| { | |
| "best_metric": 0.5973209738731384, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.635593220338983, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.00423728813559322, | |
| "grad_norm": 2.8869433403015137, | |
| "learning_rate": 5e-06, | |
| "loss": 6.9072, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.00423728813559322, | |
| "eval_loss": 1.7482346296310425, | |
| "eval_runtime": 42.7906, | |
| "eval_samples_per_second": 9.301, | |
| "eval_steps_per_second": 4.651, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.00847457627118644, | |
| "grad_norm": 2.9117305278778076, | |
| "learning_rate": 1e-05, | |
| "loss": 6.6519, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.012711864406779662, | |
| "grad_norm": 3.042799711227417, | |
| "learning_rate": 1.5e-05, | |
| "loss": 6.8339, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.01694915254237288, | |
| "grad_norm": 2.9370641708374023, | |
| "learning_rate": 2e-05, | |
| "loss": 6.8588, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.0211864406779661, | |
| "grad_norm": 2.8842549324035645, | |
| "learning_rate": 2.5e-05, | |
| "loss": 6.3375, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.025423728813559324, | |
| "grad_norm": 3.039867877960205, | |
| "learning_rate": 3e-05, | |
| "loss": 6.757, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.029661016949152543, | |
| "grad_norm": 2.889957904815674, | |
| "learning_rate": 3.5e-05, | |
| "loss": 6.3478, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.03389830508474576, | |
| "grad_norm": 2.849916934967041, | |
| "learning_rate": 4e-05, | |
| "loss": 5.7138, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.038135593220338986, | |
| "grad_norm": 2.6687538623809814, | |
| "learning_rate": 4.5e-05, | |
| "loss": 5.1771, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.0423728813559322, | |
| "grad_norm": 2.5558407306671143, | |
| "learning_rate": 5e-05, | |
| "loss": 5.3018, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.046610169491525424, | |
| "grad_norm": 2.6599059104919434, | |
| "learning_rate": 5.500000000000001e-05, | |
| "loss": 4.3197, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.05084745762711865, | |
| "grad_norm": 2.4746832847595215, | |
| "learning_rate": 6e-05, | |
| "loss": 3.7984, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.05508474576271186, | |
| "grad_norm": 2.8662729263305664, | |
| "learning_rate": 6.500000000000001e-05, | |
| "loss": 3.2409, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.059322033898305086, | |
| "grad_norm": 2.0270683765411377, | |
| "learning_rate": 7e-05, | |
| "loss": 3.0089, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.0635593220338983, | |
| "grad_norm": 2.611672878265381, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 2.484, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.06779661016949153, | |
| "grad_norm": 2.9032974243164062, | |
| "learning_rate": 8e-05, | |
| "loss": 2.8204, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.07203389830508475, | |
| "grad_norm": 2.599370002746582, | |
| "learning_rate": 8.5e-05, | |
| "loss": 2.7265, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.07627118644067797, | |
| "grad_norm": 1.805683970451355, | |
| "learning_rate": 9e-05, | |
| "loss": 2.9644, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.08050847457627118, | |
| "grad_norm": 2.769292116165161, | |
| "learning_rate": 9.5e-05, | |
| "loss": 2.6633, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.0847457627118644, | |
| "grad_norm": 3.1495392322540283, | |
| "learning_rate": 0.0001, | |
| "loss": 2.9764, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.08898305084745763, | |
| "grad_norm": 2.3471312522888184, | |
| "learning_rate": 9.999238475781957e-05, | |
| "loss": 2.4665, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.09322033898305085, | |
| "grad_norm": 2.6704769134521484, | |
| "learning_rate": 9.99695413509548e-05, | |
| "loss": 2.2892, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.09745762711864407, | |
| "grad_norm": 1.0994267463684082, | |
| "learning_rate": 9.99314767377287e-05, | |
| "loss": 2.344, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.1016949152542373, | |
| "grad_norm": 2.1697537899017334, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 2.7479, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.1059322033898305, | |
| "grad_norm": 2.7966907024383545, | |
| "learning_rate": 9.980973490458728e-05, | |
| "loss": 2.5423, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.11016949152542373, | |
| "grad_norm": 1.6926209926605225, | |
| "learning_rate": 9.972609476841367e-05, | |
| "loss": 2.7024, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.11440677966101695, | |
| "grad_norm": 2.5315053462982178, | |
| "learning_rate": 9.962730758206611e-05, | |
| "loss": 2.4407, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.11864406779661017, | |
| "grad_norm": 3.6142640113830566, | |
| "learning_rate": 9.951340343707852e-05, | |
| "loss": 2.2694, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.1228813559322034, | |
| "grad_norm": 1.3599029779434204, | |
| "learning_rate": 9.938441702975689e-05, | |
| "loss": 2.3543, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.1271186440677966, | |
| "grad_norm": 2.29304575920105, | |
| "learning_rate": 9.924038765061042e-05, | |
| "loss": 2.5706, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.13135593220338984, | |
| "grad_norm": 1.6786038875579834, | |
| "learning_rate": 9.908135917238321e-05, | |
| "loss": 2.5246, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.13559322033898305, | |
| "grad_norm": 1.260206699371338, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 1.9627, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.13983050847457626, | |
| "grad_norm": 1.5896670818328857, | |
| "learning_rate": 9.871850323926177e-05, | |
| "loss": 2.4204, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.1440677966101695, | |
| "grad_norm": 2.370612144470215, | |
| "learning_rate": 9.851478631379982e-05, | |
| "loss": 2.8077, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.1483050847457627, | |
| "grad_norm": 1.4520262479782104, | |
| "learning_rate": 9.829629131445342e-05, | |
| "loss": 2.4863, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.15254237288135594, | |
| "grad_norm": 1.3572436571121216, | |
| "learning_rate": 9.806308479691595e-05, | |
| "loss": 2.4152, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.15677966101694915, | |
| "grad_norm": 1.1999151706695557, | |
| "learning_rate": 9.781523779815179e-05, | |
| "loss": 2.3192, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.16101694915254236, | |
| "grad_norm": 2.2570760250091553, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 2.3025, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.1652542372881356, | |
| "grad_norm": 1.7871544361114502, | |
| "learning_rate": 9.727592877996585e-05, | |
| "loss": 2.0637, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.1694915254237288, | |
| "grad_norm": 1.6054717302322388, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 2.1793, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.17372881355932204, | |
| "grad_norm": 1.5231215953826904, | |
| "learning_rate": 9.667902132486009e-05, | |
| "loss": 2.2197, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.17796610169491525, | |
| "grad_norm": 1.3351085186004639, | |
| "learning_rate": 9.635919272833938e-05, | |
| "loss": 2.3261, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.18220338983050846, | |
| "grad_norm": 1.2223113775253296, | |
| "learning_rate": 9.602524267262203e-05, | |
| "loss": 2.1145, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.1864406779661017, | |
| "grad_norm": 1.6961467266082764, | |
| "learning_rate": 9.567727288213005e-05, | |
| "loss": 2.2667, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.1906779661016949, | |
| "grad_norm": 1.4688829183578491, | |
| "learning_rate": 9.53153893518325e-05, | |
| "loss": 2.572, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.19491525423728814, | |
| "grad_norm": 1.4838528633117676, | |
| "learning_rate": 9.493970231495835e-05, | |
| "loss": 2.3653, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.19915254237288135, | |
| "grad_norm": 2.620893955230713, | |
| "learning_rate": 9.45503262094184e-05, | |
| "loss": 2.1069, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.2033898305084746, | |
| "grad_norm": 1.6659537553787231, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 2.6601, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.2076271186440678, | |
| "grad_norm": 1.215427279472351, | |
| "learning_rate": 9.373098535696979e-05, | |
| "loss": 2.7816, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.211864406779661, | |
| "grad_norm": 2.305882453918457, | |
| "learning_rate": 9.330127018922194e-05, | |
| "loss": 3.6478, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.211864406779661, | |
| "eval_loss": 0.7533118724822998, | |
| "eval_runtime": 43.5309, | |
| "eval_samples_per_second": 9.143, | |
| "eval_steps_per_second": 4.571, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.21610169491525424, | |
| "grad_norm": 10.459978103637695, | |
| "learning_rate": 9.285836503510562e-05, | |
| "loss": 3.4556, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.22033898305084745, | |
| "grad_norm": 8.529092788696289, | |
| "learning_rate": 9.24024048078213e-05, | |
| "loss": 3.3025, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.2245762711864407, | |
| "grad_norm": 5.131667137145996, | |
| "learning_rate": 9.193352839727121e-05, | |
| "loss": 2.5715, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.2288135593220339, | |
| "grad_norm": 1.3166052103042603, | |
| "learning_rate": 9.145187862775209e-05, | |
| "loss": 2.4219, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.2330508474576271, | |
| "grad_norm": 2.2176835536956787, | |
| "learning_rate": 9.09576022144496e-05, | |
| "loss": 2.5179, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.23728813559322035, | |
| "grad_norm": 2.267003059387207, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 2.6671, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.24152542372881355, | |
| "grad_norm": 1.7872462272644043, | |
| "learning_rate": 8.993177550236464e-05, | |
| "loss": 2.6358, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.2457627118644068, | |
| "grad_norm": 1.1025911569595337, | |
| "learning_rate": 8.940053768033609e-05, | |
| "loss": 2.4904, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 1.2282601594924927, | |
| "learning_rate": 8.885729807284856e-05, | |
| "loss": 2.2931, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.2542372881355932, | |
| "grad_norm": 2.1205925941467285, | |
| "learning_rate": 8.83022221559489e-05, | |
| "loss": 2.4453, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.2584745762711864, | |
| "grad_norm": 1.7363345623016357, | |
| "learning_rate": 8.773547901113862e-05, | |
| "loss": 2.1719, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.2627118644067797, | |
| "grad_norm": 0.8407338857650757, | |
| "learning_rate": 8.715724127386972e-05, | |
| "loss": 2.2618, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.2669491525423729, | |
| "grad_norm": 1.0721104145050049, | |
| "learning_rate": 8.656768508095853e-05, | |
| "loss": 2.228, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.2711864406779661, | |
| "grad_norm": 1.6518641710281372, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 2.6224, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.2754237288135593, | |
| "grad_norm": 1.0283952951431274, | |
| "learning_rate": 8.535533905932738e-05, | |
| "loss": 2.2026, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.2796610169491525, | |
| "grad_norm": 0.8485298156738281, | |
| "learning_rate": 8.473291852294987e-05, | |
| "loss": 2.3136, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.2838983050847458, | |
| "grad_norm": 1.244860053062439, | |
| "learning_rate": 8.409991800312493e-05, | |
| "loss": 2.476, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.288135593220339, | |
| "grad_norm": 1.248189091682434, | |
| "learning_rate": 8.345653031794292e-05, | |
| "loss": 2.2362, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.2923728813559322, | |
| "grad_norm": 1.0057225227355957, | |
| "learning_rate": 8.280295144952536e-05, | |
| "loss": 2.514, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.2966101694915254, | |
| "grad_norm": 0.8649203777313232, | |
| "learning_rate": 8.213938048432697e-05, | |
| "loss": 2.1474, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.3008474576271186, | |
| "grad_norm": 0.601910412311554, | |
| "learning_rate": 8.146601955249188e-05, | |
| "loss": 2.0126, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.3050847457627119, | |
| "grad_norm": 0.7620223164558411, | |
| "learning_rate": 8.07830737662829e-05, | |
| "loss": 2.4381, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.3093220338983051, | |
| "grad_norm": 0.8148074746131897, | |
| "learning_rate": 8.009075115760243e-05, | |
| "loss": 2.1575, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.3135593220338983, | |
| "grad_norm": 0.7290023565292358, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 2.299, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.3177966101694915, | |
| "grad_norm": 0.8110753297805786, | |
| "learning_rate": 7.86788218175523e-05, | |
| "loss": 2.1619, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.3220338983050847, | |
| "grad_norm": 0.6752878427505493, | |
| "learning_rate": 7.795964517353735e-05, | |
| "loss": 2.2223, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.326271186440678, | |
| "grad_norm": 0.8308572173118591, | |
| "learning_rate": 7.723195175075136e-05, | |
| "loss": 1.9585, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.3305084745762712, | |
| "grad_norm": 0.8173496127128601, | |
| "learning_rate": 7.649596321166024e-05, | |
| "loss": 2.0674, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.3347457627118644, | |
| "grad_norm": 0.7854796648025513, | |
| "learning_rate": 7.575190374550272e-05, | |
| "loss": 2.0757, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.3389830508474576, | |
| "grad_norm": 0.6990013718605042, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 2.5116, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.3432203389830508, | |
| "grad_norm": 1.0307629108428955, | |
| "learning_rate": 7.424048101231686e-05, | |
| "loss": 2.2457, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.3474576271186441, | |
| "grad_norm": 0.9620252847671509, | |
| "learning_rate": 7.347357813929454e-05, | |
| "loss": 2.3633, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.3516949152542373, | |
| "grad_norm": 0.6838775277137756, | |
| "learning_rate": 7.269952498697734e-05, | |
| "loss": 2.5503, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.3559322033898305, | |
| "grad_norm": 0.6633744239807129, | |
| "learning_rate": 7.191855733945387e-05, | |
| "loss": 2.1999, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.3601694915254237, | |
| "grad_norm": 1.1613788604736328, | |
| "learning_rate": 7.113091308703498e-05, | |
| "loss": 2.5896, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.3644067796610169, | |
| "grad_norm": 0.7708940505981445, | |
| "learning_rate": 7.033683215379002e-05, | |
| "loss": 2.6019, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.3686440677966102, | |
| "grad_norm": 1.6062999963760376, | |
| "learning_rate": 6.953655642446368e-05, | |
| "loss": 2.0793, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.3728813559322034, | |
| "grad_norm": 0.7241120934486389, | |
| "learning_rate": 6.873032967079561e-05, | |
| "loss": 2.402, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.3771186440677966, | |
| "grad_norm": 1.5668368339538574, | |
| "learning_rate": 6.7918397477265e-05, | |
| "loss": 2.532, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.3813559322033898, | |
| "grad_norm": 0.7401297092437744, | |
| "learning_rate": 6.710100716628344e-05, | |
| "loss": 2.2997, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3855932203389831, | |
| "grad_norm": 0.7686449289321899, | |
| "learning_rate": 6.627840772285784e-05, | |
| "loss": 2.0515, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.3898305084745763, | |
| "grad_norm": 0.9585121273994446, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 2.215, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.3940677966101695, | |
| "grad_norm": 0.6215618848800659, | |
| "learning_rate": 6.461858523613684e-05, | |
| "loss": 2.497, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.3983050847457627, | |
| "grad_norm": 0.5972080826759338, | |
| "learning_rate": 6.378186779084995e-05, | |
| "loss": 2.0889, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.4025423728813559, | |
| "grad_norm": 0.8943988680839539, | |
| "learning_rate": 6.294095225512603e-05, | |
| "loss": 2.1595, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.4067796610169492, | |
| "grad_norm": 0.8203272223472595, | |
| "learning_rate": 6.209609477998338e-05, | |
| "loss": 1.8233, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.4110169491525424, | |
| "grad_norm": 0.5806548595428467, | |
| "learning_rate": 6.124755271719325e-05, | |
| "loss": 2.298, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.4152542372881356, | |
| "grad_norm": 0.7718837261199951, | |
| "learning_rate": 6.0395584540887963e-05, | |
| "loss": 2.623, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.4194915254237288, | |
| "grad_norm": 1.1169939041137695, | |
| "learning_rate": 5.9540449768827246e-05, | |
| "loss": 3.1457, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.423728813559322, | |
| "grad_norm": 1.5888675451278687, | |
| "learning_rate": 5.868240888334653e-05, | |
| "loss": 3.7138, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.423728813559322, | |
| "eval_loss": 0.6284539103507996, | |
| "eval_runtime": 43.533, | |
| "eval_samples_per_second": 9.142, | |
| "eval_steps_per_second": 4.571, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.4279661016949153, | |
| "grad_norm": 3.528254508972168, | |
| "learning_rate": 5.782172325201155e-05, | |
| "loss": 2.6844, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.4322033898305085, | |
| "grad_norm": 4.093210697174072, | |
| "learning_rate": 5.695865504800327e-05, | |
| "loss": 2.7143, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.4364406779661017, | |
| "grad_norm": 3.43405818939209, | |
| "learning_rate": 5.6093467170257374e-05, | |
| "loss": 2.6711, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.4406779661016949, | |
| "grad_norm": 2.003894567489624, | |
| "learning_rate": 5.522642316338268e-05, | |
| "loss": 2.3973, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.4449152542372881, | |
| "grad_norm": 1.3836109638214111, | |
| "learning_rate": 5.435778713738292e-05, | |
| "loss": 2.6954, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.4491525423728814, | |
| "grad_norm": 0.8960191011428833, | |
| "learning_rate": 5.348782368720626e-05, | |
| "loss": 2.3673, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.4533898305084746, | |
| "grad_norm": 1.0750250816345215, | |
| "learning_rate": 5.26167978121472e-05, | |
| "loss": 2.3827, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.4576271186440678, | |
| "grad_norm": 1.209937572479248, | |
| "learning_rate": 5.174497483512506e-05, | |
| "loss": 2.5107, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.461864406779661, | |
| "grad_norm": 1.4963562488555908, | |
| "learning_rate": 5.0872620321864185e-05, | |
| "loss": 2.6218, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.4661016949152542, | |
| "grad_norm": 1.4280909299850464, | |
| "learning_rate": 5e-05, | |
| "loss": 2.5591, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.4703389830508475, | |
| "grad_norm": 1.069968819618225, | |
| "learning_rate": 4.912737967813583e-05, | |
| "loss": 2.238, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.4745762711864407, | |
| "grad_norm": 0.7898651957511902, | |
| "learning_rate": 4.825502516487497e-05, | |
| "loss": 2.1143, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.4788135593220339, | |
| "grad_norm": 0.6083970069885254, | |
| "learning_rate": 4.738320218785281e-05, | |
| "loss": 2.3231, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.4830508474576271, | |
| "grad_norm": 1.1110987663269043, | |
| "learning_rate": 4.6512176312793736e-05, | |
| "loss": 2.3884, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.4872881355932203, | |
| "grad_norm": 0.694742739200592, | |
| "learning_rate": 4.564221286261709e-05, | |
| "loss": 2.2887, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.4915254237288136, | |
| "grad_norm": 0.9567046165466309, | |
| "learning_rate": 4.477357683661734e-05, | |
| "loss": 2.2007, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.4957627118644068, | |
| "grad_norm": 0.8701348900794983, | |
| "learning_rate": 4.390653282974264e-05, | |
| "loss": 1.9919, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 0.9415125250816345, | |
| "learning_rate": 4.3041344951996746e-05, | |
| "loss": 2.2197, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.5042372881355932, | |
| "grad_norm": 0.6797757744789124, | |
| "learning_rate": 4.2178276747988446e-05, | |
| "loss": 2.1456, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.5084745762711864, | |
| "grad_norm": 0.8693689703941345, | |
| "learning_rate": 4.131759111665349e-05, | |
| "loss": 2.1877, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.5127118644067796, | |
| "grad_norm": 0.6730765104293823, | |
| "learning_rate": 4.045955023117276e-05, | |
| "loss": 2.0728, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.5169491525423728, | |
| "grad_norm": 0.8778936266899109, | |
| "learning_rate": 3.960441545911204e-05, | |
| "loss": 2.304, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.5211864406779662, | |
| "grad_norm": 0.6395032405853271, | |
| "learning_rate": 3.875244728280676e-05, | |
| "loss": 2.2309, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.5254237288135594, | |
| "grad_norm": 0.7323006391525269, | |
| "learning_rate": 3.790390522001662e-05, | |
| "loss": 2.4723, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.5296610169491526, | |
| "grad_norm": 0.7824187278747559, | |
| "learning_rate": 3.705904774487396e-05, | |
| "loss": 2.2452, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.5338983050847458, | |
| "grad_norm": 0.854575514793396, | |
| "learning_rate": 3.6218132209150045e-05, | |
| "loss": 2.2604, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.538135593220339, | |
| "grad_norm": 1.17578125, | |
| "learning_rate": 3.5381414763863166e-05, | |
| "loss": 2.7593, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.5423728813559322, | |
| "grad_norm": 0.9400164484977722, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 1.8751, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.5466101694915254, | |
| "grad_norm": 0.8660760521888733, | |
| "learning_rate": 3.372159227714218e-05, | |
| "loss": 2.0508, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.5508474576271186, | |
| "grad_norm": 0.6780417561531067, | |
| "learning_rate": 3.289899283371657e-05, | |
| "loss": 2.2132, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.5550847457627118, | |
| "grad_norm": 0.8800328969955444, | |
| "learning_rate": 3.2081602522734986e-05, | |
| "loss": 2.0362, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.559322033898305, | |
| "grad_norm": 0.6752195954322815, | |
| "learning_rate": 3.12696703292044e-05, | |
| "loss": 2.4794, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.5635593220338984, | |
| "grad_norm": 0.6003003716468811, | |
| "learning_rate": 3.046344357553632e-05, | |
| "loss": 2.3735, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.5677966101694916, | |
| "grad_norm": 0.9153070449829102, | |
| "learning_rate": 2.9663167846209998e-05, | |
| "loss": 2.4439, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.5720338983050848, | |
| "grad_norm": 1.3099541664123535, | |
| "learning_rate": 2.886908691296504e-05, | |
| "loss": 1.9584, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.576271186440678, | |
| "grad_norm": 0.6542685031890869, | |
| "learning_rate": 2.8081442660546125e-05, | |
| "loss": 2.0345, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.5805084745762712, | |
| "grad_norm": 0.6798728704452515, | |
| "learning_rate": 2.7300475013022663e-05, | |
| "loss": 2.5846, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.5847457627118644, | |
| "grad_norm": 0.6184638142585754, | |
| "learning_rate": 2.6526421860705473e-05, | |
| "loss": 1.7818, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.5889830508474576, | |
| "grad_norm": 0.6993879079818726, | |
| "learning_rate": 2.575951898768315e-05, | |
| "loss": 2.295, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.5932203389830508, | |
| "grad_norm": 0.8818557262420654, | |
| "learning_rate": 2.500000000000001e-05, | |
| "loss": 2.26, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.597457627118644, | |
| "grad_norm": 0.7364823818206787, | |
| "learning_rate": 2.4248096254497288e-05, | |
| "loss": 1.9119, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.6016949152542372, | |
| "grad_norm": 0.8221387267112732, | |
| "learning_rate": 2.350403678833976e-05, | |
| "loss": 1.935, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.6059322033898306, | |
| "grad_norm": 0.719155490398407, | |
| "learning_rate": 2.2768048249248648e-05, | |
| "loss": 1.978, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.6101694915254238, | |
| "grad_norm": 0.7184621095657349, | |
| "learning_rate": 2.2040354826462668e-05, | |
| "loss": 1.9602, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.614406779661017, | |
| "grad_norm": 0.7189619541168213, | |
| "learning_rate": 2.132117818244771e-05, | |
| "loss": 2.6127, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.6186440677966102, | |
| "grad_norm": 0.5349248647689819, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 1.8459, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.6228813559322034, | |
| "grad_norm": 0.6641550660133362, | |
| "learning_rate": 1.9909248842397584e-05, | |
| "loss": 2.0695, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.6271186440677966, | |
| "grad_norm": 0.6735299229621887, | |
| "learning_rate": 1.9216926233717085e-05, | |
| "loss": 1.8445, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.6313559322033898, | |
| "grad_norm": 0.7525951862335205, | |
| "learning_rate": 1.8533980447508137e-05, | |
| "loss": 2.0764, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.635593220338983, | |
| "grad_norm": 1.8119826316833496, | |
| "learning_rate": 1.7860619515673033e-05, | |
| "loss": 2.5421, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.635593220338983, | |
| "eval_loss": 0.5973209738731384, | |
| "eval_runtime": 43.5624, | |
| "eval_samples_per_second": 9.136, | |
| "eval_steps_per_second": 4.568, | |
| "step": 150 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 5, | |
| "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": 3.81890674556928e+17, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |