropy": 0.155446, "bpb": 2.106045}, "tr": {"chunks": 4057.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.224107, "cps": 31.695312, "mean_p": 0.226118, "boundary_entropy": 0.199625, "bpb": 2.330939}, "fi": {"chunks": 4768.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.253469, "cps": 37.25, "mean_p": 0.26099, "boundary_entropy": 0.180923, "bpb": 2.485006}}, "bpb": {"en": 2.10604, "tr": 2.33094, "fi": 2.48501}, "gini_cpc": 0.036659, "gini_cps": 0.054084, "bpic": 4.512209, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0405884335167122, "fi": 1.17692530659557}, "t": 213.0} |
| {"event": "eval", "step": 2000, "phase": 1, "bpb": {"en": 2.11711, "tr": 2.34277, "fi": 2.49247}, "bytes_per_chunk": {"en": 4.611, "tr": 5.07, "fi": 4.37}, "bpb_mean": 2.31745, "frozen_inventory": false, "t": 213.5} |
| {"event": "train", "step": 2000, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.58564, "bpb": 2.2876, "chunk_loss": 0.030847, "F": 0.2197, "G": 0.23281, "chunks": 224.97, "bytes_per_chunk": 4.552, "boundary_entropy": 0.1634, "router_noise": 0.0, "grad_norm": 0.777, "lr": 0.000245, "router_trainable": true, "sec_per_step": 0.104, "alpha": 0.03, "t": 213.8} |
| {"event": "probe", "step": 2048, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3785.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.217491, "cps": 29.570312, "mean_p": 0.22754, "boundary_entropy": 0.159114, "bpb": 2.098032}, "tr": {"chunks": 4084.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.225598, "cps": 31.90625, "mean_p": 0.225373, "boundary_entropy": 0.210035, "bpb": 2.325286}, "fi": {"chunks": 4695.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.249588, "cps": 36.679688, "mean_p": 0.25861, "boundary_entropy": 0.182441, "bpb": 2.459564}}, "bpb": {"en": 2.09803, "tr": 2.32529, "fi": 2.45956}, "gini_cpc": 0.030892, "gini_cps": 0.048286, "bpic": 4.515441, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0372738237697672, "fi": 1.1475773014847617}, "t": 218.1} |
| {"event": "train", "step": 2050, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.53469, "bpb": 2.2141, "chunk_loss": 0.030846, "F": 0.2193, "G": 0.23039, "chunks": 224.56, "bytes_per_chunk": 4.56, "boundary_entropy": 0.15965, "router_noise": 0.0, "grad_norm": 0.7791, "lr": 0.0002419, "router_trainable": true, "sec_per_step": 0.094, "alpha": 0.03, "t": 218.5} |
| {"event": "train", "step": 2100, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.55581, "bpb": 2.24456, "chunk_loss": 0.030815, "F": 0.22015, "G": 0.23147, "chunks": 225.44, "bytes_per_chunk": 4.542, "boundary_entropy": 0.16258, "router_noise": 0.0, "grad_norm": 0.8178, "lr": 0.0002388, "router_trainable": true, "sec_per_step": 0.087, "alpha": 0.03, "t": 222.9} |
| {"event": "train", "step": 2150, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.53015, "bpb": 2.20755, "chunk_loss": 0.03092, "F": 0.22421, "G": 0.2347, "chunks": 229.59, "bytes_per_chunk": 4.46, "boundary_entropy": 0.16277, "router_noise": 0.0, "grad_norm": 0.7435, "lr": 0.0002357, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 227.3} |
| {"event": "probe", "step": 2200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3769.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.216572, "cps": 29.445312, "mean_p": 0.225545, "boundary_entropy": 0.154796, "bpb": 2.080707}, "tr": {"chunks": 3821.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.21107, "cps": 29.851562, "mean_p": 0.214471, "boundary_entropy": 0.202346, "bpb": 2.296945}, "fi": {"chunks": 4695.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.249588, "cps": 36.679688, "mean_p": 0.255449, "boundary_entropy": 0.173832, "bpb": 2.438581}}, "bpb": {"en": 2.08071, "tr": 2.29695, "fi": 2.43858}, "gini_cpc": 0.037917, "gini_cps": 0.050251, "bpic": 4.617989, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9745956508676575, "fi": 1.1524489482939304}, "t": 231.9} |
| {"event": "eval", "step": 2200, "phase": 1, "bpb": {"en": 2.0921, "tr": 2.30681, "fi": 2.44667}, "bytes_per_chunk": {"en": 4.571, "tr": 5.256, "fi": 4.441}, "bpb_mean": 2.28186, "frozen_inventory": false, "t": 232.6} |
| {"event": "train", "step": 2200, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.60368, "bpb": 2.31362, "chunk_loss": 0.031074, "F": 0.22253, "G": 0.23229, "chunks": 227.88, "bytes_per_chunk": 4.494, "boundary_entropy": 0.16805, "router_noise": 0.0, "grad_norm": 0.7926, "lr": 0.0002325, "router_trainable": true, "sec_per_step": 0.113, "alpha": 0.03, "t": 233.0} |
| {"event": "train", "step": 2250, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.59032, "bpb": 2.29435, "chunk_loss": 0.030844, "F": 0.21902, "G": 0.2301, "chunks": 224.28, "bytes_per_chunk": 4.566, "boundary_entropy": 0.16428, "router_noise": 0.0, "grad_norm": 0.7879, "lr": 0.0002293, "router_trainable": true, "sec_per_step": 0.091, "alpha": 0.03, "t": 237.6} |
| {"event": "train", "step": 2300, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.50971, "bpb": 2.17805, "chunk_loss": 0.030919, "F": 0.22177, "G": 0.23387, "chunks": 227.09, "bytes_per_chunk": 4.509, "boundary_entropy": 0.15983, "router_noise": 0.0, "grad_norm": 0.7242, "lr": 0.000226, "router_trainable": true, "sec_per_step": 0.093, "alpha": 0.03, "t": 242.2} |
| {"event": "train", "step": 2350, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.40926, "bpb": 2.03314, "chunk_loss": 0.030839, "F": 0.21906, "G": 0.2303, "chunks": 224.31, "bytes_per_chunk": 4.565, "boundary_entropy": 0.15211, "router_noise": 0.0, "grad_norm": 0.7078, "lr": 0.0002226, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 246.7} |
| {"event": "probe", "step": 2400, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3788.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.217664, "cps": 29.59375, "mean_p": 0.225099, "boundary_entropy": 0.143915, "bpb": 2.063838}, "tr": {"chunks": 3697.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.20422, "cps": 28.882812, "mean_p": 0.211021, "boundary_entropy": 0.187861, "bpb": 2.266365}, "fi": {"chunks": 4629.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.246079, "cps": 36.164062, "mean_p": 0.249454, "boundary_entropy": 0.170377, "bpb": 2.409703}}, "bpb": {"en": 2.06384, "tr": 2.26637, "fi": 2.4097}, "gini_cpc": 0.041778, "gini_cps": 0.051291, "bpic": 4.683176, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9382380658698224, "fi": 1.1305491484598897}, "t": 251.2} |
| {"event": "eval", "step": 2400, "phase": 1, "bpb": {"en": 2.07684, "tr": 2.27763, "fi": 2.42004}, "bytes_per_chunk": {"en": 4.561, "tr": 5.578, "fi": 4.496}, "bpb_mean": 2.25817, "frozen_inventory": false, "t": 251.8} |
| {"event": "phase_transition", "step": 2400, "phase": 2, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 252.1} |
| {"event": "train", "step": 2400, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.48219, "bpb": 2.13835, "chunk_loss": 0.030696, "F": 0.21494, "G": 0.22404, "chunks": 220.09, "bytes_per_chunk": 4.653, "boundary_entropy": 0.14799, "router_noise": 0.0, "grad_norm": 0.7398, "lr": 0.0002192, "router_trainable": true, "sec_per_step": 0.11, "alpha": 0.03, "t": 252.2} |
| {"event": "train", "step": 2450, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.50174, "bpb": 2.16656, "chunk_loss": 0.030773, "F": 0.21771, "G": 0.22876, "chunks": 222.94, "bytes_per_chunk": 4.593, "boundary_entropy": 0.15312, "router_noise": 0.0, "grad_norm": 0.7286, "lr": 0.0002158, "router_trainable": true, "sec_per_step": 0.085, "alpha": 0.03, "t": 256.4} |
| {"event": "train", "step": 2500, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.41048, "bpb": 2.03489, "chunk_loss": 0.030847, "F": 0.22, "G": 0.22934, "chunks": 225.28, "bytes_per_chunk": 4.545, "boundary_entropy": 0.14262, "router_noise": 0.0, "grad_norm": 0.7456, "lr": 0.0002123, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 260.9} |
| {"event": "train", "step": 2550, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.44289, "bpb": 2.08165, "chunk_loss": 0.030731, "F": 0.211, "G": 0.22041, "chunks": 216.06, "bytes_per_chunk": 4.739, "boundary_entropy": 0.14527, "router_noise": 0.0, "grad_norm": 0.704, "lr": 0.0002088, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 265.3} |
| {"event": "probe", "step": 2600, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3768.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.216514, "cps": 29.4375, "mean_p": 0.223811, "boundary_entropy": 0.143762, "bpb": 2.03794}, "tr": {"chunks": 3658.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.202066, "cps": 28.578125, "mean_p": 0.215009, "boundary_entropy": 0.186487, "bpb": 2.23475}, "fi": {"chunks": 4667.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.2481, "cps": 36.460938, "mean_p": 0.249718, "boundary_entropy": 0.159293, "bpb": 2.36265}}, "bpb": {"en": 2.03794, "tr": 2.23475, "fi": 2.36265}, "gini_cpc": 0.046033, "gini_cps": 0.055624, "bpic": 4.691309, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9332680018197357, "fi": 1.145880010916502}, "t": 270.0} |
| {"event": "eval", "step": 2600, "phase": 2, "bpb": {"en": 2.04982, "tr": 2.24621, "fi": 2.37291}, "bytes_per_chunk": {"en": 4.605, "tr": 5.693, "fi": 4.453}, "bpb_mean": 2.22298, "frozen_inventory": false, "t": 270.5} |
| {"event": "train", "step": 2600, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.41779, "bpb": 2.04544, "chunk_loss": 0.030647, "F": 0.21207, "G": 0.22164, "chunks": 217.16, "bytes_per_chunk": 4.715, "boundary_entropy": 0.1472, "router_noise": 0.0, "grad_norm": 0.6461, "lr": 0.0002053, "router_trainable": true, "sec_per_step": 0.112, "alpha": 0.03, "t": 270.9} |
| {"event": "train", "step": 2650, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.40463, "bpb": 2.02646, "chunk_loss": 0.030693, "F": 0.21545, "G": 0.22282, "chunks": 220.62, "bytes_per_chunk": 4.641, "boundary_entropy": 0.14118, "router_noise": 0.0, "grad_norm": 0.7032, "lr": 0.0002017, "router_trainable": true, "sec_per_step": 0.093, "alpha": 0.03, "t": 275.5} |
| {"event": "train", "step": 2700, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.51852, "bpb": 2.19077, "chunk_loss": 0.030861, "F": 0.21997, "G": 0.22873, "chunks": 225.25, "bytes_per_chunk": 4.546, "boundary_entropy": 0.14961, "router_noise": 0.0, "grad_norm": 0.7215, "lr": 0.0001981, "router_trainable": true, "sec_per_step": 0.093, "alpha": 0.03, "t": 280.2} |
| {"event": "train", "step": 2750, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.44762, "bpb": 2.08848, "chunk_loss": 0.030659, "F": 0.21167, "G": 0.2217, "chunks": 216.75, "bytes_per_chunk": 4.724, "boundary_entropy": 0.14782, "router_noise": 0.0, "grad_norm": 0.6709, "lr": 0.0001945, "router_trainable": true, "sec_per_step": 0.084, "alpha": 0.03, "t": 284.4} |
| {"event": "probe", "step": 2800, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3750.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.21548, "cps": 29.296875, "mean_p": 0.222632, "boundary_entropy": 0.138384, "bpb": 2.022505}, "tr": {"chunks": 3585.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.198033, "cps": 28.007812, "mean_p": 0.21565, "boundary_entropy": 0.186762, "bpb": 2.207642}, "fi": {"chunks": 4697.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.249694, "cps": 36.695312, "mean_p": 0.249612, "boundary_entropy": 0.15609, "bpb": 2.347231}}, "bpb": {"en": 2.0225, "tr": 2.20764, "fi": 2.34723}, "gini_cpc": 0.05193, "gini_cps": 0.061613, "bpic": 4.715093, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9190337513119372, "fi": 1.1587814363935995}, "t": 288.9} |
| {"event": "eval", "step": 2800, "phase": 2, "bpb": {"en": 2.03417, "tr": 2.21751, "fi": 2.35499}, "bytes_per_chunk": {"en": 4.642, "tr": 5.739, "fi": 4.445}, "bpb_mean": 2.20222, "frozen_inventory": false, "t": 289.5} |
| {"event": "train", "step": 2800, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.46818, "bpb": 2.11813, "chunk_loss": 0.03086, "F": 0.21704, "G": 0.22762, "chunks": 222.25, "bytes_per_chunk": 4.607, "boundary_entropy": 0.14433, "router_noise": 0.0, "grad_norm": 0.7242, "lr": 0.0001909, "router_trainable": true, "sec_per_step": 0.109, "alpha": 0.03, "t": 289.8} |
| {"event": "train", "step": 2850, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.42923, "bpb": 2.06194, "chunk_loss": 0.030467, "F": 0.20428, "G": 0.21648, "chunks": 209.19, "bytes_per_chunk": 4.895, "boundary_entropy": 0.14469, "router_noise": 0.0, "grad_norm": 0.7026, "lr": 0.0001872, "router_trainable": true, "sec_per_step": 0.082, "alpha": 0.03, "t": 293.9} |
| {"event": "train", "step": 2900, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.4479, "bpb": 2.08888, "chunk_loss": 0.030669, "F": 0.21292, "G": 0.2227, "chunks": 218.03, "bytes_per_chunk": 4.697, "boundary_entropy": 0.1409, "router_noise": 0.0, "grad_norm": 0.7336, "lr": 0.0001835, "router_trainable": true, "sec_per_step": 0.085, "alpha": 0.03, "t": 298.1} |
| {"event": "train", "step": 2950, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.51892, "bpb": 2.19133, "chunk_loss": 0.0309, "F": 0.22043, "G": 0.22961, "chunks": 225.72, "bytes_per_chunk": 4.537, "boundary_entropy": 0.14342, "router_noise": 0.0, "grad_norm": 0.7235, "lr": 0.0001799, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 302.4} |
| {"event": "probe", "step": 3000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3841.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.220709, "cps": 30.007812, "mean_p": 0.226017, "boundary_entropy": 0.128126, "bpb": 2.005745}, "tr": {"chunks": 3951.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.218251, "cps": 30.867188, "mean_p": 0.219054, "boundary_entropy": 0.18029, "bpb": 2.191159}, "fi": {"chunks": 4774.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.253788, "cps": 37.296875, "mean_p": 0.254343, "boundary_entropy": 0.14908, "bpb": 2.303989}}, "bpb": {"en": 2.00575, "tr": 2.19116, "fi": 2.30399}, "gini_cpc": 0.034199, "gini_cps": 0.049499, "bpic": 4.514722, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.988863373335228, "fi": 1.1498742385286953}, "t": 307.2} |
| {"event": "eval", "step": 3000, "phase": 2, "bpb": {"en": 2.01886, "tr": 2.19885, "fi": 2.31396}, "bytes_per_chunk": {"en": 4.532, "tr": 5.285, "fi": 4.371}, "bpb_mean": 2.17722, "frozen_inventory": false, "t": 307.7} |
| {"event": "train", "step": 3000, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.40979, "bpb": 2.0339, "chunk_loss": 0.030846, "F": 0.22214, "G": 0.22928, "chunks": 227.47, "bytes_per_chunk": 4.502, "boundary_entropy": 0.12844, "router_noise": 0.0, "grad_norm": 0.7034, "lr": 0.0001761, "router_trainable": true, "sec_per_step": 0.11, "alpha": 0.03, "t": 307.9} |
| {"event": "train", "step": 3050, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.43882, "bpb": 2.07578, "chunk_loss": 0.030844, "F": 0.22092, "G": 0.22894, "chunks": 226.22, "bytes_per_chunk": 4.527, "boundary_entropy": 0.14095, "router_noise": 0.0, "grad_norm": 0.6523, "lr": 0.0001724, "router_trainable": true, "sec_per_step": 0.097, "alpha": 0.03, "t": 312.8} |
| {"event": "train", "step": 3100, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.48649, "bpb": 2.14456, "chunk_loss": 0.031151, "F": 0.2261, "G": 0.23486, "chunks": 231.53, "bytes_per_chunk": 4.423, "boundary_entropy": 0.13699, "router_noise": 0.0, "grad_norm": 0.6754, "lr": 0.0001687, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 317.2} |
| {"event": "train", "step": 3150, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.446, "bpb": 2.08613, "chunk_loss": 0.030623, "F": 0.21344, "G": 0.22136, "chunks": 218.56, "bytes_per_chunk": 4.685, "boundary_entropy": 0.13313, "router_noise": 0.0, "grad_norm": 0.7124, "lr": 0.000165, "router_trainable": true, "sec_per_step": 0.082, "alpha": 0.03, "t": 321.4} |
| {"event": "probe", "step": 3200, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3735.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.214618, "cps": 29.179688, "mean_p": 0.222141, "boundary_entropy": 0.134563, "bpb": 1.989082}, "tr": {"chunks": 3695.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.20411, "cps": 28.867188, "mean_p": 0.21507, "boundary_entropy": 0.180942, "bpb": 2.164979}, "fi": {"chunks": 4654.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.247408, "cps": 36.359375, "mean_p": 0.246859, "boundary_entropy": 0.149775, "bpb": 2.285955}}, "bpb": {"en": 1.98908, "tr": 2.16498, "fi": 2.28595}, "gini_cpc": 0.043333, "gini_cps": 0.052907, "bpic": 4.694803, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9510369822659138, "fi": 1.1527841844225555}, "t": 325.9} |
| {"event": "eval", "step": 3200, "phase": 2, "bpb": {"en": 2.00067, "tr": 2.17313, "fi": 2.29425}, "bytes_per_chunk": {"en": 4.624, "tr": 5.649, "fi": 4.494}, "bpb_mean": 2.15602, "frozen_inventory": false, "t": 326.6} |
| {"event": "train", "step": 3200, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.45086, "bpb": 2.09315, "chunk_loss": 0.030735, "F": 0.2168, "G": 0.22543, "chunks": 222.0, "bytes_per_chunk": 4.613, "boundary_entropy": 0.13882, "router_noise": 0.0, "grad_norm": 0.6904, "lr": 0.0001613, "router_trainable": true, "sec_per_step": 0.109, "alpha": 0.03, "t": 326.8} |
| {"event": "train", "step": 3250, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.41951, "bpb": 2.04791, "chunk_loss": 0.030547, "F": 0.20935, "G": 0.21862, "chunks": 214.38, "bytes_per_chunk": 4.777, "boundary_entropy": 0.13075, "router_noise": 0.0, "grad_norm": 0.6265, "lr": 0.0001576, "router_trainable": true, "sec_per_step": 0.082, "alpha": 0.03, "t": 330.9} |
| {"event": "train", "step": 3300, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.37387, "bpb": 1.98208, "chunk_loss": 0.030681, "F": 0.21558, "G": 0.22242, "chunks": 220.75, "bytes_per_chunk": 4.639, "boundary_entropy": 0.12712, "router_noise": 0.0, "grad_norm": 0.6728, "lr": 0.0001539, "router_trainable": true, "sec_per_step": 0.082, "alpha": 0.03, "t": 335.0} |
| {"event": "train", "step": 3350, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.3627, "bpb": 1.96596, "chunk_loss": 0.03052, "F": 0.2084, "G": 0.21762, "chunks": 213.41, "bytes_per_chunk": 4.798, "boundary_entropy": 0.13433, "router_noise": 0.0, "grad_norm": 0.7062, "lr": 0.0001501, "router_trainable": true, "sec_per_step": 0.091, "alpha": 0.03, "t": 339.6} |
| {"event": "probe", "step": 3400, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3667.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.210711, "cps": 28.648438, "mean_p": 0.217364, "boundary_entropy": 0.127512, "bpb": 1.978198}, "tr": {"chunks": 3591.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.198365, "cps": 28.054688, "mean_p": 0.213106, "boundary_entropy": 0.177333, "bpb": 2.146281}, "fi": {"chunks": 4759.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.25299, "cps": 37.179688, "mean_p": 0.248765, "boundary_entropy": 0.138576, "bpb": 2.260733}}, "bpb": {"en": 1.9782, "tr": 2.14628, "fi": 2.26073}, "gini_cpc": 0.055005, "gini_cps": 0.064797, "bpic": 4.720979, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9414083887850725, "fi": 1.2006516764432533}, "t": 344.2} |
| {"event": "eval", "step": 3400, "phase": 2, "bpb": {"en": 1.98975, "tr": 2.15408, "fi": 2.26918}, "bytes_per_chunk": {"en": 4.691, "tr": 5.777, "fi": 4.401}, "bpb_mean": 2.13767, "frozen_inventory": false, "t": 344.9} |
| {"event": "train", "step": 3400, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.44955, "bpb": 2.09126, "chunk_loss": 0.030641, "F": 0.21063, "G": 0.22088, "chunks": 215.69, "bytes_per_chunk": 4.748, "boundary_entropy": 0.13378, "router_noise": 0.0, "grad_norm": 0.6994, "lr": 0.0001465, "router_trainable": true, "sec_per_step": 0.113, "alpha": 0.03, "t": 345.2} |
| {"event": "train", "step": 3450, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.41082, "bpb": 2.03538, "chunk_loss": 0.03043, "F": 0.20245, "G": 0.21388, "chunks": 207.31, "bytes_per_chunk": 4.939, "boundary_entropy": 0.13268, "router_noise": 0.0, "grad_norm": 0.6462, "lr": 0.0001428, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 349.7} |
| {"event": "train", "step": 3500, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.40447, "bpb": 2.02622, "chunk_loss": 0.030544, "F": 0.21042, "G": 0.21713, "chunks": 215.47, "bytes_per_chunk": 4.752, "boundary_entropy": 0.12649, "router_noise": 0.0, "grad_norm": 0.7565, "lr": 0.0001391, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 354.2} |
| {"event": "train", "step": 3550, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.4172, "bpb": 2.04458, "chunk_loss": 0.030613, "F": 0.211, "G": 0.22168, "chunks": 216.06, "bytes_per_chunk": 4.739, "boundary_entropy": 0.13382, "router_noise": 0.0, "grad_norm": 0.6731, "lr": 0.0001355, "router_trainable": true, "sec_per_step": 0.08, "alpha": 0.03, "t": 358.2} |
| {"event": "probe", "step": 3600, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3743.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.215078, "cps": 29.242188, "mean_p": 0.221119, "boundary_entropy": 0.123105, "bpb": 1.961852}, "tr": {"chunks": 3593.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.198475, "cps": 28.070312, "mean_p": 0.214121, "boundary_entropy": 0.174791, "bpb": 2.124453}, "fi": {"chunks": 4702.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.24996, "cps": 36.734375, "mean_p": 0.246843, "boundary_entropy": 0.141178, "bpb": 2.242234}}, "bpb": {"en": 1.96185, "tr": 2.12445, "fi": 2.24223}, "gini_cpc": 0.051729, "gini_cps": 0.061417, "bpic": 4.712743, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9228071670923214, "fi": 1.1621843808085583}, "t": 362.9} |
| {"event": "eval", "step": 3600, "phase": 2, "bpb": {"en": 1.97383, "tr": 2.13301, "fi": 2.25161}, "bytes_per_chunk": {"en": 4.616, "tr": 5.745, "fi": 4.444}, "bpb_mean": 2.11948, "frozen_inventory": false, "t": 363.4} |
| {"event": "train", "step": 3600, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.43126, "bpb": 2.06486, "chunk_loss": 0.030463, "F": 0.20044, "G": 0.21674, "chunks": 205.25, "bytes_per_chunk": 4.989, "boundary_entropy": 0.14047, "router_noise": 0.0, "grad_norm": 0.6855, "lr": 0.0001319, "router_trainable": true, "sec_per_step": 0.111, "alpha": 0.03, "t": 363.7} |
| {"event": "train", "step": 3650, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.43518, "bpb": 2.07053, "chunk_loss": 0.030735, "F": 0.21594, "G": 0.22367, "chunks": 221.12, "bytes_per_chunk": 4.631, "boundary_entropy": 0.13186, "router_noise": 0.0, "grad_norm": 0.6911, "lr": 0.0001283, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 368.2} |
| {"event": "train", "step": 3700, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.50933, "bpb": 2.17751, "chunk_loss": 0.030935, "F": 0.22092, "G": 0.23047, "chunks": 226.22, "bytes_per_chunk": 4.527, "boundary_entropy": 0.12953, "router_noise": 0.0, "grad_norm": 0.7374, "lr": 0.0001247, "router_trainable": true, "sec_per_step": 0.083, "alpha": 0.03, "t": 372.3} |
| {"event": "train", "step": 3750, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.35141, "bpb": 1.94968, "chunk_loss": 0.030749, "F": 0.21838, "G": 0.22456, "chunks": 223.62, "bytes_per_chunk": 4.579, "boundary_entropy": 0.1213, "router_noise": 0.0, "grad_norm": 0.6456, "lr": 0.0001212, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 376.7} |
| {"event": "probe", "step": 3800, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3739.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.214848, "cps": 29.210938, "mean_p": 0.221233, "boundary_entropy": 0.120933, "bpb": 1.949443}, "tr": {"chunks": 3709.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.204883, "cps": 28.976562, "mean_p": 0.217862, "boundary_entropy": 0.172178, "bpb": 2.106707}, "fi": {"chunks": 4759.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.25299, "cps": 37.179688, "mean_p": 0.249414, "boundary_entropy": 0.136744, "bpb": 2.221187}}, "bpb": {"en": 1.94944, "tr": 2.10671, "fi": 2.22119}, "gini_cpc": 0.047674, "gini_cps": 0.057344, "bpic": 4.647497, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9536190912075633, "fi": 1.1775313446155147}, "t": 381.5} |
| {"event": "eval", "step": 3800, "phase": 2, "bpb": {"en": 1.96302, "tr": 2.11659, "fi": 2.23104}, "bytes_per_chunk": {"en": 4.639, "tr": 5.599, "fi": 4.407}, "bpb_mean": 2.10355, "frozen_inventory": false, "t": 381.9} |
| {"event": "train", "step": 3800, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.39341, "bpb": 2.01026, "chunk_loss": 0.030538, "F": 0.20709, "G": 0.21798, "chunks": 212.06, "bytes_per_chunk": 4.829, "boundary_entropy": 0.12964, "router_noise": 0.0, "grad_norm": 0.6603, "lr": 0.0001177, "router_trainable": true, "sec_per_step": 0.108, "alpha": 0.03, "t": 382.2} |
| {"event": "train", "step": 3850, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.44557, "bpb": 2.08552, "chunk_loss": 0.030418, "F": 0.20251, "G": 0.21289, "chunks": 207.38, "bytes_per_chunk": 4.938, "boundary_entropy": 0.13145, "router_noise": 0.0, "grad_norm": 0.797, "lr": 0.0001142, "router_trainable": true, "sec_per_step": 0.094, "alpha": 0.03, "t": 386.8} |
| {"event": "train", "step": 3900, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.40729, "bpb": 2.03029, "chunk_loss": 0.030619, "F": 0.21378, "G": 0.22139, "chunks": 218.91, "bytes_per_chunk": 4.678, "boundary_entropy": 0.12344, "router_noise": 0.0, "grad_norm": 0.6995, "lr": 0.0001108, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 391.2} |
| {"event": "train", "step": 3950, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.30006, "bpb": 1.87559, "chunk_loss": 0.030507, "F": 0.2077, "G": 0.21475, "chunks": 212.69, "bytes_per_chunk": 4.815, "boundary_entropy": 0.12045, "router_noise": 0.0, "grad_norm": 0.6009, "lr": 0.0001074, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 395.6} |
| {"event": "probe", "step": 4000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3752.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.215595, "cps": 29.3125, "mean_p": 0.220274, "boundary_entropy": 0.116458, "bpb": 1.935511}, "tr": {"chunks": 3655.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.2019, "cps": 28.554688, "mean_p": 0.213657, "boundary_entropy": 0.167252, "bpb": 2.095572}, "fi": {"chunks": 4766.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.253362, "cps": 37.234375, "mean_p": 0.247755, "boundary_entropy": 0.132236, "bpb": 2.205276}}, "bpb": {"en": 1.93551, "tr": 2.09557, "fi": 2.20528}, "gini_cpc": 0.051141, "gini_cps": 0.060845, "bpic": 4.660478, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9364791667721791, "fi": 1.1751774383699416}, "t": 400.1} |
| {"event": "eval", "step": 4000, "phase": 2, "bpb": {"en": 1.94841, "tr": 2.10216, "fi": 2.21474}, "bytes_per_chunk": {"en": 4.613, "tr": 5.757, "fi": 4.372}, "bpb_mean": 2.08843, "frozen_inventory": false, "t": 400.7} |
| {"event": "train", "step": 4000, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.36041, "bpb": 1.96266, "chunk_loss": 0.030636, "F": 0.21365, "G": 0.22108, "chunks": 218.78, "bytes_per_chunk": 4.68, "boundary_entropy": 0.12442, "router_noise": 0.0, "grad_norm": 0.7473, "lr": 0.000104, "router_trainable": true, "sec_per_step": 0.109, "alpha": 0.03, "t": 401.0} |
| {"event": "train", "step": 4050, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.44145, "bpb": 2.07958, "chunk_loss": 0.030669, "F": 0.21463, "G": 0.22051, "chunks": 219.78, "bytes_per_chunk": 4.659, "boundary_entropy": 0.12305, "router_noise": 0.0, "grad_norm": 0.7171, "lr": 0.0001007, "router_trainable": true, "sec_per_step": 0.085, "alpha": 0.03, "t": 405.2} |
| {"event": "probe", "step": 4096, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3711.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.213239, "cps": 28.992188, "mean_p": 0.218098, "boundary_entropy": 0.115647, "bpb": 1.928508}, "tr": {"chunks": 3643.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.201237, "cps": 28.460938, "mean_p": 0.213236, "boundary_entropy": 0.169805, "bpb": 2.090935}, "fi": {"chunks": 4695.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.249588, "cps": 36.679688, "mean_p": 0.245111, "boundary_entropy": 0.132677, "bpb": 2.197451}}, "bpb": {"en": 1.92851, "tr": 2.09093, "fi": 2.19745}, "gini_cpc": 0.04854, "gini_cps": 0.058207, "bpic": 4.708441, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9437170156882307, "fi": 1.1704608154459237}, "t": 409.8} |
| {"event": "train", "step": 4100, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.37073, "bpb": 1.97754, "chunk_loss": 0.030483, "F": 0.20755, "G": 0.21259, "chunks": 212.53, "bytes_per_chunk": 4.818, "boundary_entropy": 0.11597, "router_noise": 0.0, "grad_norm": 0.6947, "lr": 9.75e-05, "router_trainable": true, "sec_per_step": 0.108, "alpha": 0.03, "t": 410.6} |
| {"event": "train", "step": 4150, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.36038, "bpb": 1.96261, "chunk_loss": 0.030529, "F": 0.20697, "G": 0.21596, "chunks": 211.94, "bytes_per_chunk": 4.832, "boundary_entropy": 0.12137, "router_noise": 0.0, "grad_norm": 0.6563, "lr": 9.43e-05, "router_trainable": true, "sec_per_step": 0.091, "alpha": 0.03, "t": 415.2} |
| {"event": "probe", "step": 4200, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3716.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.213526, "cps": 29.03125, "mean_p": 0.218317, "boundary_entropy": 0.112324, "bpb": 1.923883}, "tr": {"chunks": 3502.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.193449, "cps": 27.359375, "mean_p": 0.21009, "boundary_entropy": 0.166445, "bpb": 2.072089}, "fi": {"chunks": 4680.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.248791, "cps": 36.5625, "mean_p": 0.24453, "boundary_entropy": 0.131604, "bpb": 2.186975}}, "bpb": {"en": 1.92388, "tr": 2.07209, "fi": 2.18698}, "gini_cpc": 0.056262, "gini_cps": 0.066005, "bpic": 4.768196, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9059703929559398, "fi": 1.1651514621766723}, "t": 420.0} |
| {"event": "eval", "step": 4200, "phase": 2, "bpb": {"en": 1.93867, "tr": 2.07952, "fi": 2.19563}, "bytes_per_chunk": {"en": 4.655, "tr": 5.75, "fi": 4.463}, "bpb_mean": 2.07127, "frozen_inventory": false, "t": 420.5} |
| {"event": "phase_transition", "step": 4200, "phase": 3, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 420.8} |
| {"event": "train", "step": 4200, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.38065, "bpb": 1.99185, "chunk_loss": 0.030451, "F": 0.20184, "G": 0.21273, "chunks": 206.69, "bytes_per_chunk": 4.954, "boundary_entropy": 0.12639, "router_noise": 0.0, "grad_norm": 0.6991, "lr": 9.12e-05, "router_trainable": true, "sec_per_step": 0.113, "alpha": 0.03, "t": 420.8} |
| {"event": "train", "step": 4250, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.40896, "bpb": 2.03271, "chunk_loss": 0.030707, "F": 0.21429, "G": 0.22222, "chunks": 219.44, "bytes_per_chunk": 4.666, "boundary_entropy": 0.11887, "router_noise": 0.0, "grad_norm": 0.7075, "lr": 8.81e-05, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 425.2} |
| {"event": "train", "step": 4300, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.37141, "bpb": 1.97853, "chunk_loss": 0.030528, "F": 0.20831, "G": 0.21552, "chunks": 213.31, "bytes_per_chunk": 4.8, "boundary_entropy": 0.12005, "router_noise": 0.0, "grad_norm": 0.7053, "lr": 8.5e-05, "router_trainable": true, "sec_per_step": 0.08, "alpha": 0.03, "t": 429.3} |
| {"event": "train", "step": 4350, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.41568, "bpb": 2.0424, "chunk_loss": 0.030521, "F": 0.20599, "G": 0.21419, "chunks": 210.94, "bytes_per_chunk": 4.855, "boundary_entropy": 0.11644, "router_noise": 0.0, "grad_norm": 0.6658, "lr": 8.21e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 433.5} |
| {"event": "probe", "step": 4400, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3690.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.212032, "cps": 28.828125, "mean_p": 0.216277, "boundary_entropy": 0.111344, "bpb": 1.914898}, "tr": {"chunks": 3747.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.206982, "cps": 29.273438, "mean_p": 0.215043, "boundary_entropy": 0.166443, "bpb": 2.060257}, "fi": {"chunks": 4756.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.252831, "cps": 37.15625, "mean_p": 0.247357, "boundary_entropy": 0.126994, "bpb": 2.171307}}, "bpb": {"en": 1.9149, "tr": 2.06026, "fi": 2.17131}, "gini_cpc": 0.045495, "gini_cps": 0.058285, "bpic": 4.652834, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9761822255575482, "fi": 1.1924157850902841}, "t": 438.2} |
| {"event": "eval", "step": 4400, "phase": 3, "bpb": {"en": 1.9289, "tr": 2.06857, "fi": 2.17956}, "bytes_per_chunk": {"en": 4.671, "tr": 5.651, "fi": 4.406}, "bpb_mean": 2.05901, "frozen_inventory": false, "t": 438.8} |
| {"event": "train", "step": 4400, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.35335, "bpb": 1.95246, "chunk_loss": 0.030642, "F": 0.2135, "G": 0.22085, "chunks": 218.62, "bytes_per_chunk": 4.684, "boundary_entropy": 0.11562, "router_noise": 0.0, "grad_norm": 0.6624, "lr": 7.92e-05, "router_trainable": true, "sec_per_step": 0.11, "alpha": 0.03, "t": 439.0} |
| {"event": "train", "step": 4450, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.32335, "bpb": 1.90918, "chunk_loss": 0.030458, "F": 0.20428, "G": 0.2112, "chunks": 209.19, "bytes_per_chunk": 4.895, "boundary_entropy": 0.11301, "router_noise": 0.0, "grad_norm": 0.6913, "lr": 7.63e-05, "router_trainable": true, "sec_per_step": 0.082, "alpha": 0.03, "t": 443.2} |
| {"event": "train", "step": 4500, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.42483, "bpb": 2.0556, "chunk_loss": 0.03068, "F": 0.21555, "G": 0.22097, "chunks": 220.72, "bytes_per_chunk": 4.639, "boundary_entropy": 0.11791, "router_noise": 0.0, "grad_norm": 0.7324, "lr": 7.36e-05, "router_trainable": true, "sec_per_step": 0.081, "alpha": 0.03, "t": 447.2} |
| {"event": "train", "step": 4550, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.36815, "bpb": 1.97383, "chunk_loss": 0.030668, "F": 0.21423, "G": 0.21979, "chunks": 219.38, "bytes_per_chunk": 4.668, "boundary_entropy": 0.11429, "router_noise": 0.0, "grad_norm": 0.6881, "lr": 7.09e-05, "router_trainable": true, "sec_per_step": 0.087, "alpha": 0.03, "t": 451.5} |
| {"event": "probe", "step": 4600, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3659.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.210251, "cps": 28.585938, "mean_p": 0.215487, "boundary_entropy": 0.10887, "bpb": 1.906908}, "tr": {"chunks": 3556.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.196432, "cps": 27.78125, "mean_p": 0.211236, "boundary_entropy": 0.160659, "bpb": 2.048889}, "fi": {"chunks": 4776.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.253894, "cps": 37.3125, "mean_p": 0.246915, "boundary_entropy": 0.124307, "bpb": 2.159032}}, "bpb": {"en": 1.90691, "tr": 2.04889, "fi": 2.15903}, "gini_cpc": 0.057992, "gini_cps": 0.067829, "bpic": 4.731215, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9342710927904168, "fi": 1.2075750889710013}, "t": 456.3} |
| {"event": "eval", "step": 4600, "phase": 3, "bpb": {"en": 1.92125, "tr": 2.05782, "fi": 2.16632}, "bytes_per_chunk": {"en": 4.661, "tr": 5.78, "fi": 4.387}, "bpb_mean": 2.04846, "frozen_inventory": false, "t": 456.8} |
| {"event": "train", "step": 4600, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.39718, "bpb": 2.0157, "chunk_loss": 0.030678, "F": 0.21475, "G": 0.21845, "chunks": 219.91, "bytes_per_chunk": 4.657, "boundary_entropy": 0.10776, "router_noise": 0.0, "grad_norm": 0.6951, "lr": 6.82e-05, "router_trainable": true, "sec_per_step": 0.112, "alpha": 0.03, "t": 457.2} |
| {"event": "train", "step": 4650, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.40207, "bpb": 2.02276, "chunk_loss": 0.030523, "F": 0.20853, "G": 0.21481, "chunks": 213.53, "bytes_per_chunk": 4.796, "boundary_entropy": 0.11417, "router_noise": 0.0, "grad_norm": 0.6996, "lr": 6.57e-05, "router_trainable": true, "sec_per_step": 0.083, "alpha": 0.03, "t": 461.3} |
| {"event": "train", "step": 4700, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.42459, "bpb": 2.05525, "chunk_loss": 0.030759, "F": 0.21686, "G": 0.22371, "chunks": 222.06, "bytes_per_chunk": 4.611, "boundary_entropy": 0.11392, "router_noise": 0.0, "grad_norm": 0.7113, "lr": 6.32e-05, "router_trainable": true, "sec_per_step": 0.081, "alpha": 0.03, "t": 465.4} |
| {"event": "train", "step": 4750, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.37767, "bpb": 1.98755, "chunk_loss": 0.030527, "F": 0.20468, "G": 0.21569, "chunks": 209.59, "bytes_per_chunk": 4.886, "boundary_entropy": 0.11645, "router_noise": 0.0, "grad_norm": 0.6705, "lr": 6.08e-05, "router_trainable": true, "sec_per_step": 0.085, "alpha": 0.03, "t": 469.6} |
| {"event": "probe", "step": 4800, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3588.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.206171, "cps": 28.03125, "mean_p": 0.213229, "boundary_entropy": 0.109577, "bpb": 1.901161}, "tr": {"chunks": 3543.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.195713, "cps": 27.679688, "mean_p": 0.209368, "boundary_entropy": 0.159366, "bpb": 2.036853}, "fi": {"chunks": 4748.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.252406, "cps": 37.09375, "mean_p": 0.24662, "boundary_entropy": 0.12703, "bpb": 2.152207}}, "bpb": {"en": 1.90116, "tr": 2.03685, "fi": 2.15221}, "gini_cpc": 0.057764, "gini_cps": 0.067626, "bpic": 4.775823, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9492755316705948, "fi": 1.2242511275246517}, "t": 474.3} |
| {"event": "eval", "step": 4800, "phase": 3, "bpb": {"en": 1.91558, "tr": 2.04505, "fi": 2.16019}, "bytes_per_chunk": {"en": 4.804, "tr": 5.712, "fi": 4.408}, "bpb_mean": 2.04027, "frozen_inventory": false, "t": 475.0} |
| {"event": "train", "step": 4800, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.36906, "bpb": 1.97513, "chunk_loss": 0.030536, "F": 0.2077, "G": 0.21693, "chunks": 212.69, "bytes_per_chunk": 4.815, "boundary_entropy": 0.11669, "router_noise": 0.0, "grad_norm": 0.6518, "lr": 5.85e-05, "router_trainable": true, "sec_per_step": 0.113, "alpha": 0.03, "t": 475.3} |
| {"event": "train", "step": 4850, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.41497, "bpb": 2.04137, "chunk_loss": 0.03069, "F": 0.21811, "G": 0.22313, "chunks": 223.34, "bytes_per_chunk": 4.585, "boundary_entropy": 0.10805, "router_noise": 0.0, "grad_norm": 0.6946, "lr": 5.62e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 479.6} |
| {"event": "train", "step": 4900, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.37436, "bpb": 1.98278, "chunk_loss": 0.030463, "F": 0.20279, "G": 0.21395, "chunks": 207.66, "bytes_per_chunk": 4.931, "boundary_entropy": 0.1152, "router_noise": 0.0, "grad_norm": 0.703, "lr": 5.41e-05, "router_trainable": true, "sec_per_step": 0.085, "alpha": 0.03, "t": 483.8} |
| {"event": "train", "step": 4950, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.40256, "bpb": 2.02347, "chunk_loss": 0.030587, "F": 0.20856, "G": 0.21711, "chunks": 213.56, "bytes_per_chunk": 4.795, "boundary_entropy": 0.10838, "router_noise": 0.0, "grad_norm": 0.7118, "lr": 5.2e-05, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 488.2} |
| {"event": "probe", "step": 5000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3593.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.206459, "cps": 28.070312, "mean_p": 0.21262, "boundary_entropy": 0.106241, "bpb": 1.892908}, "tr": {"chunks": 3549.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.196045, "cps": 27.726562, "mean_p": 0.209012, "boundary_entropy": 0.157954, "bpb": 2.026019}, "fi": {"chunks": 4672.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.248365, "cps": 36.5, "mean_p": 0.244353, "boundary_entropy": 0.123532, "bpb": 2.137405}}, "bpb": {"en": 1.89291, "tr": 2.02602, "fi": 2.13741}, "gini_cpc": 0.05359, "gini_cps": 0.063371, "bpic": 4.802099, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9495598669326997, "fi": 1.2029784670258323}, "t": 492.7} |
| {"event": "eval", "step": 5000, "phase": 3, "bpb": {"en": 1.90763, "tr": 2.03331, "fi": 2.14456}, "bytes_per_chunk": {"en": 4.799, "tr": 5.714, "fi": 4.475}, "bpb_mean": 2.0285, "frozen_inventory": false, "t": 493.3} |
| {"event": "train", "step": 5000, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.36202, "bpb": 1.96498, "chunk_loss": 0.03057, "F": 0.20944, "G": 0.21853, "chunks": 214.47, "bytes_per_chunk": 4.775, "boundary_entropy": 0.11236, "router_noise": 0.0, "grad_norm": 0.6799, "lr": 5e-05, "router_trainable": true, "sec_per_step": 0.108, "alpha": 0.03, "t": 493.6} |
| {"event": "train", "step": 5050, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.32938, "bpb": 1.91789, "chunk_loss": 0.030525, "F": 0.21048, "G": 0.21608, "chunks": 215.53, "bytes_per_chunk": 4.751, "boundary_entropy": 0.10314, "router_noise": 0.0, "grad_norm": 0.6651, "lr": 4.81e-05, "router_trainable": true, "sec_per_step": 0.084, "alpha": 0.03, "t": 497.8} |
| {"event": "train", "step": 5100, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.40133, "bpb": 2.02169, "chunk_loss": 0.030518, "F": 0.20673, "G": 0.21538, "chunks": 211.69, "bytes_per_chunk": 4.837, "boundary_entropy": 0.10956, "router_noise": 0.0, "grad_norm": 0.6744, "lr": 4.63e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 502.1} |
| {"event": "train", "step": 5150, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.3647, "bpb": 1.96884, "chunk_loss": 0.030619, "F": 0.20395, "G": 0.21344, "chunks": 208.84, "bytes_per_chunk": 4.903, "boundary_entropy": 0.10658, "router_noise": 0.0, "grad_norm": 0.6794, "lr": 4.45e-05, "router_trainable": true, "sec_per_step": 0.081, "alpha": 0.03, "t": 506.2} |
| {"event": "probe", "step": 5200, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3458.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.198701, "cps": 27.015625, "mean_p": 0.207744, "boundary_entropy": 0.101045, "bpb": 1.888631}, "tr": {"chunks": 3564.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.196873, "cps": 27.84375, "mean_p": 0.210238, "boundary_entropy": 0.157605, "bpb": 2.017727}, "fi": {"chunks": 4698.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.249747, "cps": 36.703125, "mean_p": 0.243855, "boundary_entropy": 0.122818, "bpb": 2.125672}}, "bpb": {"en": 1.88863, "tr": 2.01773, "fi": 2.12567}, "gini_cpc": 0.054623, "gini_cps": 0.070535, "bpic": 4.840614, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9908006325988806, "fi": 1.2568986514326446}, "t": 511.0} |
| {"event": "eval", "step": 5200, "phase": 3, "bpb": {"en": 1.90389, "tr": 2.02605, "fi": 2.13372}, "bytes_per_chunk": {"en": 4.9, "tr": 5.659, "fi": 4.468}, "bpb_mean": 2.02122, "frozen_inventory": false, "t": 511.7} |
| {"event": "train", "step": 5200, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.34784, "bpb": 1.94452, "chunk_loss": 0.030431, "F": 0.20102, "G": 0.20869, "chunks": 205.84, "bytes_per_chunk": 4.975, "boundary_entropy": 0.10335, "router_noise": 0.0, "grad_norm": 0.7118, "lr": 4.29e-05, "router_trainable": true, "sec_per_step": 0.115, "alpha": 0.03, "t": 511.9} |
| {"event": "train", "step": 5250, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.36218, "bpb": 1.9652, "chunk_loss": 0.030424, "F": 0.20151, "G": 0.2108, "chunks": 206.34, "bytes_per_chunk": 4.963, "boundary_entropy": 0.10915, "router_noise": 0.0, "grad_norm": 0.7175, "lr": 4.14e-05, "router_trainable": true, "sec_per_step": 0.099, "alpha": 0.03, "t": 516.9} |
| {"event": "train", "step": 5300, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.33976, "bpb": 1.93286, "chunk_loss": 0.030504, "F": 0.20712, "G": 0.21444, "chunks": 212.09, "bytes_per_chunk": 4.828, "boundary_entropy": 0.10828, "router_noise": 0.0, "grad_norm": 0.7041, "lr": 3.99e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 521.2} |
| {"event": "train", "step": 5350, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.34614, "bpb": 1.94208, "chunk_loss": 0.030536, "F": 0.20755, "G": 0.21504, "chunks": 212.53, "bytes_per_chunk": 4.818, "boundary_entropy": 0.10309, "router_noise": 0.0, "grad_norm": 0.6916, "lr": 3.86e-05, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 525.6} |
| {"event": "probe", "step": 5400, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3508.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.201574, "cps": 27.40625, "mean_p": 0.208086, "boundary_entropy": 0.1002, "bpb": 1.885172}, "tr": {"chunks": 3516.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.194222, "cps": 27.46875, "mean_p": 0.209383, "boundary_entropy": 0.15915, "bpb": 2.01167}, "fi": {"chunks": 4656.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.247515, "cps": 36.375, "mean_p": 0.243414, "boundary_entropy": 0.124328, "bpb": 2.120241}}, "bpb": {"en": 1.88517, "tr": 2.01167, "fi": 2.12024}, "gini_cpc": 0.055228, "gini_cps": 0.065525, "bpic": 4.857192, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9635246959766711, "fi": 1.2279074199470978}, "t": 530.2} |
| {"event": "eval", "step": 5400, "phase": 3, "bpb": {"en": 1.89961, "tr": 2.01874, "fi": 2.12783}, "bytes_per_chunk": {"en": 4.893, "tr": 5.749, "fi": 4.503}, "bpb_mean": 2.01539, "frozen_inventory": false, "t": 530.9} |
| {"event": "train", "step": 5400, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.34257, "bpb": 1.93692, "chunk_loss": 0.030458, "F": 0.20416, "G": 0.21302, "chunks": 209.06, "bytes_per_chunk": 4.898, "boundary_entropy": 0.10649, "router_noise": 0.0, "grad_norm": 0.7093, "lr": 3.73e-05, "router_trainable": true, "sec_per_step": 0.112, "alpha": 0.03, "t": 531.2} |
| {"event": "train", "step": 5450, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.35466, "bpb": 1.95436, "chunk_loss": 0.030471, "F": 0.20361, "G": 0.21064, "chunks": 208.5, "bytes_per_chunk": 4.911, "boundary_entropy": 0.10456, "router_noise": 0.0, "grad_norm": 0.683, "lr": 3.62e-05, "router_trainable": true, "sec_per_step": 0.087, "alpha": 0.03, "t": 535.6} |
| {"event": "train", "step": 5500, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.35472, "bpb": 1.95445, "chunk_loss": 0.030592, "F": 0.2066, "G": 0.21428, "chunks": 211.56, "bytes_per_chunk": 4.84, "boundary_entropy": 0.10832, "router_noise": 0.0, "grad_norm": 0.7233, "lr": 3.51e-05, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 540.1} |
| {"event": "train", "step": 5550, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.30365, "bpb": 1.88078, "chunk_loss": 0.030461, "F": 0.20251, "G": 0.21192, "chunks": 207.38, "bytes_per_chunk": 4.938, "boundary_entropy": 0.10586, "router_noise": 0.0, "grad_norm": 0.6886, "lr": 3.41e-05, "router_trainable": true, "sec_per_step": 0.083, "alpha": 0.03, "t": 544.2} |
| {"event": "probe", "step": 5600, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3478.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.199851, "cps": 27.171875, "mean_p": 0.208413, "boundary_entropy": 0.099191, "bpb": 1.876684}, "tr": {"chunks": 3541.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.195603, "cps": 27.664062, "mean_p": 0.210952, "boundary_entropy": 0.158678, "bpb": 2.005737}, "fi": {"chunks": 4642.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.246771, "cps": 36.265625, "mean_p": 0.243223, "boundary_entropy": 0.123699, "bpb": 2.114076}}, "bpb": {"en": 1.87668, "tr": 2.00574, "fi": 2.11408}, "gini_cpc": 0.053115, "gini_cps": 0.066547, "bpic": 4.865106, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9787458145147773, "fi": 1.234774907038872}, "t": 548.9} |
| {"event": "eval", "step": 5600, "phase": 3, "bpb": {"en": 1.89198, "tr": 2.01223, "fi": 2.12157}, "bytes_per_chunk": {"en": 4.874, "tr": 5.713, "fi": 4.532}, "bpb_mean": 2.00859, "frozen_inventory": false, "t": 549.5} |
| {"event": "train", "step": 5600, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.38103, "bpb": 1.99241, "chunk_loss": 0.030699, "F": 0.21164, "G": 0.2198, "chunks": 216.72, "bytes_per_chunk": 4.725, "boundary_entropy": 0.10522, "router_noise": 0.0, "grad_norm": 0.7581, "lr": 3.33e-05, "router_trainable": true, "sec_per_step": 0.115, "alpha": 0.03, "t": 549.9} |
| {"event": "train", "step": 5650, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.35289, "bpb": 1.95181, "chunk_loss": 0.030408, "F": 0.19946, "G": 0.2098, "chunks": 204.25, "bytes_per_chunk": 5.013, "boundary_entropy": 0.10634, "router_noise": 0.0, "grad_norm": 0.7208, "lr": 3.25e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 554.3} |
| {"event": "train", "step": 5700, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.36391, "bpb": 1.9677, "chunk_loss": 0.03052, "F": 0.20694, "G": 0.21516, "chunks": 211.91, "bytes_per_chunk": 4.832, "boundary_entropy": 0.10293, "router_noise": 0.0, "grad_norm": 0.7442, "lr": 3.18e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 558.6} |
| {"event": "train", "step": 5750, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.36651, "bpb": 1.97145, "chunk_loss": 0.030526, "F": 0.20587, "G": 0.2158, "chunks": 210.81, "bytes_per_chunk": 4.857, "boundary_entropy": 0.10703, "router_noise": 0.0, "grad_norm": 0.7058, "lr": 3.13e-05, "router_trainable": true, "sec_per_step": 0.083, "alpha": 0.03, "t": 562.7} |
| {"event": "probe", "step": 5800, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3481.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.200023, "cps": 27.195312, "mean_p": 0.208651, "boundary_entropy": 0.098009, "bpb": 1.87195}, "tr": {"chunks": 3547.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.195934, "cps": 27.710938, "mean_p": 0.2091, "boundary_entropy": 0.154636, "bpb": 1.996694}, "fi": {"chunks": 4643.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.246824, "cps": 36.273438, "mean_p": 0.243733, "boundary_entropy": 0.122679, "bpb": 2.110577}}, "bpb": {"en": 1.87195, "tr": 1.99669, "fi": 2.11058}, "gini_cpc": 0.05278, "gini_cps": 0.066375, "bpic": 4.860937, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9795593039751482, "fi": 1.2339765230851147}, "t": 567.6} |
| {"event": "eval", "step": 5800, "phase": 3, "bpb": {"en": 1.88697, "tr": 2.00427, "fi": 2.11822}, "bytes_per_chunk": {"en": 4.88, "tr": 5.722, "fi": 4.552}, "bpb_mean": 2.00316, "frozen_inventory": false, "t": 568.1} |
| {"event": "train", "step": 5800, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.36218, "bpb": 1.96521, "chunk_loss": 0.030451, "F": 0.20148, "G": 0.21163, "chunks": 206.31, "bytes_per_chunk": 4.963, "boundary_entropy": 0.11001, "router_noise": 0.0, "grad_norm": 0.68, "lr": 3.08e-05, "router_trainable": true, "sec_per_step": 0.112, "alpha": 0.03, "t": 568.3} |
| {"event": "train", "step": 5850, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.28788, "bpb": 1.85802, "chunk_loss": 0.030356, "F": 0.19894, "G": 0.20821, "chunks": 203.72, "bytes_per_chunk": 5.027, "boundary_entropy": 0.10358, "router_noise": 0.0, "grad_norm": 0.6749, "lr": 3.05e-05, "router_trainable": true, "sec_per_step": 0.087, "alpha": 0.03, "t": 572.7} |
| {"event": "train", "step": 5900, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.35461, "bpb": 1.95429, "chunk_loss": 0.030429, "F": 0.2037, "G": 0.21036, "chunks": 208.59, "bytes_per_chunk": 4.909, "boundary_entropy": 0.09833, "router_noise": 0.0, "grad_norm": 0.6668, "lr": 3.02e-05, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 577.1} |
| {"event": "train", "step": 5950, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.35902, "bpb": 1.96065, "chunk_loss": 0.030455, "F": 0.20297, "G": 0.2107, "chunks": 207.84, "bytes_per_chunk": 4.927, "boundary_entropy": 0.09854, "router_noise": 0.0, "grad_norm": 0.7319, "lr": 3.01e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 581.4} |
| {"event": "train", "step": 5999, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.33177, "bpb": 1.92134, "chunk_loss": 0.030297, "F": 0.19476, "G": 0.20455, "chunks": 199.44, "bytes_per_chunk": 5.134, "boundary_entropy": 0.10597, "router_noise": 0.0, "grad_norm": 0.7249, "lr": 3e-05, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 584.9} |
| {"event": "probe", "step": 6000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3496.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.200885, "cps": 27.3125, "mean_p": 0.207238, "boundary_entropy": 0.098353, "bpb": 1.867491}, "tr": {"chunks": 3521.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.194498, "cps": 27.507812, "mean_p": 0.208309, "boundary_entropy": 0.156115, "bpb": 1.996591}, "fi": {"chunks": 4572.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.243049, "cps": 35.71875, "mean_p": 0.241148, "boundary_entropy": 0.123854, "bpb": 2.105436}}, "bpb": {"en": 1.86749, "tr": 1.99659, "fi": 2.10544}, "gini_cpc": 0.050698, "gini_cps": 0.061898, "bpic": 4.895332, "worst_lang": "fi", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9682068922669934, "fi": 1.20989319628578}, "t": 585.3} |
| {"event": "eval", "step": 6000, "phase": 3, "bpb": {"en": 1.8835, "tr": 2.00416, "fi": 2.1146}, "bytes_per_chunk": {"en": 4.912, "tr": 5.749, "fi": 4.602}, "bpb_mean": 2.00075, "frozen_inventory": false, "t": 585.9} |
| {"event": "reps_dumped", "dirs": ["/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s3/probe/reps/th0.30", "/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s3/probe/reps/th0.40", "/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s3/probe/reps/th0.50", "/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s3/probe/reps/th0.60", "/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s3/probe/reps/th0.70"], "t": 587.0} |
| {"event": "final", "run": "ali_hnet_s3", "arch": "hnet", "objective": "ratio", "condition": "A", "step": 6000, "params": 22516992, "seed": 3, "per_lang": {"en": {"cpc": 0.20088490490145378, "cps": 27.3125, "bpb": 1.8884673001084857, "bytes_per_chunk": 4.852361913223753}, "tr": {"cpc": 0.19449814947798708, "cps": 27.5078125, "bpb": 1.9947880605123576, "bytes_per_chunk": 5.749528446725447}, "fi": {"cpc": 0.24304927967678486, "cps": 35.71875, "bpb": 2.132476768853844, "bytes_per_chunk": 4.580414631802417}}, "bpb_mean": 2.0052440431582292, "gini_cpc": 0.05069827827623996, "gini_cps": 0.06189777662726148, "bpic": 4.8953317801363365, "mean_boundary_entropy": 0.10596814751625061, "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": 588.1204781532288, "t": 588.1} |
| {"event": "hub_done", "repo": "AdaptiveChunking/ali_hnet_s3", "pushed": [0, 1, 2, 4, 6, 8, 16, 32, 64, 128, 256, 512, 1000, 1024, 2048, 3000, 4096, 6000], "failed": [], "ok": true, "t": 604.1} |
|
|
