| { | |
| "best_metric": 0.34317871928215027, | |
| "best_model_checkpoint": "../../saves/LLaMA3-70B-qlora-bnb/lora/sft/A61K/checkpoint-500", | |
| "epoch": 2.9925925925925925, | |
| "eval_steps": 100, | |
| "global_step": 606, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.04938271604938271, | |
| "grad_norm": 15.14680004119873, | |
| "learning_rate": 9e-07, | |
| "loss": 12.2914, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.09876543209876543, | |
| "grad_norm": 16.71466636657715, | |
| "learning_rate": 3.6e-06, | |
| "loss": 12.285, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.14814814814814814, | |
| "grad_norm": NaN, | |
| "learning_rate": 6.3e-06, | |
| "loss": 12.0986, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.19753086419753085, | |
| "grad_norm": 15.837638854980469, | |
| "learning_rate": 9.3e-06, | |
| "loss": 11.846, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.24691358024691357, | |
| "grad_norm": 19.359682083129883, | |
| "learning_rate": 1.2299999999999999e-05, | |
| "loss": 10.9695, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.2962962962962963, | |
| "grad_norm": 33.46400451660156, | |
| "learning_rate": 1.53e-05, | |
| "loss": 9.6625, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.345679012345679, | |
| "grad_norm": 18.000560760498047, | |
| "learning_rate": 1.83e-05, | |
| "loss": 7.6763, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.3950617283950617, | |
| "grad_norm": 23.49736976623535, | |
| "learning_rate": 2.13e-05, | |
| "loss": 5.599, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.4444444444444444, | |
| "grad_norm": 24.056060791015625, | |
| "learning_rate": 2.43e-05, | |
| "loss": 2.2361, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.49382716049382713, | |
| "grad_norm": 9.21530532836914, | |
| "learning_rate": 2.7300000000000003e-05, | |
| "loss": 0.5658, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.49382716049382713, | |
| "eval_loss": 0.43361854553222656, | |
| "eval_runtime": 640.3211, | |
| "eval_samples_per_second": 0.281, | |
| "eval_steps_per_second": 0.281, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.5432098765432098, | |
| "grad_norm": 11.147759437561035, | |
| "learning_rate": 2.9999710893018065e-05, | |
| "loss": 0.3497, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.5925925925925926, | |
| "grad_norm": 8.447771072387695, | |
| "learning_rate": 2.9965031537858088e-05, | |
| "loss": 0.3664, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.6419753086419753, | |
| "grad_norm": 5.72174596786499, | |
| "learning_rate": 2.9872683924428126e-05, | |
| "loss": 0.3909, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.691358024691358, | |
| "grad_norm": 8.332149505615234, | |
| "learning_rate": 2.9723023917398982e-05, | |
| "loss": 0.391, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.7407407407407407, | |
| "grad_norm": 6.015285491943359, | |
| "learning_rate": 2.9516628236700468e-05, | |
| "loss": 0.3285, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.7901234567901234, | |
| "grad_norm": 8.486276626586914, | |
| "learning_rate": 2.9254292235111508e-05, | |
| "loss": 0.3513, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.8395061728395061, | |
| "grad_norm": 19.51106834411621, | |
| "learning_rate": 2.8937026833341146e-05, | |
| "loss": 0.382, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.8888888888888888, | |
| "grad_norm": 9.629849433898926, | |
| "learning_rate": 2.85660546244112e-05, | |
| "loss": 0.3862, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.9382716049382716, | |
| "grad_norm": 6.940722942352295, | |
| "learning_rate": 2.814280516235244e-05, | |
| "loss": 0.4923, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.9876543209876543, | |
| "grad_norm": 7.83641242980957, | |
| "learning_rate": 2.7668909453369486e-05, | |
| "loss": 0.361, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.9876543209876543, | |
| "eval_loss": 0.3438129127025604, | |
| "eval_runtime": 641.0521, | |
| "eval_samples_per_second": 0.281, | |
| "eval_steps_per_second": 0.281, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.037037037037037, | |
| "grad_norm": 8.031428337097168, | |
| "learning_rate": 2.7146193670702906e-05, | |
| "loss": 0.3862, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.0864197530864197, | |
| "grad_norm": 4.103025436401367, | |
| "learning_rate": 2.6576672117408486e-05, | |
| "loss": 0.3276, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.1358024691358024, | |
| "grad_norm": 6.102336883544922, | |
| "learning_rate": 2.5962539464171864e-05, | |
| "loss": 0.3845, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.1851851851851851, | |
| "grad_norm": 11.080580711364746, | |
| "learning_rate": 2.530616229207034e-05, | |
| "loss": 0.3431, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.2345679012345678, | |
| "grad_norm": 7.723226547241211, | |
| "learning_rate": 2.4610069972872206e-05, | |
| "loss": 0.3262, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.2839506172839505, | |
| "grad_norm": 5.732240200042725, | |
| "learning_rate": 2.3876944922016517e-05, | |
| "loss": 0.3465, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 1.3333333333333333, | |
| "grad_norm": 6.469529151916504, | |
| "learning_rate": 2.3109612261833967e-05, | |
| "loss": 0.3217, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 1.382716049382716, | |
| "grad_norm": 3.8795225620269775, | |
| "learning_rate": 2.2311028934841804e-05, | |
| "loss": 0.3276, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 1.4320987654320987, | |
| "grad_norm": 4.8721466064453125, | |
| "learning_rate": 2.1484272309065257e-05, | |
| "loss": 0.3561, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 1.4814814814814814, | |
| "grad_norm": 6.701916217803955, | |
| "learning_rate": 2.0632528319295113e-05, | |
| "loss": 0.3343, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.4814814814814814, | |
| "eval_loss": 0.3502769470214844, | |
| "eval_runtime": 641.2195, | |
| "eval_samples_per_second": 0.281, | |
| "eval_steps_per_second": 0.281, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.5308641975308643, | |
| "grad_norm": 4.695794582366943, | |
| "learning_rate": 1.9759079189979598e-05, | |
| "loss": 0.3471, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.5802469135802468, | |
| "grad_norm": 4.3271870613098145, | |
| "learning_rate": 1.8867290787060762e-05, | |
| "loss": 0.3365, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.6296296296296298, | |
| "grad_norm": 4.225152969360352, | |
| "learning_rate": 1.7960599647495552e-05, | |
| "loss": 0.3332, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.6790123456790123, | |
| "grad_norm": 5.201157093048096, | |
| "learning_rate": 1.713472257409928e-05, | |
| "loss": 0.3339, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.7283950617283952, | |
| "grad_norm": 5.604823589324951, | |
| "learning_rate": 1.6209378726328168e-05, | |
| "loss": 0.3225, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.7777777777777777, | |
| "grad_norm": 4.6840715408325195, | |
| "learning_rate": 1.5279374496481708e-05, | |
| "loss": 0.3359, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.8271604938271606, | |
| "grad_norm": 7.939080715179443, | |
| "learning_rate": 1.434829368750633e-05, | |
| "loss": 0.3253, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.876543209876543, | |
| "grad_norm": 9.724854469299316, | |
| "learning_rate": 1.3419724250982795e-05, | |
| "loss": 0.3561, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.925925925925926, | |
| "grad_norm": 5.289661407470703, | |
| "learning_rate": 1.2497244460832644e-05, | |
| "loss": 0.3276, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.9753086419753085, | |
| "grad_norm": 5.043463230133057, | |
| "learning_rate": 1.1584409124317906e-05, | |
| "loss": 0.3275, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.9753086419753085, | |
| "eval_loss": 0.3461529314517975, | |
| "eval_runtime": 641.2613, | |
| "eval_samples_per_second": 0.281, | |
| "eval_steps_per_second": 0.281, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.0246913580246915, | |
| "grad_norm": 7.275540828704834, | |
| "learning_rate": 1.0684735883470333e-05, | |
| "loss": 0.3025, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 2.074074074074074, | |
| "grad_norm": 9.59693717956543, | |
| "learning_rate": 9.80169165973814e-06, | |
| "loss": 0.2947, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 2.123456790123457, | |
| "grad_norm": 11.817838668823242, | |
| "learning_rate": 8.938679294086226e-06, | |
| "loss": 0.2596, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 2.1728395061728394, | |
| "grad_norm": 5.468794822692871, | |
| "learning_rate": 8.099024434032719e-06, | |
| "loss": 0.3097, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 2.2222222222222223, | |
| "grad_norm": 8.117734909057617, | |
| "learning_rate": 7.285962718153099e-06, | |
| "loss": 0.306, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 2.271604938271605, | |
| "grad_norm": 5.816110134124756, | |
| "learning_rate": 6.5026273074368575e-06, | |
| "loss": 0.313, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 2.3209876543209877, | |
| "grad_norm": 7.187597751617432, | |
| "learning_rate": 5.752036811544973e-06, | |
| "loss": 0.3078, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 2.3703703703703702, | |
| "grad_norm": 17.781335830688477, | |
| "learning_rate": 5.03708365649491e-06, | |
| "loss": 0.3034, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 2.419753086419753, | |
| "grad_norm": 6.449460506439209, | |
| "learning_rate": 4.3605229385984915e-06, | |
| "loss": 0.3052, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 2.4691358024691357, | |
| "grad_norm": 8.452071189880371, | |
| "learning_rate": 3.7249618076045316e-06, | |
| "loss": 0.2916, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.4691358024691357, | |
| "eval_loss": 0.34317871928215027, | |
| "eval_runtime": 641.0511, | |
| "eval_samples_per_second": 0.281, | |
| "eval_steps_per_second": 0.281, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.5185185185185186, | |
| "grad_norm": 7.553008079528809, | |
| "learning_rate": 3.1328494199585704e-06, | |
| "loss": 0.2849, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 2.567901234567901, | |
| "grad_norm": 7.099052906036377, | |
| "learning_rate": 2.5864675008951393e-06, | |
| "loss": 0.2827, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 2.617283950617284, | |
| "grad_norm": 11.348811149597168, | |
| "learning_rate": 2.0879215517319634e-06, | |
| "loss": 0.3025, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 2.6666666666666665, | |
| "grad_norm": 11.900992393493652, | |
| "learning_rate": 1.639132736248945e-06, | |
| "loss": 0.2994, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 2.7160493827160495, | |
| "grad_norm": 8.492768287658691, | |
| "learning_rate": 1.2418304774182077e-06, | |
| "loss": 0.3024, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 2.765432098765432, | |
| "grad_norm": 4.38970422744751, | |
| "learning_rate": 8.975457930137787e-07, | |
| "loss": 0.3267, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 2.814814814814815, | |
| "grad_norm": 5.739566326141357, | |
| "learning_rate": 6.076053957825411e-07, | |
| "loss": 0.2997, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 2.8641975308641974, | |
| "grad_norm": 7.524597644805908, | |
| "learning_rate": 3.7312658091147976e-07, | |
| "loss": 0.2949, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 2.9135802469135803, | |
| "grad_norm": 5.539340019226074, | |
| "learning_rate": 1.9501292049268593e-07, | |
| "loss": 0.282, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 2.962962962962963, | |
| "grad_norm": 7.615926742553711, | |
| "learning_rate": 7.395078157753921e-08, | |
| "loss": 0.291, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.962962962962963, | |
| "eval_loss": 0.34410154819488525, | |
| "eval_runtime": 641.306, | |
| "eval_samples_per_second": 0.281, | |
| "eval_steps_per_second": 0.281, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.9925925925925925, | |
| "step": 606, | |
| "total_flos": 1.4871334589840622e+19, | |
| "train_loss": 1.6824197243935992, | |
| "train_runtime": 45546.3315, | |
| "train_samples_per_second": 0.107, | |
| "train_steps_per_second": 0.013 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 606, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 100, | |
| "total_flos": 1.4871334589840622e+19, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |