{ "config": { "model": "LiquidAI/LFM2.5-Encoder-350M", "dataset": "HuggingFaceFW/fineweb-edu", "dataset_config": "sample-10BT", "split": "train", "text_column": "text", "output_dir": "/tmp/minima", "output_repo": null, "sequence_length": 256, "batch_size": 4, "gradient_accumulation": 4, "steps": 2000, "learning_rate": 5e-05, "warmup_steps": 100, "group_size": 32, "recovery_rank": 128, "teacher_topk": 32, "hidden_loss_weight": 2.0, "distill_loss_weight": 1.0, "mlm_loss_weight": 1.0, "temperature": 2.0, "seed": 42, "log_every": 50, "train_ternary_weights": false }, "duration_seconds": 667.6906430721283, "trainable_parameters": 56492032, "history": [ { "step": 1, "lr": 5e-07, "elapsed_s": 7.1458351612091064, "loss": 7.548747301101685, "hidden": 0.42470795661211014, "mlm": 2.650780439376831, "distill": 4.048550963401794, "teacher_acc": 0.7796393632888794 }, { "step": 51, "lr": 2.5500000000000003e-05, "elapsed_s": 25.010111331939697, "loss": 4.090662062168121, "hidden": 0.30477869510650635, "mlm": 1.5020031332969666, "distill": 1.9791015088558197, "teacher_acc": 0.7966673225164413 }, { "step": 101, "lr": 5e-05, "elapsed_s": 42.059675216674805, "loss": 3.8598657846450806, "hidden": 0.2699665352702141, "mlm": 1.7310407161712646, "distill": 1.5888920724391937, "teacher_acc": 0.7436814606189728 }, { "step": 151, "lr": 4.991461232516675e-05, "elapsed_s": 58.84036660194397, "loss": 2.8976023197174072, "hidden": 0.2286553680896759, "mlm": 1.2483952790498734, "distill": 1.191896229982376, "teacher_acc": 0.8049547970294952 }, { "step": 201, "lr": 4.965903258506806e-05, "elapsed_s": 75.38389277458191, "loss": 2.984165906906128, "hidden": 0.22897232323884964, "mlm": 1.3140619993209839, "distill": 1.2121593058109283, "teacher_acc": 0.8042650818824768 }, { "step": 251, "lr": 4.923500664848326e-05, "elapsed_s": 91.81105828285217, "loss": 2.715985119342804, "hidden": 0.19838016480207443, "mlm": 1.3606167435646057, "distill": 0.9586080610752106, "teacher_acc": 0.7605845779180527 }, { "step": 301, "lr": 4.864543104251587e-05, "elapsed_s": 108.18112087249756, "loss": 2.683966338634491, "hidden": 0.18938461691141129, "mlm": 1.3027366399765015, "distill": 1.002460464835167, "teacher_acc": 0.7672062069177628 }, { "step": 351, "lr": 4.789433316637644e-05, "elapsed_s": 124.41784739494324, "loss": 2.9636711478233337, "hidden": 0.19395725429058075, "mlm": 1.370901107788086, "distill": 1.204855501651764, "teacher_acc": 0.7925134599208832 }, { "step": 401, "lr": 4.698684378016222e-05, "elapsed_s": 140.6487717628479, "loss": 2.5739916563034058, "hidden": 0.1828564964234829, "mlm": 1.262264758348465, "distill": 0.9460138827562332, "teacher_acc": 0.7860687077045441 }, { "step": 451, "lr": 4.592916195656322e-05, "elapsed_s": 156.78596544265747, "loss": 2.4724334478378296, "hidden": 0.18117711693048477, "mlm": 1.1873478889465332, "distill": 0.9227313250303268, "teacher_acc": 0.8220411986112595 }, { "step": 501, "lr": 4.4728512734909844e-05, "elapsed_s": 173.03300976753235, "loss": 2.516365110874176, "hidden": 0.18044547364115715, "mlm": 1.1758874356746674, "distill": 0.9795867055654526, "teacher_acc": 0.7911238372325897 }, { "step": 551, "lr": 4.3393097766828293e-05, "elapsed_s": 189.45385837554932, "loss": 2.470432221889496, "hidden": 0.17229371145367622, "mlm": 1.2614093571901321, "distill": 0.864435464143753, "teacher_acc": 0.7440680414438248 }, { "step": 601, "lr": 4.193203929064353e-05, "elapsed_s": 205.8019621372223, "loss": 2.225749373435974, "hidden": 0.17066730186343193, "mlm": 1.0380512028932571, "distill": 0.8463635146617889, "teacher_acc": 0.7945135682821274 }, { "step": 651, "lr": 4.03553178172417e-05, "elapsed_s": 222.18287634849548, "loss": 2.412706196308136, "hidden": 0.1683909185230732, "mlm": 1.1580713838338852, "distill": 0.9178528636693954, "teacher_acc": 0.7857213616371155 }, { "step": 701, "lr": 3.867370395306068e-05, "elapsed_s": 238.73087334632874, "loss": 2.531893312931061, "hidden": 0.1756228394806385, "mlm": 1.2564030885696411, "distill": 0.9242444187402725, "teacher_acc": 0.7829260528087616 }, { "step": 751, "lr": 3.689868482592684e-05, "elapsed_s": 254.98722314834595, "loss": 2.4058578610420227, "hidden": 0.1706235148012638, "mlm": 1.092416599392891, "distill": 0.9721942096948624, "teacher_acc": 0.807672306895256 }, { "step": 801, "lr": 3.504238561632424e-05, "elapsed_s": 271.2415738105774, "loss": 2.4489212036132812, "hidden": 0.16438886150717735, "mlm": 1.249473124742508, "distill": 0.8706703633069992, "teacher_acc": 0.7654353976249695 }, { "step": 851, "lr": 3.311748673011709e-05, "elapsed_s": 287.5516035556793, "loss": 2.139587104320526, "hidden": 0.16303318738937378, "mlm": 1.0034526735544205, "distill": 0.8100681155920029, "teacher_acc": 0.8189881294965744 }, { "step": 901, "lr": 3.1137137178519985e-05, "elapsed_s": 303.8067092895508, "loss": 2.2904033064842224, "hidden": 0.1645217277109623, "mlm": 1.1553480178117752, "distill": 0.8060119301080704, "teacher_acc": 0.7967804372310638 }, { "step": 951, "lr": 2.9114864757018352e-05, "elapsed_s": 320.1048364639282, "loss": 2.320096254348755, "hidden": 0.1607908457517624, "mlm": 1.1159868836402893, "distill": 0.8825277388095856, "teacher_acc": 0.7653505355119705 }, { "step": 1001, "lr": 2.7064483636808313e-05, "elapsed_s": 336.35610246658325, "loss": 2.4431122541427612, "hidden": 0.1690278798341751, "mlm": 1.2797610759735107, "distill": 0.825295478105545, "teacher_acc": 0.7673374265432358 }, { "step": 1051, "lr": 2.5e-05, "elapsed_s": 352.62019753456116, "loss": 2.4243820905685425, "hidden": 0.15917402133345604, "mlm": 1.2654687762260437, "distill": 0.8405653089284897, "teacher_acc": 0.7821898609399796 }, { "step": 1101, "lr": 2.2935516363191693e-05, "elapsed_s": 368.93653750419617, "loss": 2.39467990398407, "hidden": 0.1649530865252018, "mlm": 1.2698808908462524, "distill": 0.7948928475379944, "teacher_acc": 0.7567200213670731 }, { "step": 1151, "lr": 2.088513524298165e-05, "elapsed_s": 385.3177971839905, "loss": 2.786853849887848, "hidden": 0.17868195474147797, "mlm": 1.416623666882515, "distill": 1.0128663033246994, "teacher_acc": 0.747254803776741 }, { "step": 1201, "lr": 1.8862862821480025e-05, "elapsed_s": 401.6415090560913, "loss": 2.155725210905075, "hidden": 0.15804878994822502, "mlm": 0.908401682972908, "distill": 0.9312258809804916, "teacher_acc": 0.8400556594133377 }, { "step": 1251, "lr": 1.6882513269882917e-05, "elapsed_s": 417.91031408309937, "loss": 2.757974684238434, "hidden": 0.16840778663754463, "mlm": 1.393098160624504, "distill": 1.0280609875917435, "teacher_acc": 0.755160465836525 }, { "step": 1301, "lr": 1.495761438367577e-05, "elapsed_s": 434.33623337745667, "loss": 2.444422423839569, "hidden": 0.15633703023195267, "mlm": 1.2066814005374908, "distill": 0.9250669926404953, "teacher_acc": 0.7845637202262878 }, { "step": 1351, "lr": 1.3101315174073162e-05, "elapsed_s": 450.75909185409546, "loss": 2.466778337955475, "hidden": 0.16462653502821922, "mlm": 1.2565617859363556, "distill": 0.880963459610939, "teacher_acc": 0.762797936797142 }, { "step": 1401, "lr": 1.1326296046939333e-05, "elapsed_s": 467.06108593940735, "loss": 2.127809077501297, "hidden": 0.1562965139746666, "mlm": 1.076869934797287, "distill": 0.7383460700511932, "teacher_acc": 0.79219089448452 }, { "step": 1451, "lr": 9.644682182758306e-06, "elapsed_s": 483.31524443626404, "loss": 2.3861597776412964, "hidden": 0.15558123216032982, "mlm": 1.28284552693367, "distill": 0.7921517193317413, "teacher_acc": 0.7626918703317642 }, { "step": 1501, "lr": 8.067960709356478e-06, "elapsed_s": 499.5990273952484, "loss": 2.2876625657081604, "hidden": 0.15779289230704308, "mlm": 1.210097774863243, "distill": 0.7619789987802505, "teacher_acc": 0.7711311429738998 }, { "step": 1551, "lr": 6.606902233171711e-06, "elapsed_s": 515.7814161777496, "loss": 1.987428605556488, "hidden": 0.1505429930984974, "mlm": 0.9722132682800293, "distill": 0.7141293436288834, "teacher_acc": 0.81315678358078 }, { "step": 1601, "lr": 5.271487265090163e-06, "elapsed_s": 531.9731659889221, "loss": 2.178088665008545, "hidden": 0.1585002951323986, "mlm": 1.1130180209875107, "distill": 0.7480701059103012, "teacher_acc": 0.7750980257987976 }, { "step": 1651, "lr": 4.070838043436786e-06, "elapsed_s": 549.6739189624786, "loss": 2.2565123438835144, "hidden": 0.15409696847200394, "mlm": 1.1320631355047226, "distill": 0.8162552565336227, "teacher_acc": 0.7810630649328232 }, { "step": 1701, "lr": 3.013156219837776e-06, "elapsed_s": 569.2521679401398, "loss": 2.5504953861236572, "hidden": 0.1535499133169651, "mlm": 1.3510537147521973, "distill": 0.8923418670892715, "teacher_acc": 0.7388543039560318 }, { "step": 1751, "lr": 2.1056668336235622e-06, "elapsed_s": 585.5749201774597, "loss": 2.126615345478058, "hidden": 0.15592899918556213, "mlm": 1.0641395151615143, "distill": 0.7506178617477417, "teacher_acc": 0.8043123483657837 }, { "step": 1801, "lr": 1.3545689574841342e-06, "elapsed_s": 602.0534913539886, "loss": 2.315759599208832, "hidden": 0.15497129783034325, "mlm": 1.1510487794876099, "distill": 0.8547682464122772, "teacher_acc": 0.7958026677370071 }, { "step": 1851, "lr": 7.649933515167407e-07, "elapsed_s": 618.2598822116852, "loss": 2.6134771704673767, "hidden": 0.16912326216697693, "mlm": 1.3344760239124298, "distill": 0.9407546371221542, "teacher_acc": 0.7435210645198822 }, { "step": 1901, "lr": 3.4096741493194197e-07, "elapsed_s": 634.6096048355103, "loss": 2.525772988796234, "hidden": 0.15609804540872574, "mlm": 1.3272225856781006, "distill": 0.8863543570041656, "teacher_acc": 0.7534686625003815 }, { "step": 1951, "lr": 8.538767483325383e-08, "elapsed_s": 650.8219337463379, "loss": 2.462930738925934, "hidden": 0.15661108866333961, "mlm": 1.2163092494010925, "distill": 0.9333993047475815, "teacher_acc": 0.7857980132102966 }, { "step": 2000, "lr": 3.417452268950072e-11, "elapsed_s": 666.7361209392548, "loss": 2.3209976255893707, "hidden": 0.15706806257367134, "mlm": 1.1457223743200302, "distill": 0.8611391484737396, "teacher_acc": 0.7896232903003693 } ] }