| { | |
| "best_metric": 0.7303625345230103, | |
| "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e3l17-l/checkpoint-21000", | |
| "epoch": 2.2397610921501707, | |
| "eval_steps": 500, | |
| "global_step": 21000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.05332764505119454, | |
| "grad_norm": 246.1711883544922, | |
| "learning_rate": 9.822241183162683e-08, | |
| "loss": 0.4515, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.05332764505119454, | |
| "eval_loss": 1.3850886821746826, | |
| "eval_runtime": 63.8439, | |
| "eval_samples_per_second": 15.46, | |
| "eval_steps_per_second": 1.942, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.10665529010238908, | |
| "grad_norm": 81.48503875732422, | |
| "learning_rate": 9.64448236632537e-08, | |
| "loss": 0.4148, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.10665529010238908, | |
| "eval_loss": 1.284144401550293, | |
| "eval_runtime": 63.853, | |
| "eval_samples_per_second": 15.457, | |
| "eval_steps_per_second": 1.942, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.1599829351535836, | |
| "grad_norm": 718.367919921875, | |
| "learning_rate": 9.466723549488054e-08, | |
| "loss": 0.3281, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.1599829351535836, | |
| "eval_loss": 1.2112621068954468, | |
| "eval_runtime": 64.127, | |
| "eval_samples_per_second": 15.391, | |
| "eval_steps_per_second": 1.934, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.21331058020477817, | |
| "grad_norm": 4.5923566818237305, | |
| "learning_rate": 9.288964732650739e-08, | |
| "loss": 0.2912, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.21331058020477817, | |
| "eval_loss": 1.159011960029602, | |
| "eval_runtime": 63.7062, | |
| "eval_samples_per_second": 15.493, | |
| "eval_steps_per_second": 1.946, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.2666382252559727, | |
| "grad_norm": 319.4483642578125, | |
| "learning_rate": 9.111205915813424e-08, | |
| "loss": 0.3073, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.2666382252559727, | |
| "eval_loss": 1.1095446348190308, | |
| "eval_runtime": 63.9925, | |
| "eval_samples_per_second": 15.424, | |
| "eval_steps_per_second": 1.938, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.3199658703071672, | |
| "grad_norm": 0.00027541385497897863, | |
| "learning_rate": 8.933447098976109e-08, | |
| "loss": 0.2446, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.3199658703071672, | |
| "eval_loss": 1.069318175315857, | |
| "eval_runtime": 64.1107, | |
| "eval_samples_per_second": 15.395, | |
| "eval_steps_per_second": 1.934, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.37329351535836175, | |
| "grad_norm": 5.2035698890686035, | |
| "learning_rate": 8.755688282138793e-08, | |
| "loss": 0.273, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.37329351535836175, | |
| "eval_loss": 1.043074369430542, | |
| "eval_runtime": 63.9447, | |
| "eval_samples_per_second": 15.435, | |
| "eval_steps_per_second": 1.939, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.42662116040955633, | |
| "grad_norm": 3.6792104244232178, | |
| "learning_rate": 8.57792946530148e-08, | |
| "loss": 0.193, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.42662116040955633, | |
| "eval_loss": 1.0195527076721191, | |
| "eval_runtime": 63.8076, | |
| "eval_samples_per_second": 15.468, | |
| "eval_steps_per_second": 1.943, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.47994880546075086, | |
| "grad_norm": 111.2447280883789, | |
| "learning_rate": 8.400170648464163e-08, | |
| "loss": 0.2419, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.47994880546075086, | |
| "eval_loss": 0.9946303963661194, | |
| "eval_runtime": 63.8873, | |
| "eval_samples_per_second": 15.449, | |
| "eval_steps_per_second": 1.941, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.5332764505119454, | |
| "grad_norm": 0.024256089702248573, | |
| "learning_rate": 8.222411831626848e-08, | |
| "loss": 0.2577, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.5332764505119454, | |
| "eval_loss": 0.9679088592529297, | |
| "eval_runtime": 62.896, | |
| "eval_samples_per_second": 15.693, | |
| "eval_steps_per_second": 1.972, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.58660409556314, | |
| "grad_norm": 0.05403488501906395, | |
| "learning_rate": 8.044653014789534e-08, | |
| "loss": 0.2468, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.58660409556314, | |
| "eval_loss": 0.9470070004463196, | |
| "eval_runtime": 64.2367, | |
| "eval_samples_per_second": 15.365, | |
| "eval_steps_per_second": 1.93, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.6399317406143344, | |
| "grad_norm": 56.23408508300781, | |
| "learning_rate": 7.866894197952219e-08, | |
| "loss": 0.2148, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.6399317406143344, | |
| "eval_loss": 0.9275079369544983, | |
| "eval_runtime": 62.7425, | |
| "eval_samples_per_second": 15.731, | |
| "eval_steps_per_second": 1.976, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.693259385665529, | |
| "grad_norm": 335.7691955566406, | |
| "learning_rate": 7.689135381114903e-08, | |
| "loss": 0.179, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.693259385665529, | |
| "eval_loss": 0.9107733964920044, | |
| "eval_runtime": 63.711, | |
| "eval_samples_per_second": 15.492, | |
| "eval_steps_per_second": 1.946, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.7465870307167235, | |
| "grad_norm": 0.0303946640342474, | |
| "learning_rate": 7.511376564277589e-08, | |
| "loss": 0.1992, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.7465870307167235, | |
| "eval_loss": 0.8946455121040344, | |
| "eval_runtime": 62.7465, | |
| "eval_samples_per_second": 15.73, | |
| "eval_steps_per_second": 1.976, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.7999146757679181, | |
| "grad_norm": 24.27122688293457, | |
| "learning_rate": 7.333617747440273e-08, | |
| "loss": 0.2264, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.7999146757679181, | |
| "eval_loss": 0.8790715932846069, | |
| "eval_runtime": 62.625, | |
| "eval_samples_per_second": 15.76, | |
| "eval_steps_per_second": 1.98, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.8532423208191127, | |
| "grad_norm": 485.13018798828125, | |
| "learning_rate": 7.155858930602957e-08, | |
| "loss": 0.2349, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.8532423208191127, | |
| "eval_loss": 0.8648595213890076, | |
| "eval_runtime": 63.4718, | |
| "eval_samples_per_second": 15.55, | |
| "eval_steps_per_second": 1.954, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.9065699658703071, | |
| "grad_norm": 39.18803787231445, | |
| "learning_rate": 6.978100113765643e-08, | |
| "loss": 0.1869, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 0.9065699658703071, | |
| "eval_loss": 0.8521347641944885, | |
| "eval_runtime": 62.649, | |
| "eval_samples_per_second": 15.754, | |
| "eval_steps_per_second": 1.979, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 0.9598976109215017, | |
| "grad_norm": 15.91203784942627, | |
| "learning_rate": 6.800341296928327e-08, | |
| "loss": 0.1859, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.9598976109215017, | |
| "eval_loss": 0.8395125269889832, | |
| "eval_runtime": 63.1398, | |
| "eval_samples_per_second": 15.632, | |
| "eval_steps_per_second": 1.964, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.0132252559726962, | |
| "grad_norm": 545.9375610351562, | |
| "learning_rate": 6.622582480091012e-08, | |
| "loss": 0.1943, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 1.0132252559726962, | |
| "eval_loss": 0.8300319910049438, | |
| "eval_runtime": 63.2047, | |
| "eval_samples_per_second": 15.616, | |
| "eval_steps_per_second": 1.962, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 1.0665529010238908, | |
| "grad_norm": 451.04473876953125, | |
| "learning_rate": 6.444823663253697e-08, | |
| "loss": 0.1419, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.0665529010238908, | |
| "eval_loss": 0.8227387070655823, | |
| "eval_runtime": 62.7357, | |
| "eval_samples_per_second": 15.733, | |
| "eval_steps_per_second": 1.977, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.1198805460750854, | |
| "grad_norm": 12.075934410095215, | |
| "learning_rate": 6.267064846416382e-08, | |
| "loss": 0.1558, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 1.1198805460750854, | |
| "eval_loss": 0.8159177303314209, | |
| "eval_runtime": 62.8297, | |
| "eval_samples_per_second": 15.709, | |
| "eval_steps_per_second": 1.974, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 1.17320819112628, | |
| "grad_norm": 363.5056457519531, | |
| "learning_rate": 6.089306029579066e-08, | |
| "loss": 0.15, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.17320819112628, | |
| "eval_loss": 0.8096186518669128, | |
| "eval_runtime": 63.3529, | |
| "eval_samples_per_second": 15.579, | |
| "eval_steps_per_second": 1.957, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.2265358361774745, | |
| "grad_norm": 6.587007646885468e-06, | |
| "learning_rate": 5.911547212741752e-08, | |
| "loss": 0.1919, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 1.2265358361774745, | |
| "eval_loss": 0.8023524284362793, | |
| "eval_runtime": 62.9393, | |
| "eval_samples_per_second": 15.682, | |
| "eval_steps_per_second": 1.97, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 1.2798634812286689, | |
| "grad_norm": 0.11630824208259583, | |
| "learning_rate": 5.7337883959044365e-08, | |
| "loss": 0.1298, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.2798634812286689, | |
| "eval_loss": 0.7956322431564331, | |
| "eval_runtime": 63.0726, | |
| "eval_samples_per_second": 15.649, | |
| "eval_steps_per_second": 1.966, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.3331911262798635, | |
| "grad_norm": 0.0007865954539738595, | |
| "learning_rate": 5.556029579067122e-08, | |
| "loss": 0.1385, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.3331911262798635, | |
| "eval_loss": 0.7920646667480469, | |
| "eval_runtime": 63.3155, | |
| "eval_samples_per_second": 15.589, | |
| "eval_steps_per_second": 1.958, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.386518771331058, | |
| "grad_norm": 0.00023369144764728844, | |
| "learning_rate": 5.378270762229806e-08, | |
| "loss": 0.1414, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 1.386518771331058, | |
| "eval_loss": 0.7884665131568909, | |
| "eval_runtime": 63.002, | |
| "eval_samples_per_second": 15.666, | |
| "eval_steps_per_second": 1.968, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 1.4398464163822526, | |
| "grad_norm": 4.277670383453369, | |
| "learning_rate": 5.200511945392492e-08, | |
| "loss": 0.1549, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 1.4398464163822526, | |
| "eval_loss": 0.7809606194496155, | |
| "eval_runtime": 63.0606, | |
| "eval_samples_per_second": 15.652, | |
| "eval_steps_per_second": 1.966, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 1.493174061433447, | |
| "grad_norm": 0.04092055931687355, | |
| "learning_rate": 5.022753128555176e-08, | |
| "loss": 0.126, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.493174061433447, | |
| "eval_loss": 0.7771912217140198, | |
| "eval_runtime": 63.2952, | |
| "eval_samples_per_second": 15.594, | |
| "eval_steps_per_second": 1.959, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.5465017064846416, | |
| "grad_norm": 213.26644897460938, | |
| "learning_rate": 4.844994311717861e-08, | |
| "loss": 0.1518, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 1.5465017064846416, | |
| "eval_loss": 0.7739039659500122, | |
| "eval_runtime": 63.1919, | |
| "eval_samples_per_second": 15.619, | |
| "eval_steps_per_second": 1.962, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 1.5998293515358362, | |
| "grad_norm": 550.1605834960938, | |
| "learning_rate": 4.667235494880546e-08, | |
| "loss": 0.1143, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.5998293515358362, | |
| "eval_loss": 0.7702728509902954, | |
| "eval_runtime": 62.9504, | |
| "eval_samples_per_second": 15.679, | |
| "eval_steps_per_second": 1.97, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.6531569965870307, | |
| "grad_norm": 6.588364124298096, | |
| "learning_rate": 4.4894766780432306e-08, | |
| "loss": 0.161, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 1.6531569965870307, | |
| "eval_loss": 0.7638877034187317, | |
| "eval_runtime": 63.3812, | |
| "eval_samples_per_second": 15.572, | |
| "eval_steps_per_second": 1.956, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 1.7064846416382253, | |
| "grad_norm": 245.7973175048828, | |
| "learning_rate": 4.311717861205916e-08, | |
| "loss": 0.1277, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.7064846416382253, | |
| "eval_loss": 0.7590143084526062, | |
| "eval_runtime": 63.157, | |
| "eval_samples_per_second": 15.628, | |
| "eval_steps_per_second": 1.963, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.75981228668942, | |
| "grad_norm": 0.013761636801064014, | |
| "learning_rate": 4.133959044368601e-08, | |
| "loss": 0.1394, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 1.75981228668942, | |
| "eval_loss": 0.7563473582267761, | |
| "eval_runtime": 63.006, | |
| "eval_samples_per_second": 15.665, | |
| "eval_steps_per_second": 1.968, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 1.8131399317406145, | |
| "grad_norm": 0.0008418822544626892, | |
| "learning_rate": 3.9562002275312854e-08, | |
| "loss": 0.1253, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 1.8131399317406145, | |
| "eval_loss": 0.7549872398376465, | |
| "eval_runtime": 63.1042, | |
| "eval_samples_per_second": 15.641, | |
| "eval_steps_per_second": 1.965, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 1.8664675767918089, | |
| "grad_norm": 58.008296966552734, | |
| "learning_rate": 3.7784414106939705e-08, | |
| "loss": 0.2146, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 1.8664675767918089, | |
| "eval_loss": 0.7488264441490173, | |
| "eval_runtime": 63.1754, | |
| "eval_samples_per_second": 15.623, | |
| "eval_steps_per_second": 1.963, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 1.9197952218430034, | |
| "grad_norm": 0.0007694943924434483, | |
| "learning_rate": 3.600682593856656e-08, | |
| "loss": 0.1267, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.9197952218430034, | |
| "eval_loss": 0.7465928792953491, | |
| "eval_runtime": 62.8126, | |
| "eval_samples_per_second": 15.713, | |
| "eval_steps_per_second": 1.974, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.9731228668941978, | |
| "grad_norm": 515.19580078125, | |
| "learning_rate": 3.4229237770193395e-08, | |
| "loss": 0.1564, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 1.9731228668941978, | |
| "eval_loss": 0.7438480854034424, | |
| "eval_runtime": 63.2698, | |
| "eval_samples_per_second": 15.6, | |
| "eval_steps_per_second": 1.96, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 2.0264505119453924, | |
| "grad_norm": 27.036834716796875, | |
| "learning_rate": 3.2451649601820246e-08, | |
| "loss": 0.1355, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 2.0264505119453924, | |
| "eval_loss": 0.7395957112312317, | |
| "eval_runtime": 63.3696, | |
| "eval_samples_per_second": 15.575, | |
| "eval_steps_per_second": 1.957, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 2.079778156996587, | |
| "grad_norm": 6.471228122711182, | |
| "learning_rate": 3.06740614334471e-08, | |
| "loss": 0.1228, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 2.079778156996587, | |
| "eval_loss": 0.7349324226379395, | |
| "eval_runtime": 62.8971, | |
| "eval_samples_per_second": 15.692, | |
| "eval_steps_per_second": 1.971, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 2.1331058020477816, | |
| "grad_norm": 0.05623076856136322, | |
| "learning_rate": 2.8896473265073946e-08, | |
| "loss": 0.1082, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.1331058020477816, | |
| "eval_loss": 0.7336575388908386, | |
| "eval_runtime": 63.3128, | |
| "eval_samples_per_second": 15.589, | |
| "eval_steps_per_second": 1.959, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.186433447098976, | |
| "grad_norm": 0.013808301649987698, | |
| "learning_rate": 2.7118885096700794e-08, | |
| "loss": 0.1353, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 2.186433447098976, | |
| "eval_loss": 0.7329619526863098, | |
| "eval_runtime": 63.2838, | |
| "eval_samples_per_second": 15.596, | |
| "eval_steps_per_second": 1.959, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 2.2397610921501707, | |
| "grad_norm": 1.4853014945983887, | |
| "learning_rate": 2.5341296928327643e-08, | |
| "loss": 0.0925, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 2.2397610921501707, | |
| "eval_loss": 0.7303625345230103, | |
| "eval_runtime": 62.7838, | |
| "eval_samples_per_second": 15.721, | |
| "eval_steps_per_second": 1.975, | |
| "step": 21000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 28128, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 7555644893162520.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |