| { | |
| "best_metric": 64.42910915934755, | |
| "best_model_checkpoint": "/scratch/camembertv2/runs/results/fquad/camembertav2-base-bf16-p2-17000/max_seq_length-896-doc_stride-128-max_answer_length-30-gradient_accumulation_steps-2-precision-fp32-learning_rate-3e-05-epochs-6-lr_scheduler-cosine-warmup_steps-0/SEED-1/checkpoint-6480", | |
| "epoch": 6.0, | |
| "eval_steps": 500, | |
| "global_step": 7776, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.07716049382716049, | |
| "grad_norm": 23.665807723999023, | |
| "learning_rate": 2.998775977415799e-05, | |
| "loss": 3.705, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.15432098765432098, | |
| "grad_norm": 20.01573944091797, | |
| "learning_rate": 2.9951059073049117e-05, | |
| "loss": 1.6565, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.23148148148148148, | |
| "grad_norm": 12.386185646057129, | |
| "learning_rate": 2.988995779332273e-05, | |
| "loss": 1.4648, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.30864197530864196, | |
| "grad_norm": 18.621143341064453, | |
| "learning_rate": 2.980455565410724e-05, | |
| "loss": 1.243, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.38580246913580246, | |
| "grad_norm": 15.412405967712402, | |
| "learning_rate": 2.96949920342655e-05, | |
| "loss": 1.1382, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.46296296296296297, | |
| "grad_norm": 18.231019973754883, | |
| "learning_rate": 2.95614457449243e-05, | |
| "loss": 1.1534, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.5401234567901234, | |
| "grad_norm": 16.341306686401367, | |
| "learning_rate": 2.940413473764923e-05, | |
| "loss": 1.0392, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.6172839506172839, | |
| "grad_norm": 12.782474517822266, | |
| "learning_rate": 2.9223315748741146e-05, | |
| "loss": 1.0394, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.6944444444444444, | |
| "grad_norm": 15.871848106384277, | |
| "learning_rate": 2.9019283880234828e-05, | |
| "loss": 0.9878, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.7716049382716049, | |
| "grad_norm": 8.874638557434082, | |
| "learning_rate": 2.879237211828353e-05, | |
| "loss": 1.0025, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.8487654320987654, | |
| "grad_norm": 19.52781867980957, | |
| "learning_rate": 2.8542950789715587e-05, | |
| "loss": 0.9396, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.9259259259259259, | |
| "grad_norm": 11.584712982177734, | |
| "learning_rate": 2.8271426957649866e-05, | |
| "loss": 0.9625, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_exact_match": 61.91969887076537, | |
| "eval_f1": 81.01142429829103, | |
| "eval_runtime": 46.1272, | |
| "eval_samples_per_second": 69.113, | |
| "eval_steps_per_second": 1.084, | |
| "step": 1296 | |
| }, | |
| { | |
| "epoch": 1.0030864197530864, | |
| "grad_norm": 7.031412601470947, | |
| "learning_rate": 2.7978243757156497e-05, | |
| "loss": 0.9411, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.0802469135802468, | |
| "grad_norm": 11.493968963623047, | |
| "learning_rate": 2.7663879672047095e-05, | |
| "loss": 0.6781, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.1574074074074074, | |
| "grad_norm": 5.261780738830566, | |
| "learning_rate": 2.732884775397477e-05, | |
| "loss": 0.6742, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.2345679012345678, | |
| "grad_norm": 13.310417175292969, | |
| "learning_rate": 2.6973694785118392e-05, | |
| "loss": 0.7004, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.3117283950617284, | |
| "grad_norm": 17.390798568725586, | |
| "learning_rate": 2.65990003858176e-05, | |
| "loss": 0.6512, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.3888888888888888, | |
| "grad_norm": 14.149763107299805, | |
| "learning_rate": 2.620537606861494e-05, | |
| "loss": 0.7074, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.4660493827160495, | |
| "grad_norm": 5.80026912689209, | |
| "learning_rate": 2.5793464240249014e-05, | |
| "loss": 0.6629, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.5432098765432098, | |
| "grad_norm": 17.484006881713867, | |
| "learning_rate": 2.536393715322732e-05, | |
| "loss": 0.6993, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.6203703703703702, | |
| "grad_norm": 21.80438804626465, | |
| "learning_rate": 2.49174958086899e-05, | |
| "loss": 0.6408, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 1.6975308641975309, | |
| "grad_norm": 19.416994094848633, | |
| "learning_rate": 2.4454868812354406e-05, | |
| "loss": 0.6592, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 1.7746913580246915, | |
| "grad_norm": 5.628990650177002, | |
| "learning_rate": 2.3976811185409607e-05, | |
| "loss": 0.622, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 1.8518518518518519, | |
| "grad_norm": 14.69774341583252, | |
| "learning_rate": 2.3484103132298082e-05, | |
| "loss": 0.645, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 1.9290123456790123, | |
| "grad_norm": 13.895576477050781, | |
| "learning_rate": 2.297754876739905e-05, | |
| "loss": 0.6746, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_exact_match": 63.26850690087829, | |
| "eval_f1": 82.69346269167056, | |
| "eval_runtime": 45.8773, | |
| "eval_samples_per_second": 69.49, | |
| "eval_steps_per_second": 1.09, | |
| "step": 2592 | |
| }, | |
| { | |
| "epoch": 2.006172839506173, | |
| "grad_norm": 7.386751651763916, | |
| "learning_rate": 2.2457974802689542e-05, | |
| "loss": 0.6472, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 2.0833333333333335, | |
| "grad_norm": 12.677722930908203, | |
| "learning_rate": 2.192622919852551e-05, | |
| "loss": 0.4365, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 2.1604938271604937, | |
| "grad_norm": 7.171934127807617, | |
| "learning_rate": 2.138317977974501e-05, | |
| "loss": 0.4287, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 2.2376543209876543, | |
| "grad_norm": 10.016780853271484, | |
| "learning_rate": 2.082971281935195e-05, | |
| "loss": 0.4462, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 2.314814814814815, | |
| "grad_norm": 21.301910400390625, | |
| "learning_rate": 2.0266731592091834e-05, | |
| "loss": 0.4425, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.3919753086419755, | |
| "grad_norm": 21.78326988220215, | |
| "learning_rate": 1.969515490028019e-05, | |
| "loss": 0.425, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 2.4691358024691357, | |
| "grad_norm": 17.772539138793945, | |
| "learning_rate": 1.9115915574289523e-05, | |
| "loss": 0.4181, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 2.5462962962962963, | |
| "grad_norm": 7.547439098358154, | |
| "learning_rate": 1.8529958950142064e-05, | |
| "loss": 0.4233, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 2.623456790123457, | |
| "grad_norm": 9.031538963317871, | |
| "learning_rate": 1.7938241326692906e-05, | |
| "loss": 0.4691, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 2.700617283950617, | |
| "grad_norm": 9.722735404968262, | |
| "learning_rate": 1.734172840492147e-05, | |
| "loss": 0.4498, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.7777777777777777, | |
| "grad_norm": 9.985281944274902, | |
| "learning_rate": 1.6741393711878455e-05, | |
| "loss": 0.4388, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 2.8549382716049383, | |
| "grad_norm": 9.514204978942871, | |
| "learning_rate": 1.6138217011860335e-05, | |
| "loss": 0.4501, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 2.932098765432099, | |
| "grad_norm": 16.88687515258789, | |
| "learning_rate": 1.5533182707404563e-05, | |
| "loss": 0.4172, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_exact_match": 63.676286072772896, | |
| "eval_f1": 82.60726439387956, | |
| "eval_runtime": 45.9653, | |
| "eval_samples_per_second": 69.357, | |
| "eval_steps_per_second": 1.088, | |
| "step": 3888 | |
| }, | |
| { | |
| "epoch": 3.009259259259259, | |
| "grad_norm": 4.132925033569336, | |
| "learning_rate": 1.4927278232714974e-05, | |
| "loss": 0.3689, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 3.0864197530864197, | |
| "grad_norm": 9.779620170593262, | |
| "learning_rate": 1.4321492442139406e-05, | |
| "loss": 0.2905, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 3.1635802469135803, | |
| "grad_norm": 7.350837230682373, | |
| "learning_rate": 1.371681399632967e-05, | |
| "loss": 0.2937, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 3.240740740740741, | |
| "grad_norm": 6.923620223999023, | |
| "learning_rate": 1.3114229748717562e-05, | |
| "loss": 0.2922, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 3.317901234567901, | |
| "grad_norm": 16.84642791748047, | |
| "learning_rate": 1.2514723134940363e-05, | |
| "loss": 0.28, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 3.3950617283950617, | |
| "grad_norm": 22.180021286010742, | |
| "learning_rate": 1.191927256784427e-05, | |
| "loss": 0.2907, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 3.4722222222222223, | |
| "grad_norm": 2.5661354064941406, | |
| "learning_rate": 1.1328849840685143e-05, | |
| "loss": 0.2806, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 3.549382716049383, | |
| "grad_norm": 11.584675788879395, | |
| "learning_rate": 1.0744418541132676e-05, | |
| "loss": 0.2963, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 3.626543209876543, | |
| "grad_norm": 5.476423740386963, | |
| "learning_rate": 1.0166932478666293e-05, | |
| "loss": 0.3199, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 3.7037037037037037, | |
| "grad_norm": 9.405366897583008, | |
| "learning_rate": 9.597334127929346e-06, | |
| "loss": 0.3107, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 3.7808641975308643, | |
| "grad_norm": 8.880900382995605, | |
| "learning_rate": 9.036553090582144e-06, | |
| "loss": 0.2991, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 3.8580246913580245, | |
| "grad_norm": 3.8892629146575928, | |
| "learning_rate": 8.485504578164017e-06, | |
| "loss": 0.2716, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 3.935185185185185, | |
| "grad_norm": 5.704967498779297, | |
| "learning_rate": 7.945087918440563e-06, | |
| "loss": 0.2688, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_exact_match": 64.2409033877039, | |
| "eval_f1": 83.135484930466, | |
| "eval_runtime": 45.9298, | |
| "eval_samples_per_second": 69.41, | |
| "eval_steps_per_second": 1.089, | |
| "step": 5184 | |
| }, | |
| { | |
| "epoch": 4.012345679012346, | |
| "grad_norm": 16.147579193115234, | |
| "learning_rate": 7.416185087673616e-06, | |
| "loss": 0.2919, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 4.089506172839506, | |
| "grad_norm": 13.380005836486816, | |
| "learning_rate": 6.899659271209459e-06, | |
| "loss": 0.2068, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 4.166666666666667, | |
| "grad_norm": 9.491084098815918, | |
| "learning_rate": 6.3963534547343126e-06, | |
| "loss": 0.2009, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 4.243827160493828, | |
| "grad_norm": 14.11040210723877, | |
| "learning_rate": 5.907089048496351e-06, | |
| "loss": 0.2124, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 4.320987654320987, | |
| "grad_norm": 12.674304962158203, | |
| "learning_rate": 5.4326645467394085e-06, | |
| "loss": 0.2173, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 4.398148148148148, | |
| "grad_norm": 5.682621955871582, | |
| "learning_rate": 4.973854224536363e-06, | |
| "loss": 0.213, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 4.4753086419753085, | |
| "grad_norm": 5.133475303649902, | |
| "learning_rate": 4.5314068741488615e-06, | |
| "loss": 0.2, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 4.552469135802469, | |
| "grad_norm": 6.370384693145752, | |
| "learning_rate": 4.1060445829758305e-06, | |
| "loss": 0.197, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 4.62962962962963, | |
| "grad_norm": 16.37765884399414, | |
| "learning_rate": 3.6984615550850894e-06, | |
| "loss": 0.2051, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 4.70679012345679, | |
| "grad_norm": 11.54761791229248, | |
| "learning_rate": 3.3093229782514023e-06, | |
| "loss": 0.1733, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 4.783950617283951, | |
| "grad_norm": 22.175281524658203, | |
| "learning_rate": 2.939263938350012e-06, | |
| "loss": 0.2003, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 4.861111111111111, | |
| "grad_norm": 1.2753137350082397, | |
| "learning_rate": 2.588888382877342e-06, | |
| "loss": 0.194, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 4.938271604938271, | |
| "grad_norm": 32.319236755371094, | |
| "learning_rate": 2.2587681352905404e-06, | |
| "loss": 0.2149, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_exact_match": 64.42910915934755, | |
| "eval_f1": 83.36016013340664, | |
| "eval_runtime": 45.8927, | |
| "eval_samples_per_second": 69.466, | |
| "eval_steps_per_second": 1.089, | |
| "step": 6480 | |
| }, | |
| { | |
| "epoch": 5.015432098765432, | |
| "grad_norm": 9.053484916687012, | |
| "learning_rate": 1.9494419617743312e-06, | |
| "loss": 0.198, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 5.092592592592593, | |
| "grad_norm": 11.60733699798584, | |
| "learning_rate": 1.6614146919584094e-06, | |
| "loss": 0.1512, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 5.169753086419753, | |
| "grad_norm": 9.327279090881348, | |
| "learning_rate": 1.3951563950202656e-06, | |
| "loss": 0.167, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 5.246913580246914, | |
| "grad_norm": 4.551391124725342, | |
| "learning_rate": 1.1511016125181445e-06, | |
| "loss": 0.1315, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 5.324074074074074, | |
| "grad_norm": 4.2411274909973145, | |
| "learning_rate": 9.296486492061334e-07, | |
| "loss": 0.1532, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 5.401234567901234, | |
| "grad_norm": 5.1193461418151855, | |
| "learning_rate": 7.311589229888083e-07, | |
| "loss": 0.1624, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 5.478395061728395, | |
| "grad_norm": 1.735378384590149, | |
| "learning_rate": 5.55956375076332e-07, | |
| "loss": 0.1688, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 5.555555555555555, | |
| "grad_norm": 3.7359869480133057, | |
| "learning_rate": 4.043269413026429e-07, | |
| "loss": 0.148, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 5.632716049382716, | |
| "grad_norm": 6.915912628173828, | |
| "learning_rate": 2.7651808546956646e-07, | |
| "loss": 0.1822, | |
| "step": 7300 | |
| }, | |
| { | |
| "epoch": 5.709876543209877, | |
| "grad_norm": 5.2128448486328125, | |
| "learning_rate": 1.727383954784373e-07, | |
| "loss": 0.163, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 5.787037037037037, | |
| "grad_norm": 3.4656639099121094, | |
| "learning_rate": 9.315724290836047e-08, | |
| "loss": 0.1716, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 5.864197530864198, | |
| "grad_norm": 21.396251678466797, | |
| "learning_rate": 3.790450659670097e-08, | |
| "loss": 0.1694, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 5.9413580246913575, | |
| "grad_norm": 14.200843811035156, | |
| "learning_rate": 7.070360672907228e-09, | |
| "loss": 0.1618, | |
| "step": 7700 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_exact_match": 64.0840652446675, | |
| "eval_f1": 83.12314115625247, | |
| "eval_runtime": 45.7565, | |
| "eval_samples_per_second": 69.673, | |
| "eval_steps_per_second": 1.093, | |
| "step": 7776 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "step": 7776, | |
| "total_flos": 2.0394634246921464e+16, | |
| "train_loss": 0.5145930189164087, | |
| "train_runtime": 3736.1381, | |
| "train_samples_per_second": 33.293, | |
| "train_steps_per_second": 2.081 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 7776, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 6, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.0394634246921464e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |