| { |
| "best_global_step": 5000, |
| "best_metric": 0.9768723268723268, |
| "best_model_checkpoint": "models/mmbert32k/feedback-detector_lora/checkpoint-5000", |
| "epoch": 4.468275245755138, |
| "eval_steps": 500, |
| "global_step": 5000, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.08936550491510277, |
| "grad_norm": 20.91827964782715, |
| "learning_rate": 5.303571428571429e-06, |
| "loss": 2.3772, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.17873100983020554, |
| "grad_norm": 13.255019187927246, |
| "learning_rate": 1.0660714285714287e-05, |
| "loss": 1.4536, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.2680965147453083, |
| "grad_norm": 6.325151443481445, |
| "learning_rate": 1.601785714285714e-05, |
| "loss": 0.7688, |
| "step": 300 |
| }, |
| { |
| "epoch": 0.3574620196604111, |
| "grad_norm": 14.112616539001465, |
| "learning_rate": 2.1375e-05, |
| "loss": 0.417, |
| "step": 400 |
| }, |
| { |
| "epoch": 0.44682752457551383, |
| "grad_norm": 7.2552032470703125, |
| "learning_rate": 2.673214285714286e-05, |
| "loss": 0.312, |
| "step": 500 |
| }, |
| { |
| "epoch": 0.44682752457551383, |
| "eval_accuracy": 0.9366834170854271, |
| "eval_f1_NEED_CLARIFICATION": 0.9595141700404858, |
| "eval_f1_SAT": 0.9993297587131368, |
| "eval_f1_WANT_DIFFERENT": 0.8411037107516651, |
| "eval_f1_WRONG_ANSWER": 0.8215417106652587, |
| "eval_f1_macro": 0.9053723375426366, |
| "eval_f1_weighted": 0.9366285248867929, |
| "eval_loss": 0.16554784774780273, |
| "eval_runtime": 6.6327, |
| "eval_samples_per_second": 450.046, |
| "eval_steps_per_second": 14.172, |
| "step": 500 |
| }, |
| { |
| "epoch": 0.5361930294906166, |
| "grad_norm": 12.230180740356445, |
| "learning_rate": 2.976762661370407e-05, |
| "loss": 0.2387, |
| "step": 600 |
| }, |
| { |
| "epoch": 0.6255585344057194, |
| "grad_norm": 0.957978367805481, |
| "learning_rate": 2.9171797418073483e-05, |
| "loss": 0.2194, |
| "step": 700 |
| }, |
| { |
| "epoch": 0.7149240393208222, |
| "grad_norm": 17.851573944091797, |
| "learning_rate": 2.85759682224429e-05, |
| "loss": 0.2051, |
| "step": 800 |
| }, |
| { |
| "epoch": 0.8042895442359249, |
| "grad_norm": 6.15651798248291, |
| "learning_rate": 2.7980139026812316e-05, |
| "loss": 0.1754, |
| "step": 900 |
| }, |
| { |
| "epoch": 0.8936550491510277, |
| "grad_norm": 16.395740509033203, |
| "learning_rate": 2.738430983118173e-05, |
| "loss": 0.1795, |
| "step": 1000 |
| }, |
| { |
| "epoch": 0.8936550491510277, |
| "eval_accuracy": 0.9571189279731993, |
| "eval_f1_NEED_CLARIFICATION": 0.9839034205231388, |
| "eval_f1_SAT": 1.0, |
| "eval_f1_WANT_DIFFERENT": 0.8676307007786429, |
| "eval_f1_WRONG_ANSWER": 0.8894977168949771, |
| "eval_f1_macro": 0.9352579595491897, |
| "eval_f1_weighted": 0.9567952614479014, |
| "eval_loss": 0.12704455852508545, |
| "eval_runtime": 4.9266, |
| "eval_samples_per_second": 605.9, |
| "eval_steps_per_second": 19.08, |
| "step": 1000 |
| }, |
| { |
| "epoch": 0.9830205540661304, |
| "grad_norm": 1.0810153484344482, |
| "learning_rate": 2.6788480635551143e-05, |
| "loss": 0.1464, |
| "step": 1100 |
| }, |
| { |
| "epoch": 1.0723860589812333, |
| "grad_norm": 1.037031888961792, |
| "learning_rate": 2.6192651439920558e-05, |
| "loss": 0.0986, |
| "step": 1200 |
| }, |
| { |
| "epoch": 1.161751563896336, |
| "grad_norm": 13.27597427368164, |
| "learning_rate": 2.5596822244289973e-05, |
| "loss": 0.1327, |
| "step": 1300 |
| }, |
| { |
| "epoch": 1.2511170688114388, |
| "grad_norm": 10.263288497924805, |
| "learning_rate": 2.5000993048659384e-05, |
| "loss": 0.1431, |
| "step": 1400 |
| }, |
| { |
| "epoch": 1.3404825737265416, |
| "grad_norm": 17.2136173248291, |
| "learning_rate": 2.44051638530288e-05, |
| "loss": 0.0957, |
| "step": 1500 |
| }, |
| { |
| "epoch": 1.3404825737265416, |
| "eval_accuracy": 0.9768844221105528, |
| "eval_f1_NEED_CLARIFICATION": 0.9940119760479041, |
| "eval_f1_SAT": 0.999664767013074, |
| "eval_f1_WANT_DIFFERENT": 0.9278996865203761, |
| "eval_f1_WRONG_ANSWER": 0.9396887159533074, |
| "eval_f1_macro": 0.9653162863836654, |
| "eval_f1_weighted": 0.9767427658694287, |
| "eval_loss": 0.07943707704544067, |
| "eval_runtime": 4.9017, |
| "eval_samples_per_second": 608.967, |
| "eval_steps_per_second": 19.177, |
| "step": 1500 |
| }, |
| { |
| "epoch": 1.4298480786416443, |
| "grad_norm": 2.8994057178497314, |
| "learning_rate": 2.3809334657398214e-05, |
| "loss": 0.0967, |
| "step": 1600 |
| }, |
| { |
| "epoch": 1.519213583556747, |
| "grad_norm": 1.162245750427246, |
| "learning_rate": 2.3213505461767625e-05, |
| "loss": 0.1492, |
| "step": 1700 |
| }, |
| { |
| "epoch": 1.6085790884718498, |
| "grad_norm": 13.979327201843262, |
| "learning_rate": 2.261767626613704e-05, |
| "loss": 0.1293, |
| "step": 1800 |
| }, |
| { |
| "epoch": 1.6979445933869526, |
| "grad_norm": 10.116354942321777, |
| "learning_rate": 2.2021847070506455e-05, |
| "loss": 0.0858, |
| "step": 1900 |
| }, |
| { |
| "epoch": 1.7873100983020556, |
| "grad_norm": 7.858895778656006, |
| "learning_rate": 2.142601787487587e-05, |
| "loss": 0.0978, |
| "step": 2000 |
| }, |
| { |
| "epoch": 1.7873100983020556, |
| "eval_accuracy": 0.9752093802345059, |
| "eval_f1_NEED_CLARIFICATION": 0.994994994994995, |
| "eval_f1_SAT": 1.0, |
| "eval_f1_WANT_DIFFERENT": 0.9229144667370645, |
| "eval_f1_WRONG_ANSWER": 0.9328214971209213, |
| "eval_f1_macro": 0.9626827397132453, |
| "eval_f1_weighted": 0.9750968232860249, |
| "eval_loss": 0.08431631326675415, |
| "eval_runtime": 4.8873, |
| "eval_samples_per_second": 610.767, |
| "eval_steps_per_second": 19.234, |
| "step": 2000 |
| }, |
| { |
| "epoch": 1.876675603217158, |
| "grad_norm": 0.5830298662185669, |
| "learning_rate": 2.083018867924528e-05, |
| "loss": 0.0922, |
| "step": 2100 |
| }, |
| { |
| "epoch": 1.966041108132261, |
| "grad_norm": 0.5619744062423706, |
| "learning_rate": 2.0234359483614696e-05, |
| "loss": 0.1096, |
| "step": 2200 |
| }, |
| { |
| "epoch": 2.0554066130473636, |
| "grad_norm": 0.08848729729652405, |
| "learning_rate": 1.9638530287984115e-05, |
| "loss": 0.0675, |
| "step": 2300 |
| }, |
| { |
| "epoch": 2.1447721179624666, |
| "grad_norm": 0.039530836045742035, |
| "learning_rate": 1.9042701092353526e-05, |
| "loss": 0.0792, |
| "step": 2400 |
| }, |
| { |
| "epoch": 2.234137622877569, |
| "grad_norm": 1.8711425065994263, |
| "learning_rate": 1.844687189672294e-05, |
| "loss": 0.0729, |
| "step": 2500 |
| }, |
| { |
| "epoch": 2.234137622877569, |
| "eval_accuracy": 0.9785594639865997, |
| "eval_f1_NEED_CLARIFICATION": 0.9949849548645938, |
| "eval_f1_SAT": 1.0, |
| "eval_f1_WANT_DIFFERENT": 0.9351464435146444, |
| "eval_f1_WRONG_ANSWER": 0.9410628019323671, |
| "eval_f1_macro": 0.9677985500779014, |
| "eval_f1_weighted": 0.9785107912077653, |
| "eval_loss": 0.08227432519197464, |
| "eval_runtime": 4.9766, |
| "eval_samples_per_second": 599.811, |
| "eval_steps_per_second": 18.889, |
| "step": 2500 |
| }, |
| { |
| "epoch": 2.323503127792672, |
| "grad_norm": 5.597414970397949, |
| "learning_rate": 1.7851042701092356e-05, |
| "loss": 0.0764, |
| "step": 2600 |
| }, |
| { |
| "epoch": 2.4128686327077746, |
| "grad_norm": 12.755281448364258, |
| "learning_rate": 1.7255213505461767e-05, |
| "loss": 0.0803, |
| "step": 2700 |
| }, |
| { |
| "epoch": 2.5022341376228776, |
| "grad_norm": 8.895140647888184, |
| "learning_rate": 1.6659384309831182e-05, |
| "loss": 0.0724, |
| "step": 2800 |
| }, |
| { |
| "epoch": 2.5915996425379806, |
| "grad_norm": 1.5834654569625854, |
| "learning_rate": 1.6063555114200597e-05, |
| "loss": 0.0575, |
| "step": 2900 |
| }, |
| { |
| "epoch": 2.680965147453083, |
| "grad_norm": 0.673122227191925, |
| "learning_rate": 1.5467725918570012e-05, |
| "loss": 0.0631, |
| "step": 3000 |
| }, |
| { |
| "epoch": 2.680965147453083, |
| "eval_accuracy": 0.9825795644891122, |
| "eval_f1_NEED_CLARIFICATION": 0.996996996996997, |
| "eval_f1_SAT": 1.0, |
| "eval_f1_WANT_DIFFERENT": 0.9477911646586346, |
| "eval_f1_WRONG_ANSWER": 0.9506545820745217, |
| "eval_f1_macro": 0.9738606859325383, |
| "eval_f1_weighted": 0.9825562768434225, |
| "eval_loss": 0.06965455412864685, |
| "eval_runtime": 4.8413, |
| "eval_samples_per_second": 616.566, |
| "eval_steps_per_second": 19.416, |
| "step": 3000 |
| }, |
| { |
| "epoch": 2.7703306523681857, |
| "grad_norm": 0.04906867444515228, |
| "learning_rate": 1.4871896722939424e-05, |
| "loss": 0.0582, |
| "step": 3100 |
| }, |
| { |
| "epoch": 2.8596961572832886, |
| "grad_norm": 2.8738303184509277, |
| "learning_rate": 1.4276067527308838e-05, |
| "loss": 0.0716, |
| "step": 3200 |
| }, |
| { |
| "epoch": 2.9490616621983916, |
| "grad_norm": 0.9897614121437073, |
| "learning_rate": 1.3680238331678252e-05, |
| "loss": 0.0701, |
| "step": 3300 |
| }, |
| { |
| "epoch": 3.038427167113494, |
| "grad_norm": 0.36938703060150146, |
| "learning_rate": 1.3084409136047666e-05, |
| "loss": 0.0467, |
| "step": 3400 |
| }, |
| { |
| "epoch": 3.127792672028597, |
| "grad_norm": 3.860152244567871, |
| "learning_rate": 1.2488579940417081e-05, |
| "loss": 0.0429, |
| "step": 3500 |
| }, |
| { |
| "epoch": 3.127792672028597, |
| "eval_accuracy": 0.9802345058626466, |
| "eval_f1_NEED_CLARIFICATION": 0.9939879759519038, |
| "eval_f1_SAT": 1.0, |
| "eval_f1_WANT_DIFFERENT": 0.940809968847352, |
| "eval_f1_WRONG_ANSWER": 0.9464459591041869, |
| "eval_f1_macro": 0.9703109759758607, |
| "eval_f1_weighted": 0.9801874251738407, |
| "eval_loss": 0.07702852040529251, |
| "eval_runtime": 4.8634, |
| "eval_samples_per_second": 613.763, |
| "eval_steps_per_second": 19.328, |
| "step": 3500 |
| }, |
| { |
| "epoch": 3.2171581769436997, |
| "grad_norm": 11.94687557220459, |
| "learning_rate": 1.1892750744786495e-05, |
| "loss": 0.0414, |
| "step": 3600 |
| }, |
| { |
| "epoch": 3.3065236818588026, |
| "grad_norm": 0.9737893342971802, |
| "learning_rate": 1.129692154915591e-05, |
| "loss": 0.0396, |
| "step": 3700 |
| }, |
| { |
| "epoch": 3.395889186773905, |
| "grad_norm": 2.069952964782715, |
| "learning_rate": 1.0701092353525323e-05, |
| "loss": 0.0575, |
| "step": 3800 |
| }, |
| { |
| "epoch": 3.485254691689008, |
| "grad_norm": 4.099809169769287, |
| "learning_rate": 1.0105263157894738e-05, |
| "loss": 0.0466, |
| "step": 3900 |
| }, |
| { |
| "epoch": 3.5746201966041107, |
| "grad_norm": 3.0518603324890137, |
| "learning_rate": 9.50943396226415e-06, |
| "loss": 0.0534, |
| "step": 4000 |
| }, |
| { |
| "epoch": 3.5746201966041107, |
| "eval_accuracy": 0.9839195979899498, |
| "eval_f1_NEED_CLARIFICATION": 0.996, |
| "eval_f1_SAT": 1.0, |
| "eval_f1_WANT_DIFFERENT": 0.9512195121951219, |
| "eval_f1_WRONG_ANSWER": 0.9561752988047809, |
| "eval_f1_macro": 0.9758487027499757, |
| "eval_f1_weighted": 0.9838829533929485, |
| "eval_loss": 0.06795033067464828, |
| "eval_runtime": 4.9963, |
| "eval_samples_per_second": 597.44, |
| "eval_steps_per_second": 18.814, |
| "step": 4000 |
| }, |
| { |
| "epoch": 3.6639857015192137, |
| "grad_norm": 0.4241458475589752, |
| "learning_rate": 8.913604766633564e-06, |
| "loss": 0.058, |
| "step": 4100 |
| }, |
| { |
| "epoch": 3.753351206434316, |
| "grad_norm": 0.6920946836471558, |
| "learning_rate": 8.31777557100298e-06, |
| "loss": 0.0426, |
| "step": 4200 |
| }, |
| { |
| "epoch": 3.842716711349419, |
| "grad_norm": 1.3809353113174438, |
| "learning_rate": 7.721946375372394e-06, |
| "loss": 0.0344, |
| "step": 4300 |
| }, |
| { |
| "epoch": 3.932082216264522, |
| "grad_norm": 8.121597290039062, |
| "learning_rate": 7.126117179741808e-06, |
| "loss": 0.0524, |
| "step": 4400 |
| }, |
| { |
| "epoch": 4.021447721179625, |
| "grad_norm": 14.424481391906738, |
| "learning_rate": 6.530287984111222e-06, |
| "loss": 0.0166, |
| "step": 4500 |
| }, |
| { |
| "epoch": 4.021447721179625, |
| "eval_accuracy": 0.9825795644891122, |
| "eval_f1_NEED_CLARIFICATION": 0.9959919839679359, |
| "eval_f1_SAT": 1.0, |
| "eval_f1_WANT_DIFFERENT": 0.947906026557712, |
| "eval_f1_WRONG_ANSWER": 0.9515331355093967, |
| "eval_f1_macro": 0.9738577865087612, |
| "eval_f1_weighted": 0.9825543419515753, |
| "eval_loss": 0.0710817500948906, |
| "eval_runtime": 4.9399, |
| "eval_samples_per_second": 604.261, |
| "eval_steps_per_second": 19.029, |
| "step": 4500 |
| }, |
| { |
| "epoch": 4.110813226094727, |
| "grad_norm": 1.4233607053756714, |
| "learning_rate": 5.934458788480636e-06, |
| "loss": 0.0212, |
| "step": 4600 |
| }, |
| { |
| "epoch": 4.200178731009831, |
| "grad_norm": 0.012283151037991047, |
| "learning_rate": 5.33862959285005e-06, |
| "loss": 0.0275, |
| "step": 4700 |
| }, |
| { |
| "epoch": 4.289544235924933, |
| "grad_norm": 0.15345770120620728, |
| "learning_rate": 4.742800397219464e-06, |
| "loss": 0.0365, |
| "step": 4800 |
| }, |
| { |
| "epoch": 4.378909740840036, |
| "grad_norm": 3.415753126144409, |
| "learning_rate": 4.146971201588878e-06, |
| "loss": 0.0224, |
| "step": 4900 |
| }, |
| { |
| "epoch": 4.468275245755138, |
| "grad_norm": 0.09987159073352814, |
| "learning_rate": 3.551142005958292e-06, |
| "loss": 0.0219, |
| "step": 5000 |
| }, |
| { |
| "epoch": 4.468275245755138, |
| "eval_accuracy": 0.9845896147403685, |
| "eval_f1_NEED_CLARIFICATION": 0.996996996996997, |
| "eval_f1_SAT": 1.0, |
| "eval_f1_WANT_DIFFERENT": 0.9535353535353536, |
| "eval_f1_WRONG_ANSWER": 0.9569569569569569, |
| "eval_f1_macro": 0.9768723268723268, |
| "eval_f1_weighted": 0.9845660553198242, |
| "eval_loss": 0.06945836544036865, |
| "eval_runtime": 4.8765, |
| "eval_samples_per_second": 612.114, |
| "eval_steps_per_second": 19.276, |
| "step": 5000 |
| } |
| ], |
| "logging_steps": 100, |
| "max_steps": 5595, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 5, |
| "save_steps": 500, |
| "stateful_callbacks": { |
| "EarlyStoppingCallback": { |
| "args": { |
| "early_stopping_patience": 3, |
| "early_stopping_threshold": 0.0 |
| }, |
| "attributes": { |
| "early_stopping_patience_counter": 0 |
| } |
| }, |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 2.766592691483443e+16, |
| "train_batch_size": 16, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
