| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 2.0, |
| "eval_steps": 500, |
| "global_step": 33218, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 6.020832078993317e-05, |
| "grad_norm": 0.38543766736984253, |
| "learning_rate": 0.0, |
| "loss": 0.0947, |
| "step": 1 |
| }, |
| { |
| "epoch": 0.030104160394966585, |
| "grad_norm": 0.10387147963047028, |
| "learning_rate": 9.013728323699423e-07, |
| "loss": 0.0776, |
| "step": 500 |
| }, |
| { |
| "epoch": 0.030104160394966585, |
| "eval_cosine_accuracy": 0.9121845364570618, |
| "eval_loss": 0.03408537805080414, |
| "eval_runtime": 49.9493, |
| "eval_samples_per_second": 378.704, |
| "eval_steps_per_second": 1.482, |
| "step": 500 |
| }, |
| { |
| "epoch": 0.06020832078993317, |
| "grad_norm": 0.0778941884636879, |
| "learning_rate": 1.8045520231213872e-06, |
| "loss": 0.0395, |
| "step": 1000 |
| }, |
| { |
| "epoch": 0.06020832078993317, |
| "eval_cosine_accuracy": 0.9305813312530518, |
| "eval_loss": 0.03211512416601181, |
| "eval_runtime": 50.2255, |
| "eval_samples_per_second": 376.622, |
| "eval_steps_per_second": 1.473, |
| "step": 1000 |
| }, |
| { |
| "epoch": 0.09031248118489975, |
| "grad_norm": 0.07489586621522903, |
| "learning_rate": 2.7077312138728323e-06, |
| "loss": 0.035, |
| "step": 1500 |
| }, |
| { |
| "epoch": 0.09031248118489975, |
| "eval_cosine_accuracy": 0.9341157078742981, |
| "eval_loss": 0.03070036694407463, |
| "eval_runtime": 49.7769, |
| "eval_samples_per_second": 380.016, |
| "eval_steps_per_second": 1.487, |
| "step": 1500 |
| }, |
| { |
| "epoch": 0.12041664157986634, |
| "grad_norm": 0.11629298329353333, |
| "learning_rate": 3.6109104046242777e-06, |
| "loss": 0.0314, |
| "step": 2000 |
| }, |
| { |
| "epoch": 0.12041664157986634, |
| "eval_cosine_accuracy": 0.9239204525947571, |
| "eval_loss": 0.032454561442136765, |
| "eval_runtime": 50.1109, |
| "eval_samples_per_second": 377.483, |
| "eval_steps_per_second": 1.477, |
| "step": 2000 |
| }, |
| { |
| "epoch": 0.15052080197483292, |
| "grad_norm": 0.08288636803627014, |
| "learning_rate": 4.514089595375723e-06, |
| "loss": 0.0283, |
| "step": 2500 |
| }, |
| { |
| "epoch": 0.15052080197483292, |
| "eval_cosine_accuracy": 0.9068829417228699, |
| "eval_loss": 0.0350261889398098, |
| "eval_runtime": 49.5334, |
| "eval_samples_per_second": 381.884, |
| "eval_steps_per_second": 1.494, |
| "step": 2500 |
| }, |
| { |
| "epoch": 0.1806249623697995, |
| "grad_norm": 0.09410930424928665, |
| "learning_rate": 5.4172687861271674e-06, |
| "loss": 0.0258, |
| "step": 3000 |
| }, |
| { |
| "epoch": 0.1806249623697995, |
| "eval_cosine_accuracy": 0.8916126489639282, |
| "eval_loss": 0.037413690239191055, |
| "eval_runtime": 50.1483, |
| "eval_samples_per_second": 377.201, |
| "eval_steps_per_second": 1.476, |
| "step": 3000 |
| }, |
| { |
| "epoch": 0.2107291227647661, |
| "grad_norm": 0.0764598697423935, |
| "learning_rate": 6.320447976878613e-06, |
| "loss": 0.0236, |
| "step": 3500 |
| }, |
| { |
| "epoch": 0.2107291227647661, |
| "eval_cosine_accuracy": 0.867868959903717, |
| "eval_loss": 0.04107580706477165, |
| "eval_runtime": 49.5476, |
| "eval_samples_per_second": 381.774, |
| "eval_steps_per_second": 1.494, |
| "step": 3500 |
| }, |
| { |
| "epoch": 0.24083328315973268, |
| "grad_norm": 0.09663169085979462, |
| "learning_rate": 7.2236271676300576e-06, |
| "loss": 0.0219, |
| "step": 4000 |
| }, |
| { |
| "epoch": 0.24083328315973268, |
| "eval_cosine_accuracy": 0.8625220656394958, |
| "eval_loss": 0.042248476296663284, |
| "eval_runtime": 50.2346, |
| "eval_samples_per_second": 376.553, |
| "eval_steps_per_second": 1.473, |
| "step": 4000 |
| }, |
| { |
| "epoch": 0.2709374435546993, |
| "grad_norm": 0.10593143850564957, |
| "learning_rate": 8.126806358381503e-06, |
| "loss": 0.0205, |
| "step": 4500 |
| }, |
| { |
| "epoch": 0.2709374435546993, |
| "eval_cosine_accuracy": 0.8380987048149109, |
| "eval_loss": 0.04543215036392212, |
| "eval_runtime": 49.5535, |
| "eval_samples_per_second": 381.729, |
| "eval_steps_per_second": 1.493, |
| "step": 4500 |
| }, |
| { |
| "epoch": 0.30104160394966584, |
| "grad_norm": 0.07745221257209778, |
| "learning_rate": 9.02998554913295e-06, |
| "loss": 0.019, |
| "step": 5000 |
| }, |
| { |
| "epoch": 0.30104160394966584, |
| "eval_cosine_accuracy": 0.8265440464019775, |
| "eval_loss": 0.04734260216355324, |
| "eval_runtime": 49.6392, |
| "eval_samples_per_second": 381.07, |
| "eval_steps_per_second": 1.491, |
| "step": 5000 |
| }, |
| { |
| "epoch": 0.33114576434463244, |
| "grad_norm": 0.09874160587787628, |
| "learning_rate": 9.933164739884392e-06, |
| "loss": 0.0179, |
| "step": 5500 |
| }, |
| { |
| "epoch": 0.33114576434463244, |
| "eval_cosine_accuracy": 0.812904953956604, |
| "eval_loss": 0.048521969467401505, |
| "eval_runtime": 50.9191, |
| "eval_samples_per_second": 371.491, |
| "eval_steps_per_second": 1.453, |
| "step": 5500 |
| }, |
| { |
| "epoch": 0.361249924739599, |
| "grad_norm": 0.08646717667579651, |
| "learning_rate": 1.0836343930635839e-05, |
| "loss": 0.017, |
| "step": 6000 |
| }, |
| { |
| "epoch": 0.361249924739599, |
| "eval_cosine_accuracy": 0.8010784387588501, |
| "eval_loss": 0.051017291843891144, |
| "eval_runtime": 49.7081, |
| "eval_samples_per_second": 380.541, |
| "eval_steps_per_second": 1.489, |
| "step": 6000 |
| }, |
| { |
| "epoch": 0.3913540851345656, |
| "grad_norm": 0.10478880256414413, |
| "learning_rate": 1.1739523121387283e-05, |
| "loss": 0.0161, |
| "step": 6500 |
| }, |
| { |
| "epoch": 0.3913540851345656, |
| "eval_cosine_accuracy": 0.7926956415176392, |
| "eval_loss": 0.05125237628817558, |
| "eval_runtime": 50.4664, |
| "eval_samples_per_second": 374.824, |
| "eval_steps_per_second": 1.466, |
| "step": 6500 |
| }, |
| { |
| "epoch": 0.4214582455295322, |
| "grad_norm": 0.09687662869691849, |
| "learning_rate": 1.2642702312138728e-05, |
| "loss": 0.0153, |
| "step": 7000 |
| }, |
| { |
| "epoch": 0.4214582455295322, |
| "eval_cosine_accuracy": 0.7693597078323364, |
| "eval_loss": 0.053376998752355576, |
| "eval_runtime": 50.2601, |
| "eval_samples_per_second": 376.362, |
| "eval_steps_per_second": 1.472, |
| "step": 7000 |
| }, |
| { |
| "epoch": 0.45156240592449876, |
| "grad_norm": 0.12575113773345947, |
| "learning_rate": 1.3545881502890174e-05, |
| "loss": 0.0145, |
| "step": 7500 |
| }, |
| { |
| "epoch": 0.45156240592449876, |
| "eval_cosine_accuracy": 0.7683628797531128, |
| "eval_loss": 0.05461883544921875, |
| "eval_runtime": 49.9602, |
| "eval_samples_per_second": 378.621, |
| "eval_steps_per_second": 1.481, |
| "step": 7500 |
| }, |
| { |
| "epoch": 0.48166656631946536, |
| "grad_norm": 0.09359786659479141, |
| "learning_rate": 1.4449060693641619e-05, |
| "loss": 0.0139, |
| "step": 8000 |
| }, |
| { |
| "epoch": 0.48166656631946536, |
| "eval_cosine_accuracy": 0.7625628709793091, |
| "eval_loss": 0.05498296022415161, |
| "eval_runtime": 49.4526, |
| "eval_samples_per_second": 382.508, |
| "eval_steps_per_second": 1.496, |
| "step": 8000 |
| }, |
| { |
| "epoch": 0.5117707267144319, |
| "grad_norm": 0.08199989050626755, |
| "learning_rate": 1.5352239884393065e-05, |
| "loss": 0.0132, |
| "step": 8500 |
| }, |
| { |
| "epoch": 0.5117707267144319, |
| "eval_cosine_accuracy": 0.7560831904411316, |
| "eval_loss": 0.055622003972530365, |
| "eval_runtime": 51.0115, |
| "eval_samples_per_second": 370.818, |
| "eval_steps_per_second": 1.451, |
| "step": 8500 |
| }, |
| { |
| "epoch": 0.5418748871093986, |
| "grad_norm": 0.08838484436273575, |
| "learning_rate": 1.625541907514451e-05, |
| "loss": 0.0128, |
| "step": 9000 |
| }, |
| { |
| "epoch": 0.5418748871093986, |
| "eval_cosine_accuracy": 0.7378675937652588, |
| "eval_loss": 0.056548163294792175, |
| "eval_runtime": 50.9508, |
| "eval_samples_per_second": 371.26, |
| "eval_steps_per_second": 1.452, |
| "step": 9000 |
| }, |
| { |
| "epoch": 0.5719790475043651, |
| "grad_norm": 0.09884914755821228, |
| "learning_rate": 1.7158598265895954e-05, |
| "loss": 0.0123, |
| "step": 9500 |
| }, |
| { |
| "epoch": 0.5719790475043651, |
| "eval_cosine_accuracy": 0.7324753999710083, |
| "eval_loss": 0.05744454637169838, |
| "eval_runtime": 49.4469, |
| "eval_samples_per_second": 382.552, |
| "eval_steps_per_second": 1.497, |
| "step": 9500 |
| }, |
| { |
| "epoch": 0.6020832078993317, |
| "grad_norm": 0.0804782584309578, |
| "learning_rate": 1.80617774566474e-05, |
| "loss": 0.0119, |
| "step": 10000 |
| }, |
| { |
| "epoch": 0.6020832078993317, |
| "eval_cosine_accuracy": 0.7181566953659058, |
| "eval_loss": 0.058585748076438904, |
| "eval_runtime": 50.1288, |
| "eval_samples_per_second": 377.348, |
| "eval_steps_per_second": 1.476, |
| "step": 10000 |
| }, |
| { |
| "epoch": 0.6321873682942982, |
| "grad_norm": 0.09254036843776703, |
| "learning_rate": 1.8964956647398844e-05, |
| "loss": 0.0114, |
| "step": 10500 |
| }, |
| { |
| "epoch": 0.6321873682942982, |
| "eval_cosine_accuracy": 0.725497305393219, |
| "eval_loss": 0.05707884952425957, |
| "eval_runtime": 49.3636, |
| "eval_samples_per_second": 383.197, |
| "eval_steps_per_second": 1.499, |
| "step": 10500 |
| }, |
| { |
| "epoch": 0.6622915286892649, |
| "grad_norm": 0.13671717047691345, |
| "learning_rate": 1.986813583815029e-05, |
| "loss": 0.0111, |
| "step": 11000 |
| }, |
| { |
| "epoch": 0.6622915286892649, |
| "eval_cosine_accuracy": 0.7044723629951477, |
| "eval_loss": 0.0592338964343071, |
| "eval_runtime": 51.6332, |
| "eval_samples_per_second": 366.353, |
| "eval_steps_per_second": 1.433, |
| "step": 11000 |
| }, |
| { |
| "epoch": 0.6923956890842314, |
| "grad_norm": 0.09350993484258652, |
| "learning_rate": 2.0771315028901736e-05, |
| "loss": 0.0108, |
| "step": 11500 |
| }, |
| { |
| "epoch": 0.6923956890842314, |
| "eval_cosine_accuracy": 0.712537944316864, |
| "eval_loss": 0.05877280980348587, |
| "eval_runtime": 49.2617, |
| "eval_samples_per_second": 383.99, |
| "eval_steps_per_second": 1.502, |
| "step": 11500 |
| }, |
| { |
| "epoch": 0.722499849479198, |
| "grad_norm": 0.08495638519525528, |
| "learning_rate": 2.167449421965318e-05, |
| "loss": 0.0104, |
| "step": 12000 |
| }, |
| { |
| "epoch": 0.722499849479198, |
| "eval_cosine_accuracy": 0.690652072429657, |
| "eval_loss": 0.06046264246106148, |
| "eval_runtime": 50.2235, |
| "eval_samples_per_second": 376.636, |
| "eval_steps_per_second": 1.473, |
| "step": 12000 |
| }, |
| { |
| "epoch": 0.7526040098741646, |
| "grad_norm": 0.08468858897686005, |
| "learning_rate": 2.2577673410404626e-05, |
| "loss": 0.01, |
| "step": 12500 |
| }, |
| { |
| "epoch": 0.7526040098741646, |
| "eval_cosine_accuracy": 0.6941410899162292, |
| "eval_loss": 0.0607633963227272, |
| "eval_runtime": 49.538, |
| "eval_samples_per_second": 381.848, |
| "eval_steps_per_second": 1.494, |
| "step": 12500 |
| }, |
| { |
| "epoch": 0.7827081702691312, |
| "grad_norm": 0.08423547446727753, |
| "learning_rate": 2.348085260115607e-05, |
| "loss": 0.0098, |
| "step": 13000 |
| }, |
| { |
| "epoch": 0.7827081702691312, |
| "eval_cosine_accuracy": 0.6935520172119141, |
| "eval_loss": 0.060512788593769073, |
| "eval_runtime": 50.0245, |
| "eval_samples_per_second": 378.135, |
| "eval_steps_per_second": 1.479, |
| "step": 13000 |
| }, |
| { |
| "epoch": 0.8128123306640977, |
| "grad_norm": 0.09347085654735565, |
| "learning_rate": 2.4384031791907515e-05, |
| "loss": 0.0095, |
| "step": 13500 |
| }, |
| { |
| "epoch": 0.8128123306640977, |
| "eval_cosine_accuracy": 0.6864833235740662, |
| "eval_loss": 0.06169861555099487, |
| "eval_runtime": 49.1632, |
| "eval_samples_per_second": 384.759, |
| "eval_steps_per_second": 1.505, |
| "step": 13500 |
| }, |
| { |
| "epoch": 0.8429164910590644, |
| "grad_norm": 0.08714573830366135, |
| "learning_rate": 2.528721098265896e-05, |
| "loss": 0.0093, |
| "step": 14000 |
| }, |
| { |
| "epoch": 0.8429164910590644, |
| "eval_cosine_accuracy": 0.6852598786354065, |
| "eval_loss": 0.06119261682033539, |
| "eval_runtime": 49.9368, |
| "eval_samples_per_second": 378.799, |
| "eval_steps_per_second": 1.482, |
| "step": 14000 |
| }, |
| { |
| "epoch": 0.873020651454031, |
| "grad_norm": 0.07207969576120377, |
| "learning_rate": 2.6190390173410404e-05, |
| "loss": 0.009, |
| "step": 14500 |
| }, |
| { |
| "epoch": 0.873020651454031, |
| "eval_cosine_accuracy": 0.6856223940849304, |
| "eval_loss": 0.062048524618148804, |
| "eval_runtime": 51.5454, |
| "eval_samples_per_second": 366.977, |
| "eval_steps_per_second": 1.436, |
| "step": 14500 |
| }, |
| { |
| "epoch": 0.9031248118489975, |
| "grad_norm": 0.06467518210411072, |
| "learning_rate": 2.709176300578035e-05, |
| "loss": 0.0088, |
| "step": 15000 |
| }, |
| { |
| "epoch": 0.9031248118489975, |
| "eval_cosine_accuracy": 0.6777833104133606, |
| "eval_loss": 0.06211797520518303, |
| "eval_runtime": 50.4574, |
| "eval_samples_per_second": 374.891, |
| "eval_steps_per_second": 1.467, |
| "step": 15000 |
| }, |
| { |
| "epoch": 0.9332289722439641, |
| "grad_norm": 0.06801892071962357, |
| "learning_rate": 2.7994942196531793e-05, |
| "loss": 0.0087, |
| "step": 15500 |
| }, |
| { |
| "epoch": 0.9332289722439641, |
| "eval_cosine_accuracy": 0.6719833016395569, |
| "eval_loss": 0.062346458435058594, |
| "eval_runtime": 49.8448, |
| "eval_samples_per_second": 379.498, |
| "eval_steps_per_second": 1.485, |
| "step": 15500 |
| }, |
| { |
| "epoch": 0.9633331326389307, |
| "grad_norm": 0.09557406604290009, |
| "learning_rate": 2.8898121387283238e-05, |
| "loss": 0.0084, |
| "step": 16000 |
| }, |
| { |
| "epoch": 0.9633331326389307, |
| "eval_cosine_accuracy": 0.6719833016395569, |
| "eval_loss": 0.06308408081531525, |
| "eval_runtime": 50.6807, |
| "eval_samples_per_second": 373.239, |
| "eval_steps_per_second": 1.46, |
| "step": 16000 |
| }, |
| { |
| "epoch": 0.9934372930338973, |
| "grad_norm": 0.09788572043180466, |
| "learning_rate": 2.9801300578034682e-05, |
| "loss": 0.0081, |
| "step": 16500 |
| }, |
| { |
| "epoch": 0.9934372930338973, |
| "eval_cosine_accuracy": 0.6685395836830139, |
| "eval_loss": 0.06304680556058884, |
| "eval_runtime": 49.8791, |
| "eval_samples_per_second": 379.237, |
| "eval_steps_per_second": 1.484, |
| "step": 16500 |
| }, |
| { |
| "epoch": 1.0235414534288638, |
| "grad_norm": 0.07311633229255676, |
| "learning_rate": 2.9765192261379144e-05, |
| "loss": 0.0079, |
| "step": 17000 |
| }, |
| { |
| "epoch": 1.0235414534288638, |
| "eval_cosine_accuracy": 0.6755629777908325, |
| "eval_loss": 0.06368044018745422, |
| "eval_runtime": 50.1499, |
| "eval_samples_per_second": 377.189, |
| "eval_steps_per_second": 1.476, |
| "step": 17000 |
| }, |
| { |
| "epoch": 1.0536456138238304, |
| "grad_norm": 0.08946435153484344, |
| "learning_rate": 2.9464156699044712e-05, |
| "loss": 0.0076, |
| "step": 17500 |
| }, |
| { |
| "epoch": 1.0536456138238304, |
| "eval_cosine_accuracy": 0.6592505574226379, |
| "eval_loss": 0.06433094292879105, |
| "eval_runtime": 51.6083, |
| "eval_samples_per_second": 366.53, |
| "eval_steps_per_second": 1.434, |
| "step": 17500 |
| }, |
| { |
| "epoch": 1.0837497742187971, |
| "grad_norm": 0.09596831351518631, |
| "learning_rate": 2.9163723207834953e-05, |
| "loss": 0.0075, |
| "step": 18000 |
| }, |
| { |
| "epoch": 1.0837497742187971, |
| "eval_cosine_accuracy": 0.6624224185943604, |
| "eval_loss": 0.06404981762170792, |
| "eval_runtime": 50.7693, |
| "eval_samples_per_second": 372.587, |
| "eval_steps_per_second": 1.458, |
| "step": 18000 |
| }, |
| { |
| "epoch": 1.1138539346137637, |
| "grad_norm": 0.06813184171915054, |
| "learning_rate": 2.8862687645500524e-05, |
| "loss": 0.0072, |
| "step": 18500 |
| }, |
| { |
| "epoch": 1.1138539346137637, |
| "eval_cosine_accuracy": 0.6661379933357239, |
| "eval_loss": 0.06414858251810074, |
| "eval_runtime": 50.2287, |
| "eval_samples_per_second": 376.597, |
| "eval_steps_per_second": 1.473, |
| "step": 18500 |
| }, |
| { |
| "epoch": 1.1439580950087302, |
| "grad_norm": 0.10846801847219467, |
| "learning_rate": 2.8561652083166093e-05, |
| "loss": 0.0072, |
| "step": 19000 |
| }, |
| { |
| "epoch": 1.1439580950087302, |
| "eval_cosine_accuracy": 0.6571661829948425, |
| "eval_loss": 0.06486046314239502, |
| "eval_runtime": 50.9943, |
| "eval_samples_per_second": 370.944, |
| "eval_steps_per_second": 1.451, |
| "step": 19000 |
| }, |
| { |
| "epoch": 1.1740622554036968, |
| "grad_norm": 0.06637932360172272, |
| "learning_rate": 2.826061652083166e-05, |
| "loss": 0.0069, |
| "step": 19500 |
| }, |
| { |
| "epoch": 1.1740622554036968, |
| "eval_cosine_accuracy": 0.665820837020874, |
| "eval_loss": 0.06439951062202454, |
| "eval_runtime": 50.2122, |
| "eval_samples_per_second": 376.722, |
| "eval_steps_per_second": 1.474, |
| "step": 19500 |
| }, |
| { |
| "epoch": 1.2041664157986633, |
| "grad_norm": 0.09676710516214371, |
| "learning_rate": 2.7959580958497233e-05, |
| "loss": 0.0068, |
| "step": 20000 |
| }, |
| { |
| "epoch": 1.2041664157986633, |
| "eval_cosine_accuracy": 0.653948962688446, |
| "eval_loss": 0.06513187289237976, |
| "eval_runtime": 50.884, |
| "eval_samples_per_second": 371.747, |
| "eval_steps_per_second": 1.454, |
| "step": 20000 |
| }, |
| { |
| "epoch": 1.23427057619363, |
| "grad_norm": 0.097003273665905, |
| "learning_rate": 2.765914746728747e-05, |
| "loss": 0.0068, |
| "step": 20500 |
| }, |
| { |
| "epoch": 1.23427057619363, |
| "eval_cosine_accuracy": 0.6686755418777466, |
| "eval_loss": 0.06469842791557312, |
| "eval_runtime": 51.9979, |
| "eval_samples_per_second": 363.784, |
| "eval_steps_per_second": 1.423, |
| "step": 20500 |
| }, |
| { |
| "epoch": 1.2643747365885964, |
| "grad_norm": 0.05733317509293556, |
| "learning_rate": 2.7358111904953038e-05, |
| "loss": 0.0066, |
| "step": 21000 |
| }, |
| { |
| "epoch": 1.2643747365885964, |
| "eval_cosine_accuracy": 0.6604739427566528, |
| "eval_loss": 0.06482899188995361, |
| "eval_runtime": 50.6939, |
| "eval_samples_per_second": 373.142, |
| "eval_steps_per_second": 1.46, |
| "step": 21000 |
| }, |
| { |
| "epoch": 1.2944788969835632, |
| "grad_norm": 0.08877195417881012, |
| "learning_rate": 2.705707634261861e-05, |
| "loss": 0.0065, |
| "step": 21500 |
| }, |
| { |
| "epoch": 1.2944788969835632, |
| "eval_cosine_accuracy": 0.6476505398750305, |
| "eval_loss": 0.0662488043308258, |
| "eval_runtime": 51.5772, |
| "eval_samples_per_second": 366.752, |
| "eval_steps_per_second": 1.435, |
| "step": 21500 |
| }, |
| { |
| "epoch": 1.3245830573785298, |
| "grad_norm": 0.07117483764886856, |
| "learning_rate": 2.6756040780284178e-05, |
| "loss": 0.0063, |
| "step": 22000 |
| }, |
| { |
| "epoch": 1.3245830573785298, |
| "eval_cosine_accuracy": 0.6538130640983582, |
| "eval_loss": 0.06583953648805618, |
| "eval_runtime": 50.9895, |
| "eval_samples_per_second": 370.978, |
| "eval_steps_per_second": 1.451, |
| "step": 22000 |
| }, |
| { |
| "epoch": 1.3546872177734963, |
| "grad_norm": 0.07409454882144928, |
| "learning_rate": 2.6455005217949746e-05, |
| "loss": 0.0062, |
| "step": 22500 |
| }, |
| { |
| "epoch": 1.3546872177734963, |
| "eval_cosine_accuracy": 0.6610177159309387, |
| "eval_loss": 0.06499389559030533, |
| "eval_runtime": 50.7896, |
| "eval_samples_per_second": 372.438, |
| "eval_steps_per_second": 1.457, |
| "step": 22500 |
| }, |
| { |
| "epoch": 1.3847913781684629, |
| "grad_norm": 0.05476192757487297, |
| "learning_rate": 2.6153969655615315e-05, |
| "loss": 0.0062, |
| "step": 23000 |
| }, |
| { |
| "epoch": 1.3847913781684629, |
| "eval_cosine_accuracy": 0.652861475944519, |
| "eval_loss": 0.06607314944267273, |
| "eval_runtime": 51.3197, |
| "eval_samples_per_second": 368.591, |
| "eval_steps_per_second": 1.442, |
| "step": 23000 |
| }, |
| { |
| "epoch": 1.4148955385634294, |
| "grad_norm": 0.07840067893266678, |
| "learning_rate": 2.5852934093280886e-05, |
| "loss": 0.0061, |
| "step": 23500 |
| }, |
| { |
| "epoch": 1.4148955385634294, |
| "eval_cosine_accuracy": 0.6539036631584167, |
| "eval_loss": 0.06463825702667236, |
| "eval_runtime": 50.7198, |
| "eval_samples_per_second": 372.951, |
| "eval_steps_per_second": 1.459, |
| "step": 23500 |
| }, |
| { |
| "epoch": 1.444999698958396, |
| "grad_norm": 0.07566720247268677, |
| "learning_rate": 2.5551898530946458e-05, |
| "loss": 0.0062, |
| "step": 24000 |
| }, |
| { |
| "epoch": 1.444999698958396, |
| "eval_cosine_accuracy": 0.6476958394050598, |
| "eval_loss": 0.06537715345621109, |
| "eval_runtime": 50.9271, |
| "eval_samples_per_second": 371.433, |
| "eval_steps_per_second": 1.453, |
| "step": 24000 |
| }, |
| { |
| "epoch": 1.4751038593533625, |
| "grad_norm": 0.06703758984804153, |
| "learning_rate": 2.5251465039736695e-05, |
| "loss": 0.0059, |
| "step": 24500 |
| }, |
| { |
| "epoch": 1.4751038593533625, |
| "eval_cosine_accuracy": 0.6491458415985107, |
| "eval_loss": 0.0658637285232544, |
| "eval_runtime": 50.4309, |
| "eval_samples_per_second": 375.088, |
| "eval_steps_per_second": 1.467, |
| "step": 24500 |
| }, |
| { |
| "epoch": 1.5052080197483293, |
| "grad_norm": 0.0920301228761673, |
| "learning_rate": 2.4950429477402267e-05, |
| "loss": 0.0059, |
| "step": 25000 |
| }, |
| { |
| "epoch": 1.5052080197483293, |
| "eval_cosine_accuracy": 0.6486474275588989, |
| "eval_loss": 0.06605885177850723, |
| "eval_runtime": 50.4688, |
| "eval_samples_per_second": 374.806, |
| "eval_steps_per_second": 1.466, |
| "step": 25000 |
| }, |
| { |
| "epoch": 1.5353121801432958, |
| "grad_norm": 0.07251323014497757, |
| "learning_rate": 2.4649393915067835e-05, |
| "loss": 0.0058, |
| "step": 25500 |
| }, |
| { |
| "epoch": 1.5353121801432958, |
| "eval_cosine_accuracy": 0.6536318063735962, |
| "eval_loss": 0.06586486846208572, |
| "eval_runtime": 52.1004, |
| "eval_samples_per_second": 363.068, |
| "eval_steps_per_second": 1.42, |
| "step": 25500 |
| }, |
| { |
| "epoch": 1.5654163405382624, |
| "grad_norm": 0.08530519902706146, |
| "learning_rate": 2.4348358352733404e-05, |
| "loss": 0.0058, |
| "step": 26000 |
| }, |
| { |
| "epoch": 1.5654163405382624, |
| "eval_cosine_accuracy": 0.6499161720275879, |
| "eval_loss": 0.06592197716236115, |
| "eval_runtime": 51.0471, |
| "eval_samples_per_second": 370.56, |
| "eval_steps_per_second": 1.45, |
| "step": 26000 |
| }, |
| { |
| "epoch": 1.595520500933229, |
| "grad_norm": 0.06739653646945953, |
| "learning_rate": 2.404792486152364e-05, |
| "loss": 0.0056, |
| "step": 26500 |
| }, |
| { |
| "epoch": 1.595520500933229, |
| "eval_cosine_accuracy": 0.6436630487442017, |
| "eval_loss": 0.06647592782974243, |
| "eval_runtime": 50.1354, |
| "eval_samples_per_second": 377.298, |
| "eval_steps_per_second": 1.476, |
| "step": 26500 |
| }, |
| { |
| "epoch": 1.6256246613281955, |
| "grad_norm": 0.07019318640232086, |
| "learning_rate": 2.3746889299189213e-05, |
| "loss": 0.0057, |
| "step": 27000 |
| }, |
| { |
| "epoch": 1.6256246613281955, |
| "eval_cosine_accuracy": 0.6358693242073059, |
| "eval_loss": 0.06627315282821655, |
| "eval_runtime": 51.0378, |
| "eval_samples_per_second": 370.627, |
| "eval_steps_per_second": 1.45, |
| "step": 27000 |
| }, |
| { |
| "epoch": 1.6557288217231623, |
| "grad_norm": 0.07720500975847244, |
| "learning_rate": 2.344585373685478e-05, |
| "loss": 0.0055, |
| "step": 27500 |
| }, |
| { |
| "epoch": 1.6557288217231623, |
| "eval_cosine_accuracy": 0.650641143321991, |
| "eval_loss": 0.06608286499977112, |
| "eval_runtime": 50.218, |
| "eval_samples_per_second": 376.678, |
| "eval_steps_per_second": 1.474, |
| "step": 27500 |
| }, |
| { |
| "epoch": 1.6858329821181286, |
| "grad_norm": 0.07247819751501083, |
| "learning_rate": 2.314481817452035e-05, |
| "loss": 0.0055, |
| "step": 28000 |
| }, |
| { |
| "epoch": 1.6858329821181286, |
| "eval_cosine_accuracy": 0.6416239738464355, |
| "eval_loss": 0.06629126518964767, |
| "eval_runtime": 50.2828, |
| "eval_samples_per_second": 376.193, |
| "eval_steps_per_second": 1.472, |
| "step": 28000 |
| }, |
| { |
| "epoch": 1.7159371425130954, |
| "grad_norm": 0.06985726207494736, |
| "learning_rate": 2.284438468331059e-05, |
| "loss": 0.0055, |
| "step": 28500 |
| }, |
| { |
| "epoch": 1.7159371425130954, |
| "eval_cosine_accuracy": 0.6505505442619324, |
| "eval_loss": 0.06659659743309021, |
| "eval_runtime": 51.8903, |
| "eval_samples_per_second": 364.539, |
| "eval_steps_per_second": 1.426, |
| "step": 28500 |
| }, |
| { |
| "epoch": 1.746041302908062, |
| "grad_norm": 0.06849692761898041, |
| "learning_rate": 2.2543349120976158e-05, |
| "loss": 0.0053, |
| "step": 29000 |
| }, |
| { |
| "epoch": 1.746041302908062, |
| "eval_cosine_accuracy": 0.6441161632537842, |
| "eval_loss": 0.06675810366868973, |
| "eval_runtime": 51.0645, |
| "eval_samples_per_second": 370.433, |
| "eval_steps_per_second": 1.449, |
| "step": 29000 |
| }, |
| { |
| "epoch": 1.7761454633030285, |
| "grad_norm": 0.06756169348955154, |
| "learning_rate": 2.2242313558641726e-05, |
| "loss": 0.0054, |
| "step": 29500 |
| }, |
| { |
| "epoch": 1.7761454633030285, |
| "eval_cosine_accuracy": 0.6515474319458008, |
| "eval_loss": 0.06594917923212051, |
| "eval_runtime": 50.7029, |
| "eval_samples_per_second": 373.076, |
| "eval_steps_per_second": 1.459, |
| "step": 29500 |
| }, |
| { |
| "epoch": 1.806249623697995, |
| "grad_norm": 0.07231052964925766, |
| "learning_rate": 2.1941880067431964e-05, |
| "loss": 0.0052, |
| "step": 30000 |
| }, |
| { |
| "epoch": 1.806249623697995, |
| "eval_cosine_accuracy": 0.6406271457672119, |
| "eval_loss": 0.06695431470870972, |
| "eval_runtime": 50.5078, |
| "eval_samples_per_second": 374.517, |
| "eval_steps_per_second": 1.465, |
| "step": 30000 |
| }, |
| { |
| "epoch": 1.8363537840929616, |
| "grad_norm": 0.07190578430891037, |
| "learning_rate": 2.1640844505097535e-05, |
| "loss": 0.0051, |
| "step": 30500 |
| }, |
| { |
| "epoch": 1.8363537840929616, |
| "eval_cosine_accuracy": 0.6470614671707153, |
| "eval_loss": 0.06647095829248428, |
| "eval_runtime": 51.5154, |
| "eval_samples_per_second": 367.191, |
| "eval_steps_per_second": 1.436, |
| "step": 30500 |
| }, |
| { |
| "epoch": 1.8664579444879283, |
| "grad_norm": 0.07759535312652588, |
| "learning_rate": 2.1339808942763107e-05, |
| "loss": 0.0052, |
| "step": 31000 |
| }, |
| { |
| "epoch": 1.8664579444879283, |
| "eval_cosine_accuracy": 0.6441161632537842, |
| "eval_loss": 0.06696631759405136, |
| "eval_runtime": 50.0831, |
| "eval_samples_per_second": 377.693, |
| "eval_steps_per_second": 1.478, |
| "step": 31000 |
| }, |
| { |
| "epoch": 1.8965621048828947, |
| "grad_norm": 0.051443614065647125, |
| "learning_rate": 2.1038773380428675e-05, |
| "loss": 0.0051, |
| "step": 31500 |
| }, |
| { |
| "epoch": 1.8965621048828947, |
| "eval_cosine_accuracy": 0.6458833813667297, |
| "eval_loss": 0.06696301698684692, |
| "eval_runtime": 50.3608, |
| "eval_samples_per_second": 375.61, |
| "eval_steps_per_second": 1.469, |
| "step": 31500 |
| }, |
| { |
| "epoch": 1.9266662652778614, |
| "grad_norm": 0.054654281586408615, |
| "learning_rate": 2.0737737818094247e-05, |
| "loss": 0.005, |
| "step": 32000 |
| }, |
| { |
| "epoch": 1.9266662652778614, |
| "eval_cosine_accuracy": 0.6497349143028259, |
| "eval_loss": 0.0671425610780716, |
| "eval_runtime": 50.0416, |
| "eval_samples_per_second": 378.005, |
| "eval_steps_per_second": 1.479, |
| "step": 32000 |
| }, |
| { |
| "epoch": 1.956770425672828, |
| "grad_norm": 0.09258880466222763, |
| "learning_rate": 2.0437304326884484e-05, |
| "loss": 0.005, |
| "step": 32500 |
| }, |
| { |
| "epoch": 1.956770425672828, |
| "eval_cosine_accuracy": 0.6404005885124207, |
| "eval_loss": 0.06720578670501709, |
| "eval_runtime": 50.6021, |
| "eval_samples_per_second": 373.819, |
| "eval_steps_per_second": 1.462, |
| "step": 32500 |
| }, |
| { |
| "epoch": 1.9868745860677945, |
| "grad_norm": 0.080738864839077, |
| "learning_rate": 2.0136268764550053e-05, |
| "loss": 0.005, |
| "step": 33000 |
| }, |
| { |
| "epoch": 1.9868745860677945, |
| "eval_cosine_accuracy": 0.6384521126747131, |
| "eval_loss": 0.06707385927438736, |
| "eval_runtime": 50.3284, |
| "eval_samples_per_second": 375.852, |
| "eval_steps_per_second": 1.47, |
| "step": 33000 |
| } |
| ], |
| "logging_steps": 500, |
| "max_steps": 66436, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 4, |
| "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": 0.0, |
| "train_batch_size": 256, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
