| {"event": "hub", "repo": "AdaptiveChunking/ali_hnet_s1", "keep_steps": [0, 1, 2, 4, 6, 8, 16, 32, 64, 128, 256, 512, 1000, 1024, 2048, 3000, 4096, 6000], "t": 2.3} |
| {"event": "start", "run": "ali_hnet_s1", "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": 2.3} |
| {"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": 2.3} |
| {"event": "probe", "step": 0, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2674.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.153652, "cps": 20.890625, "mean_p": 0.478126, "boundary_entropy": 0.685438, "bpb": 8.041379}, "tr": {"chunks": 5645.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.311827, "cps": 44.101562, "mean_p": 0.486063, "boundary_entropy": 0.686375, "bpb": 8.01701}, "fi": {"chunks": 4732.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.251555, "cps": 36.96875, "mean_p": 0.48451, "boundary_entropy": 0.686595, "bpb": 8.075294}}, "bpb": {"en": 8.04138, "tr": 8.01701, "fi": 8.07529}, "gini_cpc": 0.147064, "gini_cps": 0.151764, "bpic": 4.346947, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 2.029439514461233, "fi": 1.6371767549398106}, "t": 3.3} |
| {"event": "train", "step": 0, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 5.58509, "bpb": 8.05758, "chunk_loss": 0.030141, "F": 0.16724, "G": 0.47713, "chunks": 171.25, "bytes_per_chunk": 5.98, "boundary_entropy": 0.69022, "router_noise": 0.0, "grad_norm": 11.3983, "lr": 1e-06, "router_trainable": true, "sec_per_step": 0.007, "alpha": 0.03, "t": 3.7} |
| {"event": "probe", "step": 1, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2658.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.152732, "cps": 20.765625, "mean_p": 0.478104, "boundary_entropy": 0.685439, "bpb": 8.008088}, "tr": {"chunks": 5623.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.310612, "cps": 43.929688, "mean_p": 0.485999, "boundary_entropy": 0.686373, "bpb": 7.993602}, "fi": {"chunks": 4707.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.250226, "cps": 36.773438, "mean_p": 0.484401, "boundary_entropy": 0.686587, "bpb": 8.047689}}, "bpb": {"en": 8.00809, "tr": 7.9936, "fi": 8.04769}, "gini_cpc": 0.147502, "gini_cps": 0.152192, "bpic": 4.368032, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 2.0336990027842936, "fi": 1.638330281511238}, "t": 4.0} |
| {"event": "probe", "step": 2, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2658.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.152732, "cps": 20.765625, "mean_p": 0.478077, "boundary_entropy": 0.685441, "bpb": 7.943231}, "tr": {"chunks": 5590.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.308789, "cps": 43.671875, "mean_p": 0.485897, "boundary_entropy": 0.68637, "bpb": 7.946018}, "fi": {"chunks": 4666.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.248046, "cps": 36.453125, "mean_p": 0.484223, "boundary_entropy": 0.686575, "bpb": 7.994001}}, "bpb": {"en": 7.94323, "tr": 7.94602, "fi": 7.994}, "gini_cpc": 0.146621, "gini_cps": 0.15136, "bpic": 4.393062, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 2.021763724980295, "fi": 1.6240597181923597}, "t": 4.6} |
| {"event": "probe", "step": 4, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2616.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.150319, "cps": 20.4375, "mean_p": 0.478039, "boundary_entropy": 0.685454, "bpb": 7.727414}, "tr": {"chunks": 5457.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.301442, "cps": 42.632812, "mean_p": 0.485571, "boundary_entropy": 0.686365, "bpb": 7.790785}, "fi": {"chunks": 4537.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.241189, "cps": 35.445312, "mean_p": 0.483674, "boundary_entropy": 0.686539, "bpb": 7.814343}}, "bpb": {"en": 7.72741, "tr": 7.79078, "fi": 7.81434}, "gini_cpc": 0.145391, "gini_cps": 0.150198, "bpic": 4.498969, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 2.005348155584735, "fi": 1.6045131337851806}, "t": 5.4} |
| {"event": "probe", "step": 6, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2577.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.148078, "cps": 20.132812, "mean_p": 0.47806, "boundary_entropy": 0.685487, "bpb": 7.444462}, "tr": {"chunks": 5240.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.289455, "cps": 40.9375, "mean_p": 0.485071, "boundary_entropy": 0.686358, "bpb": 7.58975}, "fi": {"chunks": 4282.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.227633, "cps": 33.453125, "mean_p": 0.482818, "boundary_entropy": 0.686483, "bpb": 7.577043}}, "bpb": {"en": 7.44446, "tr": 7.58975, "fi": 7.57704}, "gini_cpc": 0.141696, "gini_cps": 0.146734, "bpic": 4.688983, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.9547464685488425, "fi": 1.5372499272680533}, "t": 6.3} |
| {"event": "probe", "step": 8, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2622.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.150664, "cps": 20.484375, "mean_p": 0.478833, "boundary_entropy": 0.685611, "bpb": 7.169775}, "tr": {"chunks": 5153.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.284649, "cps": 40.257812, "mean_p": 0.484936, "boundary_entropy": 0.68639, "bpb": 7.398399}, "fi": {"chunks": 4101.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.218011, "cps": 32.039062, "mean_p": 0.482314, "boundary_entropy": 0.686465, "bpb": 7.341459}}, "bpb": {"en": 7.16977, "tr": 7.3984, "fi": 7.34146}, "gini_cpc": 0.136722, "gini_cps": 0.142079, "bpic": 4.777029, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.8893004320181075, "fi": 1.4470026240793188}, "t": 7.0} |
| {"event": "probe", "step": 16, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 2600.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.1494, "cps": 20.3125, "mean_p": 0.481106, "boundary_entropy": 0.686032, "bpb": 6.515077}, "tr": {"chunks": 4392.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.242612, "cps": 34.3125, "mean_p": 0.482961, "boundary_entropy": 0.686438, "bpb": 6.898325}, "fi": {"chunks": 2928.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.155654, "cps": 22.875, "mean_p": 0.478276, "boundary_entropy": 0.686272, "bpb": 6.694056}}, "bpb": {"en": 6.51508, "tr": 6.89832, "fi": 6.69406}, "gini_cpc": 0.113466, "gini_cps": 0.12043, "bpic": 5.718952, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.6239122287423673, "fi": 1.0418614313229166}, "t": 8.2} |
| {"event": "probe", "step": 32, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3677.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.211285, "cps": 28.726562, "mean_p": 0.490434, "boundary_entropy": 0.687168, "bpb": 6.161268}, "tr": {"chunks": 3936.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.217423, "cps": 30.75, "mean_p": 0.486196, "boundary_entropy": 0.68728, "bpb": 6.577363}, "fi": {"chunks": 3146.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.167243, "cps": 24.578125, "mean_p": 0.485011, "boundary_entropy": 0.687293, "bpb": 6.341649}}, "bpb": {"en": 6.16127, "tr": 6.57736, "fi": 6.34165}, "gini_cpc": 0.056134, "gini_cps": 0.048951, "bpic": 5.272981, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0290465682194374, "fi": 0.7915481262612489}, "t": 10.0} |
| {"event": "train", "step": 50, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 4.09855, "bpb": 5.91295, "chunk_loss": 0.030482, "F": 0.17328, "G": 0.48629, "chunks": 177.44, "bytes_per_chunk": 5.771, "boundary_entropy": 0.69164, "router_noise": 0.0, "grad_norm": 2.3618, "lr": 5.1e-05, "router_trainable": true, "sec_per_step": 0.163, "alpha": 0.03, "t": 11.8} |
| {"event": "probe", "step": 64, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3736.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.214676, "cps": 29.1875, "mean_p": 0.484523, "boundary_entropy": 0.686191, "bpb": 5.429434}, "tr": {"chunks": 4462.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.246478, "cps": 34.859375, "mean_p": 0.481636, "boundary_entropy": 0.686154, "bpb": 5.921939}, "fi": {"chunks": 3983.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.211738, "cps": 31.117188, "mean_p": 0.482711, "boundary_entropy": 0.68657, "bpb": 5.65423}}, "bpb": {"en": 5.42943, "tr": 5.92194, "fi": 5.65423}, "gini_cpc": 0.034419, "gini_cps": 0.039734, "bpic": 4.657417, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.1481437529549268, "fi": 0.9863150854715401}, "t": 13.1} |
| {"event": "train", "step": 100, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 3.2075, "bpb": 4.62745, "chunk_loss": 0.032951, "F": 0.21274, "G": 0.46189, "chunks": 217.84, "bytes_per_chunk": 4.701, "boundary_entropy": 0.68352, "router_noise": 0.0, "grad_norm": 1.5489, "lr": 0.000101, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 16.2} |
| {"event": "probe", "step": 128, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3731.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.214388, "cps": 29.148438, "mean_p": 0.426789, "boundary_entropy": 0.622802, "bpb": 4.131319}, "tr": {"chunks": 4296.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.237309, "cps": 33.5625, "mean_p": 0.410802, "boundary_entropy": 0.642197, "bpb": 4.379207}, "fi": {"chunks": 4267.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.226835, "cps": 33.335938, "mean_p": 0.414446, "boundary_entropy": 0.644092, "bpb": 4.331886}}, "bpb": {"en": 4.13132, "tr": 4.37921, "fi": 4.33189}, "gini_cpc": 0.02252, "gini_cps": 0.030638, "bpic": 4.614609, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.1069107174078323, "fi": 1.0580583784236033}, "t": 18.5} |
| {"event": "train", "step": 150, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.74899, "bpb": 3.96595, "chunk_loss": 0.032028, "F": 0.20706, "G": 0.39699, "chunks": 212.03, "bytes_per_chunk": 4.829, "boundary_entropy": 0.57884, "router_noise": 0.0, "grad_norm": 1.7337, "lr": 0.000151, "router_trainable": true, "sec_per_step": 0.085, "alpha": 0.03, "t": 20.5} |
| {"event": "probe", "step": 200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3297.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.18945, "cps": 25.757812, "mean_p": 0.326504, "boundary_entropy": 0.46959, "bpb": 3.547963}, "tr": {"chunks": 4510.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.24913, "cps": 35.234375, "mean_p": 0.309049, "boundary_entropy": 0.487546, "bpb": 3.784123}, "fi": {"chunks": 3919.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.208336, "cps": 30.617188, "mean_p": 0.335608, "boundary_entropy": 0.532689, "bpb": 3.872132}}, "bpb": {"en": 3.54796, "tr": 3.78412, "fi": 3.87213}, "gini_cpc": 0.061502, "gini_cps": 0.068964, "bpic": 4.838137, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.3150163830999009, "fi": 1.0996856378295683}, "t": 24.3} |
| {"event": "eval", "step": 200, "phase": 1, "bpb": {"en": 3.55608, "tr": 3.79368, "fi": 3.86956}, "bytes_per_chunk": {"en": 5.146, "tr": 4.344, "fi": 5.133}, "bpb_mean": 3.73977, "frozen_inventory": false, "t": 24.7} |
| {"event": "train", "step": 200, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.48451, "bpb": 3.5844, "chunk_loss": 0.031111, "F": 0.19855, "G": 0.32836, "chunks": 203.31, "bytes_per_chunk": 5.037, "boundary_entropy": 0.48125, "router_noise": 0.0, "grad_norm": 1.2315, "lr": 0.000201, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 24.9} |
| {"event": "train", "step": 250, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.37042, "bpb": 3.41979, "chunk_loss": 0.030584, "F": 0.18735, "G": 0.29318, "chunks": 191.84, "bytes_per_chunk": 5.338, "boundary_entropy": 0.4022, "router_noise": 0.0, "grad_norm": 1.9, "lr": 0.000251, "router_trainable": true, "sec_per_step": 0.08, "alpha": 0.03, "t": 28.9} |
| {"event": "probe", "step": 256, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3124.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.179509, "cps": 24.40625, "mean_p": 0.274857, "boundary_entropy": 0.36695, "bpb": 3.27455}, "tr": {"chunks": 3771.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.208308, "cps": 29.460938, "mean_p": 0.2704, "boundary_entropy": 0.394112, "bpb": 3.514713}, "fi": {"chunks": 3392.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.18032, "cps": 26.5, "mean_p": 0.284604, "boundary_entropy": 0.439168, "bpb": 3.650937}}, "bpb": {"en": 3.27455, "tr": 3.51471, "fi": 3.65094}, "gini_cpc": 0.033793, "gini_cps": 0.04193, "bpic": 5.514922, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.1604303422944098, "fi": 1.0045164545685294}, "t": 29.7} |
| {"event": "train", "step": 300, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.23683, "bpb": 3.22706, "chunk_loss": 0.030421, "F": 0.18463, "G": 0.27341, "chunks": 189.06, "bytes_per_chunk": 5.416, "boundary_entropy": 0.34817, "router_noise": 0.0, "grad_norm": 2.3794, "lr": 0.0003, "router_trainable": true, "sec_per_step": 0.112, "alpha": 0.03, "t": 34.5} |
| {"event": "train", "step": 350, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.19083, "bpb": 3.1607, "chunk_loss": 0.030468, "F": 0.18628, "G": 0.28164, "chunks": 190.75, "bytes_per_chunk": 5.368, "boundary_entropy": 0.36141, "router_noise": 0.0, "grad_norm": 1.3944, "lr": 0.0002999, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 38.9} |
| {"event": "probe", "step": 400, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3273.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.188071, "cps": 25.570312, "mean_p": 0.252341, "boundary_entropy": 0.280551, "bpb": 2.865337}, "tr": {"chunks": 3425.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.189195, "cps": 26.757812, "mean_p": 0.285184, "boundary_entropy": 0.389521, "bpb": 3.129532}, "fi": {"chunks": 3375.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.179416, "cps": 26.367188, "mean_p": 0.284077, "boundary_entropy": 0.412654, "bpb": 3.350044}}, "bpb": {"en": 2.86534, "tr": 3.12953, "fi": 3.35004}, "gini_cpc": 0.011711, "gini_cps": 0.01006, "bpic": 5.632086, "worst_lang": "tr", "cpc_parity_vs_en": {"en": 1.0, "tr": 1.0059772035697756, "fi": 0.9539816226839603}, "t": 42.6} |
| {"event": "eval", "step": 400, "phase": 1, "bpb": {"en": 2.87853, "tr": 3.13647, "fi": 3.34248}, "bytes_per_chunk": {"en": 4.963, "tr": 5.604, "fi": 5.716}, "bpb_mean": 3.11916, "frozen_inventory": false, "t": 43.0} |
| {"event": "train", "step": 400, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.02179, "bpb": 2.91682, "chunk_loss": 0.030586, "F": 0.19589, "G": 0.26014, "chunks": 200.59, "bytes_per_chunk": 5.105, "boundary_entropy": 0.28743, "router_noise": 0.0, "grad_norm": 1.1661, "lr": 0.0002998, "router_trainable": true, "sec_per_step": 0.087, "alpha": 0.03, "t": 43.2} |
| {"event": "train", "step": 450, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.06103, "bpb": 2.97344, "chunk_loss": 0.030569, "F": 0.19583, "G": 0.25422, "chunks": 200.53, "bytes_per_chunk": 5.106, "boundary_entropy": 0.2688, "router_noise": 0.0, "grad_norm": 1.1356, "lr": 0.0002995, "router_trainable": true, "sec_per_step": 0.083, "alpha": 0.03, "t": 47.4} |
| {"event": "train", "step": 500, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 2.02455, "bpb": 2.92081, "chunk_loss": 0.030447, "F": 0.19186, "G": 0.24986, "chunks": 196.47, "bytes_per_chunk": 5.212, "boundary_entropy": 0.26705, "router_noise": 0.0, "grad_norm": 1.1377, "lr": 0.0002992, "router_trainable": true, "sec_per_step": 0.081, "alpha": 0.03, "t": 51.4} |
| {"event": "probe", "step": 512, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3373.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.193817, "cps": 26.351562, "mean_p": 0.241709, "boundary_entropy": 0.236225, "bpb": 2.686646}, "tr": {"chunks": 3318.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.183285, "cps": 25.921875, "mean_p": 0.285974, "boundary_entropy": 0.381188, "bpb": 2.984598}, "fi": {"chunks": 3271.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.173888, "cps": 25.554688, "mean_p": 0.254018, "boundary_entropy": 0.353041, "bpb": 3.221906}}, "bpb": {"en": 2.68665, "tr": 2.9846, "fi": 3.22191}, "gini_cpc": 0.024114, "gini_cps": 0.006826, "bpic": 5.69484, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9456569294598279, "fi": 0.8971735050469978}, "t": 52.6} |
| {"event": "train", "step": 550, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.95272, "bpb": 2.81718, "chunk_loss": 0.03073, "F": 0.20273, "G": 0.25393, "chunks": 207.59, "bytes_per_chunk": 4.933, "boundary_entropy": 0.25034, "router_noise": 0.0, "grad_norm": 1.376, "lr": 0.0002987, "router_trainable": true, "sec_per_step": 0.108, "alpha": 0.03, "t": 56.8} |
| {"event": "probe", "step": 600, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3296.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.189393, "cps": 25.75, "mean_p": 0.223152, "boundary_entropy": 0.20068, "bpb": 2.587371}, "tr": {"chunks": 3321.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.18345, "cps": 25.945312, "mean_p": 0.284507, "boundary_entropy": 0.372299, "bpb": 2.891488}, "fi": {"chunks": 3237.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.17208, "cps": 25.289062, "mean_p": 0.236009, "boundary_entropy": 0.318962, "bpb": 3.132179}}, "bpb": {"en": 2.58737, "tr": 2.89149, "fi": 3.13218}, "gini_cpc": 0.02118, "gini_cps": 0.005683, "bpic": 5.757256, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9686240352535035, "fi": 0.9085895407974988}, "t": 61.0} |
| {"event": "eval", "step": 600, "phase": 1, "bpb": {"en": 2.60032, "tr": 2.90244, "fi": 3.13019}, "bytes_per_chunk": {"en": 4.95, "tr": 5.601, "fi": 5.717}, "bpb_mean": 2.87765, "frozen_inventory": false, "t": 61.4} |
| {"event": "train", "step": 600, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.92255, "bpb": 2.77366, "chunk_loss": 0.030509, "F": 0.19443, "G": 0.24358, "chunks": 199.09, "bytes_per_chunk": 5.143, "boundary_entropy": 0.23671, "router_noise": 0.0, "grad_norm": 1.2417, "lr": 0.0002982, "router_trainable": true, "sec_per_step": 0.098, "alpha": 0.03, "t": 61.7} |
| {"event": "train", "step": 650, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.86123, "bpb": 2.68518, "chunk_loss": 0.030428, "F": 0.19293, "G": 0.24675, "chunks": 197.56, "bytes_per_chunk": 5.183, "boundary_entropy": 0.24, "router_noise": 0.0, "grad_norm": 1.1862, "lr": 0.0002975, "router_trainable": true, "sec_per_step": 0.094, "alpha": 0.03, "t": 66.4} |
| {"event": "train", "step": 700, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.87485, "bpb": 2.70483, "chunk_loss": 0.030387, "F": 0.18951, "G": 0.25739, "chunks": 194.06, "bytes_per_chunk": 5.277, "boundary_entropy": 0.27087, "router_noise": 0.0, "grad_norm": 1.0775, "lr": 0.0002967, "router_trainable": true, "sec_per_step": 0.064, "alpha": 0.03, "t": 69.6} |
| {"event": "train", "step": 750, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.77815, "bpb": 2.56533, "chunk_loss": 0.030398, "F": 0.19476, "G": 0.23664, "chunks": 199.44, "bytes_per_chunk": 5.134, "boundary_entropy": 0.21162, "router_noise": 0.0, "grad_norm": 1.0383, "lr": 0.0002959, "router_trainable": true, "sec_per_step": 0.09, "alpha": 0.03, "t": 74.1} |
| {"event": "probe", "step": 800, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3363.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.193243, "cps": 26.273438, "mean_p": 0.217036, "boundary_entropy": 0.170083, "bpb": 2.424518}, "tr": {"chunks": 3189.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.176159, "cps": 24.914062, "mean_p": 0.271259, "boundary_entropy": 0.346812, "bpb": 2.710948}, "fi": {"chunks": 3294.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.17511, "cps": 25.734375, "mean_p": 0.241204, "boundary_entropy": 0.299546, "bpb": 2.980158}}, "bpb": {"en": 2.42452, "tr": 2.71095, "fi": 2.98016}, "gini_cpc": 0.0222, "gini_cps": 0.011781, "bpic": 5.761934, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9115935018089323, "fi": 0.9061685062799677}, "t": 77.5} |
| {"event": "eval", "step": 800, "phase": 1, "bpb": {"en": 2.44043, "tr": 2.72196, "fi": 2.97901}, "bytes_per_chunk": {"en": 4.89, "tr": 5.895, "fi": 5.729}, "bpb_mean": 2.7138, "frozen_inventory": false, "t": 78.0} |
| {"event": "train", "step": 800, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.79564, "bpb": 2.59056, "chunk_loss": 0.030404, "F": 0.19287, "G": 0.2404, "chunks": 197.5, "bytes_per_chunk": 5.185, "boundary_entropy": 0.2258, "router_noise": 0.0, "grad_norm": 1.0089, "lr": 0.0002949, "router_trainable": true, "sec_per_step": 0.085, "alpha": 0.03, "t": 78.4} |
| {"event": "train", "step": 850, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.7256, "bpb": 2.48952, "chunk_loss": 0.030368, "F": 0.19653, "G": 0.22, "chunks": 201.25, "bytes_per_chunk": 5.088, "boundary_entropy": 0.16971, "router_noise": 0.0, "grad_norm": 1.1487, "lr": 0.0002938, "router_trainable": true, "sec_per_step": 0.077, "alpha": 0.03, "t": 82.2} |
| {"event": "train", "step": 900, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.70349, "bpb": 2.45762, "chunk_loss": 0.030548, "F": 0.19879, "G": 0.23166, "chunks": 203.56, "bytes_per_chunk": 5.03, "boundary_entropy": 0.19138, "router_noise": 0.0, "grad_norm": 0.9838, "lr": 0.0002927, "router_trainable": true, "sec_per_step": 0.081, "alpha": 0.03, "t": 86.3} |
| {"event": "train", "step": 950, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.71236, "bpb": 2.47042, "chunk_loss": 0.030359, "F": 0.19568, "G": 0.22503, "chunks": 200.38, "bytes_per_chunk": 5.11, "boundary_entropy": 0.17428, "router_noise": 0.0, "grad_norm": 1.1, "lr": 0.0002914, "router_trainable": true, "sec_per_step": 0.08, "alpha": 0.03, "t": 90.3} |
| {"event": "probe", "step": 1000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3358.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.192955, "cps": 26.234375, "mean_p": 0.214841, "boundary_entropy": 0.165911, "bpb": 2.330884}, "tr": {"chunks": 3162.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.174667, "cps": 24.703125, "mean_p": 0.258541, "boundary_entropy": 0.323095, "bpb": 2.621655}, "fi": {"chunks": 3196.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.169901, "cps": 24.96875, "mean_p": 0.228486, "boundary_entropy": 0.275674, "bpb": 2.865323}}, "bpb": {"en": 2.33088, "tr": 2.62165, "fi": 2.86532}, "gini_cpc": 0.028594, "gini_cps": 0.013449, "bpic": 5.839028, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.905221254447739, "fi": 0.8805181563927863}, "t": 94.5} |
| {"event": "eval", "step": 1000, "phase": 1, "bpb": {"en": 2.34535, "tr": 2.62882, "fi": 2.86576}, "bytes_per_chunk": {"en": 4.894, "tr": 5.877, "fi": 5.958}, "bpb_mean": 2.61331, "frozen_inventory": false, "t": 95.1} |
| {"event": "train", "step": 1000, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.71839, "bpb": 2.47911, "chunk_loss": 0.030375, "F": 0.19482, "G": 0.22976, "chunks": 199.5, "bytes_per_chunk": 5.133, "boundary_entropy": 0.19166, "router_noise": 0.0, "grad_norm": 0.9869, "lr": 0.0002901, "router_trainable": true, "sec_per_step": 0.102, "alpha": 0.03, "t": 95.4} |
| {"event": "probe", "step": 1024, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3346.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.192266, "cps": 26.140625, "mean_p": 0.211739, "boundary_entropy": 0.150974, "bpb": 2.325549}, "tr": {"chunks": 3192.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.176324, "cps": 24.9375, "mean_p": 0.259632, "boundary_entropy": 0.324051, "bpb": 2.600288}, "fi": {"chunks": 3223.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.171336, "cps": 25.179688, "mean_p": 0.226786, "boundary_entropy": 0.269754, "bpb": 2.825315}}, "bpb": {"en": 2.32555, "tr": 2.60029, "fi": 2.82532}, "gini_cpc": 0.025843, "gini_cps": 0.010518, "bpic": 5.812109, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9170869526930625, "fi": 0.8911413699866509}, "t": 97.3} |
| {"event": "train", "step": 1050, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.73908, "bpb": 2.50896, "chunk_loss": 0.030273, "F": 0.18982, "G": 0.22379, "chunks": 194.38, "bytes_per_chunk": 5.268, "boundary_entropy": 0.18701, "router_noise": 0.0, "grad_norm": 1.0398, "lr": 0.0002886, "router_trainable": true, "sec_per_step": 0.083, "alpha": 0.03, "t": 99.6} |
| {"event": "train", "step": 1100, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.71721, "bpb": 2.47742, "chunk_loss": 0.030388, "F": 0.19788, "G": 0.22535, "chunks": 202.62, "bytes_per_chunk": 5.054, "boundary_entropy": 0.16843, "router_noise": 0.0, "grad_norm": 1.0599, "lr": 0.0002871, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 103.1} |
| {"event": "train", "step": 1150, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.72884, "bpb": 2.49419, "chunk_loss": 0.030449, "F": 0.19556, "G": 0.23857, "chunks": 200.25, "bytes_per_chunk": 5.114, "boundary_entropy": 0.20727, "router_noise": 0.0, "grad_norm": 0.9273, "lr": 0.0002855, "router_trainable": true, "sec_per_step": 0.063, "alpha": 0.03, "t": 106.3} |
| {"event": "probe", "step": 1200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3315.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.190484, "cps": 25.898438, "mean_p": 0.205986, "boundary_entropy": 0.143728, "bpb": 2.269169}, "tr": {"chunks": 3190.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.176214, "cps": 24.921875, "mean_p": 0.247745, "boundary_entropy": 0.297794, "bpb": 2.539007}, "fi": {"chunks": 3129.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.166339, "cps": 24.445312, "mean_p": 0.21671, "boundary_entropy": 0.245932, "bpb": 2.743606}}, "bpb": {"en": 2.26917, "tr": 2.53901, "fi": 2.74361}, "gini_cpc": 0.030199, "gini_cps": 0.012871, "bpic": 5.888727, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9250830404100417, "fi": 0.8732412993167807}, "t": 110.8} |
| {"event": "eval", "step": 1200, "phase": 1, "bpb": {"en": 2.28178, "tr": 2.54583, "fi": 2.74826}, "bytes_per_chunk": {"en": 4.921, "tr": 5.94, "fi": 6.031}, "bpb_mean": 2.52529, "frozen_inventory": false, "t": 111.3} |
| {"event": "train", "step": 1200, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.62403, "bpb": 2.34298, "chunk_loss": 0.030336, "F": 0.19385, "G": 0.22685, "chunks": 198.5, "bytes_per_chunk": 5.159, "boundary_entropy": 0.18028, "router_noise": 0.0, "grad_norm": 0.9238, "lr": 0.0002837, "router_trainable": true, "sec_per_step": 0.109, "alpha": 0.03, "t": 111.8} |
| {"event": "train", "step": 1250, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.66516, "bpb": 2.40232, "chunk_loss": 0.030326, "F": 0.19147, "G": 0.22769, "chunks": 196.06, "bytes_per_chunk": 5.223, "boundary_entropy": 0.18988, "router_noise": 0.0, "grad_norm": 0.8786, "lr": 0.0002819, "router_trainable": true, "sec_per_step": 0.082, "alpha": 0.03, "t": 115.9} |
| {"event": "train", "step": 1300, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.59137, "bpb": 2.29587, "chunk_loss": 0.030359, "F": 0.19724, "G": 0.22384, "chunks": 201.97, "bytes_per_chunk": 5.07, "boundary_entropy": 0.1579, "router_noise": 0.0, "grad_norm": 0.8282, "lr": 0.00028, "router_trainable": true, "sec_per_step": 0.087, "alpha": 0.03, "t": 120.2} |
| {"event": "train", "step": 1350, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.61603, "bpb": 2.33143, "chunk_loss": 0.030325, "F": 0.19635, "G": 0.21283, "chunks": 201.06, "bytes_per_chunk": 5.093, "boundary_entropy": 0.13741, "router_noise": 0.0, "grad_norm": 0.978, "lr": 0.000278, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 123.7} |
| {"event": "probe", "step": 1400, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3308.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.190082, "cps": 25.84375, "mean_p": 0.202512, "boundary_entropy": 0.133905, "bpb": 2.208175}, "tr": {"chunks": 3158.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.174446, "cps": 24.671875, "mean_p": 0.250445, "boundary_entropy": 0.295973, "bpb": 2.47158}, "fi": {"chunks": 3139.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.16687, "cps": 24.523438, "mean_p": 0.212276, "boundary_entropy": 0.239764, "bpb": 2.681357}}, "bpb": {"en": 2.20818, "tr": 2.47158, "fi": 2.68136}, "gini_cpc": 0.02912, "gini_cps": 0.01173, "bpic": 5.906507, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9177411254329234, "fi": 0.8778858552043535}, "t": 128.7} |
| {"event": "eval", "step": 1400, "phase": 1, "bpb": {"en": 2.22198, "tr": 2.48087, "fi": 2.68714}, "bytes_per_chunk": {"en": 4.935, "tr": 5.99, "fi": 6.024}, "bpb_mean": 2.46333, "frozen_inventory": false, "t": 129.1} |
| {"event": "train", "step": 1400, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.58496, "bpb": 2.28662, "chunk_loss": 0.030256, "F": 0.19196, "G": 0.21741, "chunks": 196.56, "bytes_per_chunk": 5.21, "boundary_entropy": 0.16124, "router_noise": 0.0, "grad_norm": 0.8757, "lr": 0.0002759, "router_trainable": true, "sec_per_step": 0.113, "alpha": 0.03, "t": 129.4} |
| {"event": "train", "step": 1450, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.55947, "bpb": 2.24984, "chunk_loss": 0.030412, "F": 0.19913, "G": 0.22595, "chunks": 203.91, "bytes_per_chunk": 5.022, "boundary_entropy": 0.16583, "router_noise": 0.0, "grad_norm": 0.858, "lr": 0.0002738, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 132.9} |
| {"event": "train", "step": 1500, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.55232, "bpb": 2.23952, "chunk_loss": 0.03025, "F": 0.19122, "G": 0.21436, "chunks": 195.81, "bytes_per_chunk": 5.229, "boundary_entropy": 0.15409, "router_noise": 0.0, "grad_norm": 1.0657, "lr": 0.0002715, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 136.5} |
| {"event": "train", "step": 1550, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.54415, "bpb": 2.22774, "chunk_loss": 0.030368, "F": 0.19675, "G": 0.22122, "chunks": 201.47, "bytes_per_chunk": 5.083, "boundary_entropy": 0.15217, "router_noise": 0.0, "grad_norm": 0.8513, "lr": 0.0002692, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 140.9} |
| {"event": "probe", "step": 1600, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3301.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.18968, "cps": 25.789062, "mean_p": 0.200742, "boundary_entropy": 0.128618, "bpb": 2.17038}, "tr": {"chunks": 3160.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.174557, "cps": 24.6875, "mean_p": 0.246489, "boundary_entropy": 0.293705, "bpb": 2.427883}, "fi": {"chunks": 3172.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.168625, "cps": 24.78125, "mean_p": 0.218431, "boundary_entropy": 0.235053, "bpb": 2.614407}}, "bpb": {"en": 2.17038, "tr": 2.42788, "fi": 2.61441}, "gini_cpc": 0.026342, "gini_cps": 0.009758, "bpic": 5.889339, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9202697084773733, "fi": 0.8889961723395583}, "t": 144.9} |
| {"event": "eval", "step": 1600, "phase": 1, "bpb": {"en": 2.18566, "tr": 2.43797, "fi": 2.62277}, "bytes_per_chunk": {"en": 4.937, "tr": 6.108, "fi": 6.0}, "bpb_mean": 2.41547, "frozen_inventory": false, "t": 145.2} |
| {"event": "train", "step": 1600, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.52477, "bpb": 2.19978, "chunk_loss": 0.030245, "F": 0.19012, "G": 0.22143, "chunks": 194.69, "bytes_per_chunk": 5.26, "boundary_entropy": 0.17003, "router_noise": 0.0, "grad_norm": 0.8465, "lr": 0.0002668, "router_trainable": true, "sec_per_step": 0.092, "alpha": 0.03, "t": 145.5} |
| {"event": "train", "step": 1650, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.57888, "bpb": 2.27785, "chunk_loss": 0.030291, "F": 0.19522, "G": 0.21803, "chunks": 199.91, "bytes_per_chunk": 5.122, "boundary_entropy": 0.14965, "router_noise": 0.0, "grad_norm": 0.859, "lr": 0.0002643, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 149.2} |
| {"event": "train", "step": 1700, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.56947, "bpb": 2.26426, "chunk_loss": 0.030262, "F": 0.19125, "G": 0.22213, "chunks": 195.84, "bytes_per_chunk": 5.229, "boundary_entropy": 0.1605, "router_noise": 0.0, "grad_norm": 0.8502, "lr": 0.0002618, "router_trainable": true, "sec_per_step": 0.076, "alpha": 0.03, "t": 153.0} |
| {"event": "train", "step": 1750, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.54073, "bpb": 2.22281, "chunk_loss": 0.030322, "F": 0.19626, "G": 0.2157, "chunks": 200.97, "bytes_per_chunk": 5.095, "boundary_entropy": 0.12998, "router_noise": 0.0, "grad_norm": 0.7631, "lr": 0.0002591, "router_trainable": true, "sec_per_step": 0.065, "alpha": 0.03, "t": 156.3} |
| {"event": "probe", "step": 1800, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3320.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.190772, "cps": 25.9375, "mean_p": 0.200129, "boundary_entropy": 0.117423, "bpb": 2.135638}, "tr": {"chunks": 3327.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.183782, "cps": 25.992188, "mean_p": 0.248511, "boundary_entropy": 0.275704, "bpb": 2.373389}, "fi": {"chunks": 3226.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.171495, "cps": 25.203125, "mean_p": 0.213281, "boundary_entropy": 0.226587, "bpb": 2.582022}}, "bpb": {"en": 2.13564, "tr": 2.37339, "fi": 2.58202}, "gini_cpc": 0.023534, "gini_cps": 0.00682, "bpic": 5.746176, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9633592814610372, "fi": 0.8989561670209625}, "t": 160.1} |
| {"event": "eval", "step": 1800, "phase": 1, "bpb": {"en": 2.15213, "tr": 2.38087, "fi": 2.58942}, "bytes_per_chunk": {"en": 4.926, "tr": 5.813, "fi": 5.948}, "bpb_mean": 2.37414, "frozen_inventory": false, "t": 160.7} |
| {"event": "train", "step": 1800, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.59155, "bpb": 2.29613, "chunk_loss": 0.030354, "F": 0.19571, "G": 0.22339, "chunks": 200.41, "bytes_per_chunk": 5.11, "boundary_entropy": 0.15759, "router_noise": 0.0, "grad_norm": 0.8952, "lr": 0.0002564, "router_trainable": true, "sec_per_step": 0.096, "alpha": 0.03, "t": 161.1} |
| {"event": "train", "step": 1850, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.53817, "bpb": 2.21911, "chunk_loss": 0.030317, "F": 0.19449, "G": 0.21774, "chunks": 199.16, "bytes_per_chunk": 5.142, "boundary_entropy": 0.14567, "router_noise": 0.0, "grad_norm": 0.7707, "lr": 0.0002537, "router_trainable": true, "sec_per_step": 0.079, "alpha": 0.03, "t": 165.1} |
| {"event": "train", "step": 1900, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.58229, "bpb": 2.28276, "chunk_loss": 0.030316, "F": 0.19534, "G": 0.21615, "chunks": 200.03, "bytes_per_chunk": 5.119, "boundary_entropy": 0.13652, "router_noise": 0.0, "grad_norm": 0.7476, "lr": 0.0002508, "router_trainable": true, "sec_per_step": 0.087, "alpha": 0.03, "t": 169.4} |
| {"event": "train", "step": 1950, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.53961, "bpb": 2.22119, "chunk_loss": 0.03031, "F": 0.19376, "G": 0.21428, "chunks": 198.41, "bytes_per_chunk": 5.161, "boundary_entropy": 0.13274, "router_noise": 0.0, "grad_norm": 0.7664, "lr": 0.0002479, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 173.1} |
| {"event": "probe", "step": 2000, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3284.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.188703, "cps": 25.65625, "mean_p": 0.195771, "boundary_entropy": 0.10924, "bpb": 2.106718}, "tr": {"chunks": 3204.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.176987, "cps": 25.03125, "mean_p": 0.240039, "boundary_entropy": 0.265927, "bpb": 2.343858}, "fi": {"chunks": 3188.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.169475, "cps": 24.90625, "mean_p": 0.209703, "boundary_entropy": 0.215741, "bpb": 2.510134}}, "bpb": {"en": 2.10672, "tr": 2.34386, "fi": 2.51013}, "gini_cpc": 0.023952, "gini_cps": 0.006614, "bpic": 5.863167, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9379138039650362, "fi": 0.898105593100572}, "t": 178.4} |
| {"event": "eval", "step": 2000, "phase": 1, "bpb": {"en": 2.1227, "tr": 2.35, "fi": 2.51834}, "bytes_per_chunk": {"en": 4.967, "tr": 5.982, "fi": 5.967}, "bpb_mean": 2.33035, "frozen_inventory": false, "t": 178.7} |
| {"event": "train", "step": 2000, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.56416, "bpb": 2.25661, "chunk_loss": 0.030244, "F": 0.19046, "G": 0.21323, "chunks": 195.03, "bytes_per_chunk": 5.25, "boundary_entropy": 0.14492, "router_noise": 0.0, "grad_norm": 0.822, "lr": 0.000245, "router_trainable": true, "sec_per_step": 0.118, "alpha": 0.03, "t": 179.0} |
| {"event": "probe", "step": 2048, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3303.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.189795, "cps": 25.804688, "mean_p": 0.1981, "boundary_entropy": 0.104418, "bpb": 2.094304}, "tr": {"chunks": 3204.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.176987, "cps": 25.03125, "mean_p": 0.243951, "boundary_entropy": 0.268734, "bpb": 2.337709}, "fi": {"chunks": 3190.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.169582, "cps": 24.921875, "mean_p": 0.208622, "boundary_entropy": 0.210466, "bpb": 2.504412}}, "bpb": {"en": 2.0943, "tr": 2.33771, "fi": 2.50441}, "gini_cpc": 0.025124, "gini_cps": 0.007769, "bpic": 5.850469, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9325185989164937, "fi": 0.8934995664845453}, "t": 183.2} |
| {"event": "train", "step": 2050, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.62092, "bpb": 2.33849, "chunk_loss": 0.030278, "F": 0.19016, "G": 0.22015, "chunks": 194.72, "bytes_per_chunk": 5.259, "boundary_entropy": 0.1503, "router_noise": 0.0, "grad_norm": 0.9949, "lr": 0.0002419, "router_trainable": true, "sec_per_step": 0.093, "alpha": 0.03, "t": 183.6} |
| {"event": "train", "step": 2100, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.49996, "bpb": 2.16399, "chunk_loss": 0.030316, "F": 0.19571, "G": 0.21721, "chunks": 200.41, "bytes_per_chunk": 5.11, "boundary_entropy": 0.1361, "router_noise": 0.0, "grad_norm": 0.7816, "lr": 0.0002388, "router_trainable": true, "sec_per_step": 0.092, "alpha": 0.03, "t": 188.2} |
| {"event": "train", "step": 2150, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.51581, "bpb": 2.18685, "chunk_loss": 0.030216, "F": 0.18945, "G": 0.20846, "chunks": 194.0, "bytes_per_chunk": 5.278, "boundary_entropy": 0.12763, "router_noise": 0.0, "grad_norm": 0.7266, "lr": 0.0002357, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 192.6} |
| {"event": "probe", "step": 2200, "phase": 1, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3315.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.190484, "cps": 25.898438, "mean_p": 0.198404, "boundary_entropy": 0.098872, "bpb": 2.07283}, "tr": {"chunks": 3230.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.178423, "cps": 25.234375, "mean_p": 0.238737, "boundary_entropy": 0.260135, "bpb": 2.295122}, "fi": {"chunks": 3181.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.169103, "cps": 24.851562, "mean_p": 0.202258, "boundary_entropy": 0.194158, "bpb": 2.474101}}, "bpb": {"en": 2.07283, "tr": 2.29512, "fi": 2.4741}, "gini_cpc": 0.026494, "gini_cps": 0.009185, "bpic": 5.833025, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9366828277506065, "fi": 0.8877534589730519}, "t": 196.8} |
| {"event": "eval", "step": 2200, "phase": 1, "bpb": {"en": 2.09028, "tr": 2.30273, "fi": 2.48357}, "bytes_per_chunk": {"en": 4.932, "tr": 5.879, "fi": 5.976}, "bpb_mean": 2.2922, "frozen_inventory": false, "t": 197.3} |
| {"event": "train", "step": 2200, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.48577, "bpb": 2.14351, "chunk_loss": 0.030336, "F": 0.19705, "G": 0.21446, "chunks": 201.78, "bytes_per_chunk": 5.075, "boundary_entropy": 0.11924, "router_noise": 0.0, "grad_norm": 0.6772, "lr": 0.0002325, "router_trainable": true, "sec_per_step": 0.099, "alpha": 0.03, "t": 197.6} |
| {"event": "train", "step": 2250, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.51704, "bpb": 2.18862, "chunk_loss": 0.030229, "F": 0.19034, "G": 0.21161, "chunks": 194.91, "bytes_per_chunk": 5.254, "boundary_entropy": 0.12865, "router_noise": 0.0, "grad_norm": 0.7606, "lr": 0.0002293, "router_trainable": true, "sec_per_step": 0.075, "alpha": 0.03, "t": 201.3} |
| {"event": "train", "step": 2300, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.49959, "bpb": 2.16346, "chunk_loss": 0.030295, "F": 0.1962, "G": 0.21216, "chunks": 200.91, "bytes_per_chunk": 5.097, "boundary_entropy": 0.11419, "router_noise": 0.0, "grad_norm": 0.7514, "lr": 0.000226, "router_trainable": true, "sec_per_step": 0.065, "alpha": 0.03, "t": 204.6} |
| {"event": "train", "step": 2350, "arch": "hnet", "phase": 1, "objective": "global", "lm_loss": 1.50122, "bpb": 2.16581, "chunk_loss": 0.030169, "F": 0.18451, "G": 0.22008, "chunks": 188.94, "bytes_per_chunk": 5.42, "boundary_entropy": 0.15807, "router_noise": 0.0, "grad_norm": 0.8619, "lr": 0.0002226, "router_trainable": true, "sec_per_step": 0.087, "alpha": 0.03, "t": 208.9} |
| {"event": "probe", "step": 2400, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3312.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.190312, "cps": 25.875, "mean_p": 0.196765, "boundary_entropy": 0.094788, "bpb": 2.05299}, "tr": {"chunks": 3203.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.176932, "cps": 25.023438, "mean_p": 0.239442, "boundary_entropy": 0.258622, "bpb": 2.265854}, "fi": {"chunks": 3226.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.171495, "cps": 25.203125, "mean_p": 0.211212, "boundary_entropy": 0.204783, "bpb": 2.445447}}, "bpb": {"en": 2.05299, "tr": 2.26585, "fi": 2.44545}, "gini_cpc": 0.023285, "gini_cps": 0.00746, "bpic": 5.824043, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9296943236247974, "fi": 0.9011275587287426}, "t": 213.0} |
| {"event": "eval", "step": 2400, "phase": 1, "bpb": {"en": 2.068, "tr": 2.2728, "fi": 2.45608}, "bytes_per_chunk": {"en": 4.929, "tr": 5.955, "fi": 5.926}, "bpb_mean": 2.26563, "frozen_inventory": false, "t": 213.5} |
| {"event": "phase_transition", "step": 2400, "phase": 2, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 213.7} |
| {"event": "train", "step": 2400, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.45802, "bpb": 2.10348, "chunk_loss": 0.030395, "F": 0.20117, "G": 0.21725, "chunks": 206.0, "bytes_per_chunk": 4.971, "boundary_entropy": 0.11748, "router_noise": 0.0, "grad_norm": 0.797, "lr": 0.0002192, "router_trainable": true, "sec_per_step": 0.097, "alpha": 0.03, "t": 213.8} |
| {"event": "train", "step": 2450, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.63941, "bpb": 2.36517, "chunk_loss": 0.030315, "F": 0.19598, "G": 0.21551, "chunks": 200.69, "bytes_per_chunk": 5.102, "boundary_entropy": 0.12068, "router_noise": 0.0, "grad_norm": 0.7318, "lr": 0.0002158, "router_trainable": true, "sec_per_step": 0.085, "alpha": 0.03, "t": 218.0} |
| {"event": "train", "step": 2500, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.49557, "bpb": 2.15765, "chunk_loss": 0.030258, "F": 0.19339, "G": 0.20758, "chunks": 198.03, "bytes_per_chunk": 5.171, "boundary_entropy": 0.11151, "router_noise": 0.0, "grad_norm": 0.7394, "lr": 0.0002123, "router_trainable": true, "sec_per_step": 0.088, "alpha": 0.03, "t": 222.4} |
| {"event": "train", "step": 2550, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.48285, "bpb": 2.1393, "chunk_loss": 0.030233, "F": 0.18945, "G": 0.2142, "chunks": 194.0, "bytes_per_chunk": 5.278, "boundary_entropy": 0.13429, "router_noise": 0.0, "grad_norm": 0.6972, "lr": 0.0002088, "router_trainable": true, "sec_per_step": 0.08, "alpha": 0.03, "t": 226.4} |
| {"event": "probe", "step": 2600, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3263.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.187496, "cps": 25.492188, "mean_p": 0.193674, "boundary_entropy": 0.088614, "bpb": 2.031911}, "tr": {"chunks": 3095.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.170966, "cps": 24.179688, "mean_p": 0.22835, "boundary_entropy": 0.243587, "bpb": 2.248741}, "fi": {"chunks": 3080.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.163734, "cps": 24.0625, "mean_p": 0.1933, "boundary_entropy": 0.179456, "bpb": 2.415101}}, "bpb": {"en": 2.03191, "tr": 2.24874, "fi": 2.4151}, "gini_cpc": 0.030337, "gini_cps": 0.012926, "bpic": 6.011019, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9118368689100841, "fi": 0.8732646486389369}, "t": 230.8} |
| {"event": "eval", "step": 2600, "phase": 2, "bpb": {"en": 2.04663, "tr": 2.25366, "fi": 2.42458}, "bytes_per_chunk": {"en": 4.989, "tr": 6.119, "fi": 6.064}, "bpb_mean": 2.24162, "frozen_inventory": false, "t": 231.2} |
| {"event": "train", "step": 2600, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.51181, "bpb": 2.18109, "chunk_loss": 0.030199, "F": 0.18518, "G": 0.20632, "chunks": 189.62, "bytes_per_chunk": 5.4, "boundary_entropy": 0.122, "router_noise": 0.0, "grad_norm": 0.751, "lr": 0.0002053, "router_trainable": true, "sec_per_step": 0.102, "alpha": 0.03, "t": 231.5} |
| {"event": "train", "step": 2650, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.46659, "bpb": 2.11585, "chunk_loss": 0.030243, "F": 0.18823, "G": 0.21546, "chunks": 192.75, "bytes_per_chunk": 5.313, "boundary_entropy": 0.13854, "router_noise": 0.0, "grad_norm": 0.7293, "lr": 0.0002017, "router_trainable": true, "sec_per_step": 0.08, "alpha": 0.03, "t": 235.5} |
| {"event": "train", "step": 2700, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.54835, "bpb": 2.2338, "chunk_loss": 0.030277, "F": 0.19244, "G": 0.21383, "chunks": 197.06, "bytes_per_chunk": 5.196, "boundary_entropy": 0.1224, "router_noise": 0.0, "grad_norm": 0.8449, "lr": 0.0001981, "router_trainable": true, "sec_per_step": 0.081, "alpha": 0.03, "t": 239.5} |
| {"event": "train", "step": 2750, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.48021, "bpb": 2.1355, "chunk_loss": 0.030241, "F": 0.19113, "G": 0.21376, "chunks": 195.72, "bytes_per_chunk": 5.232, "boundary_entropy": 0.12293, "router_noise": 0.0, "grad_norm": 0.7362, "lr": 0.0001945, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 244.0} |
| {"event": "probe", "step": 2800, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3279.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.188416, "cps": 25.617188, "mean_p": 0.192747, "boundary_entropy": 0.086255, "bpb": 2.014392}, "tr": {"chunks": 3125.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.172623, "cps": 24.414062, "mean_p": 0.22532, "boundary_entropy": 0.231705, "bpb": 2.220755}, "fi": {"chunks": 3134.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.166605, "cps": 24.484375, "mean_p": 0.197248, "boundary_entropy": 0.178641, "bpb": 2.376574}}, "bpb": {"en": 2.01439, "tr": 2.22076, "fi": 2.37657}, "gini_cpc": 0.027558, "gini_cps": 0.010764, "bpic": 5.947997, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9161828833574515, "fi": 0.8842392979949877}, "t": 248.8} |
| {"event": "eval", "step": 2800, "phase": 2, "bpb": {"en": 2.03102, "tr": 2.22827, "fi": 2.38572}, "bytes_per_chunk": {"en": 4.97, "tr": 6.117, "fi": 6.005}, "bpb_mean": 2.21501, "frozen_inventory": false, "t": 249.2} |
| {"event": "train", "step": 2800, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.45802, "bpb": 2.10348, "chunk_loss": 0.030291, "F": 0.19595, "G": 0.20884, "chunks": 200.66, "bytes_per_chunk": 5.103, "boundary_entropy": 0.10528, "router_noise": 0.0, "grad_norm": 0.6789, "lr": 0.0001909, "router_trainable": true, "sec_per_step": 0.114, "alpha": 0.03, "t": 249.7} |
| {"event": "train", "step": 2850, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.54932, "bpb": 2.23519, "chunk_loss": 0.030284, "F": 0.19241, "G": 0.2089, "chunks": 197.03, "bytes_per_chunk": 5.197, "boundary_entropy": 0.1094, "router_noise": 0.0, "grad_norm": 0.7597, "lr": 0.0001872, "router_trainable": true, "sec_per_step": 0.092, "alpha": 0.03, "t": 254.2} |
| {"event": "train", "step": 2900, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.42129, "bpb": 2.05049, "chunk_loss": 0.030244, "F": 0.19101, "G": 0.21286, "chunks": 195.59, "bytes_per_chunk": 5.235, "boundary_entropy": 0.12519, "router_noise": 0.0, "grad_norm": 0.7039, "lr": 0.0001835, "router_trainable": true, "sec_per_step": 0.079, "alpha": 0.03, "t": 258.2} |
| {"event": "train", "step": 2950, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.36397, "bpb": 1.96779, "chunk_loss": 0.030388, "F": 0.20233, "G": 0.21417, "chunks": 207.19, "bytes_per_chunk": 4.942, "boundary_entropy": 0.09802, "router_noise": 0.0, "grad_norm": 0.6482, "lr": 0.0001799, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 261.8} |
| {"event": "probe", "step": 3000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3281.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.188531, "cps": 25.632812, "mean_p": 0.193047, "boundary_entropy": 0.083806, "bpb": 1.999928}, "tr": {"chunks": 3104.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.171463, "cps": 24.25, "mean_p": 0.227046, "boundary_entropy": 0.235271, "bpb": 2.199837}, "fi": {"chunks": 3128.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.166286, "cps": 24.4375, "mean_p": 0.197681, "boundary_entropy": 0.176688, "bpb": 2.354968}}, "bpb": {"en": 1.99993, "tr": 2.19984, "fi": 2.35497}, "gini_cpc": 0.028179, "gini_cps": 0.012404, "bpic": 5.963629, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9094714095203438, "fi": 0.8820084599381411}, "t": 265.5} |
| {"event": "eval", "step": 3000, "phase": 2, "bpb": {"en": 2.01576, "tr": 2.2069, "fi": 2.36339}, "bytes_per_chunk": {"en": 4.969, "tr": 6.13, "fi": 6.005}, "bpb_mean": 2.19535, "frozen_inventory": false, "t": 266.0} |
| {"event": "train", "step": 3000, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.45859, "bpb": 2.1043, "chunk_loss": 0.030226, "F": 0.18988, "G": 0.20983, "chunks": 194.44, "bytes_per_chunk": 5.266, "boundary_entropy": 0.11881, "router_noise": 0.0, "grad_norm": 0.7921, "lr": 0.0001761, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 266.3} |
| {"event": "train", "step": 3050, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.41667, "bpb": 2.04383, "chunk_loss": 0.030283, "F": 0.19586, "G": 0.21055, "chunks": 200.56, "bytes_per_chunk": 5.106, "boundary_entropy": 0.1034, "router_noise": 0.0, "grad_norm": 0.7154, "lr": 0.0001724, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 269.9} |
| {"event": "train", "step": 3100, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.51071, "bpb": 2.1795, "chunk_loss": 0.030203, "F": 0.18805, "G": 0.20857, "chunks": 192.56, "bytes_per_chunk": 5.318, "boundary_entropy": 0.11306, "router_noise": 0.0, "grad_norm": 0.8548, "lr": 0.0001687, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 273.4} |
| {"event": "train", "step": 3150, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.42043, "bpb": 2.04924, "chunk_loss": 0.030191, "F": 0.18781, "G": 0.2035, "chunks": 192.31, "bytes_per_chunk": 5.325, "boundary_entropy": 0.10236, "router_noise": 0.0, "grad_norm": 0.7061, "lr": 0.000165, "router_trainable": true, "sec_per_step": 0.062, "alpha": 0.03, "t": 276.5} |
| {"event": "probe", "step": 3200, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3270.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.187899, "cps": 25.546875, "mean_p": 0.19295, "boundary_entropy": 0.077731, "bpb": 1.978173}, "tr": {"chunks": 3124.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.172568, "cps": 24.40625, "mean_p": 0.226919, "boundary_entropy": 0.233524, "bpb": 2.169949}, "fi": {"chunks": 3129.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.166339, "cps": 24.445312, "mean_p": 0.19672, "boundary_entropy": 0.174532, "bpb": 2.33197}}, "bpb": {"en": 1.97817, "tr": 2.16995, "fi": 2.33197}, "gini_cpc": 0.027284, "gini_cps": 0.010221, "bpic": 5.957366, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9184105021875335, "fi": 0.8852583814174705}, "t": 280.6} |
| {"event": "eval", "step": 3200, "phase": 2, "bpb": {"en": 1.99367, "tr": 2.17867, "fi": 2.3437}, "bytes_per_chunk": {"en": 4.986, "tr": 6.057, "fi": 5.993}, "bpb_mean": 2.17201, "frozen_inventory": false, "t": 281.0} |
| {"event": "train", "step": 3200, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.34414, "bpb": 1.93919, "chunk_loss": 0.030257, "F": 0.1955, "G": 0.20463, "chunks": 200.19, "bytes_per_chunk": 5.115, "boundary_entropy": 0.08565, "router_noise": 0.0, "grad_norm": 0.6943, "lr": 0.0001613, "router_trainable": true, "sec_per_step": 0.099, "alpha": 0.03, "t": 281.5} |
| {"event": "train", "step": 3250, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.34947, "bpb": 1.94688, "chunk_loss": 0.030294, "F": 0.19443, "G": 0.21007, "chunks": 199.09, "bytes_per_chunk": 5.143, "boundary_entropy": 0.10304, "router_noise": 0.0, "grad_norm": 0.6584, "lr": 0.0001576, "router_trainable": true, "sec_per_step": 0.079, "alpha": 0.03, "t": 285.4} |
| {"event": "train", "step": 3300, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.47737, "bpb": 2.1314, "chunk_loss": 0.030316, "F": 0.19617, "G": 0.21117, "chunks": 200.88, "bytes_per_chunk": 5.098, "boundary_entropy": 0.10276, "router_noise": 0.0, "grad_norm": 0.7167, "lr": 0.0001539, "router_trainable": true, "sec_per_step": 0.077, "alpha": 0.03, "t": 289.3} |
| {"event": "train", "step": 3350, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.42522, "bpb": 2.05616, "chunk_loss": 0.030237, "F": 0.19217, "G": 0.20782, "chunks": 196.78, "bytes_per_chunk": 5.204, "boundary_entropy": 0.10429, "router_noise": 0.0, "grad_norm": 0.721, "lr": 0.0001501, "router_trainable": true, "sec_per_step": 0.08, "alpha": 0.03, "t": 293.3} |
| {"event": "probe", "step": 3400, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3250.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.186749, "cps": 25.390625, "mean_p": 0.191864, "boundary_entropy": 0.076289, "bpb": 1.961778}, "tr": {"chunks": 3085.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.170414, "cps": 24.101562, "mean_p": 0.220803, "boundary_entropy": 0.223283, "bpb": 2.14885}, "fi": {"chunks": 3097.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.164638, "cps": 24.195312, "mean_p": 0.192981, "boundary_entropy": 0.167655, "bpb": 2.311005}}, "bpb": {"en": 1.96178, "tr": 2.14885, "fi": 2.31101}, "gini_cpc": 0.02825, "gini_cps": 0.011662, "bpic": 6.014843, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9125262706138803, "fi": 0.8815969543188723}, "t": 297.4} |
| {"event": "eval", "step": 3400, "phase": 2, "bpb": {"en": 1.97994, "tr": 2.15666, "fi": 2.31922}, "bytes_per_chunk": {"en": 4.989, "tr": 6.134, "fi": 6.029}, "bpb_mean": 2.15194, "frozen_inventory": false, "t": 297.9} |
| {"event": "train", "step": 3400, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.38543, "bpb": 1.99875, "chunk_loss": 0.030271, "F": 0.1947, "G": 0.2063, "chunks": 199.38, "bytes_per_chunk": 5.136, "boundary_entropy": 0.09228, "router_noise": 0.0, "grad_norm": 0.7163, "lr": 0.0001465, "router_trainable": true, "sec_per_step": 0.099, "alpha": 0.03, "t": 298.2} |
| {"event": "train", "step": 3450, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.53928, "bpb": 2.22071, "chunk_loss": 0.030302, "F": 0.19342, "G": 0.21054, "chunks": 198.06, "bytes_per_chunk": 5.17, "boundary_entropy": 0.10621, "router_noise": 0.0, "grad_norm": 0.7588, "lr": 0.0001428, "router_trainable": true, "sec_per_step": 0.065, "alpha": 0.03, "t": 301.5} |
| {"event": "train", "step": 3500, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.41507, "bpb": 2.04151, "chunk_loss": 0.030253, "F": 0.19257, "G": 0.20807, "chunks": 197.19, "bytes_per_chunk": 5.193, "boundary_entropy": 0.10107, "router_noise": 0.0, "grad_norm": 0.7104, "lr": 0.0001391, "router_trainable": true, "sec_per_step": 0.065, "alpha": 0.03, "t": 304.7} |
| {"event": "train", "step": 3550, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.41078, "bpb": 2.03533, "chunk_loss": 0.030215, "F": 0.19144, "G": 0.20532, "chunks": 196.03, "bytes_per_chunk": 5.224, "boundary_entropy": 0.10156, "router_noise": 0.0, "grad_norm": 0.6819, "lr": 0.0001355, "router_trainable": true, "sec_per_step": 0.071, "alpha": 0.03, "t": 308.3} |
| {"event": "probe", "step": 3600, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3275.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.188186, "cps": 25.585938, "mean_p": 0.193201, "boundary_entropy": 0.077258, "bpb": 1.946868}, "tr": {"chunks": 3151.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.17406, "cps": 24.617188, "mean_p": 0.228881, "boundary_entropy": 0.229191, "bpb": 2.140499}, "fi": {"chunks": 3137.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.166764, "cps": 24.507812, "mean_p": 0.197978, "boundary_entropy": 0.172435, "bpb": 2.29085}}, "bpb": {"en": 1.94687, "tr": 2.1405, "fi": 2.29085}, "gini_cpc": 0.026996, "gini_cps": 0.00962, "bpic": 5.932448, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9249338370385238, "fi": 0.8861667507358248}, "t": 311.4} |
| {"event": "eval", "step": 3600, "phase": 2, "bpb": {"en": 1.96281, "tr": 2.14745, "fi": 2.30064}, "bytes_per_chunk": {"en": 4.964, "tr": 5.992, "fi": 5.969}, "bpb_mean": 2.13697, "frozen_inventory": false, "t": 311.8} |
| {"event": "train", "step": 3600, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.43237, "bpb": 2.06647, "chunk_loss": 0.030237, "F": 0.19128, "G": 0.21493, "chunks": 195.88, "bytes_per_chunk": 5.228, "boundary_entropy": 0.12057, "router_noise": 0.0, "grad_norm": 0.7292, "lr": 0.0001319, "router_trainable": true, "sec_per_step": 0.077, "alpha": 0.03, "t": 312.1} |
| {"event": "train", "step": 3650, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.36031, "bpb": 1.96251, "chunk_loss": 0.030281, "F": 0.19617, "G": 0.20601, "chunks": 200.88, "bytes_per_chunk": 5.098, "boundary_entropy": 0.08504, "router_noise": 0.0, "grad_norm": 0.6861, "lr": 0.0001283, "router_trainable": true, "sec_per_step": 0.073, "alpha": 0.03, "t": 315.7} |
| {"event": "train", "step": 3700, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.46343, "bpb": 2.11128, "chunk_loss": 0.030181, "F": 0.18521, "G": 0.20562, "chunks": 189.66, "bytes_per_chunk": 5.399, "boundary_entropy": 0.11075, "router_noise": 0.0, "grad_norm": 0.7695, "lr": 0.0001247, "router_trainable": true, "sec_per_step": 0.075, "alpha": 0.03, "t": 319.5} |
| {"event": "train", "step": 3750, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.44836, "bpb": 2.08955, "chunk_loss": 0.030445, "F": 0.19907, "G": 0.21526, "chunks": 203.84, "bytes_per_chunk": 5.023, "boundary_entropy": 0.10731, "router_noise": 0.0, "grad_norm": 0.7744, "lr": 0.0001212, "router_trainable": true, "sec_per_step": 0.085, "alpha": 0.03, "t": 323.7} |
| {"event": "probe", "step": 3800, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3271.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.187956, "cps": 25.554688, "mean_p": 0.192301, "boundary_entropy": 0.072262, "bpb": 1.93168}, "tr": {"chunks": 3163.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.174722, "cps": 24.710938, "mean_p": 0.220152, "boundary_entropy": 0.217548, "bpb": 2.125734}, "fi": {"chunks": 3219.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.171123, "cps": 25.148438, "mean_p": 0.197169, "boundary_entropy": 0.164522, "bpb": 2.271604}}, "bpb": {"en": 1.93168, "tr": 2.12573, "fi": 2.2716}, "gini_cpc": 0.021023, "gini_cps": 0.007459, "bpic": 5.877137, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9295916553993755, "fi": 0.910442807478748}, "t": 327.8} |
| {"event": "eval", "step": 3800, "phase": 2, "bpb": {"en": 1.94934, "tr": 2.13246, "fi": 2.28223}, "bytes_per_chunk": {"en": 4.968, "tr": 6.027, "fi": 5.89}, "bpb_mean": 2.12135, "frozen_inventory": false, "t": 328.2} |
| {"event": "train", "step": 3800, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.342, "bpb": 1.9361, "chunk_loss": 0.030261, "F": 0.19492, "G": 0.20386, "chunks": 199.59, "bytes_per_chunk": 5.13, "boundary_entropy": 0.08352, "router_noise": 0.0, "grad_norm": 0.7332, "lr": 0.0001177, "router_trainable": true, "sec_per_step": 0.096, "alpha": 0.03, "t": 328.5} |
| {"event": "train", "step": 3850, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.35874, "bpb": 1.96025, "chunk_loss": 0.030279, "F": 0.19617, "G": 0.20666, "chunks": 200.88, "bytes_per_chunk": 5.098, "boundary_entropy": 0.08851, "router_noise": 0.0, "grad_norm": 0.7029, "lr": 0.0001142, "router_trainable": true, "sec_per_step": 0.07, "alpha": 0.03, "t": 332.0} |
| {"event": "train", "step": 3900, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.42346, "bpb": 2.05362, "chunk_loss": 0.030227, "F": 0.19135, "G": 0.20936, "chunks": 195.94, "bytes_per_chunk": 5.226, "boundary_entropy": 0.10234, "router_noise": 0.0, "grad_norm": 0.7104, "lr": 0.0001108, "router_trainable": true, "sec_per_step": 0.062, "alpha": 0.03, "t": 335.1} |
| {"event": "train", "step": 3950, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.42537, "bpb": 2.05637, "chunk_loss": 0.030275, "F": 0.19574, "G": 0.20438, "chunks": 200.44, "bytes_per_chunk": 5.109, "boundary_entropy": 0.08349, "router_noise": 0.0, "grad_norm": 0.7408, "lr": 0.0001074, "router_trainable": true, "sec_per_step": 0.066, "alpha": 0.03, "t": 338.4} |
| {"event": "probe", "step": 4000, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3282.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.188588, "cps": 25.640625, "mean_p": 0.192016, "boundary_entropy": 0.072066, "bpb": 1.922586}, "tr": {"chunks": 3164.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.174778, "cps": 24.71875, "mean_p": 0.220816, "boundary_entropy": 0.215256, "bpb": 2.106659}, "fi": {"chunks": 3169.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.168465, "cps": 24.757812, "mean_p": 0.192645, "boundary_entropy": 0.157753, "bpb": 2.265242}}, "bpb": {"en": 1.92259, "tr": 2.10666, "fi": 2.26524}, "gini_cpc": 0.025225, "gini_cps": 0.008182, "bpic": 5.900364, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9267689327200508, "fi": 0.8932970488600304}, "t": 342.1} |
| {"event": "eval", "step": 4000, "phase": 2, "bpb": {"en": 1.93997, "tr": 2.11298, "fi": 2.27516}, "bytes_per_chunk": {"en": 4.966, "tr": 6.015, "fi": 5.973}, "bpb_mean": 2.10937, "frozen_inventory": false, "t": 342.8} |
| {"event": "train", "step": 4000, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.36639, "bpb": 1.97129, "chunk_loss": 0.0302, "F": 0.18787, "G": 0.20589, "chunks": 192.38, "bytes_per_chunk": 5.323, "boundary_entropy": 0.1072, "router_noise": 0.0, "grad_norm": 0.7477, "lr": 0.000104, "router_trainable": true, "sec_per_step": 0.093, "alpha": 0.03, "t": 343.1} |
| {"event": "train", "step": 4050, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.49171, "bpb": 2.15208, "chunk_loss": 0.030257, "F": 0.18845, "G": 0.20847, "chunks": 192.97, "bytes_per_chunk": 5.307, "boundary_entropy": 0.1057, "router_noise": 0.0, "grad_norm": 0.75, "lr": 0.0001007, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 346.7} |
| {"event": "probe", "step": 4096, "phase": 2, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3281.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.188531, "cps": 25.632812, "mean_p": 0.192145, "boundary_entropy": 0.069664, "bpb": 1.917746}, "tr": {"chunks": 3123.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.172513, "cps": 24.398438, "mean_p": 0.219615, "boundary_entropy": 0.212568, "bpb": 2.100109}, "fi": {"chunks": 3171.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.168572, "cps": 24.773438, "mean_p": 0.195176, "boundary_entropy": 0.160258, "bpb": 2.250079}}, "bpb": {"en": 1.91775, "tr": 2.10011, "fi": 2.25008}, "gini_cpc": 0.025124, "gini_cps": 0.011001, "bpic": 5.925013, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9150384059059387, "fi": 0.8941332565421501}, "t": 350.5} |
| {"event": "train", "step": 4100, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.43703, "bpb": 2.07319, "chunk_loss": 0.030246, "F": 0.19125, "G": 0.21148, "chunks": 195.84, "bytes_per_chunk": 5.229, "boundary_entropy": 0.10921, "router_noise": 0.0, "grad_norm": 0.7358, "lr": 9.75e-05, "router_trainable": true, "sec_per_step": 0.084, "alpha": 0.03, "t": 351.0} |
| {"event": "train", "step": 4150, "arch": "hnet", "phase": 2, "objective": "global", "lm_loss": 1.37443, "bpb": 1.98289, "chunk_loss": 0.030363, "F": 0.19919, "G": 0.21123, "chunks": 203.97, "bytes_per_chunk": 5.02, "boundary_entropy": 0.08747, "router_noise": 0.0, "grad_norm": 0.7008, "lr": 9.43e-05, "router_trainable": true, "sec_per_step": 0.067, "alpha": 0.03, "t": 354.3} |
| {"event": "probe", "step": 4200, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3263.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.187496, "cps": 25.492188, "mean_p": 0.191097, "boundary_entropy": 0.068763, "bpb": 1.911659}, "tr": {"chunks": 3125.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.172623, "cps": 24.414062, "mean_p": 0.223081, "boundary_entropy": 0.217803, "bpb": 2.097396}, "fi": {"chunks": 3116.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.165648, "cps": 24.34375, "mean_p": 0.190527, "boundary_entropy": 0.156839, "bpb": 2.247176}}, "bpb": {"en": 1.91166, "tr": 2.0974, "fi": 2.24718}, "gini_cpc": 0.027704, "gini_cps": 0.010311, "bpic": 5.969276, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9206753522920882, "fi": 0.8834716380386128}, "t": 358.5} |
| {"event": "eval", "step": 4200, "phase": 2, "bpb": {"en": 1.92791, "tr": 2.10486, "fi": 2.25384}, "bytes_per_chunk": {"en": 4.982, "tr": 6.071, "fi": 6.023}, "bpb_mean": 2.09554, "frozen_inventory": false, "t": 359.2} |
| {"event": "phase_transition", "step": 4200, "phase": 3, "objective": "global", "router_trainable": true, "encoder_trainable": true, "frozen_router_params": 0, "t": 359.6} |
| {"event": "train", "step": 4200, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.45227, "bpb": 2.09518, "chunk_loss": 0.030311, "F": 0.19641, "G": 0.20999, "chunks": 201.12, "bytes_per_chunk": 5.091, "boundary_entropy": 0.09147, "router_noise": 0.0, "grad_norm": 0.7559, "lr": 9.12e-05, "router_trainable": true, "sec_per_step": 0.108, "alpha": 0.03, "t": 359.7} |
| {"event": "train", "step": 4250, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.34596, "bpb": 1.94181, "chunk_loss": 0.030214, "F": 0.19092, "G": 0.20233, "chunks": 195.5, "bytes_per_chunk": 5.238, "boundary_entropy": 0.08668, "router_noise": 0.0, "grad_norm": 0.7407, "lr": 8.81e-05, "router_trainable": true, "sec_per_step": 0.075, "alpha": 0.03, "t": 363.5} |
| {"event": "train", "step": 4300, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.35454, "bpb": 1.95419, "chunk_loss": 0.030175, "F": 0.18652, "G": 0.20871, "chunks": 191.0, "bytes_per_chunk": 5.361, "boundary_entropy": 0.11199, "router_noise": 0.0, "grad_norm": 0.7154, "lr": 8.5e-05, "router_trainable": true, "sec_per_step": 0.075, "alpha": 0.03, "t": 367.2} |
| {"event": "train", "step": 4350, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.3304, "bpb": 1.91937, "chunk_loss": 0.030201, "F": 0.18927, "G": 0.2071, "chunks": 193.81, "bytes_per_chunk": 5.283, "boundary_entropy": 0.09664, "router_noise": 0.0, "grad_norm": 0.7344, "lr": 8.21e-05, "router_trainable": true, "sec_per_step": 0.082, "alpha": 0.03, "t": 371.3} |
| {"event": "probe", "step": 4400, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3260.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.187324, "cps": 25.46875, "mean_p": 0.190465, "boundary_entropy": 0.066153, "bpb": 1.90411}, "tr": {"chunks": 3188.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.176103, "cps": 24.90625, "mean_p": 0.223626, "boundary_entropy": 0.213214, "bpb": 2.081347}, "fi": {"chunks": 3153.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.167615, "cps": 24.632812, "mean_p": 0.191728, "boundary_entropy": 0.151596, "bpb": 2.231904}}, "bpb": {"en": 1.90411, "tr": 2.08135, "fi": 2.2319}, "gini_cpc": 0.024743, "gini_cps": 0.00743, "bpic": 5.908968, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9401004951557024, "fi": 0.8947848194813569}, "t": 375.1} |
| {"event": "eval", "step": 4400, "phase": 3, "bpb": {"en": 1.92171, "tr": 2.08659, "fi": 2.23838}, "bytes_per_chunk": {"en": 4.985, "tr": 5.946, "fi": 5.987}, "bpb_mean": 2.08223, "frozen_inventory": false, "t": 375.5} |
| {"event": "train", "step": 4400, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.37562, "bpb": 1.9846, "chunk_loss": 0.03031, "F": 0.19827, "G": 0.20517, "chunks": 203.03, "bytes_per_chunk": 5.044, "boundary_entropy": 0.07356, "router_noise": 0.0, "grad_norm": 0.8271, "lr": 7.92e-05, "router_trainable": true, "sec_per_step": 0.089, "alpha": 0.03, "t": 375.8} |
| {"event": "train", "step": 4450, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.39458, "bpb": 2.01196, "chunk_loss": 0.030192, "F": 0.1893, "G": 0.20607, "chunks": 193.84, "bytes_per_chunk": 5.283, "boundary_entropy": 0.09384, "router_noise": 0.0, "grad_norm": 0.7387, "lr": 7.63e-05, "router_trainable": true, "sec_per_step": 0.079, "alpha": 0.03, "t": 379.7} |
| {"event": "train", "step": 4500, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.41339, "bpb": 2.03909, "chunk_loss": 0.030213, "F": 0.18939, "G": 0.202, "chunks": 193.94, "bytes_per_chunk": 5.28, "boundary_entropy": 0.08825, "router_noise": 0.0, "grad_norm": 0.7384, "lr": 7.36e-05, "router_trainable": true, "sec_per_step": 0.059, "alpha": 0.03, "t": 382.6} |
| {"event": "train", "step": 4550, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.32433, "bpb": 1.91061, "chunk_loss": 0.030188, "F": 0.18561, "G": 0.20301, "chunks": 190.06, "bytes_per_chunk": 5.388, "boundary_entropy": 0.09985, "router_noise": 0.0, "grad_norm": 0.6888, "lr": 7.09e-05, "router_trainable": true, "sec_per_step": 0.074, "alpha": 0.03, "t": 386.4} |
| {"event": "probe", "step": 4600, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3263.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.187496, "cps": 25.492188, "mean_p": 0.191035, "boundary_entropy": 0.067416, "bpb": 1.891321}, "tr": {"chunks": 3147.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.173839, "cps": 24.585938, "mean_p": 0.22178, "boundary_entropy": 0.21041, "bpb": 2.06848}, "fi": {"chunks": 3144.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.167136, "cps": 24.5625, "mean_p": 0.192231, "boundary_entropy": 0.149256, "bpb": 2.217126}}, "bpb": {"en": 1.89132, "tr": 2.06848, "fi": 2.21713}, "gini_cpc": 0.025684, "gini_cps": 0.008304, "bpic": 5.938036, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9271569067722244, "fi": 0.891410407571694}, "t": 390.3} |
| {"event": "eval", "step": 4600, "phase": 3, "bpb": {"en": 1.90966, "tr": 2.0761, "fi": 2.22564}, "bytes_per_chunk": {"en": 4.987, "tr": 6.028, "fi": 5.959}, "bpb_mean": 2.07046, "frozen_inventory": false, "t": 390.8} |
| {"event": "train", "step": 4600, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.3907, "bpb": 2.00635, "chunk_loss": 0.030214, "F": 0.18954, "G": 0.20108, "chunks": 194.09, "bytes_per_chunk": 5.276, "boundary_entropy": 0.08646, "router_noise": 0.0, "grad_norm": 0.7219, "lr": 6.82e-05, "router_trainable": true, "sec_per_step": 0.094, "alpha": 0.03, "t": 391.1} |
| {"event": "train", "step": 4650, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.39644, "bpb": 2.01463, "chunk_loss": 0.030254, "F": 0.19464, "G": 0.20887, "chunks": 199.31, "bytes_per_chunk": 5.138, "boundary_entropy": 0.09248, "router_noise": 0.0, "grad_norm": 0.7776, "lr": 6.57e-05, "router_trainable": true, "sec_per_step": 0.087, "alpha": 0.03, "t": 395.4} |
| {"event": "train", "step": 4700, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.36938, "bpb": 1.9756, "chunk_loss": 0.030226, "F": 0.19067, "G": 0.20802, "chunks": 195.25, "bytes_per_chunk": 5.245, "boundary_entropy": 0.10217, "router_noise": 0.0, "grad_norm": 0.703, "lr": 6.32e-05, "router_trainable": true, "sec_per_step": 0.072, "alpha": 0.03, "t": 399.0} |
| {"event": "train", "step": 4750, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.45915, "bpb": 2.10511, "chunk_loss": 0.030247, "F": 0.18854, "G": 0.20599, "chunks": 193.06, "bytes_per_chunk": 5.304, "boundary_entropy": 0.10009, "router_noise": 0.0, "grad_norm": 0.8293, "lr": 6.08e-05, "router_trainable": true, "sec_per_step": 0.078, "alpha": 0.03, "t": 402.8} |
| {"event": "probe", "step": 4800, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3261.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.187381, "cps": 25.476562, "mean_p": 0.190424, "boundary_entropy": 0.065643, "bpb": 1.882894}, "tr": {"chunks": 3127.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.172734, "cps": 24.429688, "mean_p": 0.217948, "boundary_entropy": 0.205635, "bpb": 2.058196}, "fi": {"chunks": 3169.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.168465, "cps": 24.757812, "mean_p": 0.191761, "boundary_entropy": 0.150097, "bpb": 2.205164}}, "bpb": {"en": 1.88289, "tr": 2.0582, "fi": 2.20516}, "gini_cpc": 0.023858, "gini_cps": 0.009347, "bpic": 5.936172, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.921829604195272, "fi": 0.8990496517505735}, "t": 407.0} |
| {"event": "eval", "step": 4800, "phase": 3, "bpb": {"en": 1.90061, "tr": 2.06477, "fi": 2.21239}, "bytes_per_chunk": {"en": 4.985, "tr": 6.057, "fi": 5.96}, "bpb_mean": 2.05926, "frozen_inventory": false, "t": 407.4} |
| {"event": "train", "step": 4800, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.34667, "bpb": 1.94283, "chunk_loss": 0.0302, "F": 0.19131, "G": 0.20093, "chunks": 195.91, "bytes_per_chunk": 5.227, "boundary_entropy": 0.07852, "router_noise": 0.0, "grad_norm": 0.7112, "lr": 5.85e-05, "router_trainable": true, "sec_per_step": 0.096, "alpha": 0.03, "t": 407.6} |
| {"event": "train", "step": 4850, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.34122, "bpb": 1.93497, "chunk_loss": 0.030312, "F": 0.19818, "G": 0.20963, "chunks": 202.94, "bytes_per_chunk": 5.046, "boundary_entropy": 0.08468, "router_noise": 0.0, "grad_norm": 0.6805, "lr": 5.62e-05, "router_trainable": true, "sec_per_step": 0.083, "alpha": 0.03, "t": 411.8} |
| {"event": "train", "step": 4900, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.40085, "bpb": 2.021, "chunk_loss": 0.030229, "F": 0.19177, "G": 0.20621, "chunks": 196.38, "bytes_per_chunk": 5.215, "boundary_entropy": 0.09077, "router_noise": 0.0, "grad_norm": 0.8153, "lr": 5.41e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 416.1} |
| {"event": "train", "step": 4950, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.32595, "bpb": 1.91294, "chunk_loss": 0.030401, "F": 0.19336, "G": 0.21277, "chunks": 198.0, "bytes_per_chunk": 5.172, "boundary_entropy": 0.10342, "router_noise": 0.0, "grad_norm": 0.7381, "lr": 5.2e-05, "router_trainable": true, "sec_per_step": 0.077, "alpha": 0.03, "t": 419.9} |
| {"event": "probe", "step": 5000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3255.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.187037, "cps": 25.429688, "mean_p": 0.190254, "boundary_entropy": 0.063887, "bpb": 1.875042}, "tr": {"chunks": 3137.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.173286, "cps": 24.507812, "mean_p": 0.21676, "boundary_entropy": 0.202916, "bpb": 2.047164}, "fi": {"chunks": 3151.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.167508, "cps": 24.617188, "mean_p": 0.190257, "boundary_entropy": 0.144864, "bpb": 2.192922}}, "bpb": {"en": 1.87504, "tr": 2.04716, "fi": 2.19292}, "gini_cpc": 0.024665, "gini_cps": 0.008243, "bpic": 5.944881, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9264822313484717, "fi": 0.8955908482805065}, "t": 425.3} |
| {"event": "eval", "step": 5000, "phase": 3, "bpb": {"en": 1.89331, "tr": 2.05382, "fi": 2.20161}, "bytes_per_chunk": {"en": 4.991, "tr": 6.037, "fi": 5.975}, "bpb_mean": 2.04958, "frozen_inventory": false, "t": 425.8} |
| {"event": "train", "step": 5000, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.44161, "bpb": 2.0798, "chunk_loss": 0.030225, "F": 0.18842, "G": 0.20525, "chunks": 192.94, "bytes_per_chunk": 5.307, "boundary_entropy": 0.09666, "router_noise": 0.0, "grad_norm": 0.8116, "lr": 5e-05, "router_trainable": true, "sec_per_step": 0.126, "alpha": 0.03, "t": 426.2} |
| {"event": "train", "step": 5050, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.28489, "bpb": 1.85371, "chunk_loss": 0.030233, "F": 0.194, "G": 0.20542, "chunks": 198.66, "bytes_per_chunk": 5.155, "boundary_entropy": 0.0856, "router_noise": 0.0, "grad_norm": 0.7245, "lr": 4.81e-05, "router_trainable": true, "sec_per_step": 0.077, "alpha": 0.03, "t": 430.0} |
| {"event": "train", "step": 5100, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.34309, "bpb": 1.93767, "chunk_loss": 0.030153, "F": 0.18634, "G": 0.19832, "chunks": 190.81, "bytes_per_chunk": 5.367, "boundary_entropy": 0.08519, "router_noise": 0.0, "grad_norm": 0.7187, "lr": 4.63e-05, "router_trainable": true, "sec_per_step": 0.08, "alpha": 0.03, "t": 434.0} |
| {"event": "train", "step": 5150, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.37411, "bpb": 1.98242, "chunk_loss": 0.030316, "F": 0.19443, "G": 0.20811, "chunks": 199.09, "bytes_per_chunk": 5.143, "boundary_entropy": 0.09185, "router_noise": 0.0, "grad_norm": 0.7438, "lr": 4.45e-05, "router_trainable": true, "sec_per_step": 0.083, "alpha": 0.03, "t": 438.2} |
| {"event": "probe", "step": 5200, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3248.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.186634, "cps": 25.375, "mean_p": 0.18979, "boundary_entropy": 0.063563, "bpb": 1.869575}, "tr": {"chunks": 3105.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.171519, "cps": 24.257812, "mean_p": 0.216568, "boundary_entropy": 0.201286, "bpb": 2.04168}, "fi": {"chunks": 3138.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.166817, "cps": 24.515625, "mean_p": 0.188678, "boundary_entropy": 0.142262, "bpb": 2.187555}}, "bpb": {"en": 1.86958, "tr": 2.04168, "fi": 2.18755}, "gini_cpc": 0.025166, "gini_cps": 0.010045, "bpic": 5.977452, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9190077053608674, "fi": 0.8938181215633972}, "t": 442.2} |
| {"event": "eval", "step": 5200, "phase": 3, "bpb": {"en": 1.888, "tr": 2.04783, "fi": 2.19602}, "bytes_per_chunk": {"en": 4.996, "tr": 6.088, "fi": 5.994}, "bpb_mean": 2.04395, "frozen_inventory": false, "t": 442.6} |
| {"event": "train", "step": 5200, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.32677, "bpb": 1.91413, "chunk_loss": 0.030263, "F": 0.19385, "G": 0.20426, "chunks": 198.5, "bytes_per_chunk": 5.159, "boundary_entropy": 0.08236, "router_noise": 0.0, "grad_norm": 0.7311, "lr": 4.29e-05, "router_trainable": true, "sec_per_step": 0.093, "alpha": 0.03, "t": 442.8} |
| {"event": "train", "step": 5250, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.31112, "bpb": 1.89155, "chunk_loss": 0.030243, "F": 0.19128, "G": 0.20737, "chunks": 195.88, "bytes_per_chunk": 5.228, "boundary_entropy": 0.09505, "router_noise": 0.0, "grad_norm": 0.7091, "lr": 4.14e-05, "router_trainable": true, "sec_per_step": 0.075, "alpha": 0.03, "t": 446.6} |
| {"event": "train", "step": 5300, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.31142, "bpb": 1.89197, "chunk_loss": 0.030221, "F": 0.19293, "G": 0.20261, "chunks": 197.56, "bytes_per_chunk": 5.183, "boundary_entropy": 0.08003, "router_noise": 0.0, "grad_norm": 0.6826, "lr": 3.99e-05, "router_trainable": true, "sec_per_step": 0.075, "alpha": 0.03, "t": 450.3} |
| {"event": "train", "step": 5350, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.31394, "bpb": 1.89561, "chunk_loss": 0.030227, "F": 0.19196, "G": 0.20331, "chunks": 196.56, "bytes_per_chunk": 5.21, "boundary_entropy": 0.08133, "router_noise": 0.0, "grad_norm": 0.7348, "lr": 3.86e-05, "router_trainable": true, "sec_per_step": 0.086, "alpha": 0.03, "t": 454.6} |
| {"event": "probe", "step": 5400, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3256.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.187094, "cps": 25.4375, "mean_p": 0.189707, "boundary_entropy": 0.06418, "bpb": 1.865075}, "tr": {"chunks": 3149.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.173949, "cps": 24.601562, "mean_p": 0.219328, "boundary_entropy": 0.203634, "bpb": 2.032449}, "fi": {"chunks": 3164.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.168199, "cps": 24.71875, "mean_p": 0.191394, "boundary_entropy": 0.145167, "bpb": 2.179837}}, "bpb": {"en": 1.86507, "tr": 2.03245, "fi": 2.17984}, "gini_cpc": 0.023801, "gini_cps": 0.007455, "bpic": 5.928728, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9297406792228092, "fi": 0.8990095710897369}, "t": 459.7} |
| {"event": "eval", "step": 5400, "phase": 3, "bpb": {"en": 1.88276, "tr": 2.03972, "fi": 2.18817}, "bytes_per_chunk": {"en": 4.99, "tr": 6.025, "fi": 5.953}, "bpb_mean": 2.03688, "frozen_inventory": false, "t": 460.1} |
| {"event": "train", "step": 5400, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.38099, "bpb": 1.99235, "chunk_loss": 0.030247, "F": 0.19058, "G": 0.21345, "chunks": 195.16, "bytes_per_chunk": 5.247, "boundary_entropy": 0.11074, "router_noise": 0.0, "grad_norm": 0.8584, "lr": 3.73e-05, "router_trainable": true, "sec_per_step": 0.115, "alpha": 0.03, "t": 460.3} |
| {"event": "train", "step": 5450, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.3212, "bpb": 1.90608, "chunk_loss": 0.030226, "F": 0.19232, "G": 0.20468, "chunks": 196.94, "bytes_per_chunk": 5.2, "boundary_entropy": 0.08402, "router_noise": 0.0, "grad_norm": 0.7324, "lr": 3.62e-05, "router_trainable": true, "sec_per_step": 0.081, "alpha": 0.03, "t": 464.4} |
| {"event": "train", "step": 5500, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.31724, "bpb": 1.90037, "chunk_loss": 0.030293, "F": 0.19717, "G": 0.20952, "chunks": 201.91, "bytes_per_chunk": 5.072, "boundary_entropy": 0.08371, "router_noise": 0.0, "grad_norm": 0.7156, "lr": 3.51e-05, "router_trainable": true, "sec_per_step": 0.069, "alpha": 0.03, "t": 467.9} |
| {"event": "train", "step": 5550, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.31127, "bpb": 1.89177, "chunk_loss": 0.030258, "F": 0.19189, "G": 0.20608, "chunks": 196.5, "bytes_per_chunk": 5.211, "boundary_entropy": 0.09124, "router_noise": 0.0, "grad_norm": 0.6986, "lr": 3.41e-05, "router_trainable": true, "sec_per_step": 0.064, "alpha": 0.03, "t": 471.1} |
| {"event": "probe", "step": 5600, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3275.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.188186, "cps": 25.585938, "mean_p": 0.19101, "boundary_entropy": 0.063778, "bpb": 1.859443}, "tr": {"chunks": 3177.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.175496, "cps": 24.820312, "mean_p": 0.21873, "boundary_entropy": 0.202147, "bpb": 2.029862}, "fi": {"chunks": 3179.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.168997, "cps": 24.835938, "mean_p": 0.191827, "boundary_entropy": 0.146244, "bpb": 2.172883}}, "bpb": {"en": 1.85944, "tr": 2.02986, "fi": 2.17288}, "gini_cpc": 0.024016, "gini_cps": 0.006784, "bpic": 5.890562, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9325657887246557, "fi": 0.898031272103662}, "t": 474.4} |
| {"event": "eval", "step": 5600, "phase": 3, "bpb": {"en": 1.8779, "tr": 2.03559, "fi": 2.18073}, "bytes_per_chunk": {"en": 4.971, "tr": 5.938, "fi": 5.934}, "bpb_mean": 2.03141, "frozen_inventory": false, "t": 474.8} |
| {"event": "train", "step": 5600, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.36778, "bpb": 1.97329, "chunk_loss": 0.030191, "F": 0.18817, "G": 0.20505, "chunks": 192.69, "bytes_per_chunk": 5.314, "boundary_entropy": 0.09662, "router_noise": 0.0, "grad_norm": 0.7425, "lr": 3.33e-05, "router_trainable": true, "sec_per_step": 0.082, "alpha": 0.03, "t": 475.2} |
| {"event": "train", "step": 5650, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.32057, "bpb": 1.90517, "chunk_loss": 0.0302, "F": 0.18744, "G": 0.20019, "chunks": 191.94, "bytes_per_chunk": 5.335, "boundary_entropy": 0.0889, "router_noise": 0.0, "grad_norm": 0.7627, "lr": 3.25e-05, "router_trainable": true, "sec_per_step": 0.081, "alpha": 0.03, "t": 479.2} |
| {"event": "train", "step": 5700, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.44989, "bpb": 2.09176, "chunk_loss": 0.030185, "F": 0.18579, "G": 0.20399, "chunks": 190.25, "bytes_per_chunk": 5.382, "boundary_entropy": 0.10031, "router_noise": 0.0, "grad_norm": 0.8479, "lr": 3.18e-05, "router_trainable": true, "sec_per_step": 0.076, "alpha": 0.03, "t": 483.1} |
| {"event": "train", "step": 5750, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.27815, "bpb": 1.84398, "chunk_loss": 0.030319, "F": 0.19662, "G": 0.21032, "chunks": 201.34, "bytes_per_chunk": 5.086, "boundary_entropy": 0.08883, "router_noise": 0.0, "grad_norm": 0.7455, "lr": 3.13e-05, "router_trainable": true, "sec_per_step": 0.078, "alpha": 0.03, "t": 486.9} |
| {"event": "probe", "step": 5800, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3261.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.187381, "cps": 25.476562, "mean_p": 0.189762, "boundary_entropy": 0.063071, "bpb": 1.854553}, "tr": {"chunks": 3146.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.173783, "cps": 24.578125, "mean_p": 0.218364, "boundary_entropy": 0.201589, "bpb": 2.023461}, "fi": {"chunks": 3152.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.167562, "cps": 24.625, "mean_p": 0.189846, "boundary_entropy": 0.142613, "bpb": 2.169015}}, "bpb": {"en": 1.85455, "tr": 2.02346, "fi": 2.16902}, "gini_cpc": 0.024991, "gini_cps": 0.00802, "bpic": 5.93493, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9274307434596502, "fi": 0.8942267284057456}, "t": 491.1} |
| {"event": "eval", "step": 5800, "phase": 3, "bpb": {"en": 1.87285, "tr": 2.03079, "fi": 2.1776}, "bytes_per_chunk": {"en": 4.985, "tr": 6.026, "fi": 5.973}, "bpb_mean": 2.02708, "frozen_inventory": false, "t": 491.5} |
| {"event": "train", "step": 5800, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.26046, "bpb": 1.81846, "chunk_loss": 0.03025, "F": 0.19363, "G": 0.20689, "chunks": 198.28, "bytes_per_chunk": 5.164, "boundary_entropy": 0.08847, "router_noise": 0.0, "grad_norm": 0.7793, "lr": 3.08e-05, "router_trainable": true, "sec_per_step": 0.1, "alpha": 0.03, "t": 492.0} |
| {"event": "train", "step": 5850, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.37973, "bpb": 1.99053, "chunk_loss": 0.03026, "F": 0.18942, "G": 0.20906, "chunks": 193.97, "bytes_per_chunk": 5.279, "boundary_entropy": 0.10616, "router_noise": 0.0, "grad_norm": 0.7654, "lr": 3.05e-05, "router_trainable": true, "sec_per_step": 0.076, "alpha": 0.03, "t": 495.8} |
| {"event": "train", "step": 5900, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.36459, "bpb": 1.96869, "chunk_loss": 0.030261, "F": 0.19315, "G": 0.20671, "chunks": 197.78, "bytes_per_chunk": 5.177, "boundary_entropy": 0.09004, "router_noise": 0.0, "grad_norm": 0.7438, "lr": 3.02e-05, "router_trainable": true, "sec_per_step": 0.068, "alpha": 0.03, "t": 499.2} |
| {"event": "train", "step": 5950, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.30937, "bpb": 1.88903, "chunk_loss": 0.030177, "F": 0.1861, "G": 0.20104, "chunks": 190.56, "bytes_per_chunk": 5.374, "boundary_entropy": 0.09425, "router_noise": 0.0, "grad_norm": 0.7396, "lr": 3.01e-05, "router_trainable": true, "sec_per_step": 0.08, "alpha": 0.03, "t": 503.2} |
| {"event": "train", "step": 5999, "arch": "hnet", "phase": 3, "objective": "global", "lm_loss": 1.36009, "bpb": 1.96219, "chunk_loss": 0.030262, "F": 0.19427, "G": 0.20644, "chunks": 198.94, "bytes_per_chunk": 5.147, "boundary_entropy": 0.0852, "router_noise": 0.0, "grad_norm": 0.7574, "lr": 3e-05, "router_trainable": true, "sec_per_step": 0.077, "alpha": 0.03, "t": 507.0} |
| {"event": "probe", "step": 6000, "phase": 3, "objective": "global", "router_trainable": true, "per_lang": {"en": {"chunks": 3261.0, "chars": 17403.0, "bytes": 17418.0, "cpc": 0.187381, "cps": 25.476562, "mean_p": 0.190147, "boundary_entropy": 0.062355, "bpb": 1.853007}, "tr": {"chunks": 3124.0, "chars": 18103.0, "bytes": 19742.0, "cpc": 0.172568, "cps": 24.40625, "mean_p": 0.216311, "boundary_entropy": 0.200598, "bpb": 2.018696}, "fi": {"chunks": 3143.0, "chars": 18811.0, "bytes": 19572.0, "cpc": 0.167083, "cps": 24.554688, "mean_p": 0.189809, "boundary_entropy": 0.141131, "bpb": 2.163495}}, "bpb": {"en": 1.85301, "tr": 2.0187, "fi": 2.16349}, "gini_cpc": 0.025676, "gini_cps": 0.009586, "bpic": 5.95424, "worst_lang": "en", "cpc_parity_vs_en": {"en": 1.0, "tr": 0.9209452137851072, "fi": 0.891673416046719}, "t": 507.4} |
| {"event": "eval", "step": 6000, "phase": 3, "bpb": {"en": 1.87028, "tr": 2.02556, "fi": 2.17116}, "bytes_per_chunk": {"en": 4.985, "tr": 6.07, "fi": 5.985}, "bpb_mean": 2.02233, "frozen_inventory": false, "t": 507.8} |
| {"event": "reps_dumped", "dirs": ["/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s1/probe/reps/th0.30", "/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s1/probe/reps/th0.40", "/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s1/probe/reps/th0.50", "/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s1/probe/reps/th0.60", "/local/scratch/sas245/beetle-hnet/runs/ali_hnet_s1/probe/reps/th0.70"], "t": 508.9} |
| {"event": "final", "run": "ali_hnet_s1", "arch": "hnet", "objective": "ratio", "condition": "A", "step": 6000, "params": 22516992, "seed": 1, "per_lang": {"en": {"cpc": 0.18738148595069815, "cps": 25.4765625, "bpb": 1.877256191804282, "bytes_per_chunk": 4.972806859462587}, "tr": {"cpc": 0.17256808263823675, "cps": 24.40625, "bpb": 2.0177143367726122, "bytes_per_chunk": 6.104961632063717}, "fi": {"cpc": 0.1670830896815693, "cps": 24.5546875, "bpb": 2.190488376552282, "bytes_per_chunk": 5.963917643044022}}, "bpb_mean": 2.028486301709725, "gini_cpc": 0.025676329477999138, "gini_cps": 0.009585782255807418, "bpic": 5.95424013434089, "mean_boundary_entropy": 0.08520237356424332, "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": 509.7798869609833, "t": 509.8} |
| {"event": "hub_done", "repo": "AdaptiveChunking/ali_hnet_s1", "pushed": [0, 1, 2, 4, 6, 8, 16, 32, 64, 128, 256, 512, 1000, 1024, 2048, 3000, 4096, 6000], "failed": [], "ok": true, "t": 525.2} |
|
|
