Zenith-7B / checkpoint-1900 /trainer_state.json
gagan3012's picture
Upload folder using huggingface_hub
0af2b68 verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.8187890540831717,
"eval_steps": 100,
"global_step": 1900,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 5e-05,
"loss": 1.5475,
"step": 1
},
{
"epoch": 0.0,
"learning_rate": 4.999997705921923e-05,
"loss": 1.4738,
"step": 2
},
{
"epoch": 0.0,
"learning_rate": 4.9999908236919016e-05,
"loss": 1.709,
"step": 3
},
{
"epoch": 0.0,
"learning_rate": 4.999979353322567e-05,
"loss": 1.3682,
"step": 4
},
{
"epoch": 0.0,
"learning_rate": 4.99996329483497e-05,
"loss": 1.4525,
"step": 5
},
{
"epoch": 0.0,
"learning_rate": 4.999942648258581e-05,
"loss": 1.2902,
"step": 6
},
{
"epoch": 0.0,
"learning_rate": 4.999917413631294e-05,
"loss": 1.2581,
"step": 7
},
{
"epoch": 0.0,
"learning_rate": 4.9998875909994206e-05,
"loss": 1.1363,
"step": 8
},
{
"epoch": 0.0,
"learning_rate": 4.9998531804176926e-05,
"loss": 1.0479,
"step": 9
},
{
"epoch": 0.0,
"learning_rate": 4.999814181949263e-05,
"loss": 1.0362,
"step": 10
},
{
"epoch": 0.0,
"learning_rate": 4.999770595665704e-05,
"loss": 0.9415,
"step": 11
},
{
"epoch": 0.01,
"learning_rate": 4.999722421647006e-05,
"loss": 1.0849,
"step": 12
},
{
"epoch": 0.01,
"learning_rate": 4.999669659981584e-05,
"loss": 0.8538,
"step": 13
},
{
"epoch": 0.01,
"learning_rate": 4.999612310766268e-05,
"loss": 0.9404,
"step": 14
},
{
"epoch": 0.01,
"learning_rate": 4.9995503741063096e-05,
"loss": 0.9645,
"step": 15
},
{
"epoch": 0.01,
"learning_rate": 4.999483850115379e-05,
"loss": 1.0898,
"step": 16
},
{
"epoch": 0.01,
"learning_rate": 4.9994127389155623e-05,
"loss": 0.9545,
"step": 17
},
{
"epoch": 0.01,
"learning_rate": 4.9993370406373705e-05,
"loss": 0.8432,
"step": 18
},
{
"epoch": 0.01,
"learning_rate": 4.999256755419729e-05,
"loss": 0.9946,
"step": 19
},
{
"epoch": 0.01,
"learning_rate": 4.999171883409982e-05,
"loss": 0.7908,
"step": 20
},
{
"epoch": 0.01,
"learning_rate": 4.999082424763891e-05,
"loss": 0.8971,
"step": 21
},
{
"epoch": 0.01,
"learning_rate": 4.998988379645638e-05,
"loss": 0.85,
"step": 22
},
{
"epoch": 0.01,
"learning_rate": 4.998889748227819e-05,
"loss": 0.8626,
"step": 23
},
{
"epoch": 0.01,
"learning_rate": 4.9987865306914495e-05,
"loss": 0.8335,
"step": 24
},
{
"epoch": 0.01,
"learning_rate": 4.9986787272259596e-05,
"loss": 0.9314,
"step": 25
},
{
"epoch": 0.01,
"learning_rate": 4.998566338029198e-05,
"loss": 0.8645,
"step": 26
},
{
"epoch": 0.01,
"learning_rate": 4.998449363307428e-05,
"loss": 0.9316,
"step": 27
},
{
"epoch": 0.01,
"learning_rate": 4.998327803275328e-05,
"loss": 0.805,
"step": 28
},
{
"epoch": 0.01,
"learning_rate": 4.998201658155994e-05,
"loss": 1.0418,
"step": 29
},
{
"epoch": 0.01,
"learning_rate": 4.998070928180935e-05,
"loss": 0.649,
"step": 30
},
{
"epoch": 0.01,
"learning_rate": 4.997935613590075e-05,
"loss": 0.8,
"step": 31
},
{
"epoch": 0.01,
"learning_rate": 4.997795714631751e-05,
"loss": 0.8014,
"step": 32
},
{
"epoch": 0.01,
"learning_rate": 4.997651231562715e-05,
"loss": 0.7854,
"step": 33
},
{
"epoch": 0.01,
"learning_rate": 4.9975021646481315e-05,
"loss": 0.8782,
"step": 34
},
{
"epoch": 0.02,
"learning_rate": 4.997348514161576e-05,
"loss": 0.7564,
"step": 35
},
{
"epoch": 0.02,
"learning_rate": 4.997190280385039e-05,
"loss": 0.9438,
"step": 36
},
{
"epoch": 0.02,
"learning_rate": 4.997027463608921e-05,
"loss": 0.9248,
"step": 37
},
{
"epoch": 0.02,
"learning_rate": 4.996860064132033e-05,
"loss": 0.76,
"step": 38
},
{
"epoch": 0.02,
"learning_rate": 4.9966880822615965e-05,
"loss": 0.8834,
"step": 39
},
{
"epoch": 0.02,
"learning_rate": 4.996511518313244e-05,
"loss": 0.8928,
"step": 40
},
{
"epoch": 0.02,
"learning_rate": 4.996330372611017e-05,
"loss": 0.8857,
"step": 41
},
{
"epoch": 0.02,
"learning_rate": 4.996144645487364e-05,
"loss": 0.7975,
"step": 42
},
{
"epoch": 0.02,
"learning_rate": 4.995954337283145e-05,
"loss": 0.7815,
"step": 43
},
{
"epoch": 0.02,
"learning_rate": 4.9957594483476236e-05,
"loss": 0.8869,
"step": 44
},
{
"epoch": 0.02,
"learning_rate": 4.995559979038474e-05,
"loss": 0.9173,
"step": 45
},
{
"epoch": 0.02,
"learning_rate": 4.995355929721773e-05,
"loss": 0.9258,
"step": 46
},
{
"epoch": 0.02,
"learning_rate": 4.995147300772006e-05,
"loss": 0.896,
"step": 47
},
{
"epoch": 0.02,
"learning_rate": 4.994934092572061e-05,
"loss": 0.893,
"step": 48
},
{
"epoch": 0.02,
"learning_rate": 4.994716305513232e-05,
"loss": 0.8257,
"step": 49
},
{
"epoch": 0.02,
"learning_rate": 4.994493939995214e-05,
"loss": 0.8759,
"step": 50
},
{
"epoch": 0.02,
"learning_rate": 4.994266996426107e-05,
"loss": 0.9216,
"step": 51
},
{
"epoch": 0.02,
"learning_rate": 4.994035475222412e-05,
"loss": 0.9097,
"step": 52
},
{
"epoch": 0.02,
"learning_rate": 4.9937993768090306e-05,
"loss": 1.015,
"step": 53
},
{
"epoch": 0.02,
"learning_rate": 4.993558701619266e-05,
"loss": 0.8633,
"step": 54
},
{
"epoch": 0.02,
"learning_rate": 4.9933134500948205e-05,
"loss": 0.928,
"step": 55
},
{
"epoch": 0.02,
"learning_rate": 4.993063622685794e-05,
"loss": 0.8069,
"step": 56
},
{
"epoch": 0.02,
"learning_rate": 4.9928092198506866e-05,
"loss": 0.885,
"step": 57
},
{
"epoch": 0.02,
"learning_rate": 4.992550242056393e-05,
"loss": 0.7342,
"step": 58
},
{
"epoch": 0.03,
"learning_rate": 4.992286689778207e-05,
"loss": 0.8593,
"step": 59
},
{
"epoch": 0.03,
"learning_rate": 4.992018563499814e-05,
"loss": 0.9458,
"step": 60
},
{
"epoch": 0.03,
"learning_rate": 4.991745863713299e-05,
"loss": 0.9511,
"step": 61
},
{
"epoch": 0.03,
"learning_rate": 4.991468590919135e-05,
"loss": 0.8605,
"step": 62
},
{
"epoch": 0.03,
"learning_rate": 4.9911867456261916e-05,
"loss": 0.76,
"step": 63
},
{
"epoch": 0.03,
"learning_rate": 4.9909003283517283e-05,
"loss": 0.8037,
"step": 64
},
{
"epoch": 0.03,
"learning_rate": 4.990609339621397e-05,
"loss": 0.8945,
"step": 65
},
{
"epoch": 0.03,
"learning_rate": 4.990313779969237e-05,
"loss": 0.7774,
"step": 66
},
{
"epoch": 0.03,
"learning_rate": 4.990013649937679e-05,
"loss": 0.8737,
"step": 67
},
{
"epoch": 0.03,
"learning_rate": 4.989708950077541e-05,
"loss": 0.8914,
"step": 68
},
{
"epoch": 0.03,
"learning_rate": 4.9893996809480246e-05,
"loss": 0.8682,
"step": 69
},
{
"epoch": 0.03,
"learning_rate": 4.9890858431167226e-05,
"loss": 0.7745,
"step": 70
},
{
"epoch": 0.03,
"learning_rate": 4.988767437159608e-05,
"loss": 1.0306,
"step": 71
},
{
"epoch": 0.03,
"learning_rate": 4.988444463661041e-05,
"loss": 0.904,
"step": 72
},
{
"epoch": 0.03,
"learning_rate": 4.98811692321376e-05,
"loss": 0.8478,
"step": 73
},
{
"epoch": 0.03,
"learning_rate": 4.9877848164188904e-05,
"loss": 0.7117,
"step": 74
},
{
"epoch": 0.03,
"learning_rate": 4.987448143885933e-05,
"loss": 0.6789,
"step": 75
},
{
"epoch": 0.03,
"learning_rate": 4.987106906232773e-05,
"loss": 0.9385,
"step": 76
},
{
"epoch": 0.03,
"learning_rate": 4.9867611040856695e-05,
"loss": 1.049,
"step": 77
},
{
"epoch": 0.03,
"learning_rate": 4.9864107380792594e-05,
"loss": 0.8914,
"step": 78
},
{
"epoch": 0.03,
"learning_rate": 4.986055808856558e-05,
"loss": 0.8211,
"step": 79
},
{
"epoch": 0.03,
"learning_rate": 4.9856963170689525e-05,
"loss": 0.9007,
"step": 80
},
{
"epoch": 0.03,
"learning_rate": 4.9853322633762045e-05,
"loss": 0.832,
"step": 81
},
{
"epoch": 0.04,
"learning_rate": 4.984963648446449e-05,
"loss": 0.807,
"step": 82
},
{
"epoch": 0.04,
"learning_rate": 4.984590472956191e-05,
"loss": 0.7031,
"step": 83
},
{
"epoch": 0.04,
"learning_rate": 4.984212737590304e-05,
"loss": 0.9261,
"step": 84
},
{
"epoch": 0.04,
"learning_rate": 4.983830443042033e-05,
"loss": 0.7759,
"step": 85
},
{
"epoch": 0.04,
"learning_rate": 4.983443590012989e-05,
"loss": 0.7147,
"step": 86
},
{
"epoch": 0.04,
"learning_rate": 4.9830521792131484e-05,
"loss": 0.8554,
"step": 87
},
{
"epoch": 0.04,
"learning_rate": 4.982656211360852e-05,
"loss": 0.8052,
"step": 88
},
{
"epoch": 0.04,
"learning_rate": 4.982255687182806e-05,
"loss": 0.8408,
"step": 89
},
{
"epoch": 0.04,
"learning_rate": 4.981850607414077e-05,
"loss": 0.7086,
"step": 90
},
{
"epoch": 0.04,
"learning_rate": 4.981440972798092e-05,
"loss": 0.7808,
"step": 91
},
{
"epoch": 0.04,
"learning_rate": 4.981026784086639e-05,
"loss": 0.8632,
"step": 92
},
{
"epoch": 0.04,
"learning_rate": 4.980608042039862e-05,
"loss": 0.8798,
"step": 93
},
{
"epoch": 0.04,
"learning_rate": 4.980184747426263e-05,
"loss": 0.9253,
"step": 94
},
{
"epoch": 0.04,
"learning_rate": 4.9797569010226995e-05,
"loss": 0.8305,
"step": 95
},
{
"epoch": 0.04,
"learning_rate": 4.979324503614381e-05,
"loss": 0.9474,
"step": 96
},
{
"epoch": 0.04,
"learning_rate": 4.97888755599487e-05,
"loss": 0.8102,
"step": 97
},
{
"epoch": 0.04,
"learning_rate": 4.978446058966081e-05,
"loss": 0.6445,
"step": 98
},
{
"epoch": 0.04,
"learning_rate": 4.978000013338277e-05,
"loss": 0.8123,
"step": 99
},
{
"epoch": 0.04,
"learning_rate": 4.977549419930067e-05,
"loss": 0.9368,
"step": 100
},
{
"epoch": 0.04,
"eval_loss": 0.8984958529472351,
"eval_runtime": 10.1685,
"eval_samples_per_second": 3.639,
"eval_steps_per_second": 0.492,
"step": 100
},
{
"epoch": 0.04,
"learning_rate": 4.97709427956841e-05,
"loss": 0.7673,
"step": 101
},
{
"epoch": 0.04,
"learning_rate": 4.976634593088608e-05,
"loss": 0.8571,
"step": 102
},
{
"epoch": 0.04,
"learning_rate": 4.9761703613343045e-05,
"loss": 0.9388,
"step": 103
},
{
"epoch": 0.04,
"learning_rate": 4.975701585157489e-05,
"loss": 0.7665,
"step": 104
},
{
"epoch": 0.05,
"learning_rate": 4.9752282654184876e-05,
"loss": 0.9379,
"step": 105
},
{
"epoch": 0.05,
"learning_rate": 4.974750402985966e-05,
"loss": 0.9475,
"step": 106
},
{
"epoch": 0.05,
"learning_rate": 4.974267998736928e-05,
"loss": 0.9757,
"step": 107
},
{
"epoch": 0.05,
"learning_rate": 4.973781053556711e-05,
"loss": 0.6865,
"step": 108
},
{
"epoch": 0.05,
"learning_rate": 4.973289568338989e-05,
"loss": 0.9761,
"step": 109
},
{
"epoch": 0.05,
"learning_rate": 4.9727935439857634e-05,
"loss": 0.9132,
"step": 110
},
{
"epoch": 0.05,
"learning_rate": 4.972292981407372e-05,
"loss": 0.774,
"step": 111
},
{
"epoch": 0.05,
"learning_rate": 4.971787881522477e-05,
"loss": 1.0217,
"step": 112
},
{
"epoch": 0.05,
"learning_rate": 4.9712782452580695e-05,
"loss": 0.9341,
"step": 113
},
{
"epoch": 0.05,
"learning_rate": 4.970764073549467e-05,
"loss": 0.7135,
"step": 114
},
{
"epoch": 0.05,
"learning_rate": 4.970245367340307e-05,
"loss": 0.7177,
"step": 115
},
{
"epoch": 0.05,
"learning_rate": 4.969722127582553e-05,
"loss": 0.9312,
"step": 116
},
{
"epoch": 0.05,
"learning_rate": 4.9691943552364885e-05,
"loss": 0.8677,
"step": 117
},
{
"epoch": 0.05,
"learning_rate": 4.968662051270711e-05,
"loss": 0.7371,
"step": 118
},
{
"epoch": 0.05,
"learning_rate": 4.968125216662142e-05,
"loss": 0.7248,
"step": 119
},
{
"epoch": 0.05,
"learning_rate": 4.9675838523960114e-05,
"loss": 0.7281,
"step": 120
},
{
"epoch": 0.05,
"learning_rate": 4.967037959465865e-05,
"loss": 0.8678,
"step": 121
},
{
"epoch": 0.05,
"learning_rate": 4.96648753887356e-05,
"loss": 0.8477,
"step": 122
},
{
"epoch": 0.05,
"learning_rate": 4.965932591629262e-05,
"loss": 0.7801,
"step": 123
},
{
"epoch": 0.05,
"learning_rate": 4.965373118751446e-05,
"loss": 0.8289,
"step": 124
},
{
"epoch": 0.05,
"learning_rate": 4.9648091212668904e-05,
"loss": 0.7771,
"step": 125
},
{
"epoch": 0.05,
"learning_rate": 4.964240600210679e-05,
"loss": 0.7933,
"step": 126
},
{
"epoch": 0.05,
"learning_rate": 4.963667556626198e-05,
"loss": 0.8119,
"step": 127
},
{
"epoch": 0.06,
"learning_rate": 4.963089991565131e-05,
"loss": 0.9761,
"step": 128
},
{
"epoch": 0.06,
"learning_rate": 4.9625079060874634e-05,
"loss": 0.7456,
"step": 129
},
{
"epoch": 0.06,
"learning_rate": 4.961921301261474e-05,
"loss": 0.7955,
"step": 130
},
{
"epoch": 0.06,
"learning_rate": 4.961330178163736e-05,
"loss": 0.8379,
"step": 131
},
{
"epoch": 0.06,
"learning_rate": 4.960734537879117e-05,
"loss": 0.793,
"step": 132
},
{
"epoch": 0.06,
"learning_rate": 4.960134381500771e-05,
"loss": 0.8465,
"step": 133
},
{
"epoch": 0.06,
"learning_rate": 4.959529710130144e-05,
"loss": 0.8813,
"step": 134
},
{
"epoch": 0.06,
"learning_rate": 4.958920524876967e-05,
"loss": 0.8148,
"step": 135
},
{
"epoch": 0.06,
"learning_rate": 4.9583068268592544e-05,
"loss": 0.8524,
"step": 136
},
{
"epoch": 0.06,
"learning_rate": 4.957688617203302e-05,
"loss": 0.6524,
"step": 137
},
{
"epoch": 0.06,
"learning_rate": 4.957065897043689e-05,
"loss": 0.8021,
"step": 138
},
{
"epoch": 0.06,
"learning_rate": 4.9564386675232674e-05,
"loss": 0.9442,
"step": 139
},
{
"epoch": 0.06,
"learning_rate": 4.955806929793171e-05,
"loss": 0.8701,
"step": 140
},
{
"epoch": 0.06,
"learning_rate": 4.9551706850128024e-05,
"loss": 0.9975,
"step": 141
},
{
"epoch": 0.06,
"learning_rate": 4.954529934349839e-05,
"loss": 0.7725,
"step": 142
},
{
"epoch": 0.06,
"learning_rate": 4.953884678980225e-05,
"loss": 0.7544,
"step": 143
},
{
"epoch": 0.06,
"learning_rate": 4.9532349200881745e-05,
"loss": 0.6881,
"step": 144
},
{
"epoch": 0.06,
"learning_rate": 4.9525806588661656e-05,
"loss": 0.8211,
"step": 145
},
{
"epoch": 0.06,
"learning_rate": 4.951921896514938e-05,
"loss": 0.7736,
"step": 146
},
{
"epoch": 0.06,
"learning_rate": 4.9512586342434956e-05,
"loss": 0.9176,
"step": 147
},
{
"epoch": 0.06,
"learning_rate": 4.950590873269098e-05,
"loss": 0.7639,
"step": 148
},
{
"epoch": 0.06,
"learning_rate": 4.94991861481726e-05,
"loss": 0.8447,
"step": 149
},
{
"epoch": 0.06,
"learning_rate": 4.949241860121755e-05,
"loss": 0.7537,
"step": 150
},
{
"epoch": 0.07,
"learning_rate": 4.948560610424604e-05,
"loss": 0.7948,
"step": 151
},
{
"epoch": 0.07,
"learning_rate": 4.94787486697608e-05,
"loss": 0.7734,
"step": 152
},
{
"epoch": 0.07,
"learning_rate": 4.9471846310347006e-05,
"loss": 0.8796,
"step": 153
},
{
"epoch": 0.07,
"learning_rate": 4.946489903867231e-05,
"loss": 0.7196,
"step": 154
},
{
"epoch": 0.07,
"learning_rate": 4.945790686748678e-05,
"loss": 0.7676,
"step": 155
},
{
"epoch": 0.07,
"learning_rate": 4.945086980962288e-05,
"loss": 0.7697,
"step": 156
},
{
"epoch": 0.07,
"learning_rate": 4.9443787877995464e-05,
"loss": 0.8223,
"step": 157
},
{
"epoch": 0.07,
"learning_rate": 4.943666108560173e-05,
"loss": 0.8416,
"step": 158
},
{
"epoch": 0.07,
"learning_rate": 4.942948944552122e-05,
"loss": 0.8536,
"step": 159
},
{
"epoch": 0.07,
"learning_rate": 4.942227297091577e-05,
"loss": 1.0185,
"step": 160
},
{
"epoch": 0.07,
"learning_rate": 4.94150116750295e-05,
"loss": 0.8932,
"step": 161
},
{
"epoch": 0.07,
"learning_rate": 4.94077055711888e-05,
"loss": 0.8254,
"step": 162
},
{
"epoch": 0.07,
"learning_rate": 4.940035467280229e-05,
"loss": 0.7546,
"step": 163
},
{
"epoch": 0.07,
"learning_rate": 4.9392958993360794e-05,
"loss": 0.8746,
"step": 164
},
{
"epoch": 0.07,
"learning_rate": 4.9385518546437325e-05,
"loss": 0.9356,
"step": 165
},
{
"epoch": 0.07,
"learning_rate": 4.937803334568706e-05,
"loss": 0.8459,
"step": 166
},
{
"epoch": 0.07,
"learning_rate": 4.9370503404847304e-05,
"loss": 0.7772,
"step": 167
},
{
"epoch": 0.07,
"learning_rate": 4.9362928737737465e-05,
"loss": 0.9711,
"step": 168
},
{
"epoch": 0.07,
"learning_rate": 4.935530935825907e-05,
"loss": 0.849,
"step": 169
},
{
"epoch": 0.07,
"learning_rate": 4.934764528039566e-05,
"loss": 0.759,
"step": 170
},
{
"epoch": 0.07,
"learning_rate": 4.933993651821282e-05,
"loss": 0.7088,
"step": 171
},
{
"epoch": 0.07,
"learning_rate": 4.933218308585819e-05,
"loss": 0.6487,
"step": 172
},
{
"epoch": 0.07,
"learning_rate": 4.9324384997561315e-05,
"loss": 0.7925,
"step": 173
},
{
"epoch": 0.07,
"learning_rate": 4.931654226763375e-05,
"loss": 0.922,
"step": 174
},
{
"epoch": 0.08,
"learning_rate": 4.9308654910468964e-05,
"loss": 0.8621,
"step": 175
},
{
"epoch": 0.08,
"learning_rate": 4.930072294054233e-05,
"loss": 0.8429,
"step": 176
},
{
"epoch": 0.08,
"learning_rate": 4.929274637241108e-05,
"loss": 0.7388,
"step": 177
},
{
"epoch": 0.08,
"learning_rate": 4.9284725220714324e-05,
"loss": 0.8557,
"step": 178
},
{
"epoch": 0.08,
"learning_rate": 4.927665950017297e-05,
"loss": 0.8447,
"step": 179
},
{
"epoch": 0.08,
"learning_rate": 4.926854922558975e-05,
"loss": 0.8104,
"step": 180
},
{
"epoch": 0.08,
"learning_rate": 4.9260394411849134e-05,
"loss": 0.9483,
"step": 181
},
{
"epoch": 0.08,
"learning_rate": 4.9252195073917335e-05,
"loss": 0.7172,
"step": 182
},
{
"epoch": 0.08,
"learning_rate": 4.9243951226842305e-05,
"loss": 0.7127,
"step": 183
},
{
"epoch": 0.08,
"learning_rate": 4.9235662885753666e-05,
"loss": 0.7191,
"step": 184
},
{
"epoch": 0.08,
"learning_rate": 4.92273300658627e-05,
"loss": 0.7975,
"step": 185
},
{
"epoch": 0.08,
"learning_rate": 4.921895278246231e-05,
"loss": 0.7981,
"step": 186
},
{
"epoch": 0.08,
"learning_rate": 4.921053105092701e-05,
"loss": 0.7595,
"step": 187
},
{
"epoch": 0.08,
"learning_rate": 4.9202064886712906e-05,
"loss": 0.7376,
"step": 188
},
{
"epoch": 0.08,
"learning_rate": 4.919355430535761e-05,
"loss": 0.8999,
"step": 189
},
{
"epoch": 0.08,
"learning_rate": 4.9184999322480285e-05,
"loss": 0.8945,
"step": 190
},
{
"epoch": 0.08,
"learning_rate": 4.917639995378156e-05,
"loss": 0.7347,
"step": 191
},
{
"epoch": 0.08,
"learning_rate": 4.916775621504354e-05,
"loss": 0.8662,
"step": 192
},
{
"epoch": 0.08,
"learning_rate": 4.915906812212976e-05,
"loss": 0.6963,
"step": 193
},
{
"epoch": 0.08,
"learning_rate": 4.9150335690985136e-05,
"loss": 0.7317,
"step": 194
},
{
"epoch": 0.08,
"learning_rate": 4.9141558937635985e-05,
"loss": 0.8358,
"step": 195
},
{
"epoch": 0.08,
"learning_rate": 4.913273787818995e-05,
"loss": 0.7879,
"step": 196
},
{
"epoch": 0.08,
"learning_rate": 4.912387252883598e-05,
"loss": 0.8523,
"step": 197
},
{
"epoch": 0.09,
"learning_rate": 4.911496290584433e-05,
"loss": 0.8389,
"step": 198
},
{
"epoch": 0.09,
"learning_rate": 4.910600902556649e-05,
"loss": 0.7412,
"step": 199
},
{
"epoch": 0.09,
"learning_rate": 4.909701090443518e-05,
"loss": 0.8253,
"step": 200
},
{
"epoch": 0.09,
"eval_loss": 0.8050532341003418,
"eval_runtime": 10.2691,
"eval_samples_per_second": 3.603,
"eval_steps_per_second": 0.487,
"step": 200
},
{
"epoch": 0.09,
"learning_rate": 4.908796855896432e-05,
"loss": 0.7399,
"step": 201
},
{
"epoch": 0.09,
"learning_rate": 4.907888200574898e-05,
"loss": 0.9344,
"step": 202
},
{
"epoch": 0.09,
"learning_rate": 4.906975126146538e-05,
"loss": 0.8562,
"step": 203
},
{
"epoch": 0.09,
"learning_rate": 4.906057634287082e-05,
"loss": 0.7759,
"step": 204
},
{
"epoch": 0.09,
"learning_rate": 4.905135726680369e-05,
"loss": 0.7058,
"step": 205
},
{
"epoch": 0.09,
"learning_rate": 4.9042094050183415e-05,
"loss": 0.9686,
"step": 206
},
{
"epoch": 0.09,
"learning_rate": 4.9032786710010424e-05,
"loss": 0.9007,
"step": 207
},
{
"epoch": 0.09,
"learning_rate": 4.902343526336614e-05,
"loss": 0.8106,
"step": 208
},
{
"epoch": 0.09,
"learning_rate": 4.901403972741291e-05,
"loss": 0.8317,
"step": 209
},
{
"epoch": 0.09,
"learning_rate": 4.900460011939402e-05,
"loss": 0.9889,
"step": 210
},
{
"epoch": 0.09,
"learning_rate": 4.899511645663361e-05,
"loss": 0.7061,
"step": 211
},
{
"epoch": 0.09,
"learning_rate": 4.898558875653671e-05,
"loss": 0.9084,
"step": 212
},
{
"epoch": 0.09,
"learning_rate": 4.897601703658914e-05,
"loss": 0.8296,
"step": 213
},
{
"epoch": 0.09,
"learning_rate": 4.896640131435751e-05,
"loss": 0.7512,
"step": 214
},
{
"epoch": 0.09,
"learning_rate": 4.8956741607489215e-05,
"loss": 0.7808,
"step": 215
},
{
"epoch": 0.09,
"learning_rate": 4.8947037933712335e-05,
"loss": 0.8168,
"step": 216
},
{
"epoch": 0.09,
"learning_rate": 4.893729031083567e-05,
"loss": 0.6904,
"step": 217
},
{
"epoch": 0.09,
"learning_rate": 4.8927498756748664e-05,
"loss": 0.8154,
"step": 218
},
{
"epoch": 0.09,
"learning_rate": 4.891766328942138e-05,
"loss": 0.732,
"step": 219
},
{
"epoch": 0.09,
"learning_rate": 4.8907783926904495e-05,
"loss": 0.833,
"step": 220
},
{
"epoch": 0.1,
"learning_rate": 4.889786068732921e-05,
"loss": 0.7308,
"step": 221
},
{
"epoch": 0.1,
"learning_rate": 4.8887893588907304e-05,
"loss": 0.8428,
"step": 222
},
{
"epoch": 0.1,
"learning_rate": 4.8877882649930995e-05,
"loss": 0.735,
"step": 223
},
{
"epoch": 0.1,
"learning_rate": 4.8867827888773e-05,
"loss": 0.6815,
"step": 224
},
{
"epoch": 0.1,
"learning_rate": 4.885772932388642e-05,
"loss": 0.8329,
"step": 225
},
{
"epoch": 0.1,
"learning_rate": 4.884758697380481e-05,
"loss": 0.7601,
"step": 226
},
{
"epoch": 0.1,
"learning_rate": 4.8837400857142015e-05,
"loss": 0.7283,
"step": 227
},
{
"epoch": 0.1,
"learning_rate": 4.882717099259224e-05,
"loss": 0.7835,
"step": 228
},
{
"epoch": 0.1,
"learning_rate": 4.8816897398929975e-05,
"loss": 0.7538,
"step": 229
},
{
"epoch": 0.1,
"learning_rate": 4.880658009500995e-05,
"loss": 0.6243,
"step": 230
},
{
"epoch": 0.1,
"learning_rate": 4.8796219099767136e-05,
"loss": 0.7844,
"step": 231
},
{
"epoch": 0.1,
"learning_rate": 4.878581443221668e-05,
"loss": 0.8362,
"step": 232
},
{
"epoch": 0.1,
"learning_rate": 4.877536611145388e-05,
"loss": 0.7069,
"step": 233
},
{
"epoch": 0.1,
"learning_rate": 4.876487415665413e-05,
"loss": 0.7603,
"step": 234
},
{
"epoch": 0.1,
"learning_rate": 4.875433858707293e-05,
"loss": 0.8936,
"step": 235
},
{
"epoch": 0.1,
"learning_rate": 4.874375942204583e-05,
"loss": 0.7754,
"step": 236
},
{
"epoch": 0.1,
"learning_rate": 4.8733136680988355e-05,
"loss": 0.7515,
"step": 237
},
{
"epoch": 0.1,
"learning_rate": 4.8722470383396024e-05,
"loss": 0.7706,
"step": 238
},
{
"epoch": 0.1,
"learning_rate": 4.871176054884431e-05,
"loss": 0.7938,
"step": 239
},
{
"epoch": 0.1,
"learning_rate": 4.870100719698854e-05,
"loss": 0.7918,
"step": 240
},
{
"epoch": 0.1,
"learning_rate": 4.8690210347563975e-05,
"loss": 0.9421,
"step": 241
},
{
"epoch": 0.1,
"learning_rate": 4.867937002038564e-05,
"loss": 0.7249,
"step": 242
},
{
"epoch": 0.1,
"learning_rate": 4.866848623534839e-05,
"loss": 0.7552,
"step": 243
},
{
"epoch": 0.11,
"learning_rate": 4.865755901242682e-05,
"loss": 0.7415,
"step": 244
},
{
"epoch": 0.11,
"learning_rate": 4.864658837167526e-05,
"loss": 0.7475,
"step": 245
},
{
"epoch": 0.11,
"learning_rate": 4.863557433322771e-05,
"loss": 0.8791,
"step": 246
},
{
"epoch": 0.11,
"learning_rate": 4.862451691729783e-05,
"loss": 0.8837,
"step": 247
},
{
"epoch": 0.11,
"learning_rate": 4.861341614417887e-05,
"loss": 0.7218,
"step": 248
},
{
"epoch": 0.11,
"learning_rate": 4.860227203424367e-05,
"loss": 0.8089,
"step": 249
},
{
"epoch": 0.11,
"learning_rate": 4.8591084607944595e-05,
"loss": 0.9958,
"step": 250
},
{
"epoch": 0.11,
"learning_rate": 4.85798538858135e-05,
"loss": 0.861,
"step": 251
},
{
"epoch": 0.11,
"learning_rate": 4.856857988846172e-05,
"loss": 0.783,
"step": 252
},
{
"epoch": 0.11,
"learning_rate": 4.855726263657999e-05,
"loss": 0.8853,
"step": 253
},
{
"epoch": 0.11,
"learning_rate": 4.8545902150938436e-05,
"loss": 0.7947,
"step": 254
},
{
"epoch": 0.11,
"learning_rate": 4.8534498452386543e-05,
"loss": 0.7206,
"step": 255
},
{
"epoch": 0.11,
"learning_rate": 4.8523051561853084e-05,
"loss": 0.8859,
"step": 256
},
{
"epoch": 0.11,
"learning_rate": 4.851156150034611e-05,
"loss": 0.745,
"step": 257
},
{
"epoch": 0.11,
"learning_rate": 4.850002828895289e-05,
"loss": 0.9045,
"step": 258
},
{
"epoch": 0.11,
"learning_rate": 4.8488451948839905e-05,
"loss": 0.6362,
"step": 259
},
{
"epoch": 0.11,
"learning_rate": 4.847683250125277e-05,
"loss": 0.8622,
"step": 260
},
{
"epoch": 0.11,
"learning_rate": 4.8465169967516235e-05,
"loss": 0.7037,
"step": 261
},
{
"epoch": 0.11,
"learning_rate": 4.845346436903408e-05,
"loss": 0.9182,
"step": 262
},
{
"epoch": 0.11,
"learning_rate": 4.844171572728919e-05,
"loss": 1.0684,
"step": 263
},
{
"epoch": 0.11,
"learning_rate": 4.842992406384338e-05,
"loss": 0.7909,
"step": 264
},
{
"epoch": 0.11,
"learning_rate": 4.8418089400337444e-05,
"loss": 0.8534,
"step": 265
},
{
"epoch": 0.11,
"learning_rate": 4.840621175849112e-05,
"loss": 0.8092,
"step": 266
},
{
"epoch": 0.12,
"learning_rate": 4.839429116010297e-05,
"loss": 0.7109,
"step": 267
},
{
"epoch": 0.12,
"learning_rate": 4.838232762705044e-05,
"loss": 0.5593,
"step": 268
},
{
"epoch": 0.12,
"learning_rate": 4.8370321181289754e-05,
"loss": 0.6169,
"step": 269
},
{
"epoch": 0.12,
"learning_rate": 4.835827184485587e-05,
"loss": 0.7919,
"step": 270
},
{
"epoch": 0.12,
"learning_rate": 4.834617963986251e-05,
"loss": 0.8294,
"step": 271
},
{
"epoch": 0.12,
"learning_rate": 4.833404458850203e-05,
"loss": 0.9067,
"step": 272
},
{
"epoch": 0.12,
"learning_rate": 4.832186671304543e-05,
"loss": 0.8037,
"step": 273
},
{
"epoch": 0.12,
"learning_rate": 4.8309646035842316e-05,
"loss": 0.7649,
"step": 274
},
{
"epoch": 0.12,
"learning_rate": 4.829738257932084e-05,
"loss": 0.736,
"step": 275
},
{
"epoch": 0.12,
"learning_rate": 4.8285076365987646e-05,
"loss": 0.6717,
"step": 276
},
{
"epoch": 0.12,
"learning_rate": 4.827272741842789e-05,
"loss": 0.8744,
"step": 277
},
{
"epoch": 0.12,
"learning_rate": 4.826033575930512e-05,
"loss": 0.8303,
"step": 278
},
{
"epoch": 0.12,
"learning_rate": 4.824790141136127e-05,
"loss": 0.8506,
"step": 279
},
{
"epoch": 0.12,
"learning_rate": 4.823542439741666e-05,
"loss": 0.7496,
"step": 280
},
{
"epoch": 0.12,
"learning_rate": 4.822290474036987e-05,
"loss": 0.8215,
"step": 281
},
{
"epoch": 0.12,
"learning_rate": 4.821034246319774e-05,
"loss": 0.8849,
"step": 282
},
{
"epoch": 0.12,
"learning_rate": 4.819773758895538e-05,
"loss": 0.7652,
"step": 283
},
{
"epoch": 0.12,
"learning_rate": 4.818509014077602e-05,
"loss": 0.7457,
"step": 284
},
{
"epoch": 0.12,
"learning_rate": 4.817240014187105e-05,
"loss": 0.7701,
"step": 285
},
{
"epoch": 0.12,
"learning_rate": 4.8159667615529954e-05,
"loss": 0.7975,
"step": 286
},
{
"epoch": 0.12,
"learning_rate": 4.814689258512025e-05,
"loss": 0.7421,
"step": 287
},
{
"epoch": 0.12,
"learning_rate": 4.813407507408748e-05,
"loss": 0.771,
"step": 288
},
{
"epoch": 0.12,
"learning_rate": 4.8121215105955145e-05,
"loss": 0.7896,
"step": 289
},
{
"epoch": 0.12,
"learning_rate": 4.8108312704324654e-05,
"loss": 0.6527,
"step": 290
},
{
"epoch": 0.13,
"learning_rate": 4.8095367892875295e-05,
"loss": 0.7029,
"step": 291
},
{
"epoch": 0.13,
"learning_rate": 4.80823806953642e-05,
"loss": 0.7148,
"step": 292
},
{
"epoch": 0.13,
"learning_rate": 4.80693511356263e-05,
"loss": 0.7192,
"step": 293
},
{
"epoch": 0.13,
"learning_rate": 4.8056279237574234e-05,
"loss": 0.8897,
"step": 294
},
{
"epoch": 0.13,
"learning_rate": 4.8043165025198375e-05,
"loss": 0.9168,
"step": 295
},
{
"epoch": 0.13,
"learning_rate": 4.803000852256675e-05,
"loss": 0.6315,
"step": 296
},
{
"epoch": 0.13,
"learning_rate": 4.801680975382499e-05,
"loss": 0.8135,
"step": 297
},
{
"epoch": 0.13,
"learning_rate": 4.800356874319629e-05,
"loss": 0.8978,
"step": 298
},
{
"epoch": 0.13,
"learning_rate": 4.79902855149814e-05,
"loss": 0.8826,
"step": 299
},
{
"epoch": 0.13,
"learning_rate": 4.797696009355851e-05,
"loss": 0.8169,
"step": 300
},
{
"epoch": 0.13,
"eval_loss": 0.795865535736084,
"eval_runtime": 10.328,
"eval_samples_per_second": 3.582,
"eval_steps_per_second": 0.484,
"step": 300
},
{
"epoch": 0.13,
"learning_rate": 4.796359250338328e-05,
"loss": 0.7584,
"step": 301
},
{
"epoch": 0.13,
"learning_rate": 4.795018276898874e-05,
"loss": 0.7721,
"step": 302
},
{
"epoch": 0.13,
"learning_rate": 4.793673091498527e-05,
"loss": 0.7665,
"step": 303
},
{
"epoch": 0.13,
"learning_rate": 4.7923236966060566e-05,
"loss": 0.8482,
"step": 304
},
{
"epoch": 0.13,
"learning_rate": 4.7909700946979555e-05,
"loss": 0.6749,
"step": 305
},
{
"epoch": 0.13,
"learning_rate": 4.7896122882584374e-05,
"loss": 0.5408,
"step": 306
},
{
"epoch": 0.13,
"learning_rate": 4.7882502797794346e-05,
"loss": 0.8174,
"step": 307
},
{
"epoch": 0.13,
"learning_rate": 4.7868840717605914e-05,
"loss": 0.8283,
"step": 308
},
{
"epoch": 0.13,
"learning_rate": 4.785513666709256e-05,
"loss": 0.7986,
"step": 309
},
{
"epoch": 0.13,
"learning_rate": 4.784139067140483e-05,
"loss": 0.7339,
"step": 310
},
{
"epoch": 0.13,
"learning_rate": 4.7827602755770224e-05,
"loss": 0.6881,
"step": 311
},
{
"epoch": 0.13,
"learning_rate": 4.7813772945493185e-05,
"loss": 0.7879,
"step": 312
},
{
"epoch": 0.13,
"learning_rate": 4.779990126595505e-05,
"loss": 0.9105,
"step": 313
},
{
"epoch": 0.14,
"learning_rate": 4.778598774261399e-05,
"loss": 0.6961,
"step": 314
},
{
"epoch": 0.14,
"learning_rate": 4.777203240100497e-05,
"loss": 0.7347,
"step": 315
},
{
"epoch": 0.14,
"learning_rate": 4.775803526673971e-05,
"loss": 0.8124,
"step": 316
},
{
"epoch": 0.14,
"learning_rate": 4.7743996365506614e-05,
"loss": 0.8755,
"step": 317
},
{
"epoch": 0.14,
"learning_rate": 4.772991572307076e-05,
"loss": 0.6714,
"step": 318
},
{
"epoch": 0.14,
"learning_rate": 4.771579336527383e-05,
"loss": 0.7593,
"step": 319
},
{
"epoch": 0.14,
"learning_rate": 4.7701629318034034e-05,
"loss": 0.7754,
"step": 320
},
{
"epoch": 0.14,
"learning_rate": 4.768742360734614e-05,
"loss": 0.8489,
"step": 321
},
{
"epoch": 0.14,
"learning_rate": 4.7673176259281336e-05,
"loss": 0.7423,
"step": 322
},
{
"epoch": 0.14,
"learning_rate": 4.7658887299987265e-05,
"loss": 0.9569,
"step": 323
},
{
"epoch": 0.14,
"learning_rate": 4.7644556755687905e-05,
"loss": 0.6688,
"step": 324
},
{
"epoch": 0.14,
"learning_rate": 4.763018465268356e-05,
"loss": 0.688,
"step": 325
},
{
"epoch": 0.14,
"learning_rate": 4.7615771017350826e-05,
"loss": 0.8766,
"step": 326
},
{
"epoch": 0.14,
"learning_rate": 4.76013158761425e-05,
"loss": 0.815,
"step": 327
},
{
"epoch": 0.14,
"learning_rate": 4.758681925558756e-05,
"loss": 0.6949,
"step": 328
},
{
"epoch": 0.14,
"learning_rate": 4.75722811822911e-05,
"loss": 0.6397,
"step": 329
},
{
"epoch": 0.14,
"learning_rate": 4.755770168293432e-05,
"loss": 0.8287,
"step": 330
},
{
"epoch": 0.14,
"learning_rate": 4.7543080784274414e-05,
"loss": 0.831,
"step": 331
},
{
"epoch": 0.14,
"learning_rate": 4.752841851314457e-05,
"loss": 0.8885,
"step": 332
},
{
"epoch": 0.14,
"learning_rate": 4.751371489645391e-05,
"loss": 0.7274,
"step": 333
},
{
"epoch": 0.14,
"learning_rate": 4.7498969961187424e-05,
"loss": 0.6892,
"step": 334
},
{
"epoch": 0.14,
"learning_rate": 4.748418373440594e-05,
"loss": 0.7592,
"step": 335
},
{
"epoch": 0.14,
"learning_rate": 4.7469356243246066e-05,
"loss": 0.7485,
"step": 336
},
{
"epoch": 0.15,
"learning_rate": 4.745448751492014e-05,
"loss": 0.7914,
"step": 337
},
{
"epoch": 0.15,
"learning_rate": 4.7439577576716175e-05,
"loss": 0.9029,
"step": 338
},
{
"epoch": 0.15,
"learning_rate": 4.742462645599783e-05,
"loss": 0.6633,
"step": 339
},
{
"epoch": 0.15,
"learning_rate": 4.740963418020433e-05,
"loss": 0.7987,
"step": 340
},
{
"epoch": 0.15,
"learning_rate": 4.739460077685043e-05,
"loss": 0.9075,
"step": 341
},
{
"epoch": 0.15,
"learning_rate": 4.737952627352639e-05,
"loss": 0.7466,
"step": 342
},
{
"epoch": 0.15,
"learning_rate": 4.736441069789786e-05,
"loss": 0.6737,
"step": 343
},
{
"epoch": 0.15,
"learning_rate": 4.734925407770589e-05,
"loss": 0.7996,
"step": 344
},
{
"epoch": 0.15,
"learning_rate": 4.733405644076688e-05,
"loss": 0.8677,
"step": 345
},
{
"epoch": 0.15,
"learning_rate": 4.731881781497246e-05,
"loss": 0.8364,
"step": 346
},
{
"epoch": 0.15,
"learning_rate": 4.7303538228289504e-05,
"loss": 0.7339,
"step": 347
},
{
"epoch": 0.15,
"learning_rate": 4.728821770876008e-05,
"loss": 0.7552,
"step": 348
},
{
"epoch": 0.15,
"learning_rate": 4.727285628450136e-05,
"loss": 0.6216,
"step": 349
},
{
"epoch": 0.15,
"learning_rate": 4.725745398370558e-05,
"loss": 0.7541,
"step": 350
},
{
"epoch": 0.15,
"learning_rate": 4.724201083464e-05,
"loss": 0.7895,
"step": 351
},
{
"epoch": 0.15,
"learning_rate": 4.722652686564687e-05,
"loss": 0.8574,
"step": 352
},
{
"epoch": 0.15,
"learning_rate": 4.721100210514333e-05,
"loss": 0.7342,
"step": 353
},
{
"epoch": 0.15,
"learning_rate": 4.719543658162139e-05,
"loss": 0.676,
"step": 354
},
{
"epoch": 0.15,
"learning_rate": 4.717983032364787e-05,
"loss": 0.7245,
"step": 355
},
{
"epoch": 0.15,
"learning_rate": 4.716418335986434e-05,
"loss": 0.7016,
"step": 356
},
{
"epoch": 0.15,
"learning_rate": 4.7148495718987096e-05,
"loss": 0.8397,
"step": 357
},
{
"epoch": 0.15,
"learning_rate": 4.7132767429807076e-05,
"loss": 0.9152,
"step": 358
},
{
"epoch": 0.15,
"learning_rate": 4.711699852118981e-05,
"loss": 0.7737,
"step": 359
},
{
"epoch": 0.16,
"learning_rate": 4.710118902207541e-05,
"loss": 0.7912,
"step": 360
},
{
"epoch": 0.16,
"learning_rate": 4.708533896147842e-05,
"loss": 0.711,
"step": 361
},
{
"epoch": 0.16,
"learning_rate": 4.706944836848789e-05,
"loss": 0.9113,
"step": 362
},
{
"epoch": 0.16,
"learning_rate": 4.7053517272267214e-05,
"loss": 0.736,
"step": 363
},
{
"epoch": 0.16,
"learning_rate": 4.703754570205413e-05,
"loss": 0.8455,
"step": 364
},
{
"epoch": 0.16,
"learning_rate": 4.7021533687160676e-05,
"loss": 0.9134,
"step": 365
},
{
"epoch": 0.16,
"learning_rate": 4.7005481256973097e-05,
"loss": 0.9046,
"step": 366
},
{
"epoch": 0.16,
"learning_rate": 4.698938844095181e-05,
"loss": 0.9463,
"step": 367
},
{
"epoch": 0.16,
"learning_rate": 4.6973255268631366e-05,
"loss": 0.7204,
"step": 368
},
{
"epoch": 0.16,
"learning_rate": 4.6957081769620356e-05,
"loss": 0.7333,
"step": 369
},
{
"epoch": 0.16,
"learning_rate": 4.694086797360141e-05,
"loss": 0.8651,
"step": 370
},
{
"epoch": 0.16,
"learning_rate": 4.692461391033109e-05,
"loss": 0.8249,
"step": 371
},
{
"epoch": 0.16,
"learning_rate": 4.690831960963988e-05,
"loss": 0.6987,
"step": 372
},
{
"epoch": 0.16,
"learning_rate": 4.6891985101432084e-05,
"loss": 0.6567,
"step": 373
},
{
"epoch": 0.16,
"learning_rate": 4.687561041568582e-05,
"loss": 0.7595,
"step": 374
},
{
"epoch": 0.16,
"learning_rate": 4.685919558245293e-05,
"loss": 0.8601,
"step": 375
},
{
"epoch": 0.16,
"learning_rate": 4.684274063185894e-05,
"loss": 0.6274,
"step": 376
},
{
"epoch": 0.16,
"learning_rate": 4.682624559410301e-05,
"loss": 0.6681,
"step": 377
},
{
"epoch": 0.16,
"learning_rate": 4.680971049945786e-05,
"loss": 0.7932,
"step": 378
},
{
"epoch": 0.16,
"learning_rate": 4.679313537826973e-05,
"loss": 0.7183,
"step": 379
},
{
"epoch": 0.16,
"learning_rate": 4.677652026095831e-05,
"loss": 0.6897,
"step": 380
},
{
"epoch": 0.16,
"learning_rate": 4.675986517801672e-05,
"loss": 0.7922,
"step": 381
},
{
"epoch": 0.16,
"learning_rate": 4.674317016001138e-05,
"loss": 0.8051,
"step": 382
},
{
"epoch": 0.17,
"learning_rate": 4.672643523758206e-05,
"loss": 0.6802,
"step": 383
},
{
"epoch": 0.17,
"learning_rate": 4.6709660441441705e-05,
"loss": 0.7451,
"step": 384
},
{
"epoch": 0.17,
"learning_rate": 4.6692845802376494e-05,
"loss": 0.8396,
"step": 385
},
{
"epoch": 0.17,
"learning_rate": 4.66759913512457e-05,
"loss": 0.9847,
"step": 386
},
{
"epoch": 0.17,
"learning_rate": 4.665909711898164e-05,
"loss": 0.8155,
"step": 387
},
{
"epoch": 0.17,
"learning_rate": 4.664216313658969e-05,
"loss": 0.7643,
"step": 388
},
{
"epoch": 0.17,
"learning_rate": 4.6625189435148146e-05,
"loss": 0.7872,
"step": 389
},
{
"epoch": 0.17,
"learning_rate": 4.6608176045808206e-05,
"loss": 0.7877,
"step": 390
},
{
"epoch": 0.17,
"learning_rate": 4.659112299979389e-05,
"loss": 0.5715,
"step": 391
},
{
"epoch": 0.17,
"learning_rate": 4.657403032840203e-05,
"loss": 0.7811,
"step": 392
},
{
"epoch": 0.17,
"learning_rate": 4.6556898063002164e-05,
"loss": 0.5627,
"step": 393
},
{
"epoch": 0.17,
"learning_rate": 4.653972623503649e-05,
"loss": 0.6729,
"step": 394
},
{
"epoch": 0.17,
"learning_rate": 4.6522514876019816e-05,
"loss": 0.6761,
"step": 395
},
{
"epoch": 0.17,
"learning_rate": 4.6505264017539516e-05,
"loss": 0.7337,
"step": 396
},
{
"epoch": 0.17,
"learning_rate": 4.648797369125543e-05,
"loss": 0.7986,
"step": 397
},
{
"epoch": 0.17,
"learning_rate": 4.647064392889985e-05,
"loss": 0.7348,
"step": 398
},
{
"epoch": 0.17,
"learning_rate": 4.6453274762277435e-05,
"loss": 0.8767,
"step": 399
},
{
"epoch": 0.17,
"learning_rate": 4.643586622326516e-05,
"loss": 0.7493,
"step": 400
},
{
"epoch": 0.17,
"eval_loss": 0.7891781330108643,
"eval_runtime": 10.3744,
"eval_samples_per_second": 3.566,
"eval_steps_per_second": 0.482,
"step": 400
},
{
"epoch": 0.17,
"learning_rate": 4.641841834381228e-05,
"loss": 0.9408,
"step": 401
},
{
"epoch": 0.17,
"learning_rate": 4.640093115594022e-05,
"loss": 0.7631,
"step": 402
},
{
"epoch": 0.17,
"learning_rate": 4.638340469174256e-05,
"loss": 0.8535,
"step": 403
},
{
"epoch": 0.17,
"learning_rate": 4.636583898338497e-05,
"loss": 0.7969,
"step": 404
},
{
"epoch": 0.17,
"learning_rate": 4.6348234063105124e-05,
"loss": 0.6849,
"step": 405
},
{
"epoch": 0.17,
"learning_rate": 4.633058996321268e-05,
"loss": 0.698,
"step": 406
},
{
"epoch": 0.18,
"learning_rate": 4.631290671608919e-05,
"loss": 0.7716,
"step": 407
},
{
"epoch": 0.18,
"learning_rate": 4.629518435418805e-05,
"loss": 0.7092,
"step": 408
},
{
"epoch": 0.18,
"learning_rate": 4.627742291003445e-05,
"loss": 0.6969,
"step": 409
},
{
"epoch": 0.18,
"learning_rate": 4.62596224162253e-05,
"loss": 0.6474,
"step": 410
},
{
"epoch": 0.18,
"learning_rate": 4.624178290542917e-05,
"loss": 0.6569,
"step": 411
},
{
"epoch": 0.18,
"learning_rate": 4.622390441038627e-05,
"loss": 0.8783,
"step": 412
},
{
"epoch": 0.18,
"learning_rate": 4.620598696390831e-05,
"loss": 0.7998,
"step": 413
},
{
"epoch": 0.18,
"learning_rate": 4.61880305988785e-05,
"loss": 0.7375,
"step": 414
},
{
"epoch": 0.18,
"learning_rate": 4.61700353482515e-05,
"loss": 0.7961,
"step": 415
},
{
"epoch": 0.18,
"learning_rate": 4.615200124505331e-05,
"loss": 0.7244,
"step": 416
},
{
"epoch": 0.18,
"learning_rate": 4.613392832238124e-05,
"loss": 0.8139,
"step": 417
},
{
"epoch": 0.18,
"learning_rate": 4.611581661340386e-05,
"loss": 0.7845,
"step": 418
},
{
"epoch": 0.18,
"learning_rate": 4.609766615136089e-05,
"loss": 0.7492,
"step": 419
},
{
"epoch": 0.18,
"learning_rate": 4.607947696956321e-05,
"loss": 0.7373,
"step": 420
},
{
"epoch": 0.18,
"learning_rate": 4.606124910139272e-05,
"loss": 0.7485,
"step": 421
},
{
"epoch": 0.18,
"learning_rate": 4.604298258030237e-05,
"loss": 0.568,
"step": 422
},
{
"epoch": 0.18,
"learning_rate": 4.6024677439816e-05,
"loss": 0.8351,
"step": 423
},
{
"epoch": 0.18,
"learning_rate": 4.600633371352836e-05,
"loss": 0.6451,
"step": 424
},
{
"epoch": 0.18,
"learning_rate": 4.598795143510499e-05,
"loss": 0.7567,
"step": 425
},
{
"epoch": 0.18,
"learning_rate": 4.59695306382822e-05,
"loss": 0.728,
"step": 426
},
{
"epoch": 0.18,
"learning_rate": 4.595107135686699e-05,
"loss": 0.8118,
"step": 427
},
{
"epoch": 0.18,
"learning_rate": 4.5932573624737e-05,
"loss": 0.7179,
"step": 428
},
{
"epoch": 0.18,
"learning_rate": 4.5914037475840385e-05,
"loss": 1.0522,
"step": 429
},
{
"epoch": 0.19,
"learning_rate": 4.589546294419589e-05,
"loss": 0.8786,
"step": 430
},
{
"epoch": 0.19,
"learning_rate": 4.587685006389262e-05,
"loss": 0.6876,
"step": 431
},
{
"epoch": 0.19,
"learning_rate": 4.58581988690901e-05,
"loss": 0.6063,
"step": 432
},
{
"epoch": 0.19,
"learning_rate": 4.583950939401819e-05,
"loss": 0.7466,
"step": 433
},
{
"epoch": 0.19,
"learning_rate": 4.5820781672976955e-05,
"loss": 0.7848,
"step": 434
},
{
"epoch": 0.19,
"learning_rate": 4.58020157403367e-05,
"loss": 0.837,
"step": 435
},
{
"epoch": 0.19,
"learning_rate": 4.578321163053783e-05,
"loss": 0.7999,
"step": 436
},
{
"epoch": 0.19,
"learning_rate": 4.576436937809081e-05,
"loss": 0.7283,
"step": 437
},
{
"epoch": 0.19,
"learning_rate": 4.5745489017576124e-05,
"loss": 0.8953,
"step": 438
},
{
"epoch": 0.19,
"learning_rate": 4.5726570583644205e-05,
"loss": 0.7776,
"step": 439
},
{
"epoch": 0.19,
"learning_rate": 4.570761411101533e-05,
"loss": 0.7233,
"step": 440
},
{
"epoch": 0.19,
"learning_rate": 4.5688619634479604e-05,
"loss": 0.8226,
"step": 441
},
{
"epoch": 0.19,
"learning_rate": 4.566958718889688e-05,
"loss": 0.6558,
"step": 442
},
{
"epoch": 0.19,
"learning_rate": 4.565051680919669e-05,
"loss": 0.7776,
"step": 443
},
{
"epoch": 0.19,
"learning_rate": 4.563140853037817e-05,
"loss": 0.6725,
"step": 444
},
{
"epoch": 0.19,
"learning_rate": 4.561226238751005e-05,
"loss": 0.7942,
"step": 445
},
{
"epoch": 0.19,
"learning_rate": 4.5593078415730515e-05,
"loss": 0.7764,
"step": 446
},
{
"epoch": 0.19,
"learning_rate": 4.55738566502472e-05,
"loss": 0.9007,
"step": 447
},
{
"epoch": 0.19,
"learning_rate": 4.5554597126337076e-05,
"loss": 0.7221,
"step": 448
},
{
"epoch": 0.19,
"learning_rate": 4.553529987934643e-05,
"loss": 0.809,
"step": 449
},
{
"epoch": 0.19,
"learning_rate": 4.551596494469077e-05,
"loss": 0.7276,
"step": 450
},
{
"epoch": 0.19,
"learning_rate": 4.549659235785478e-05,
"loss": 0.7519,
"step": 451
},
{
"epoch": 0.19,
"learning_rate": 4.547718215439224e-05,
"loss": 0.7447,
"step": 452
},
{
"epoch": 0.2,
"learning_rate": 4.545773436992598e-05,
"loss": 0.6342,
"step": 453
},
{
"epoch": 0.2,
"learning_rate": 4.543824904014777e-05,
"loss": 0.6875,
"step": 454
},
{
"epoch": 0.2,
"learning_rate": 4.5418726200818316e-05,
"loss": 0.7269,
"step": 455
},
{
"epoch": 0.2,
"learning_rate": 4.539916588776715e-05,
"loss": 0.7012,
"step": 456
},
{
"epoch": 0.2,
"learning_rate": 4.5379568136892567e-05,
"loss": 0.8766,
"step": 457
},
{
"epoch": 0.2,
"learning_rate": 4.5359932984161605e-05,
"loss": 0.7997,
"step": 458
},
{
"epoch": 0.2,
"learning_rate": 4.534026046560992e-05,
"loss": 0.709,
"step": 459
},
{
"epoch": 0.2,
"learning_rate": 4.5320550617341726e-05,
"loss": 0.7729,
"step": 460
},
{
"epoch": 0.2,
"learning_rate": 4.530080347552979e-05,
"loss": 0.7965,
"step": 461
},
{
"epoch": 0.2,
"learning_rate": 4.528101907641529e-05,
"loss": 0.8593,
"step": 462
},
{
"epoch": 0.2,
"learning_rate": 4.5261197456307794e-05,
"loss": 0.7322,
"step": 463
},
{
"epoch": 0.2,
"learning_rate": 4.524133865158518e-05,
"loss": 0.7928,
"step": 464
},
{
"epoch": 0.2,
"learning_rate": 4.522144269869355e-05,
"loss": 0.7176,
"step": 465
},
{
"epoch": 0.2,
"learning_rate": 4.5201509634147224e-05,
"loss": 0.7317,
"step": 466
},
{
"epoch": 0.2,
"learning_rate": 4.518153949452858e-05,
"loss": 0.8163,
"step": 467
},
{
"epoch": 0.2,
"learning_rate": 4.51615323164881e-05,
"loss": 0.7572,
"step": 468
},
{
"epoch": 0.2,
"learning_rate": 4.5141488136744174e-05,
"loss": 0.7409,
"step": 469
},
{
"epoch": 0.2,
"learning_rate": 4.512140699208315e-05,
"loss": 0.8279,
"step": 470
},
{
"epoch": 0.2,
"learning_rate": 4.51012889193592e-05,
"loss": 0.6798,
"step": 471
},
{
"epoch": 0.2,
"learning_rate": 4.508113395549426e-05,
"loss": 0.6618,
"step": 472
},
{
"epoch": 0.2,
"learning_rate": 4.506094213747798e-05,
"loss": 0.8773,
"step": 473
},
{
"epoch": 0.2,
"learning_rate": 4.504071350236766e-05,
"loss": 0.8261,
"step": 474
},
{
"epoch": 0.2,
"learning_rate": 4.5020448087288126e-05,
"loss": 0.6215,
"step": 475
},
{
"epoch": 0.21,
"learning_rate": 4.5000145929431756e-05,
"loss": 0.8396,
"step": 476
},
{
"epoch": 0.21,
"learning_rate": 4.497980706605833e-05,
"loss": 0.7632,
"step": 477
},
{
"epoch": 0.21,
"learning_rate": 4.495943153449501e-05,
"loss": 0.7404,
"step": 478
},
{
"epoch": 0.21,
"learning_rate": 4.4939019372136224e-05,
"loss": 0.8391,
"step": 479
},
{
"epoch": 0.21,
"learning_rate": 4.491857061644367e-05,
"loss": 0.7786,
"step": 480
},
{
"epoch": 0.21,
"learning_rate": 4.489808530494617e-05,
"loss": 0.7974,
"step": 481
},
{
"epoch": 0.21,
"learning_rate": 4.487756347523965e-05,
"loss": 0.8695,
"step": 482
},
{
"epoch": 0.21,
"learning_rate": 4.4857005164987044e-05,
"loss": 0.6575,
"step": 483
},
{
"epoch": 0.21,
"learning_rate": 4.483641041191826e-05,
"loss": 0.8302,
"step": 484
},
{
"epoch": 0.21,
"learning_rate": 4.481577925383007e-05,
"loss": 0.8316,
"step": 485
},
{
"epoch": 0.21,
"learning_rate": 4.479511172858607e-05,
"loss": 0.8911,
"step": 486
},
{
"epoch": 0.21,
"learning_rate": 4.4774407874116584e-05,
"loss": 0.7963,
"step": 487
},
{
"epoch": 0.21,
"learning_rate": 4.475366772841862e-05,
"loss": 0.7823,
"step": 488
},
{
"epoch": 0.21,
"learning_rate": 4.47328913295558e-05,
"loss": 0.8895,
"step": 489
},
{
"epoch": 0.21,
"learning_rate": 4.4712078715658254e-05,
"loss": 0.7182,
"step": 490
},
{
"epoch": 0.21,
"learning_rate": 4.46912299249226e-05,
"loss": 0.8522,
"step": 491
},
{
"epoch": 0.21,
"learning_rate": 4.467034499561185e-05,
"loss": 0.7836,
"step": 492
},
{
"epoch": 0.21,
"learning_rate": 4.46494239660553e-05,
"loss": 0.6369,
"step": 493
},
{
"epoch": 0.21,
"learning_rate": 4.462846687464857e-05,
"loss": 0.7994,
"step": 494
},
{
"epoch": 0.21,
"learning_rate": 4.4607473759853384e-05,
"loss": 0.7834,
"step": 495
},
{
"epoch": 0.21,
"learning_rate": 4.458644466019764e-05,
"loss": 0.741,
"step": 496
},
{
"epoch": 0.21,
"learning_rate": 4.456537961427526e-05,
"loss": 0.7608,
"step": 497
},
{
"epoch": 0.21,
"learning_rate": 4.454427866074612e-05,
"loss": 0.8057,
"step": 498
},
{
"epoch": 0.22,
"learning_rate": 4.452314183833601e-05,
"loss": 0.8335,
"step": 499
},
{
"epoch": 0.22,
"learning_rate": 4.450196918583655e-05,
"loss": 0.7366,
"step": 500
},
{
"epoch": 0.22,
"eval_loss": 0.7860936522483826,
"eval_runtime": 10.1703,
"eval_samples_per_second": 3.638,
"eval_steps_per_second": 0.492,
"step": 500
},
{
"epoch": 0.22,
"learning_rate": 4.448076074210511e-05,
"loss": 0.703,
"step": 501
},
{
"epoch": 0.22,
"learning_rate": 4.445951654606476e-05,
"loss": 0.8081,
"step": 502
},
{
"epoch": 0.22,
"learning_rate": 4.443823663670416e-05,
"loss": 0.5891,
"step": 503
},
{
"epoch": 0.22,
"learning_rate": 4.441692105307754e-05,
"loss": 0.7695,
"step": 504
},
{
"epoch": 0.22,
"learning_rate": 4.43955698343046e-05,
"loss": 0.7281,
"step": 505
},
{
"epoch": 0.22,
"learning_rate": 4.4374183019570404e-05,
"loss": 0.8165,
"step": 506
},
{
"epoch": 0.22,
"learning_rate": 4.435276064812539e-05,
"loss": 0.9928,
"step": 507
},
{
"epoch": 0.22,
"learning_rate": 4.4331302759285225e-05,
"loss": 0.698,
"step": 508
},
{
"epoch": 0.22,
"learning_rate": 4.4309809392430776e-05,
"loss": 0.6933,
"step": 509
},
{
"epoch": 0.22,
"learning_rate": 4.4288280587007994e-05,
"loss": 0.7442,
"step": 510
},
{
"epoch": 0.22,
"learning_rate": 4.4266716382527904e-05,
"loss": 0.9552,
"step": 511
},
{
"epoch": 0.22,
"learning_rate": 4.424511681856648e-05,
"loss": 0.7013,
"step": 512
},
{
"epoch": 0.22,
"learning_rate": 4.422348193476458e-05,
"loss": 0.648,
"step": 513
},
{
"epoch": 0.22,
"learning_rate": 4.4201811770827905e-05,
"loss": 0.6661,
"step": 514
},
{
"epoch": 0.22,
"learning_rate": 4.418010636652689e-05,
"loss": 0.8251,
"step": 515
},
{
"epoch": 0.22,
"learning_rate": 4.415836576169664e-05,
"loss": 0.6331,
"step": 516
},
{
"epoch": 0.22,
"learning_rate": 4.413658999623689e-05,
"loss": 0.79,
"step": 517
},
{
"epoch": 0.22,
"learning_rate": 4.4114779110111866e-05,
"loss": 0.7234,
"step": 518
},
{
"epoch": 0.22,
"learning_rate": 4.409293314335028e-05,
"loss": 0.8827,
"step": 519
},
{
"epoch": 0.22,
"learning_rate": 4.40710521360452e-05,
"loss": 0.7483,
"step": 520
},
{
"epoch": 0.22,
"learning_rate": 4.404913612835404e-05,
"loss": 0.8601,
"step": 521
},
{
"epoch": 0.22,
"learning_rate": 4.402718516049842e-05,
"loss": 0.7445,
"step": 522
},
{
"epoch": 0.23,
"learning_rate": 4.4005199272764105e-05,
"loss": 0.8318,
"step": 523
},
{
"epoch": 0.23,
"learning_rate": 4.3983178505501e-05,
"loss": 0.8368,
"step": 524
},
{
"epoch": 0.23,
"learning_rate": 4.3961122899122975e-05,
"loss": 0.7101,
"step": 525
},
{
"epoch": 0.23,
"learning_rate": 4.3939032494107855e-05,
"loss": 0.7236,
"step": 526
},
{
"epoch": 0.23,
"learning_rate": 4.391690733099735e-05,
"loss": 0.7242,
"step": 527
},
{
"epoch": 0.23,
"learning_rate": 4.389474745039691e-05,
"loss": 0.7646,
"step": 528
},
{
"epoch": 0.23,
"learning_rate": 4.387255289297576e-05,
"loss": 0.7416,
"step": 529
},
{
"epoch": 0.23,
"learning_rate": 4.385032369946673e-05,
"loss": 0.8072,
"step": 530
},
{
"epoch": 0.23,
"learning_rate": 4.382805991066622e-05,
"loss": 0.7637,
"step": 531
},
{
"epoch": 0.23,
"learning_rate": 4.380576156743412e-05,
"loss": 0.8242,
"step": 532
},
{
"epoch": 0.23,
"learning_rate": 4.378342871069376e-05,
"loss": 0.7322,
"step": 533
},
{
"epoch": 0.23,
"learning_rate": 4.376106138143178e-05,
"loss": 0.7279,
"step": 534
},
{
"epoch": 0.23,
"learning_rate": 4.373865962069811e-05,
"loss": 0.8358,
"step": 535
},
{
"epoch": 0.23,
"learning_rate": 4.371622346960584e-05,
"loss": 0.8739,
"step": 536
},
{
"epoch": 0.23,
"learning_rate": 4.3693752969331224e-05,
"loss": 0.6559,
"step": 537
},
{
"epoch": 0.23,
"learning_rate": 4.367124816111351e-05,
"loss": 0.7295,
"step": 538
},
{
"epoch": 0.23,
"learning_rate": 4.364870908625493e-05,
"loss": 0.8219,
"step": 539
},
{
"epoch": 0.23,
"learning_rate": 4.36261357861206e-05,
"loss": 0.7993,
"step": 540
},
{
"epoch": 0.23,
"learning_rate": 4.360352830213846e-05,
"loss": 0.7536,
"step": 541
},
{
"epoch": 0.23,
"learning_rate": 4.358088667579917e-05,
"loss": 0.8467,
"step": 542
},
{
"epoch": 0.23,
"learning_rate": 4.355821094865606e-05,
"loss": 0.8671,
"step": 543
},
{
"epoch": 0.23,
"learning_rate": 4.3535501162325034e-05,
"loss": 0.7898,
"step": 544
},
{
"epoch": 0.23,
"learning_rate": 4.351275735848451e-05,
"loss": 0.8136,
"step": 545
},
{
"epoch": 0.24,
"learning_rate": 4.348997957887535e-05,
"loss": 0.8536,
"step": 546
},
{
"epoch": 0.24,
"learning_rate": 4.346716786530075e-05,
"loss": 0.9403,
"step": 547
},
{
"epoch": 0.24,
"learning_rate": 4.344432225962619e-05,
"loss": 0.7021,
"step": 548
},
{
"epoch": 0.24,
"learning_rate": 4.342144280377936e-05,
"loss": 0.7304,
"step": 549
},
{
"epoch": 0.24,
"learning_rate": 4.339852953975005e-05,
"loss": 0.7546,
"step": 550
},
{
"epoch": 0.24,
"learning_rate": 4.337558250959013e-05,
"loss": 0.7948,
"step": 551
},
{
"epoch": 0.24,
"learning_rate": 4.335260175541342e-05,
"loss": 0.8696,
"step": 552
},
{
"epoch": 0.24,
"learning_rate": 4.332958731939562e-05,
"loss": 0.8357,
"step": 553
},
{
"epoch": 0.24,
"learning_rate": 4.330653924377428e-05,
"loss": 0.7318,
"step": 554
},
{
"epoch": 0.24,
"learning_rate": 4.328345757084866e-05,
"loss": 0.7246,
"step": 555
},
{
"epoch": 0.24,
"learning_rate": 4.32603423429797e-05,
"loss": 0.7,
"step": 556
},
{
"epoch": 0.24,
"learning_rate": 4.3237193602589885e-05,
"loss": 0.7644,
"step": 557
},
{
"epoch": 0.24,
"learning_rate": 4.3214011392163244e-05,
"loss": 0.7991,
"step": 558
},
{
"epoch": 0.24,
"learning_rate": 4.319079575424523e-05,
"loss": 0.8297,
"step": 559
},
{
"epoch": 0.24,
"learning_rate": 4.3167546731442605e-05,
"loss": 0.6975,
"step": 560
},
{
"epoch": 0.24,
"learning_rate": 4.3144264366423445e-05,
"loss": 0.7131,
"step": 561
},
{
"epoch": 0.24,
"learning_rate": 4.3120948701917e-05,
"loss": 0.7741,
"step": 562
},
{
"epoch": 0.24,
"learning_rate": 4.3097599780713625e-05,
"loss": 0.8072,
"step": 563
},
{
"epoch": 0.24,
"learning_rate": 4.307421764566473e-05,
"loss": 0.7055,
"step": 564
},
{
"epoch": 0.24,
"learning_rate": 4.3050802339682665e-05,
"loss": 0.8643,
"step": 565
},
{
"epoch": 0.24,
"learning_rate": 4.302735390574066e-05,
"loss": 0.6969,
"step": 566
},
{
"epoch": 0.24,
"learning_rate": 4.3003872386872744e-05,
"loss": 0.7327,
"step": 567
},
{
"epoch": 0.24,
"learning_rate": 4.2980357826173665e-05,
"loss": 0.8235,
"step": 568
},
{
"epoch": 0.25,
"learning_rate": 4.295681026679883e-05,
"loss": 0.7095,
"step": 569
},
{
"epoch": 0.25,
"learning_rate": 4.2933229751964175e-05,
"loss": 0.7745,
"step": 570
},
{
"epoch": 0.25,
"learning_rate": 4.290961632494614e-05,
"loss": 0.7294,
"step": 571
},
{
"epoch": 0.25,
"learning_rate": 4.2885970029081555e-05,
"loss": 0.7303,
"step": 572
},
{
"epoch": 0.25,
"learning_rate": 4.2862290907767585e-05,
"loss": 0.7418,
"step": 573
},
{
"epoch": 0.25,
"learning_rate": 4.283857900446163e-05,
"loss": 0.8628,
"step": 574
},
{
"epoch": 0.25,
"learning_rate": 4.2814834362681255e-05,
"loss": 0.7715,
"step": 575
},
{
"epoch": 0.25,
"learning_rate": 4.279105702600412e-05,
"loss": 0.675,
"step": 576
},
{
"epoch": 0.25,
"learning_rate": 4.276724703806786e-05,
"loss": 0.7764,
"step": 577
},
{
"epoch": 0.25,
"learning_rate": 4.2743404442570065e-05,
"loss": 0.6354,
"step": 578
},
{
"epoch": 0.25,
"learning_rate": 4.2719529283268156e-05,
"loss": 0.8398,
"step": 579
},
{
"epoch": 0.25,
"learning_rate": 4.269562160397931e-05,
"loss": 0.836,
"step": 580
},
{
"epoch": 0.25,
"learning_rate": 4.2671681448580404e-05,
"loss": 0.6738,
"step": 581
},
{
"epoch": 0.25,
"learning_rate": 4.264770886100789e-05,
"loss": 0.7566,
"step": 582
},
{
"epoch": 0.25,
"learning_rate": 4.2623703885257774e-05,
"loss": 0.8301,
"step": 583
},
{
"epoch": 0.25,
"learning_rate": 4.259966656538548e-05,
"loss": 0.7379,
"step": 584
},
{
"epoch": 0.25,
"learning_rate": 4.2575596945505794e-05,
"loss": 0.6814,
"step": 585
},
{
"epoch": 0.25,
"learning_rate": 4.255149506979279e-05,
"loss": 0.7451,
"step": 586
},
{
"epoch": 0.25,
"learning_rate": 4.2527360982479735e-05,
"loss": 0.6818,
"step": 587
},
{
"epoch": 0.25,
"learning_rate": 4.250319472785902e-05,
"loss": 0.8787,
"step": 588
},
{
"epoch": 0.25,
"learning_rate": 4.247899635028206e-05,
"loss": 0.7882,
"step": 589
},
{
"epoch": 0.25,
"learning_rate": 4.245476589415923e-05,
"loss": 0.7852,
"step": 590
},
{
"epoch": 0.25,
"learning_rate": 4.243050340395977e-05,
"loss": 0.8433,
"step": 591
},
{
"epoch": 0.26,
"learning_rate": 4.240620892421172e-05,
"loss": 0.8116,
"step": 592
},
{
"epoch": 0.26,
"learning_rate": 4.238188249950185e-05,
"loss": 0.7848,
"step": 593
},
{
"epoch": 0.26,
"learning_rate": 4.235752417447549e-05,
"loss": 0.8715,
"step": 594
},
{
"epoch": 0.26,
"learning_rate": 4.233313399383659e-05,
"loss": 0.6482,
"step": 595
},
{
"epoch": 0.26,
"learning_rate": 4.230871200234754e-05,
"loss": 0.6538,
"step": 596
},
{
"epoch": 0.26,
"learning_rate": 4.2284258244829076e-05,
"loss": 0.7536,
"step": 597
},
{
"epoch": 0.26,
"learning_rate": 4.2259772766160276e-05,
"loss": 0.7742,
"step": 598
},
{
"epoch": 0.26,
"learning_rate": 4.2235255611278425e-05,
"loss": 0.7807,
"step": 599
},
{
"epoch": 0.26,
"learning_rate": 4.221070682517894e-05,
"loss": 0.7651,
"step": 600
},
{
"epoch": 0.26,
"eval_loss": 0.7837547063827515,
"eval_runtime": 10.1825,
"eval_samples_per_second": 3.634,
"eval_steps_per_second": 0.491,
"step": 600
},
{
"epoch": 0.26,
"learning_rate": 4.2186126452915256e-05,
"loss": 0.8788,
"step": 601
},
{
"epoch": 0.26,
"learning_rate": 4.216151453959885e-05,
"loss": 0.7653,
"step": 602
},
{
"epoch": 0.26,
"learning_rate": 4.2136871130399015e-05,
"loss": 0.8086,
"step": 603
},
{
"epoch": 0.26,
"learning_rate": 4.211219627054288e-05,
"loss": 0.6593,
"step": 604
},
{
"epoch": 0.26,
"learning_rate": 4.2087490005315293e-05,
"loss": 0.864,
"step": 605
},
{
"epoch": 0.26,
"learning_rate": 4.206275238005873e-05,
"loss": 0.7781,
"step": 606
},
{
"epoch": 0.26,
"learning_rate": 4.203798344017324e-05,
"loss": 0.8283,
"step": 607
},
{
"epoch": 0.26,
"learning_rate": 4.201318323111631e-05,
"loss": 0.8298,
"step": 608
},
{
"epoch": 0.26,
"learning_rate": 4.1988351798402846e-05,
"loss": 0.7659,
"step": 609
},
{
"epoch": 0.26,
"learning_rate": 4.1963489187605034e-05,
"loss": 0.8664,
"step": 610
},
{
"epoch": 0.26,
"learning_rate": 4.193859544435229e-05,
"loss": 0.7306,
"step": 611
},
{
"epoch": 0.26,
"learning_rate": 4.191367061433118e-05,
"loss": 0.7797,
"step": 612
},
{
"epoch": 0.26,
"learning_rate": 4.18887147432853e-05,
"loss": 0.6415,
"step": 613
},
{
"epoch": 0.26,
"learning_rate": 4.186372787701522e-05,
"loss": 0.8717,
"step": 614
},
{
"epoch": 0.27,
"learning_rate": 4.183871006137841e-05,
"loss": 0.7534,
"step": 615
},
{
"epoch": 0.27,
"learning_rate": 4.181366134228912e-05,
"loss": 0.7723,
"step": 616
},
{
"epoch": 0.27,
"learning_rate": 4.1788581765718316e-05,
"loss": 0.7216,
"step": 617
},
{
"epoch": 0.27,
"learning_rate": 4.176347137769362e-05,
"loss": 0.8546,
"step": 618
},
{
"epoch": 0.27,
"learning_rate": 4.1738330224299165e-05,
"loss": 0.7798,
"step": 619
},
{
"epoch": 0.27,
"learning_rate": 4.171315835167558e-05,
"loss": 0.7227,
"step": 620
},
{
"epoch": 0.27,
"learning_rate": 4.1687955806019864e-05,
"loss": 0.6509,
"step": 621
},
{
"epoch": 0.27,
"learning_rate": 4.166272263358528e-05,
"loss": 0.8394,
"step": 622
},
{
"epoch": 0.27,
"learning_rate": 4.163745888068134e-05,
"loss": 0.7652,
"step": 623
},
{
"epoch": 0.27,
"learning_rate": 4.161216459367366e-05,
"loss": 0.7998,
"step": 624
},
{
"epoch": 0.27,
"learning_rate": 4.158683981898389e-05,
"loss": 0.801,
"step": 625
},
{
"epoch": 0.27,
"learning_rate": 4.156148460308964e-05,
"loss": 0.6806,
"step": 626
},
{
"epoch": 0.27,
"learning_rate": 4.153609899252439e-05,
"loss": 0.7841,
"step": 627
},
{
"epoch": 0.27,
"learning_rate": 4.15106830338774e-05,
"loss": 0.7259,
"step": 628
},
{
"epoch": 0.27,
"learning_rate": 4.148523677379361e-05,
"loss": 0.6297,
"step": 629
},
{
"epoch": 0.27,
"learning_rate": 4.14597602589736e-05,
"loss": 0.8397,
"step": 630
},
{
"epoch": 0.27,
"learning_rate": 4.143425353617346e-05,
"loss": 0.7199,
"step": 631
},
{
"epoch": 0.27,
"learning_rate": 4.1408716652204716e-05,
"loss": 0.759,
"step": 632
},
{
"epoch": 0.27,
"learning_rate": 4.138314965393425e-05,
"loss": 0.6946,
"step": 633
},
{
"epoch": 0.27,
"learning_rate": 4.135755258828422e-05,
"loss": 0.7367,
"step": 634
},
{
"epoch": 0.27,
"learning_rate": 4.133192550223196e-05,
"loss": 0.7487,
"step": 635
},
{
"epoch": 0.27,
"learning_rate": 4.130626844280989e-05,
"loss": 0.7692,
"step": 636
},
{
"epoch": 0.27,
"learning_rate": 4.128058145710547e-05,
"loss": 0.6799,
"step": 637
},
{
"epoch": 0.27,
"learning_rate": 4.125486459226104e-05,
"loss": 0.7387,
"step": 638
},
{
"epoch": 0.28,
"learning_rate": 4.12291178954738e-05,
"loss": 0.8323,
"step": 639
},
{
"epoch": 0.28,
"learning_rate": 4.12033414139957e-05,
"loss": 0.7543,
"step": 640
},
{
"epoch": 0.28,
"learning_rate": 4.117753519513336e-05,
"loss": 0.6675,
"step": 641
},
{
"epoch": 0.28,
"learning_rate": 4.115169928624794e-05,
"loss": 0.7366,
"step": 642
},
{
"epoch": 0.28,
"learning_rate": 4.1125833734755114e-05,
"loss": 0.7067,
"step": 643
},
{
"epoch": 0.28,
"learning_rate": 4.109993858812499e-05,
"loss": 0.8237,
"step": 644
},
{
"epoch": 0.28,
"learning_rate": 4.107401389388193e-05,
"loss": 0.7359,
"step": 645
},
{
"epoch": 0.28,
"learning_rate": 4.104805969960456e-05,
"loss": 0.7633,
"step": 646
},
{
"epoch": 0.28,
"learning_rate": 4.1022076052925637e-05,
"loss": 0.7281,
"step": 647
},
{
"epoch": 0.28,
"learning_rate": 4.099606300153198e-05,
"loss": 0.7322,
"step": 648
},
{
"epoch": 0.28,
"learning_rate": 4.097002059316436e-05,
"loss": 0.905,
"step": 649
},
{
"epoch": 0.28,
"learning_rate": 4.0943948875617433e-05,
"loss": 0.6952,
"step": 650
},
{
"epoch": 0.28,
"learning_rate": 4.0917847896739644e-05,
"loss": 0.7417,
"step": 651
},
{
"epoch": 0.28,
"learning_rate": 4.089171770443313e-05,
"loss": 0.7173,
"step": 652
},
{
"epoch": 0.28,
"learning_rate": 4.0865558346653664e-05,
"loss": 0.8249,
"step": 653
},
{
"epoch": 0.28,
"learning_rate": 4.0839369871410524e-05,
"loss": 0.7726,
"step": 654
},
{
"epoch": 0.28,
"learning_rate": 4.081315232676646e-05,
"loss": 0.7356,
"step": 655
},
{
"epoch": 0.28,
"learning_rate": 4.0786905760837514e-05,
"loss": 0.7732,
"step": 656
},
{
"epoch": 0.28,
"learning_rate": 4.0760630221793036e-05,
"loss": 0.9094,
"step": 657
},
{
"epoch": 0.28,
"learning_rate": 4.073432575785554e-05,
"loss": 0.8432,
"step": 658
},
{
"epoch": 0.28,
"learning_rate": 4.0707992417300616e-05,
"loss": 0.7868,
"step": 659
},
{
"epoch": 0.28,
"learning_rate": 4.068163024845686e-05,
"loss": 0.8738,
"step": 660
},
{
"epoch": 0.28,
"learning_rate": 4.065523929970578e-05,
"loss": 0.7762,
"step": 661
},
{
"epoch": 0.29,
"learning_rate": 4.062881961948167e-05,
"loss": 0.9031,
"step": 662
},
{
"epoch": 0.29,
"learning_rate": 4.06023712562716e-05,
"loss": 0.8335,
"step": 663
},
{
"epoch": 0.29,
"learning_rate": 4.057589425861525e-05,
"loss": 0.7018,
"step": 664
},
{
"epoch": 0.29,
"learning_rate": 4.054938867510485e-05,
"loss": 0.785,
"step": 665
},
{
"epoch": 0.29,
"learning_rate": 4.0522854554385116e-05,
"loss": 0.7981,
"step": 666
},
{
"epoch": 0.29,
"learning_rate": 4.049629194515312e-05,
"loss": 0.8843,
"step": 667
},
{
"epoch": 0.29,
"learning_rate": 4.046970089615822e-05,
"loss": 0.737,
"step": 668
},
{
"epoch": 0.29,
"learning_rate": 4.044308145620197e-05,
"loss": 0.8855,
"step": 669
},
{
"epoch": 0.29,
"learning_rate": 4.0416433674138024e-05,
"loss": 0.7986,
"step": 670
},
{
"epoch": 0.29,
"learning_rate": 4.038975759887207e-05,
"loss": 0.8071,
"step": 671
},
{
"epoch": 0.29,
"learning_rate": 4.0363053279361696e-05,
"loss": 0.8288,
"step": 672
},
{
"epoch": 0.29,
"learning_rate": 4.033632076461634e-05,
"loss": 0.8119,
"step": 673
},
{
"epoch": 0.29,
"learning_rate": 4.0309560103697184e-05,
"loss": 0.7528,
"step": 674
},
{
"epoch": 0.29,
"learning_rate": 4.028277134571706e-05,
"loss": 0.7865,
"step": 675
},
{
"epoch": 0.29,
"learning_rate": 4.025595453984038e-05,
"loss": 0.7289,
"step": 676
},
{
"epoch": 0.29,
"learning_rate": 4.022910973528301e-05,
"loss": 0.8222,
"step": 677
},
{
"epoch": 0.29,
"learning_rate": 4.020223698131221e-05,
"loss": 0.8173,
"step": 678
},
{
"epoch": 0.29,
"learning_rate": 4.017533632724656e-05,
"loss": 0.7644,
"step": 679
},
{
"epoch": 0.29,
"learning_rate": 4.0148407822455794e-05,
"loss": 0.8409,
"step": 680
},
{
"epoch": 0.29,
"learning_rate": 4.012145151636079e-05,
"loss": 0.665,
"step": 681
},
{
"epoch": 0.29,
"learning_rate": 4.009446745843346e-05,
"loss": 0.8076,
"step": 682
},
{
"epoch": 0.29,
"learning_rate": 4.006745569819662e-05,
"loss": 0.7214,
"step": 683
},
{
"epoch": 0.29,
"learning_rate": 4.004041628522394e-05,
"loss": 0.7404,
"step": 684
},
{
"epoch": 0.3,
"learning_rate": 4.001334926913985e-05,
"loss": 0.6325,
"step": 685
},
{
"epoch": 0.3,
"learning_rate": 3.998625469961942e-05,
"loss": 0.7213,
"step": 686
},
{
"epoch": 0.3,
"learning_rate": 3.9959132626388294e-05,
"loss": 0.6753,
"step": 687
},
{
"epoch": 0.3,
"learning_rate": 3.9931983099222595e-05,
"loss": 0.8213,
"step": 688
},
{
"epoch": 0.3,
"learning_rate": 3.990480616794884e-05,
"loss": 0.6437,
"step": 689
},
{
"epoch": 0.3,
"learning_rate": 3.987760188244382e-05,
"loss": 0.8144,
"step": 690
},
{
"epoch": 0.3,
"learning_rate": 3.985037029263454e-05,
"loss": 0.6676,
"step": 691
},
{
"epoch": 0.3,
"learning_rate": 3.982311144849813e-05,
"loss": 0.7082,
"step": 692
},
{
"epoch": 0.3,
"learning_rate": 3.97958254000617e-05,
"loss": 0.7795,
"step": 693
},
{
"epoch": 0.3,
"learning_rate": 3.976851219740234e-05,
"loss": 0.88,
"step": 694
},
{
"epoch": 0.3,
"learning_rate": 3.974117189064691e-05,
"loss": 0.7574,
"step": 695
},
{
"epoch": 0.3,
"learning_rate": 3.971380452997208e-05,
"loss": 0.7663,
"step": 696
},
{
"epoch": 0.3,
"learning_rate": 3.9686410165604124e-05,
"loss": 0.6625,
"step": 697
},
{
"epoch": 0.3,
"learning_rate": 3.965898884781889e-05,
"loss": 0.8158,
"step": 698
},
{
"epoch": 0.3,
"learning_rate": 3.963154062694169e-05,
"loss": 0.7573,
"step": 699
},
{
"epoch": 0.3,
"learning_rate": 3.9604065553347224e-05,
"loss": 0.6229,
"step": 700
},
{
"epoch": 0.3,
"eval_loss": 0.7801283001899719,
"eval_runtime": 10.3197,
"eval_samples_per_second": 3.585,
"eval_steps_per_second": 0.485,
"step": 700
},
{
"epoch": 0.3,
"learning_rate": 3.957656367745947e-05,
"loss": 0.6979,
"step": 701
},
{
"epoch": 0.3,
"learning_rate": 3.954903504975156e-05,
"loss": 0.7313,
"step": 702
},
{
"epoch": 0.3,
"learning_rate": 3.952147972074578e-05,
"loss": 0.723,
"step": 703
},
{
"epoch": 0.3,
"learning_rate": 3.949389774101337e-05,
"loss": 0.7319,
"step": 704
},
{
"epoch": 0.3,
"learning_rate": 3.946628916117452e-05,
"loss": 0.651,
"step": 705
},
{
"epoch": 0.3,
"learning_rate": 3.9438654031898204e-05,
"loss": 1.0151,
"step": 706
},
{
"epoch": 0.3,
"learning_rate": 3.941099240390215e-05,
"loss": 0.6635,
"step": 707
},
{
"epoch": 0.31,
"learning_rate": 3.9383304327952696e-05,
"loss": 0.7306,
"step": 708
},
{
"epoch": 0.31,
"learning_rate": 3.935558985486474e-05,
"loss": 0.7589,
"step": 709
},
{
"epoch": 0.31,
"learning_rate": 3.9327849035501604e-05,
"loss": 0.7903,
"step": 710
},
{
"epoch": 0.31,
"learning_rate": 3.9300081920774966e-05,
"loss": 0.7878,
"step": 711
},
{
"epoch": 0.31,
"learning_rate": 3.9272288561644786e-05,
"loss": 0.8944,
"step": 712
},
{
"epoch": 0.31,
"learning_rate": 3.9244469009119165e-05,
"loss": 1.1365,
"step": 713
},
{
"epoch": 0.31,
"learning_rate": 3.921662331425429e-05,
"loss": 0.747,
"step": 714
},
{
"epoch": 0.31,
"learning_rate": 3.918875152815431e-05,
"loss": 0.8827,
"step": 715
},
{
"epoch": 0.31,
"learning_rate": 3.916085370197127e-05,
"loss": 0.7942,
"step": 716
},
{
"epoch": 0.31,
"learning_rate": 3.9132929886905e-05,
"loss": 0.8337,
"step": 717
},
{
"epoch": 0.31,
"learning_rate": 3.910498013420304e-05,
"loss": 0.7707,
"step": 718
},
{
"epoch": 0.31,
"learning_rate": 3.907700449516051e-05,
"loss": 0.8093,
"step": 719
},
{
"epoch": 0.31,
"learning_rate": 3.9049003021120055e-05,
"loss": 0.7472,
"step": 720
},
{
"epoch": 0.31,
"learning_rate": 3.9020975763471734e-05,
"loss": 0.9015,
"step": 721
},
{
"epoch": 0.31,
"learning_rate": 3.8992922773652915e-05,
"loss": 0.6248,
"step": 722
},
{
"epoch": 0.31,
"learning_rate": 3.89648441031482e-05,
"loss": 0.815,
"step": 723
},
{
"epoch": 0.31,
"learning_rate": 3.893673980348932e-05,
"loss": 0.6683,
"step": 724
},
{
"epoch": 0.31,
"learning_rate": 3.890860992625504e-05,
"loss": 0.7272,
"step": 725
},
{
"epoch": 0.31,
"learning_rate": 3.8880454523071064e-05,
"loss": 0.7075,
"step": 726
},
{
"epoch": 0.31,
"learning_rate": 3.885227364560995e-05,
"loss": 0.6867,
"step": 727
},
{
"epoch": 0.31,
"learning_rate": 3.882406734559101e-05,
"loss": 0.7838,
"step": 728
},
{
"epoch": 0.31,
"learning_rate": 3.879583567478019e-05,
"loss": 0.8641,
"step": 729
},
{
"epoch": 0.31,
"learning_rate": 3.876757868499003e-05,
"loss": 0.6906,
"step": 730
},
{
"epoch": 0.32,
"learning_rate": 3.873929642807953e-05,
"loss": 0.8569,
"step": 731
},
{
"epoch": 0.32,
"learning_rate": 3.871098895595404e-05,
"loss": 0.7502,
"step": 732
},
{
"epoch": 0.32,
"learning_rate": 3.8682656320565204e-05,
"loss": 0.8316,
"step": 733
},
{
"epoch": 0.32,
"learning_rate": 3.865429857391084e-05,
"loss": 0.7405,
"step": 734
},
{
"epoch": 0.32,
"learning_rate": 3.8625915768034873e-05,
"loss": 0.756,
"step": 735
},
{
"epoch": 0.32,
"learning_rate": 3.859750795502718e-05,
"loss": 0.7492,
"step": 736
},
{
"epoch": 0.32,
"learning_rate": 3.8569075187023575e-05,
"loss": 0.8525,
"step": 737
},
{
"epoch": 0.32,
"learning_rate": 3.8540617516205624e-05,
"loss": 0.9735,
"step": 738
},
{
"epoch": 0.32,
"learning_rate": 3.851213499480064e-05,
"loss": 0.7838,
"step": 739
},
{
"epoch": 0.32,
"learning_rate": 3.8483627675081524e-05,
"loss": 0.9609,
"step": 740
},
{
"epoch": 0.32,
"learning_rate": 3.8455095609366693e-05,
"loss": 0.8497,
"step": 741
},
{
"epoch": 0.32,
"learning_rate": 3.8426538850019964e-05,
"loss": 0.7637,
"step": 742
},
{
"epoch": 0.32,
"learning_rate": 3.83979574494505e-05,
"loss": 0.6037,
"step": 743
},
{
"epoch": 0.32,
"learning_rate": 3.836935146011266e-05,
"loss": 0.8975,
"step": 744
},
{
"epoch": 0.32,
"learning_rate": 3.834072093450594e-05,
"loss": 0.6858,
"step": 745
},
{
"epoch": 0.32,
"learning_rate": 3.831206592517488e-05,
"loss": 0.8235,
"step": 746
},
{
"epoch": 0.32,
"learning_rate": 3.828338648470895e-05,
"loss": 0.7814,
"step": 747
},
{
"epoch": 0.32,
"learning_rate": 3.825468266574243e-05,
"loss": 0.7573,
"step": 748
},
{
"epoch": 0.32,
"learning_rate": 3.8225954520954364e-05,
"loss": 0.755,
"step": 749
},
{
"epoch": 0.32,
"learning_rate": 3.819720210306845e-05,
"loss": 0.7047,
"step": 750
},
{
"epoch": 0.32,
"learning_rate": 3.816842546485292e-05,
"loss": 0.7636,
"step": 751
},
{
"epoch": 0.32,
"learning_rate": 3.813962465912045e-05,
"loss": 0.8064,
"step": 752
},
{
"epoch": 0.32,
"learning_rate": 3.811079973872809e-05,
"loss": 0.8318,
"step": 753
},
{
"epoch": 0.32,
"learning_rate": 3.808195075657711e-05,
"loss": 0.9713,
"step": 754
},
{
"epoch": 0.33,
"learning_rate": 3.805307776561299e-05,
"loss": 0.7643,
"step": 755
},
{
"epoch": 0.33,
"learning_rate": 3.802418081882524e-05,
"loss": 0.7808,
"step": 756
},
{
"epoch": 0.33,
"learning_rate": 3.799525996924733e-05,
"loss": 0.7491,
"step": 757
},
{
"epoch": 0.33,
"learning_rate": 3.796631526995663e-05,
"loss": 0.8423,
"step": 758
},
{
"epoch": 0.33,
"learning_rate": 3.793734677407423e-05,
"loss": 0.7888,
"step": 759
},
{
"epoch": 0.33,
"learning_rate": 3.790835453476495e-05,
"loss": 0.77,
"step": 760
},
{
"epoch": 0.33,
"learning_rate": 3.7879338605237146e-05,
"loss": 0.7197,
"step": 761
},
{
"epoch": 0.33,
"learning_rate": 3.7850299038742673e-05,
"loss": 0.7985,
"step": 762
},
{
"epoch": 0.33,
"learning_rate": 3.782123588857675e-05,
"loss": 0.8538,
"step": 763
},
{
"epoch": 0.33,
"learning_rate": 3.779214920807788e-05,
"loss": 0.7602,
"step": 764
},
{
"epoch": 0.33,
"learning_rate": 3.7763039050627767e-05,
"loss": 0.8949,
"step": 765
},
{
"epoch": 0.33,
"learning_rate": 3.773390546965119e-05,
"loss": 0.7133,
"step": 766
},
{
"epoch": 0.33,
"learning_rate": 3.7704748518615915e-05,
"loss": 0.854,
"step": 767
},
{
"epoch": 0.33,
"learning_rate": 3.767556825103259e-05,
"loss": 0.6997,
"step": 768
},
{
"epoch": 0.33,
"learning_rate": 3.7646364720454675e-05,
"loss": 0.9395,
"step": 769
},
{
"epoch": 0.33,
"learning_rate": 3.761713798047831e-05,
"loss": 0.6775,
"step": 770
},
{
"epoch": 0.33,
"learning_rate": 3.758788808474224e-05,
"loss": 0.8521,
"step": 771
},
{
"epoch": 0.33,
"learning_rate": 3.755861508692769e-05,
"loss": 0.8262,
"step": 772
},
{
"epoch": 0.33,
"learning_rate": 3.75293190407583e-05,
"loss": 0.7014,
"step": 773
},
{
"epoch": 0.33,
"learning_rate": 3.7500000000000003e-05,
"loss": 0.6883,
"step": 774
},
{
"epoch": 0.33,
"learning_rate": 3.747065801846093e-05,
"loss": 0.8857,
"step": 775
},
{
"epoch": 0.33,
"learning_rate": 3.744129314999133e-05,
"loss": 0.7464,
"step": 776
},
{
"epoch": 0.33,
"learning_rate": 3.741190544848343e-05,
"loss": 0.7916,
"step": 777
},
{
"epoch": 0.34,
"learning_rate": 3.7382494967871386e-05,
"loss": 0.7719,
"step": 778
},
{
"epoch": 0.34,
"learning_rate": 3.735306176213114e-05,
"loss": 0.6991,
"step": 779
},
{
"epoch": 0.34,
"learning_rate": 3.732360588528036e-05,
"loss": 0.6965,
"step": 780
},
{
"epoch": 0.34,
"learning_rate": 3.72941273913783e-05,
"loss": 0.8097,
"step": 781
},
{
"epoch": 0.34,
"learning_rate": 3.726462633452574e-05,
"loss": 0.7607,
"step": 782
},
{
"epoch": 0.34,
"learning_rate": 3.723510276886487e-05,
"loss": 0.7492,
"step": 783
},
{
"epoch": 0.34,
"learning_rate": 3.720555674857916e-05,
"loss": 0.7734,
"step": 784
},
{
"epoch": 0.34,
"learning_rate": 3.7175988327893317e-05,
"loss": 0.8125,
"step": 785
},
{
"epoch": 0.34,
"learning_rate": 3.714639756107317e-05,
"loss": 0.7121,
"step": 786
},
{
"epoch": 0.34,
"learning_rate": 3.711678450242554e-05,
"loss": 0.6998,
"step": 787
},
{
"epoch": 0.34,
"learning_rate": 3.7087149206298154e-05,
"loss": 0.7277,
"step": 788
},
{
"epoch": 0.34,
"learning_rate": 3.7057491727079554e-05,
"loss": 0.6993,
"step": 789
},
{
"epoch": 0.34,
"learning_rate": 3.702781211919901e-05,
"loss": 0.7536,
"step": 790
},
{
"epoch": 0.34,
"learning_rate": 3.6998110437126384e-05,
"loss": 0.7398,
"step": 791
},
{
"epoch": 0.34,
"learning_rate": 3.6968386735372055e-05,
"loss": 0.7772,
"step": 792
},
{
"epoch": 0.34,
"learning_rate": 3.6938641068486826e-05,
"loss": 0.6984,
"step": 793
},
{
"epoch": 0.34,
"learning_rate": 3.69088734910618e-05,
"loss": 0.6349,
"step": 794
},
{
"epoch": 0.34,
"learning_rate": 3.6879084057728304e-05,
"loss": 0.7508,
"step": 795
},
{
"epoch": 0.34,
"learning_rate": 3.684927282315774e-05,
"loss": 0.7468,
"step": 796
},
{
"epoch": 0.34,
"learning_rate": 3.681943984206158e-05,
"loss": 0.655,
"step": 797
},
{
"epoch": 0.34,
"learning_rate": 3.678958516919114e-05,
"loss": 0.7449,
"step": 798
},
{
"epoch": 0.34,
"learning_rate": 3.6759708859337614e-05,
"loss": 0.9635,
"step": 799
},
{
"epoch": 0.34,
"learning_rate": 3.6729810967331856e-05,
"loss": 0.8153,
"step": 800
},
{
"epoch": 0.34,
"eval_loss": 0.7773626446723938,
"eval_runtime": 10.1729,
"eval_samples_per_second": 3.637,
"eval_steps_per_second": 0.492,
"step": 800
},
{
"epoch": 0.35,
"learning_rate": 3.6699891548044336e-05,
"loss": 0.8183,
"step": 801
},
{
"epoch": 0.35,
"learning_rate": 3.6669950656385065e-05,
"loss": 0.7709,
"step": 802
},
{
"epoch": 0.35,
"learning_rate": 3.663998834730341e-05,
"loss": 0.6341,
"step": 803
},
{
"epoch": 0.35,
"learning_rate": 3.66100046757881e-05,
"loss": 0.826,
"step": 804
},
{
"epoch": 0.35,
"learning_rate": 3.657999969686702e-05,
"loss": 0.7063,
"step": 805
},
{
"epoch": 0.35,
"learning_rate": 3.654997346560719e-05,
"loss": 0.6664,
"step": 806
},
{
"epoch": 0.35,
"learning_rate": 3.651992603711462e-05,
"loss": 0.8019,
"step": 807
},
{
"epoch": 0.35,
"learning_rate": 3.648985746653424e-05,
"loss": 0.7915,
"step": 808
},
{
"epoch": 0.35,
"learning_rate": 3.645976780904975e-05,
"loss": 0.776,
"step": 809
},
{
"epoch": 0.35,
"learning_rate": 3.6429657119883585e-05,
"loss": 0.8638,
"step": 810
},
{
"epoch": 0.35,
"learning_rate": 3.639952545429676e-05,
"loss": 0.7179,
"step": 811
},
{
"epoch": 0.35,
"learning_rate": 3.636937286758878e-05,
"loss": 0.705,
"step": 812
},
{
"epoch": 0.35,
"learning_rate": 3.633919941509757e-05,
"loss": 0.8691,
"step": 813
},
{
"epoch": 0.35,
"learning_rate": 3.630900515219932e-05,
"loss": 0.792,
"step": 814
},
{
"epoch": 0.35,
"learning_rate": 3.6278790134308446e-05,
"loss": 0.8007,
"step": 815
},
{
"epoch": 0.35,
"learning_rate": 3.6248554416877414e-05,
"loss": 0.8518,
"step": 816
},
{
"epoch": 0.35,
"learning_rate": 3.621829805539671e-05,
"loss": 0.7176,
"step": 817
},
{
"epoch": 0.35,
"learning_rate": 3.618802110539471e-05,
"loss": 0.6979,
"step": 818
},
{
"epoch": 0.35,
"learning_rate": 3.6157723622437544e-05,
"loss": 0.7744,
"step": 819
},
{
"epoch": 0.35,
"learning_rate": 3.612740566212906e-05,
"loss": 0.7161,
"step": 820
},
{
"epoch": 0.35,
"learning_rate": 3.6097067280110665e-05,
"loss": 0.8242,
"step": 821
},
{
"epoch": 0.35,
"learning_rate": 3.606670853206125e-05,
"loss": 0.7249,
"step": 822
},
{
"epoch": 0.35,
"learning_rate": 3.60363294736971e-05,
"loss": 0.7814,
"step": 823
},
{
"epoch": 0.36,
"learning_rate": 3.6005930160771755e-05,
"loss": 0.74,
"step": 824
},
{
"epoch": 0.36,
"learning_rate": 3.5975510649075915e-05,
"loss": 0.8421,
"step": 825
},
{
"epoch": 0.36,
"learning_rate": 3.594507099443739e-05,
"loss": 0.7091,
"step": 826
},
{
"epoch": 0.36,
"learning_rate": 3.591461125272093e-05,
"loss": 0.8217,
"step": 827
},
{
"epoch": 0.36,
"learning_rate": 3.5884131479828155e-05,
"loss": 0.797,
"step": 828
},
{
"epoch": 0.36,
"learning_rate": 3.585363173169744e-05,
"loss": 0.8585,
"step": 829
},
{
"epoch": 0.36,
"learning_rate": 3.582311206430383e-05,
"loss": 0.7903,
"step": 830
},
{
"epoch": 0.36,
"learning_rate": 3.579257253365894e-05,
"loss": 0.9192,
"step": 831
},
{
"epoch": 0.36,
"learning_rate": 3.57620131958108e-05,
"loss": 0.8264,
"step": 832
},
{
"epoch": 0.36,
"learning_rate": 3.5731434106843836e-05,
"loss": 1.0079,
"step": 833
},
{
"epoch": 0.36,
"learning_rate": 3.57008353228787e-05,
"loss": 0.7485,
"step": 834
},
{
"epoch": 0.36,
"learning_rate": 3.567021690007218e-05,
"loss": 0.7599,
"step": 835
},
{
"epoch": 0.36,
"learning_rate": 3.563957889461713e-05,
"loss": 0.6383,
"step": 836
},
{
"epoch": 0.36,
"learning_rate": 3.560892136274232e-05,
"loss": 0.6461,
"step": 837
},
{
"epoch": 0.36,
"learning_rate": 3.5578244360712373e-05,
"loss": 0.7783,
"step": 838
},
{
"epoch": 0.36,
"learning_rate": 3.5547547944827644e-05,
"loss": 0.7189,
"step": 839
},
{
"epoch": 0.36,
"learning_rate": 3.55168321714241e-05,
"loss": 0.7878,
"step": 840
},
{
"epoch": 0.36,
"learning_rate": 3.548609709687326e-05,
"loss": 0.8456,
"step": 841
},
{
"epoch": 0.36,
"learning_rate": 3.545534277758205e-05,
"loss": 0.8697,
"step": 842
},
{
"epoch": 0.36,
"learning_rate": 3.542456926999271e-05,
"loss": 0.9097,
"step": 843
},
{
"epoch": 0.36,
"learning_rate": 3.539377663058271e-05,
"loss": 0.7874,
"step": 844
},
{
"epoch": 0.36,
"learning_rate": 3.536296491586463e-05,
"loss": 0.8056,
"step": 845
},
{
"epoch": 0.36,
"learning_rate": 3.533213418238605e-05,
"loss": 0.7522,
"step": 846
},
{
"epoch": 0.37,
"learning_rate": 3.530128448672944e-05,
"loss": 0.7501,
"step": 847
},
{
"epoch": 0.37,
"learning_rate": 3.527041588551212e-05,
"loss": 0.7787,
"step": 848
},
{
"epoch": 0.37,
"learning_rate": 3.523952843538604e-05,
"loss": 0.7066,
"step": 849
},
{
"epoch": 0.37,
"learning_rate": 3.520862219303781e-05,
"loss": 0.7545,
"step": 850
},
{
"epoch": 0.37,
"learning_rate": 3.5177697215188475e-05,
"loss": 0.8294,
"step": 851
},
{
"epoch": 0.37,
"learning_rate": 3.51467535585935e-05,
"loss": 0.7757,
"step": 852
},
{
"epoch": 0.37,
"learning_rate": 3.5115791280042606e-05,
"loss": 0.7384,
"step": 853
},
{
"epoch": 0.37,
"learning_rate": 3.508481043635971e-05,
"loss": 0.8955,
"step": 854
},
{
"epoch": 0.37,
"learning_rate": 3.505381108440279e-05,
"loss": 0.6803,
"step": 855
},
{
"epoch": 0.37,
"learning_rate": 3.502279328106377e-05,
"loss": 0.8034,
"step": 856
},
{
"epoch": 0.37,
"learning_rate": 3.49917570832685e-05,
"loss": 0.8621,
"step": 857
},
{
"epoch": 0.37,
"learning_rate": 3.496070254797652e-05,
"loss": 0.782,
"step": 858
},
{
"epoch": 0.37,
"learning_rate": 3.4929629732181064e-05,
"loss": 0.7774,
"step": 859
},
{
"epoch": 0.37,
"learning_rate": 3.489853869290889e-05,
"loss": 0.7672,
"step": 860
},
{
"epoch": 0.37,
"learning_rate": 3.486742948722024e-05,
"loss": 0.7518,
"step": 861
},
{
"epoch": 0.37,
"learning_rate": 3.4836302172208645e-05,
"loss": 0.9012,
"step": 862
},
{
"epoch": 0.37,
"learning_rate": 3.480515680500092e-05,
"loss": 0.8715,
"step": 863
},
{
"epoch": 0.37,
"learning_rate": 3.477399344275698e-05,
"loss": 0.7544,
"step": 864
},
{
"epoch": 0.37,
"learning_rate": 3.474281214266977e-05,
"loss": 0.6645,
"step": 865
},
{
"epoch": 0.37,
"learning_rate": 3.471161296196516e-05,
"loss": 0.9114,
"step": 866
},
{
"epoch": 0.37,
"learning_rate": 3.468039595790184e-05,
"loss": 0.7907,
"step": 867
},
{
"epoch": 0.37,
"learning_rate": 3.4649161187771206e-05,
"loss": 0.9009,
"step": 868
},
{
"epoch": 0.37,
"learning_rate": 3.461790870889725e-05,
"loss": 0.6086,
"step": 869
},
{
"epoch": 0.37,
"learning_rate": 3.458663857863649e-05,
"loss": 0.8777,
"step": 870
},
{
"epoch": 0.38,
"learning_rate": 3.455535085437781e-05,
"loss": 0.6795,
"step": 871
},
{
"epoch": 0.38,
"learning_rate": 3.4524045593542386e-05,
"loss": 0.6594,
"step": 872
},
{
"epoch": 0.38,
"learning_rate": 3.449272285358361e-05,
"loss": 0.6547,
"step": 873
},
{
"epoch": 0.38,
"learning_rate": 3.446138269198692e-05,
"loss": 0.8464,
"step": 874
},
{
"epoch": 0.38,
"learning_rate": 3.4430025166269744e-05,
"loss": 0.7462,
"step": 875
},
{
"epoch": 0.38,
"learning_rate": 3.4398650333981354e-05,
"loss": 0.8541,
"step": 876
},
{
"epoch": 0.38,
"learning_rate": 3.436725825270283e-05,
"loss": 0.8847,
"step": 877
},
{
"epoch": 0.38,
"learning_rate": 3.433584898004685e-05,
"loss": 0.6683,
"step": 878
},
{
"epoch": 0.38,
"learning_rate": 3.43044225736577e-05,
"loss": 0.7771,
"step": 879
},
{
"epoch": 0.38,
"learning_rate": 3.427297909121107e-05,
"loss": 0.7601,
"step": 880
},
{
"epoch": 0.38,
"learning_rate": 3.4241518590414e-05,
"loss": 0.8268,
"step": 881
},
{
"epoch": 0.38,
"learning_rate": 3.421004112900477e-05,
"loss": 0.7158,
"step": 882
},
{
"epoch": 0.38,
"learning_rate": 3.417854676475279e-05,
"loss": 0.7148,
"step": 883
},
{
"epoch": 0.38,
"learning_rate": 3.4147035555458477e-05,
"loss": 0.7187,
"step": 884
},
{
"epoch": 0.38,
"learning_rate": 3.4115507558953174e-05,
"loss": 0.8365,
"step": 885
},
{
"epoch": 0.38,
"learning_rate": 3.408396283309903e-05,
"loss": 0.7431,
"step": 886
},
{
"epoch": 0.38,
"learning_rate": 3.4052401435788885e-05,
"loss": 0.7332,
"step": 887
},
{
"epoch": 0.38,
"learning_rate": 3.4020823424946205e-05,
"loss": 0.8885,
"step": 888
},
{
"epoch": 0.38,
"learning_rate": 3.3989228858524914e-05,
"loss": 0.6457,
"step": 889
},
{
"epoch": 0.38,
"learning_rate": 3.3957617794509344e-05,
"loss": 0.7648,
"step": 890
},
{
"epoch": 0.38,
"learning_rate": 3.392599029091409e-05,
"loss": 1.0406,
"step": 891
},
{
"epoch": 0.38,
"learning_rate": 3.389434640578392e-05,
"loss": 0.7927,
"step": 892
},
{
"epoch": 0.38,
"learning_rate": 3.386268619719366e-05,
"loss": 0.8087,
"step": 893
},
{
"epoch": 0.39,
"learning_rate": 3.383100972324812e-05,
"loss": 0.7671,
"step": 894
},
{
"epoch": 0.39,
"learning_rate": 3.379931704208193e-05,
"loss": 0.8655,
"step": 895
},
{
"epoch": 0.39,
"learning_rate": 3.3767608211859495e-05,
"loss": 0.7414,
"step": 896
},
{
"epoch": 0.39,
"learning_rate": 3.373588329077482e-05,
"loss": 0.7659,
"step": 897
},
{
"epoch": 0.39,
"learning_rate": 3.370414233705147e-05,
"loss": 0.6748,
"step": 898
},
{
"epoch": 0.39,
"learning_rate": 3.367238540894243e-05,
"loss": 0.6292,
"step": 899
},
{
"epoch": 0.39,
"learning_rate": 3.3640612564729996e-05,
"loss": 0.6958,
"step": 900
},
{
"epoch": 0.39,
"eval_loss": 0.775384783744812,
"eval_runtime": 10.201,
"eval_samples_per_second": 3.627,
"eval_steps_per_second": 0.49,
"step": 900
},
{
"epoch": 0.39,
"learning_rate": 3.3608823862725675e-05,
"loss": 0.6352,
"step": 901
},
{
"epoch": 0.39,
"learning_rate": 3.357701936127008e-05,
"loss": 0.7488,
"step": 902
},
{
"epoch": 0.39,
"learning_rate": 3.354519911873282e-05,
"loss": 0.6704,
"step": 903
},
{
"epoch": 0.39,
"learning_rate": 3.351336319351238e-05,
"loss": 0.8388,
"step": 904
},
{
"epoch": 0.39,
"learning_rate": 3.348151164403605e-05,
"loss": 0.7291,
"step": 905
},
{
"epoch": 0.39,
"learning_rate": 3.344964452875978e-05,
"loss": 0.7755,
"step": 906
},
{
"epoch": 0.39,
"learning_rate": 3.34177619061681e-05,
"loss": 0.8571,
"step": 907
},
{
"epoch": 0.39,
"learning_rate": 3.3385863834773975e-05,
"loss": 0.767,
"step": 908
},
{
"epoch": 0.39,
"learning_rate": 3.335395037311874e-05,
"loss": 0.8152,
"step": 909
},
{
"epoch": 0.39,
"learning_rate": 3.332202157977198e-05,
"loss": 0.6353,
"step": 910
},
{
"epoch": 0.39,
"learning_rate": 3.329007751333142e-05,
"loss": 0.9138,
"step": 911
},
{
"epoch": 0.39,
"learning_rate": 3.3258118232422787e-05,
"loss": 0.8154,
"step": 912
},
{
"epoch": 0.39,
"learning_rate": 3.322614379569975e-05,
"loss": 0.7702,
"step": 913
},
{
"epoch": 0.39,
"learning_rate": 3.319415426184381e-05,
"loss": 0.7814,
"step": 914
},
{
"epoch": 0.39,
"learning_rate": 3.3162149689564135e-05,
"loss": 0.788,
"step": 915
},
{
"epoch": 0.39,
"learning_rate": 3.3130130137597546e-05,
"loss": 0.932,
"step": 916
},
{
"epoch": 0.4,
"learning_rate": 3.309809566470829e-05,
"loss": 0.8445,
"step": 917
},
{
"epoch": 0.4,
"learning_rate": 3.3066046329688056e-05,
"loss": 0.8815,
"step": 918
},
{
"epoch": 0.4,
"learning_rate": 3.303398219135578e-05,
"loss": 0.861,
"step": 919
},
{
"epoch": 0.4,
"learning_rate": 3.300190330855756e-05,
"loss": 0.7331,
"step": 920
},
{
"epoch": 0.4,
"learning_rate": 3.2969809740166583e-05,
"loss": 0.7393,
"step": 921
},
{
"epoch": 0.4,
"learning_rate": 3.293770154508296e-05,
"loss": 0.8168,
"step": 922
},
{
"epoch": 0.4,
"learning_rate": 3.2905578782233665e-05,
"loss": 0.8113,
"step": 923
},
{
"epoch": 0.4,
"learning_rate": 3.287344151057239e-05,
"loss": 0.7994,
"step": 924
},
{
"epoch": 0.4,
"learning_rate": 3.2841289789079465e-05,
"loss": 0.7081,
"step": 925
},
{
"epoch": 0.4,
"learning_rate": 3.280912367676173e-05,
"loss": 0.6252,
"step": 926
},
{
"epoch": 0.4,
"learning_rate": 3.2776943232652465e-05,
"loss": 0.7764,
"step": 927
},
{
"epoch": 0.4,
"learning_rate": 3.27447485158112e-05,
"loss": 0.7369,
"step": 928
},
{
"epoch": 0.4,
"learning_rate": 3.271253958532372e-05,
"loss": 0.7788,
"step": 929
},
{
"epoch": 0.4,
"learning_rate": 3.268031650030186e-05,
"loss": 0.7606,
"step": 930
},
{
"epoch": 0.4,
"learning_rate": 3.264807931988342e-05,
"loss": 0.9283,
"step": 931
},
{
"epoch": 0.4,
"learning_rate": 3.261582810323211e-05,
"loss": 0.7162,
"step": 932
},
{
"epoch": 0.4,
"learning_rate": 3.258356290953736e-05,
"loss": 0.8287,
"step": 933
},
{
"epoch": 0.4,
"learning_rate": 3.2551283798014285e-05,
"loss": 0.7492,
"step": 934
},
{
"epoch": 0.4,
"learning_rate": 3.25189908279035e-05,
"loss": 0.7372,
"step": 935
},
{
"epoch": 0.4,
"learning_rate": 3.248668405847111e-05,
"loss": 0.6945,
"step": 936
},
{
"epoch": 0.4,
"learning_rate": 3.24543635490085e-05,
"loss": 0.6628,
"step": 937
},
{
"epoch": 0.4,
"learning_rate": 3.24220293588323e-05,
"loss": 0.7582,
"step": 938
},
{
"epoch": 0.4,
"learning_rate": 3.2389681547284224e-05,
"loss": 0.8779,
"step": 939
},
{
"epoch": 0.41,
"learning_rate": 3.2357320173731e-05,
"loss": 0.6828,
"step": 940
},
{
"epoch": 0.41,
"learning_rate": 3.2324945297564236e-05,
"loss": 0.7707,
"step": 941
},
{
"epoch": 0.41,
"learning_rate": 3.229255697820034e-05,
"loss": 0.8004,
"step": 942
},
{
"epoch": 0.41,
"learning_rate": 3.2260155275080363e-05,
"loss": 0.6871,
"step": 943
},
{
"epoch": 0.41,
"learning_rate": 3.222774024766995e-05,
"loss": 0.9631,
"step": 944
},
{
"epoch": 0.41,
"learning_rate": 3.219531195545918e-05,
"loss": 0.7378,
"step": 945
},
{
"epoch": 0.41,
"learning_rate": 3.216287045796247e-05,
"loss": 0.8937,
"step": 946
},
{
"epoch": 0.41,
"learning_rate": 3.213041581471849e-05,
"loss": 0.7671,
"step": 947
},
{
"epoch": 0.41,
"learning_rate": 3.209794808529004e-05,
"loss": 0.6912,
"step": 948
},
{
"epoch": 0.41,
"learning_rate": 3.20654673292639e-05,
"loss": 0.8395,
"step": 949
},
{
"epoch": 0.41,
"learning_rate": 3.2032973606250804e-05,
"loss": 0.8311,
"step": 950
},
{
"epoch": 0.41,
"learning_rate": 3.200046697588526e-05,
"loss": 0.7332,
"step": 951
},
{
"epoch": 0.41,
"learning_rate": 3.196794749782544e-05,
"loss": 0.6453,
"step": 952
},
{
"epoch": 0.41,
"learning_rate": 3.1935415231753155e-05,
"loss": 0.7365,
"step": 953
},
{
"epoch": 0.41,
"learning_rate": 3.190287023737364e-05,
"loss": 0.8021,
"step": 954
},
{
"epoch": 0.41,
"learning_rate": 3.187031257441549e-05,
"loss": 0.7917,
"step": 955
},
{
"epoch": 0.41,
"learning_rate": 3.183774230263058e-05,
"loss": 0.813,
"step": 956
},
{
"epoch": 0.41,
"learning_rate": 3.18051594817939e-05,
"loss": 0.768,
"step": 957
},
{
"epoch": 0.41,
"learning_rate": 3.1772564171703475e-05,
"loss": 0.7123,
"step": 958
},
{
"epoch": 0.41,
"learning_rate": 3.173995643218025e-05,
"loss": 0.836,
"step": 959
},
{
"epoch": 0.41,
"learning_rate": 3.1707336323068e-05,
"loss": 0.7068,
"step": 960
},
{
"epoch": 0.41,
"learning_rate": 3.167470390423317e-05,
"loss": 0.6678,
"step": 961
},
{
"epoch": 0.41,
"learning_rate": 3.164205923556483e-05,
"loss": 0.699,
"step": 962
},
{
"epoch": 0.41,
"learning_rate": 3.1609402376974495e-05,
"loss": 0.8364,
"step": 963
},
{
"epoch": 0.42,
"learning_rate": 3.157673338839609e-05,
"loss": 0.6276,
"step": 964
},
{
"epoch": 0.42,
"learning_rate": 3.154405232978577e-05,
"loss": 0.627,
"step": 965
},
{
"epoch": 0.42,
"learning_rate": 3.151135926112187e-05,
"loss": 0.6027,
"step": 966
},
{
"epoch": 0.42,
"learning_rate": 3.147865424240474e-05,
"loss": 0.6496,
"step": 967
},
{
"epoch": 0.42,
"learning_rate": 3.144593733365667e-05,
"loss": 0.7317,
"step": 968
},
{
"epoch": 0.42,
"learning_rate": 3.14132085949218e-05,
"loss": 0.7783,
"step": 969
},
{
"epoch": 0.42,
"learning_rate": 3.138046808626591e-05,
"loss": 0.7837,
"step": 970
},
{
"epoch": 0.42,
"learning_rate": 3.134771586777647e-05,
"loss": 0.7783,
"step": 971
},
{
"epoch": 0.42,
"learning_rate": 3.131495199956237e-05,
"loss": 0.7335,
"step": 972
},
{
"epoch": 0.42,
"learning_rate": 3.128217654175393e-05,
"loss": 0.7112,
"step": 973
},
{
"epoch": 0.42,
"learning_rate": 3.124938955450269e-05,
"loss": 0.8552,
"step": 974
},
{
"epoch": 0.42,
"learning_rate": 3.1216591097981394e-05,
"loss": 0.7491,
"step": 975
},
{
"epoch": 0.42,
"learning_rate": 3.118378123238383e-05,
"loss": 0.8029,
"step": 976
},
{
"epoch": 0.42,
"learning_rate": 3.1150960017924685e-05,
"loss": 0.6775,
"step": 977
},
{
"epoch": 0.42,
"learning_rate": 3.111812751483952e-05,
"loss": 0.8973,
"step": 978
},
{
"epoch": 0.42,
"learning_rate": 3.1085283783384586e-05,
"loss": 0.7634,
"step": 979
},
{
"epoch": 0.42,
"learning_rate": 3.1052428883836764e-05,
"loss": 0.8199,
"step": 980
},
{
"epoch": 0.42,
"learning_rate": 3.1019562876493397e-05,
"loss": 0.6868,
"step": 981
},
{
"epoch": 0.42,
"learning_rate": 3.0986685821672264e-05,
"loss": 0.6868,
"step": 982
},
{
"epoch": 0.42,
"learning_rate": 3.095379777971135e-05,
"loss": 0.8666,
"step": 983
},
{
"epoch": 0.42,
"learning_rate": 3.092089881096888e-05,
"loss": 0.6998,
"step": 984
},
{
"epoch": 0.42,
"learning_rate": 3.088798897582308e-05,
"loss": 0.6144,
"step": 985
},
{
"epoch": 0.42,
"learning_rate": 3.085506833467214e-05,
"loss": 0.9385,
"step": 986
},
{
"epoch": 0.43,
"learning_rate": 3.082213694793407e-05,
"loss": 0.7249,
"step": 987
},
{
"epoch": 0.43,
"learning_rate": 3.0789194876046614e-05,
"loss": 0.8582,
"step": 988
},
{
"epoch": 0.43,
"learning_rate": 3.075624217946712e-05,
"loss": 0.7414,
"step": 989
},
{
"epoch": 0.43,
"learning_rate": 3.072327891867243e-05,
"loss": 0.8248,
"step": 990
},
{
"epoch": 0.43,
"learning_rate": 3.069030515415878e-05,
"loss": 0.7254,
"step": 991
},
{
"epoch": 0.43,
"learning_rate": 3.0657320946441686e-05,
"loss": 0.6965,
"step": 992
},
{
"epoch": 0.43,
"learning_rate": 3.0624326356055824e-05,
"loss": 0.7979,
"step": 993
},
{
"epoch": 0.43,
"learning_rate": 3.059132144355494e-05,
"loss": 0.6902,
"step": 994
},
{
"epoch": 0.43,
"learning_rate": 3.0558306269511686e-05,
"loss": 0.7549,
"step": 995
},
{
"epoch": 0.43,
"learning_rate": 3.052528089451759e-05,
"loss": 0.757,
"step": 996
},
{
"epoch": 0.43,
"learning_rate": 3.0492245379182877e-05,
"loss": 0.7158,
"step": 997
},
{
"epoch": 0.43,
"learning_rate": 3.045919978413639e-05,
"loss": 0.7439,
"step": 998
},
{
"epoch": 0.43,
"learning_rate": 3.0426144170025456e-05,
"loss": 0.7263,
"step": 999
},
{
"epoch": 0.43,
"learning_rate": 3.0393078597515827e-05,
"loss": 0.8714,
"step": 1000
},
{
"epoch": 0.43,
"eval_loss": 0.7731763124465942,
"eval_runtime": 10.1599,
"eval_samples_per_second": 3.642,
"eval_steps_per_second": 0.492,
"step": 1000
},
{
"epoch": 0.43,
"learning_rate": 3.0360003127291494e-05,
"loss": 0.7357,
"step": 1001
},
{
"epoch": 0.43,
"learning_rate": 3.0326917820054622e-05,
"loss": 0.6703,
"step": 1002
},
{
"epoch": 0.43,
"learning_rate": 3.029382273652544e-05,
"loss": 0.7843,
"step": 1003
},
{
"epoch": 0.43,
"learning_rate": 3.0260717937442113e-05,
"loss": 0.8579,
"step": 1004
},
{
"epoch": 0.43,
"learning_rate": 3.022760348356063e-05,
"loss": 0.8031,
"step": 1005
},
{
"epoch": 0.43,
"learning_rate": 3.019447943565471e-05,
"loss": 0.763,
"step": 1006
},
{
"epoch": 0.43,
"learning_rate": 3.0161345854515672e-05,
"loss": 0.9019,
"step": 1007
},
{
"epoch": 0.43,
"learning_rate": 3.012820280095234e-05,
"loss": 0.7761,
"step": 1008
},
{
"epoch": 0.43,
"learning_rate": 3.0095050335790903e-05,
"loss": 0.6707,
"step": 1009
},
{
"epoch": 0.44,
"learning_rate": 3.006188851987485e-05,
"loss": 0.7175,
"step": 1010
},
{
"epoch": 0.44,
"learning_rate": 3.0028717414064805e-05,
"loss": 0.8008,
"step": 1011
},
{
"epoch": 0.44,
"learning_rate": 2.9995537079238467e-05,
"loss": 0.7714,
"step": 1012
},
{
"epoch": 0.44,
"learning_rate": 2.9962347576290444e-05,
"loss": 0.6841,
"step": 1013
},
{
"epoch": 0.44,
"learning_rate": 2.99291489661322e-05,
"loss": 0.8408,
"step": 1014
},
{
"epoch": 0.44,
"learning_rate": 2.9895941309691883e-05,
"loss": 0.8279,
"step": 1015
},
{
"epoch": 0.44,
"learning_rate": 2.9862724667914265e-05,
"loss": 0.7386,
"step": 1016
},
{
"epoch": 0.44,
"learning_rate": 2.982949910176061e-05,
"loss": 0.7823,
"step": 1017
},
{
"epoch": 0.44,
"learning_rate": 2.9796264672208535e-05,
"loss": 0.7185,
"step": 1018
},
{
"epoch": 0.44,
"learning_rate": 2.9763021440251954e-05,
"loss": 0.7604,
"step": 1019
},
{
"epoch": 0.44,
"learning_rate": 2.9729769466900918e-05,
"loss": 0.8062,
"step": 1020
},
{
"epoch": 0.44,
"learning_rate": 2.9696508813181534e-05,
"loss": 0.6379,
"step": 1021
},
{
"epoch": 0.44,
"learning_rate": 2.9663239540135818e-05,
"loss": 0.6967,
"step": 1022
},
{
"epoch": 0.44,
"learning_rate": 2.9629961708821618e-05,
"loss": 0.695,
"step": 1023
},
{
"epoch": 0.44,
"learning_rate": 2.9596675380312507e-05,
"loss": 0.8031,
"step": 1024
},
{
"epoch": 0.44,
"learning_rate": 2.9563380615697605e-05,
"loss": 0.7674,
"step": 1025
},
{
"epoch": 0.44,
"learning_rate": 2.953007747608158e-05,
"loss": 0.8916,
"step": 1026
},
{
"epoch": 0.44,
"learning_rate": 2.9496766022584398e-05,
"loss": 0.8159,
"step": 1027
},
{
"epoch": 0.44,
"learning_rate": 2.9463446316341346e-05,
"loss": 0.7979,
"step": 1028
},
{
"epoch": 0.44,
"learning_rate": 2.943011841850281e-05,
"loss": 0.729,
"step": 1029
},
{
"epoch": 0.44,
"learning_rate": 2.9396782390234245e-05,
"loss": 0.9,
"step": 1030
},
{
"epoch": 0.44,
"learning_rate": 2.9363438292716004e-05,
"loss": 0.8511,
"step": 1031
},
{
"epoch": 0.44,
"learning_rate": 2.933008618714327e-05,
"loss": 0.923,
"step": 1032
},
{
"epoch": 0.45,
"learning_rate": 2.9296726134725887e-05,
"loss": 0.8954,
"step": 1033
},
{
"epoch": 0.45,
"learning_rate": 2.926335819668832e-05,
"loss": 0.8655,
"step": 1034
},
{
"epoch": 0.45,
"learning_rate": 2.922998243426951e-05,
"loss": 0.8657,
"step": 1035
},
{
"epoch": 0.45,
"learning_rate": 2.919659890872271e-05,
"loss": 0.8267,
"step": 1036
},
{
"epoch": 0.45,
"learning_rate": 2.9163207681315476e-05,
"loss": 0.6383,
"step": 1037
},
{
"epoch": 0.45,
"learning_rate": 2.912980881332945e-05,
"loss": 0.7495,
"step": 1038
},
{
"epoch": 0.45,
"learning_rate": 2.909640236606035e-05,
"loss": 0.748,
"step": 1039
},
{
"epoch": 0.45,
"learning_rate": 2.906298840081775e-05,
"loss": 0.8319,
"step": 1040
},
{
"epoch": 0.45,
"learning_rate": 2.9029566978925055e-05,
"loss": 0.8384,
"step": 1041
},
{
"epoch": 0.45,
"learning_rate": 2.899613816171935e-05,
"loss": 0.8418,
"step": 1042
},
{
"epoch": 0.45,
"learning_rate": 2.8962702010551285e-05,
"loss": 0.6754,
"step": 1043
},
{
"epoch": 0.45,
"learning_rate": 2.8929258586784974e-05,
"loss": 0.8605,
"step": 1044
},
{
"epoch": 0.45,
"learning_rate": 2.8895807951797866e-05,
"loss": 0.7574,
"step": 1045
},
{
"epoch": 0.45,
"learning_rate": 2.8862350166980673e-05,
"loss": 0.6856,
"step": 1046
},
{
"epoch": 0.45,
"learning_rate": 2.8828885293737206e-05,
"loss": 0.676,
"step": 1047
},
{
"epoch": 0.45,
"learning_rate": 2.8795413393484288e-05,
"loss": 0.7304,
"step": 1048
},
{
"epoch": 0.45,
"learning_rate": 2.8761934527651642e-05,
"loss": 0.8223,
"step": 1049
},
{
"epoch": 0.45,
"learning_rate": 2.8728448757681775e-05,
"loss": 0.9364,
"step": 1050
},
{
"epoch": 0.45,
"learning_rate": 2.869495614502986e-05,
"loss": 0.8179,
"step": 1051
},
{
"epoch": 0.45,
"learning_rate": 2.8661456751163633e-05,
"loss": 0.8226,
"step": 1052
},
{
"epoch": 0.45,
"learning_rate": 2.8627950637563283e-05,
"loss": 0.8599,
"step": 1053
},
{
"epoch": 0.45,
"learning_rate": 2.8594437865721303e-05,
"loss": 0.8034,
"step": 1054
},
{
"epoch": 0.45,
"learning_rate": 2.8560918497142443e-05,
"loss": 0.7842,
"step": 1055
},
{
"epoch": 0.46,
"learning_rate": 2.852739259334354e-05,
"loss": 0.8364,
"step": 1056
},
{
"epoch": 0.46,
"learning_rate": 2.8493860215853413e-05,
"loss": 0.6798,
"step": 1057
},
{
"epoch": 0.46,
"learning_rate": 2.8460321426212793e-05,
"loss": 0.6865,
"step": 1058
},
{
"epoch": 0.46,
"learning_rate": 2.842677628597415e-05,
"loss": 0.7019,
"step": 1059
},
{
"epoch": 0.46,
"learning_rate": 2.839322485670163e-05,
"loss": 0.769,
"step": 1060
},
{
"epoch": 0.46,
"learning_rate": 2.83596671999709e-05,
"loss": 0.7563,
"step": 1061
},
{
"epoch": 0.46,
"learning_rate": 2.8326103377369073e-05,
"loss": 0.7969,
"step": 1062
},
{
"epoch": 0.46,
"learning_rate": 2.829253345049458e-05,
"loss": 0.7958,
"step": 1063
},
{
"epoch": 0.46,
"learning_rate": 2.825895748095703e-05,
"loss": 0.7038,
"step": 1064
},
{
"epoch": 0.46,
"learning_rate": 2.8225375530377162e-05,
"loss": 0.7533,
"step": 1065
},
{
"epoch": 0.46,
"learning_rate": 2.8191787660386653e-05,
"loss": 0.7245,
"step": 1066
},
{
"epoch": 0.46,
"learning_rate": 2.8158193932628068e-05,
"loss": 0.6946,
"step": 1067
},
{
"epoch": 0.46,
"learning_rate": 2.812459440875471e-05,
"loss": 0.7737,
"step": 1068
},
{
"epoch": 0.46,
"learning_rate": 2.8090989150430523e-05,
"loss": 0.6086,
"step": 1069
},
{
"epoch": 0.46,
"learning_rate": 2.8057378219329983e-05,
"loss": 0.7691,
"step": 1070
},
{
"epoch": 0.46,
"learning_rate": 2.802376167713796e-05,
"loss": 0.6517,
"step": 1071
},
{
"epoch": 0.46,
"learning_rate": 2.7990139585549646e-05,
"loss": 0.7769,
"step": 1072
},
{
"epoch": 0.46,
"learning_rate": 2.7956512006270387e-05,
"loss": 0.7431,
"step": 1073
},
{
"epoch": 0.46,
"learning_rate": 2.7922879001015632e-05,
"loss": 0.7643,
"step": 1074
},
{
"epoch": 0.46,
"learning_rate": 2.788924063151076e-05,
"loss": 0.8355,
"step": 1075
},
{
"epoch": 0.46,
"learning_rate": 2.7855596959491015e-05,
"loss": 0.6971,
"step": 1076
},
{
"epoch": 0.46,
"learning_rate": 2.782194804670136e-05,
"loss": 0.8161,
"step": 1077
},
{
"epoch": 0.46,
"learning_rate": 2.77882939548964e-05,
"loss": 0.6974,
"step": 1078
},
{
"epoch": 0.46,
"learning_rate": 2.77546347458402e-05,
"loss": 0.7395,
"step": 1079
},
{
"epoch": 0.47,
"learning_rate": 2.7720970481306258e-05,
"loss": 0.671,
"step": 1080
},
{
"epoch": 0.47,
"learning_rate": 2.7687301223077332e-05,
"loss": 0.64,
"step": 1081
},
{
"epoch": 0.47,
"learning_rate": 2.7653627032945346e-05,
"loss": 0.7831,
"step": 1082
},
{
"epoch": 0.47,
"learning_rate": 2.761994797271129e-05,
"loss": 0.755,
"step": 1083
},
{
"epoch": 0.47,
"learning_rate": 2.7586264104185055e-05,
"loss": 0.656,
"step": 1084
},
{
"epoch": 0.47,
"learning_rate": 2.75525754891854e-05,
"loss": 0.7147,
"step": 1085
},
{
"epoch": 0.47,
"learning_rate": 2.7518882189539762e-05,
"loss": 0.7379,
"step": 1086
},
{
"epoch": 0.47,
"learning_rate": 2.74851842670842e-05,
"loss": 0.6838,
"step": 1087
},
{
"epoch": 0.47,
"learning_rate": 2.7451481783663246e-05,
"loss": 0.8398,
"step": 1088
},
{
"epoch": 0.47,
"learning_rate": 2.741777480112979e-05,
"loss": 0.7378,
"step": 1089
},
{
"epoch": 0.47,
"learning_rate": 2.7384063381345017e-05,
"loss": 0.9643,
"step": 1090
},
{
"epoch": 0.47,
"learning_rate": 2.7350347586178194e-05,
"loss": 0.8147,
"step": 1091
},
{
"epoch": 0.47,
"learning_rate": 2.7316627477506688e-05,
"loss": 0.6773,
"step": 1092
},
{
"epoch": 0.47,
"learning_rate": 2.7282903117215724e-05,
"loss": 0.7694,
"step": 1093
},
{
"epoch": 0.47,
"learning_rate": 2.724917456719837e-05,
"loss": 0.82,
"step": 1094
},
{
"epoch": 0.47,
"learning_rate": 2.7215441889355364e-05,
"loss": 0.642,
"step": 1095
},
{
"epoch": 0.47,
"learning_rate": 2.718170514559502e-05,
"loss": 0.7938,
"step": 1096
},
{
"epoch": 0.47,
"learning_rate": 2.7147964397833127e-05,
"loss": 0.8053,
"step": 1097
},
{
"epoch": 0.47,
"learning_rate": 2.7114219707992798e-05,
"loss": 0.7755,
"step": 1098
},
{
"epoch": 0.47,
"learning_rate": 2.7080471138004405e-05,
"loss": 0.7722,
"step": 1099
},
{
"epoch": 0.47,
"learning_rate": 2.7046718749805422e-05,
"loss": 0.8225,
"step": 1100
},
{
"epoch": 0.47,
"eval_loss": 0.7722584009170532,
"eval_runtime": 10.1954,
"eval_samples_per_second": 3.629,
"eval_steps_per_second": 0.49,
"step": 1100
},
{
"epoch": 0.47,
"learning_rate": 2.7012962605340363e-05,
"loss": 0.8115,
"step": 1101
},
{
"epoch": 0.47,
"learning_rate": 2.6979202766560585e-05,
"loss": 0.7637,
"step": 1102
},
{
"epoch": 0.48,
"learning_rate": 2.694543929542427e-05,
"loss": 0.7145,
"step": 1103
},
{
"epoch": 0.48,
"learning_rate": 2.6911672253896247e-05,
"loss": 0.7826,
"step": 1104
},
{
"epoch": 0.48,
"learning_rate": 2.6877901703947895e-05,
"loss": 0.7221,
"step": 1105
},
{
"epoch": 0.48,
"learning_rate": 2.6844127707557036e-05,
"loss": 0.7527,
"step": 1106
},
{
"epoch": 0.48,
"learning_rate": 2.681035032670782e-05,
"loss": 0.8327,
"step": 1107
},
{
"epoch": 0.48,
"learning_rate": 2.6776569623390614e-05,
"loss": 0.8694,
"step": 1108
},
{
"epoch": 0.48,
"learning_rate": 2.6742785659601866e-05,
"loss": 1.0126,
"step": 1109
},
{
"epoch": 0.48,
"learning_rate": 2.6708998497344007e-05,
"loss": 0.6732,
"step": 1110
},
{
"epoch": 0.48,
"learning_rate": 2.6675208198625367e-05,
"loss": 0.907,
"step": 1111
},
{
"epoch": 0.48,
"learning_rate": 2.6641414825460004e-05,
"loss": 0.7191,
"step": 1112
},
{
"epoch": 0.48,
"learning_rate": 2.6607618439867633e-05,
"loss": 0.6305,
"step": 1113
},
{
"epoch": 0.48,
"learning_rate": 2.6573819103873475e-05,
"loss": 0.7972,
"step": 1114
},
{
"epoch": 0.48,
"learning_rate": 2.6540016879508204e-05,
"loss": 0.7695,
"step": 1115
},
{
"epoch": 0.48,
"learning_rate": 2.650621182880776e-05,
"loss": 0.745,
"step": 1116
},
{
"epoch": 0.48,
"learning_rate": 2.6472404013813283e-05,
"loss": 0.6017,
"step": 1117
},
{
"epoch": 0.48,
"learning_rate": 2.6438593496571e-05,
"loss": 0.6953,
"step": 1118
},
{
"epoch": 0.48,
"learning_rate": 2.6404780339132067e-05,
"loss": 0.7556,
"step": 1119
},
{
"epoch": 0.48,
"learning_rate": 2.6370964603552517e-05,
"loss": 0.675,
"step": 1120
},
{
"epoch": 0.48,
"learning_rate": 2.6337146351893087e-05,
"loss": 0.844,
"step": 1121
},
{
"epoch": 0.48,
"learning_rate": 2.6303325646219157e-05,
"loss": 0.7474,
"step": 1122
},
{
"epoch": 0.48,
"learning_rate": 2.626950254860059e-05,
"loss": 0.6728,
"step": 1123
},
{
"epoch": 0.48,
"learning_rate": 2.623567712111165e-05,
"loss": 0.6939,
"step": 1124
},
{
"epoch": 0.48,
"learning_rate": 2.6201849425830877e-05,
"loss": 0.8752,
"step": 1125
},
{
"epoch": 0.49,
"learning_rate": 2.6168019524840963e-05,
"loss": 0.7942,
"step": 1126
},
{
"epoch": 0.49,
"learning_rate": 2.6134187480228666e-05,
"loss": 0.7685,
"step": 1127
},
{
"epoch": 0.49,
"learning_rate": 2.610035335408466e-05,
"loss": 0.8832,
"step": 1128
},
{
"epoch": 0.49,
"learning_rate": 2.6066517208503453e-05,
"loss": 0.7798,
"step": 1129
},
{
"epoch": 0.49,
"learning_rate": 2.603267910558324e-05,
"loss": 0.7087,
"step": 1130
},
{
"epoch": 0.49,
"learning_rate": 2.5998839107425834e-05,
"loss": 0.6909,
"step": 1131
},
{
"epoch": 0.49,
"learning_rate": 2.596499727613651e-05,
"loss": 0.6266,
"step": 1132
},
{
"epoch": 0.49,
"learning_rate": 2.593115367382391e-05,
"loss": 0.8294,
"step": 1133
},
{
"epoch": 0.49,
"learning_rate": 2.589730836259992e-05,
"loss": 0.8246,
"step": 1134
},
{
"epoch": 0.49,
"learning_rate": 2.5863461404579575e-05,
"loss": 0.8762,
"step": 1135
},
{
"epoch": 0.49,
"learning_rate": 2.582961286188094e-05,
"loss": 0.7673,
"step": 1136
},
{
"epoch": 0.49,
"learning_rate": 2.579576279662495e-05,
"loss": 1.0006,
"step": 1137
},
{
"epoch": 0.49,
"learning_rate": 2.576191127093538e-05,
"loss": 0.7947,
"step": 1138
},
{
"epoch": 0.49,
"learning_rate": 2.5728058346938644e-05,
"loss": 0.7919,
"step": 1139
},
{
"epoch": 0.49,
"learning_rate": 2.5694204086763766e-05,
"loss": 0.7844,
"step": 1140
},
{
"epoch": 0.49,
"learning_rate": 2.5660348552542174e-05,
"loss": 0.7466,
"step": 1141
},
{
"epoch": 0.49,
"learning_rate": 2.5626491806407683e-05,
"loss": 0.7931,
"step": 1142
},
{
"epoch": 0.49,
"learning_rate": 2.5592633910496288e-05,
"loss": 0.6076,
"step": 1143
},
{
"epoch": 0.49,
"learning_rate": 2.5558774926946127e-05,
"loss": 0.7598,
"step": 1144
},
{
"epoch": 0.49,
"learning_rate": 2.5524914917897325e-05,
"loss": 0.6587,
"step": 1145
},
{
"epoch": 0.49,
"learning_rate": 2.5491053945491866e-05,
"loss": 0.7519,
"step": 1146
},
{
"epoch": 0.49,
"learning_rate": 2.5457192071873546e-05,
"loss": 0.7876,
"step": 1147
},
{
"epoch": 0.49,
"learning_rate": 2.542332935918777e-05,
"loss": 0.7594,
"step": 1148
},
{
"epoch": 0.5,
"learning_rate": 2.5389465869581512e-05,
"loss": 0.8365,
"step": 1149
},
{
"epoch": 0.5,
"learning_rate": 2.5355601665203167e-05,
"loss": 0.7155,
"step": 1150
},
{
"epoch": 0.5,
"learning_rate": 2.5321736808202433e-05,
"loss": 0.8049,
"step": 1151
},
{
"epoch": 0.5,
"learning_rate": 2.528787136073022e-05,
"loss": 0.7455,
"step": 1152
},
{
"epoch": 0.5,
"learning_rate": 2.5254005384938496e-05,
"loss": 0.7744,
"step": 1153
},
{
"epoch": 0.5,
"learning_rate": 2.5220138942980225e-05,
"loss": 0.7778,
"step": 1154
},
{
"epoch": 0.5,
"learning_rate": 2.5186272097009216e-05,
"loss": 0.8424,
"step": 1155
},
{
"epoch": 0.5,
"learning_rate": 2.5152404909180023e-05,
"loss": 0.7876,
"step": 1156
},
{
"epoch": 0.5,
"learning_rate": 2.5118537441647815e-05,
"loss": 0.7195,
"step": 1157
},
{
"epoch": 0.5,
"learning_rate": 2.5084669756568296e-05,
"loss": 0.8405,
"step": 1158
},
{
"epoch": 0.5,
"learning_rate": 2.505080191609755e-05,
"loss": 0.5727,
"step": 1159
},
{
"epoch": 0.5,
"learning_rate": 2.5016933982391954e-05,
"loss": 0.7626,
"step": 1160
},
{
"epoch": 0.5,
"learning_rate": 2.4983066017608048e-05,
"loss": 0.7927,
"step": 1161
},
{
"epoch": 0.5,
"learning_rate": 2.494919808390246e-05,
"loss": 0.9519,
"step": 1162
},
{
"epoch": 0.5,
"learning_rate": 2.4915330243431713e-05,
"loss": 0.6401,
"step": 1163
},
{
"epoch": 0.5,
"learning_rate": 2.488146255835219e-05,
"loss": 0.7766,
"step": 1164
},
{
"epoch": 0.5,
"learning_rate": 2.484759509081998e-05,
"loss": 0.8009,
"step": 1165
},
{
"epoch": 0.5,
"learning_rate": 2.4813727902990786e-05,
"loss": 0.676,
"step": 1166
},
{
"epoch": 0.5,
"learning_rate": 2.4779861057019784e-05,
"loss": 0.7808,
"step": 1167
},
{
"epoch": 0.5,
"learning_rate": 2.474599461506151e-05,
"loss": 0.6557,
"step": 1168
},
{
"epoch": 0.5,
"learning_rate": 2.471212863926979e-05,
"loss": 0.7341,
"step": 1169
},
{
"epoch": 0.5,
"learning_rate": 2.467826319179757e-05,
"loss": 0.7867,
"step": 1170
},
{
"epoch": 0.5,
"learning_rate": 2.4644398334796842e-05,
"loss": 0.8283,
"step": 1171
},
{
"epoch": 0.51,
"learning_rate": 2.4610534130418487e-05,
"loss": 0.8258,
"step": 1172
},
{
"epoch": 0.51,
"learning_rate": 2.457667064081224e-05,
"loss": 0.7862,
"step": 1173
},
{
"epoch": 0.51,
"learning_rate": 2.4542807928126466e-05,
"loss": 0.6769,
"step": 1174
},
{
"epoch": 0.51,
"learning_rate": 2.4508946054508136e-05,
"loss": 0.8143,
"step": 1175
},
{
"epoch": 0.51,
"learning_rate": 2.4475085082102685e-05,
"loss": 0.8902,
"step": 1176
},
{
"epoch": 0.51,
"learning_rate": 2.4441225073053875e-05,
"loss": 0.6589,
"step": 1177
},
{
"epoch": 0.51,
"learning_rate": 2.440736608950372e-05,
"loss": 0.7613,
"step": 1178
},
{
"epoch": 0.51,
"learning_rate": 2.4373508193592326e-05,
"loss": 0.7117,
"step": 1179
},
{
"epoch": 0.51,
"learning_rate": 2.4339651447457832e-05,
"loss": 0.7171,
"step": 1180
},
{
"epoch": 0.51,
"learning_rate": 2.4305795913236246e-05,
"loss": 0.6979,
"step": 1181
},
{
"epoch": 0.51,
"learning_rate": 2.4271941653061355e-05,
"loss": 0.724,
"step": 1182
},
{
"epoch": 0.51,
"learning_rate": 2.4238088729064627e-05,
"loss": 0.871,
"step": 1183
},
{
"epoch": 0.51,
"learning_rate": 2.4204237203375056e-05,
"loss": 0.8349,
"step": 1184
},
{
"epoch": 0.51,
"learning_rate": 2.4170387138119072e-05,
"loss": 0.7212,
"step": 1185
},
{
"epoch": 0.51,
"learning_rate": 2.413653859542042e-05,
"loss": 0.7457,
"step": 1186
},
{
"epoch": 0.51,
"learning_rate": 2.4102691637400083e-05,
"loss": 0.7321,
"step": 1187
},
{
"epoch": 0.51,
"learning_rate": 2.4068846326176096e-05,
"loss": 0.7771,
"step": 1188
},
{
"epoch": 0.51,
"learning_rate": 2.403500272386349e-05,
"loss": 0.8963,
"step": 1189
},
{
"epoch": 0.51,
"learning_rate": 2.400116089257417e-05,
"loss": 0.7219,
"step": 1190
},
{
"epoch": 0.51,
"learning_rate": 2.3967320894416765e-05,
"loss": 0.7724,
"step": 1191
},
{
"epoch": 0.51,
"learning_rate": 2.3933482791496556e-05,
"loss": 0.6775,
"step": 1192
},
{
"epoch": 0.51,
"learning_rate": 2.389964664591534e-05,
"loss": 0.8789,
"step": 1193
},
{
"epoch": 0.51,
"learning_rate": 2.3865812519771337e-05,
"loss": 0.8096,
"step": 1194
},
{
"epoch": 0.51,
"learning_rate": 2.3831980475159043e-05,
"loss": 0.7285,
"step": 1195
},
{
"epoch": 0.52,
"learning_rate": 2.3798150574169135e-05,
"loss": 0.6603,
"step": 1196
},
{
"epoch": 0.52,
"learning_rate": 2.3764322878888354e-05,
"loss": 0.7982,
"step": 1197
},
{
"epoch": 0.52,
"learning_rate": 2.3730497451399417e-05,
"loss": 0.8034,
"step": 1198
},
{
"epoch": 0.52,
"learning_rate": 2.3696674353780853e-05,
"loss": 0.6944,
"step": 1199
},
{
"epoch": 0.52,
"learning_rate": 2.3662853648106916e-05,
"loss": 0.638,
"step": 1200
},
{
"epoch": 0.52,
"eval_loss": 0.7718954682350159,
"eval_runtime": 10.1807,
"eval_samples_per_second": 3.634,
"eval_steps_per_second": 0.491,
"step": 1200
},
{
"epoch": 0.52,
"learning_rate": 2.362903539644749e-05,
"loss": 0.7747,
"step": 1201
},
{
"epoch": 0.52,
"learning_rate": 2.359521966086794e-05,
"loss": 0.7324,
"step": 1202
},
{
"epoch": 0.52,
"learning_rate": 2.356140650342901e-05,
"loss": 0.7613,
"step": 1203
},
{
"epoch": 0.52,
"learning_rate": 2.352759598618672e-05,
"loss": 0.7137,
"step": 1204
},
{
"epoch": 0.52,
"learning_rate": 2.3493788171192242e-05,
"loss": 0.6523,
"step": 1205
},
{
"epoch": 0.52,
"learning_rate": 2.3459983120491808e-05,
"loss": 0.7382,
"step": 1206
},
{
"epoch": 0.52,
"learning_rate": 2.342618089612652e-05,
"loss": 0.7613,
"step": 1207
},
{
"epoch": 0.52,
"learning_rate": 2.3392381560132376e-05,
"loss": 0.6524,
"step": 1208
},
{
"epoch": 0.52,
"learning_rate": 2.3358585174539998e-05,
"loss": 0.8847,
"step": 1209
},
{
"epoch": 0.52,
"learning_rate": 2.332479180137464e-05,
"loss": 0.683,
"step": 1210
},
{
"epoch": 0.52,
"learning_rate": 2.3291001502655992e-05,
"loss": 0.9041,
"step": 1211
},
{
"epoch": 0.52,
"learning_rate": 2.3257214340398143e-05,
"loss": 0.6465,
"step": 1212
},
{
"epoch": 0.52,
"learning_rate": 2.32234303766094e-05,
"loss": 0.7058,
"step": 1213
},
{
"epoch": 0.52,
"learning_rate": 2.318964967329218e-05,
"loss": 0.7959,
"step": 1214
},
{
"epoch": 0.52,
"learning_rate": 2.315587229244297e-05,
"loss": 0.7706,
"step": 1215
},
{
"epoch": 0.52,
"learning_rate": 2.3122098296052114e-05,
"loss": 0.8048,
"step": 1216
},
{
"epoch": 0.52,
"learning_rate": 2.3088327746103765e-05,
"loss": 0.7886,
"step": 1217
},
{
"epoch": 0.52,
"learning_rate": 2.3054560704575733e-05,
"loss": 0.7493,
"step": 1218
},
{
"epoch": 0.53,
"learning_rate": 2.302079723343942e-05,
"loss": 0.8783,
"step": 1219
},
{
"epoch": 0.53,
"learning_rate": 2.2987037394659647e-05,
"loss": 0.7896,
"step": 1220
},
{
"epoch": 0.53,
"learning_rate": 2.2953281250194573e-05,
"loss": 0.8975,
"step": 1221
},
{
"epoch": 0.53,
"learning_rate": 2.29195288619956e-05,
"loss": 0.7124,
"step": 1222
},
{
"epoch": 0.53,
"learning_rate": 2.2885780292007208e-05,
"loss": 0.7599,
"step": 1223
},
{
"epoch": 0.53,
"learning_rate": 2.2852035602166886e-05,
"loss": 0.6605,
"step": 1224
},
{
"epoch": 0.53,
"learning_rate": 2.281829485440498e-05,
"loss": 0.8076,
"step": 1225
},
{
"epoch": 0.53,
"learning_rate": 2.278455811064464e-05,
"loss": 0.7865,
"step": 1226
},
{
"epoch": 0.53,
"learning_rate": 2.2750825432801633e-05,
"loss": 0.7041,
"step": 1227
},
{
"epoch": 0.53,
"learning_rate": 2.2717096882784275e-05,
"loss": 0.7624,
"step": 1228
},
{
"epoch": 0.53,
"learning_rate": 2.2683372522493318e-05,
"loss": 0.8957,
"step": 1229
},
{
"epoch": 0.53,
"learning_rate": 2.2649652413821812e-05,
"loss": 0.9792,
"step": 1230
},
{
"epoch": 0.53,
"learning_rate": 2.2615936618655e-05,
"loss": 0.7144,
"step": 1231
},
{
"epoch": 0.53,
"learning_rate": 2.258222519887021e-05,
"loss": 0.723,
"step": 1232
},
{
"epoch": 0.53,
"learning_rate": 2.254851821633676e-05,
"loss": 0.8101,
"step": 1233
},
{
"epoch": 0.53,
"learning_rate": 2.2514815732915804e-05,
"loss": 0.6864,
"step": 1234
},
{
"epoch": 0.53,
"learning_rate": 2.2481117810460247e-05,
"loss": 0.9122,
"step": 1235
},
{
"epoch": 0.53,
"learning_rate": 2.2447424510814603e-05,
"loss": 0.6056,
"step": 1236
},
{
"epoch": 0.53,
"learning_rate": 2.241373589581495e-05,
"loss": 0.8158,
"step": 1237
},
{
"epoch": 0.53,
"learning_rate": 2.2380052027288723e-05,
"loss": 0.7081,
"step": 1238
},
{
"epoch": 0.53,
"learning_rate": 2.2346372967054653e-05,
"loss": 0.8738,
"step": 1239
},
{
"epoch": 0.53,
"learning_rate": 2.231269877692267e-05,
"loss": 0.6473,
"step": 1240
},
{
"epoch": 0.53,
"learning_rate": 2.227902951869375e-05,
"loss": 0.7287,
"step": 1241
},
{
"epoch": 0.54,
"learning_rate": 2.2245365254159808e-05,
"loss": 0.6395,
"step": 1242
},
{
"epoch": 0.54,
"learning_rate": 2.221170604510361e-05,
"loss": 0.7358,
"step": 1243
},
{
"epoch": 0.54,
"learning_rate": 2.217805195329864e-05,
"loss": 0.7528,
"step": 1244
},
{
"epoch": 0.54,
"learning_rate": 2.214440304050899e-05,
"loss": 0.6483,
"step": 1245
},
{
"epoch": 0.54,
"learning_rate": 2.2110759368489242e-05,
"loss": 0.632,
"step": 1246
},
{
"epoch": 0.54,
"learning_rate": 2.2077120998984373e-05,
"loss": 0.8643,
"step": 1247
},
{
"epoch": 0.54,
"learning_rate": 2.204348799372962e-05,
"loss": 0.7454,
"step": 1248
},
{
"epoch": 0.54,
"learning_rate": 2.2009860414450363e-05,
"loss": 0.8798,
"step": 1249
},
{
"epoch": 0.54,
"learning_rate": 2.1976238322862043e-05,
"loss": 0.7703,
"step": 1250
},
{
"epoch": 0.54,
"learning_rate": 2.1942621780670023e-05,
"loss": 0.7722,
"step": 1251
},
{
"epoch": 0.54,
"learning_rate": 2.1909010849569483e-05,
"loss": 0.8465,
"step": 1252
},
{
"epoch": 0.54,
"learning_rate": 2.1875405591245292e-05,
"loss": 0.7531,
"step": 1253
},
{
"epoch": 0.54,
"learning_rate": 2.1841806067371938e-05,
"loss": 0.7937,
"step": 1254
},
{
"epoch": 0.54,
"learning_rate": 2.1808212339613352e-05,
"loss": 0.7941,
"step": 1255
},
{
"epoch": 0.54,
"learning_rate": 2.1774624469622844e-05,
"loss": 0.9063,
"step": 1256
},
{
"epoch": 0.54,
"learning_rate": 2.1741042519042968e-05,
"loss": 0.705,
"step": 1257
},
{
"epoch": 0.54,
"learning_rate": 2.1707466549505427e-05,
"loss": 0.7492,
"step": 1258
},
{
"epoch": 0.54,
"learning_rate": 2.1673896622630936e-05,
"loss": 0.6582,
"step": 1259
},
{
"epoch": 0.54,
"learning_rate": 2.1640332800029105e-05,
"loss": 0.7255,
"step": 1260
},
{
"epoch": 0.54,
"learning_rate": 2.1606775143298377e-05,
"loss": 0.8809,
"step": 1261
},
{
"epoch": 0.54,
"learning_rate": 2.1573223714025852e-05,
"loss": 0.8045,
"step": 1262
},
{
"epoch": 0.54,
"learning_rate": 2.1539678573787216e-05,
"loss": 0.7698,
"step": 1263
},
{
"epoch": 0.54,
"learning_rate": 2.1506139784146583e-05,
"loss": 0.7583,
"step": 1264
},
{
"epoch": 0.55,
"learning_rate": 2.147260740665647e-05,
"loss": 0.7882,
"step": 1265
},
{
"epoch": 0.55,
"learning_rate": 2.143908150285756e-05,
"loss": 0.6991,
"step": 1266
},
{
"epoch": 0.55,
"learning_rate": 2.1405562134278703e-05,
"loss": 0.811,
"step": 1267
},
{
"epoch": 0.55,
"learning_rate": 2.1372049362436723e-05,
"loss": 0.746,
"step": 1268
},
{
"epoch": 0.55,
"learning_rate": 2.133854324883637e-05,
"loss": 0.7879,
"step": 1269
},
{
"epoch": 0.55,
"learning_rate": 2.130504385497015e-05,
"loss": 0.6192,
"step": 1270
},
{
"epoch": 0.55,
"learning_rate": 2.1271551242318224e-05,
"loss": 0.7478,
"step": 1271
},
{
"epoch": 0.55,
"learning_rate": 2.123806547234836e-05,
"loss": 0.6913,
"step": 1272
},
{
"epoch": 0.55,
"learning_rate": 2.1204586606515718e-05,
"loss": 0.7214,
"step": 1273
},
{
"epoch": 0.55,
"learning_rate": 2.11711147062628e-05,
"loss": 0.6534,
"step": 1274
},
{
"epoch": 0.55,
"learning_rate": 2.1137649833019326e-05,
"loss": 0.6412,
"step": 1275
},
{
"epoch": 0.55,
"learning_rate": 2.110419204820214e-05,
"loss": 0.6778,
"step": 1276
},
{
"epoch": 0.55,
"learning_rate": 2.107074141321504e-05,
"loss": 0.8778,
"step": 1277
},
{
"epoch": 0.55,
"learning_rate": 2.103729798944872e-05,
"loss": 0.7754,
"step": 1278
},
{
"epoch": 0.55,
"learning_rate": 2.100386183828065e-05,
"loss": 0.8971,
"step": 1279
},
{
"epoch": 0.55,
"learning_rate": 2.0970433021074947e-05,
"loss": 0.9264,
"step": 1280
},
{
"epoch": 0.55,
"learning_rate": 2.0937011599182264e-05,
"loss": 0.6872,
"step": 1281
},
{
"epoch": 0.55,
"learning_rate": 2.0903597633939654e-05,
"loss": 0.9361,
"step": 1282
},
{
"epoch": 0.55,
"learning_rate": 2.087019118667055e-05,
"loss": 0.6723,
"step": 1283
},
{
"epoch": 0.55,
"learning_rate": 2.0836792318684533e-05,
"loss": 0.6403,
"step": 1284
},
{
"epoch": 0.55,
"learning_rate": 2.080340109127729e-05,
"loss": 0.6656,
"step": 1285
},
{
"epoch": 0.55,
"learning_rate": 2.0770017565730493e-05,
"loss": 0.7191,
"step": 1286
},
{
"epoch": 0.55,
"learning_rate": 2.073664180331168e-05,
"loss": 0.9078,
"step": 1287
},
{
"epoch": 0.56,
"learning_rate": 2.070327386527412e-05,
"loss": 0.7715,
"step": 1288
},
{
"epoch": 0.56,
"learning_rate": 2.066991381285674e-05,
"loss": 0.8872,
"step": 1289
},
{
"epoch": 0.56,
"learning_rate": 2.0636561707284e-05,
"loss": 0.6862,
"step": 1290
},
{
"epoch": 0.56,
"learning_rate": 2.060321760976576e-05,
"loss": 0.7982,
"step": 1291
},
{
"epoch": 0.56,
"learning_rate": 2.0569881581497187e-05,
"loss": 0.6791,
"step": 1292
},
{
"epoch": 0.56,
"learning_rate": 2.0536553683658656e-05,
"loss": 0.7163,
"step": 1293
},
{
"epoch": 0.56,
"learning_rate": 2.0503233977415608e-05,
"loss": 0.9087,
"step": 1294
},
{
"epoch": 0.56,
"learning_rate": 2.0469922523918432e-05,
"loss": 0.8373,
"step": 1295
},
{
"epoch": 0.56,
"learning_rate": 2.043661938430239e-05,
"loss": 0.77,
"step": 1296
},
{
"epoch": 0.56,
"learning_rate": 2.0403324619687502e-05,
"loss": 0.7141,
"step": 1297
},
{
"epoch": 0.56,
"learning_rate": 2.037003829117839e-05,
"loss": 0.7237,
"step": 1298
},
{
"epoch": 0.56,
"learning_rate": 2.0336760459864194e-05,
"loss": 1.0225,
"step": 1299
},
{
"epoch": 0.56,
"learning_rate": 2.0303491186818475e-05,
"loss": 0.8124,
"step": 1300
},
{
"epoch": 0.56,
"eval_loss": 0.7701284289360046,
"eval_runtime": 10.1766,
"eval_samples_per_second": 3.636,
"eval_steps_per_second": 0.491,
"step": 1300
},
{
"epoch": 0.56,
"learning_rate": 2.0270230533099084e-05,
"loss": 0.7718,
"step": 1301
},
{
"epoch": 0.56,
"learning_rate": 2.023697855974805e-05,
"loss": 0.7646,
"step": 1302
},
{
"epoch": 0.56,
"learning_rate": 2.0203735327791468e-05,
"loss": 0.7243,
"step": 1303
},
{
"epoch": 0.56,
"learning_rate": 2.01705008982394e-05,
"loss": 0.8042,
"step": 1304
},
{
"epoch": 0.56,
"learning_rate": 2.013727533208574e-05,
"loss": 0.6866,
"step": 1305
},
{
"epoch": 0.56,
"learning_rate": 2.0104058690308126e-05,
"loss": 0.7828,
"step": 1306
},
{
"epoch": 0.56,
"learning_rate": 2.0070851033867807e-05,
"loss": 0.7398,
"step": 1307
},
{
"epoch": 0.56,
"learning_rate": 2.0037652423709558e-05,
"loss": 0.8089,
"step": 1308
},
{
"epoch": 0.56,
"learning_rate": 2.0004462920761542e-05,
"loss": 0.7142,
"step": 1309
},
{
"epoch": 0.56,
"learning_rate": 1.997128258593519e-05,
"loss": 0.8548,
"step": 1310
},
{
"epoch": 0.56,
"learning_rate": 1.9938111480125157e-05,
"loss": 0.7748,
"step": 1311
},
{
"epoch": 0.57,
"learning_rate": 1.9904949664209106e-05,
"loss": 0.7832,
"step": 1312
},
{
"epoch": 0.57,
"learning_rate": 1.987179719904767e-05,
"loss": 0.739,
"step": 1313
},
{
"epoch": 0.57,
"learning_rate": 1.983865414548433e-05,
"loss": 0.7557,
"step": 1314
},
{
"epoch": 0.57,
"learning_rate": 1.9805520564345293e-05,
"loss": 0.7588,
"step": 1315
},
{
"epoch": 0.57,
"learning_rate": 1.977239651643938e-05,
"loss": 0.8288,
"step": 1316
},
{
"epoch": 0.57,
"learning_rate": 1.9739282062557886e-05,
"loss": 0.7496,
"step": 1317
},
{
"epoch": 0.57,
"learning_rate": 1.9706177263474563e-05,
"loss": 0.6867,
"step": 1318
},
{
"epoch": 0.57,
"learning_rate": 1.9673082179945384e-05,
"loss": 0.7227,
"step": 1319
},
{
"epoch": 0.57,
"learning_rate": 1.9639996872708512e-05,
"loss": 0.6423,
"step": 1320
},
{
"epoch": 0.57,
"learning_rate": 1.9606921402484176e-05,
"loss": 0.777,
"step": 1321
},
{
"epoch": 0.57,
"learning_rate": 1.9573855829974547e-05,
"loss": 0.7567,
"step": 1322
},
{
"epoch": 0.57,
"learning_rate": 1.9540800215863624e-05,
"loss": 0.7351,
"step": 1323
},
{
"epoch": 0.57,
"learning_rate": 1.9507754620817133e-05,
"loss": 0.59,
"step": 1324
},
{
"epoch": 0.57,
"learning_rate": 1.9474719105482416e-05,
"loss": 0.7092,
"step": 1325
},
{
"epoch": 0.57,
"learning_rate": 1.944169373048832e-05,
"loss": 0.6918,
"step": 1326
},
{
"epoch": 0.57,
"learning_rate": 1.9408678556445077e-05,
"loss": 0.8209,
"step": 1327
},
{
"epoch": 0.57,
"learning_rate": 1.937567364394417e-05,
"loss": 0.8695,
"step": 1328
},
{
"epoch": 0.57,
"learning_rate": 1.934267905355832e-05,
"loss": 0.6717,
"step": 1329
},
{
"epoch": 0.57,
"learning_rate": 1.9309694845841225e-05,
"loss": 0.7827,
"step": 1330
},
{
"epoch": 0.57,
"learning_rate": 1.9276721081327576e-05,
"loss": 0.8805,
"step": 1331
},
{
"epoch": 0.57,
"learning_rate": 1.9243757820532887e-05,
"loss": 0.8435,
"step": 1332
},
{
"epoch": 0.57,
"learning_rate": 1.9210805123953395e-05,
"loss": 0.8245,
"step": 1333
},
{
"epoch": 0.57,
"learning_rate": 1.917786305206594e-05,
"loss": 0.9219,
"step": 1334
},
{
"epoch": 0.58,
"learning_rate": 1.9144931665327868e-05,
"loss": 0.8257,
"step": 1335
},
{
"epoch": 0.58,
"learning_rate": 1.9112011024176922e-05,
"loss": 0.9376,
"step": 1336
},
{
"epoch": 0.58,
"learning_rate": 1.9079101189031124e-05,
"loss": 0.5985,
"step": 1337
},
{
"epoch": 0.58,
"learning_rate": 1.9046202220288658e-05,
"loss": 0.7893,
"step": 1338
},
{
"epoch": 0.58,
"learning_rate": 1.9013314178327745e-05,
"loss": 0.7461,
"step": 1339
},
{
"epoch": 0.58,
"learning_rate": 1.8980437123506606e-05,
"loss": 0.8159,
"step": 1340
},
{
"epoch": 0.58,
"learning_rate": 1.894757111616325e-05,
"loss": 0.7261,
"step": 1341
},
{
"epoch": 0.58,
"learning_rate": 1.8914716216615416e-05,
"loss": 0.715,
"step": 1342
},
{
"epoch": 0.58,
"learning_rate": 1.8881872485160485e-05,
"loss": 0.8674,
"step": 1343
},
{
"epoch": 0.58,
"learning_rate": 1.884903998207532e-05,
"loss": 0.6603,
"step": 1344
},
{
"epoch": 0.58,
"learning_rate": 1.8816218767616185e-05,
"loss": 0.6988,
"step": 1345
},
{
"epoch": 0.58,
"learning_rate": 1.8783408902018605e-05,
"loss": 0.7279,
"step": 1346
},
{
"epoch": 0.58,
"learning_rate": 1.8750610445497314e-05,
"loss": 0.7453,
"step": 1347
},
{
"epoch": 0.58,
"learning_rate": 1.871782345824608e-05,
"loss": 0.9567,
"step": 1348
},
{
"epoch": 0.58,
"learning_rate": 1.868504800043763e-05,
"loss": 0.6403,
"step": 1349
},
{
"epoch": 0.58,
"learning_rate": 1.8652284132223534e-05,
"loss": 0.8241,
"step": 1350
},
{
"epoch": 0.58,
"learning_rate": 1.8619531913734097e-05,
"loss": 0.8506,
"step": 1351
},
{
"epoch": 0.58,
"learning_rate": 1.8586791405078218e-05,
"loss": 0.7239,
"step": 1352
},
{
"epoch": 0.58,
"learning_rate": 1.8554062666343332e-05,
"loss": 0.729,
"step": 1353
},
{
"epoch": 0.58,
"learning_rate": 1.8521345757595265e-05,
"loss": 0.8423,
"step": 1354
},
{
"epoch": 0.58,
"learning_rate": 1.8488640738878136e-05,
"loss": 0.7938,
"step": 1355
},
{
"epoch": 0.58,
"learning_rate": 1.8455947670214226e-05,
"loss": 0.6189,
"step": 1356
},
{
"epoch": 0.58,
"learning_rate": 1.8423266611603918e-05,
"loss": 0.6484,
"step": 1357
},
{
"epoch": 0.59,
"learning_rate": 1.839059762302551e-05,
"loss": 0.7402,
"step": 1358
},
{
"epoch": 0.59,
"learning_rate": 1.8357940764435182e-05,
"loss": 0.7406,
"step": 1359
},
{
"epoch": 0.59,
"learning_rate": 1.832529609576683e-05,
"loss": 0.7281,
"step": 1360
},
{
"epoch": 0.59,
"learning_rate": 1.8292663676932005e-05,
"loss": 0.7911,
"step": 1361
},
{
"epoch": 0.59,
"learning_rate": 1.8260043567819757e-05,
"loss": 0.6116,
"step": 1362
},
{
"epoch": 0.59,
"learning_rate": 1.8227435828296524e-05,
"loss": 0.688,
"step": 1363
},
{
"epoch": 0.59,
"learning_rate": 1.8194840518206104e-05,
"loss": 0.8594,
"step": 1364
},
{
"epoch": 0.59,
"learning_rate": 1.816225769736942e-05,
"loss": 0.7366,
"step": 1365
},
{
"epoch": 0.59,
"learning_rate": 1.8129687425584513e-05,
"loss": 0.6557,
"step": 1366
},
{
"epoch": 0.59,
"learning_rate": 1.8097129762626364e-05,
"loss": 0.8365,
"step": 1367
},
{
"epoch": 0.59,
"learning_rate": 1.806458476824685e-05,
"loss": 0.7672,
"step": 1368
},
{
"epoch": 0.59,
"learning_rate": 1.8032052502174564e-05,
"loss": 0.7921,
"step": 1369
},
{
"epoch": 0.59,
"learning_rate": 1.7999533024114757e-05,
"loss": 0.7398,
"step": 1370
},
{
"epoch": 0.59,
"learning_rate": 1.79670263937492e-05,
"loss": 0.7764,
"step": 1371
},
{
"epoch": 0.59,
"learning_rate": 1.79345326707361e-05,
"loss": 0.8442,
"step": 1372
},
{
"epoch": 0.59,
"learning_rate": 1.790205191470997e-05,
"loss": 0.7946,
"step": 1373
},
{
"epoch": 0.59,
"learning_rate": 1.7869584185281503e-05,
"loss": 0.7367,
"step": 1374
},
{
"epoch": 0.59,
"learning_rate": 1.7837129542037533e-05,
"loss": 0.7633,
"step": 1375
},
{
"epoch": 0.59,
"learning_rate": 1.780468804454083e-05,
"loss": 0.7926,
"step": 1376
},
{
"epoch": 0.59,
"learning_rate": 1.7772259752330056e-05,
"loss": 0.8245,
"step": 1377
},
{
"epoch": 0.59,
"learning_rate": 1.7739844724919636e-05,
"loss": 0.8456,
"step": 1378
},
{
"epoch": 0.59,
"learning_rate": 1.7707443021799664e-05,
"loss": 0.9084,
"step": 1379
},
{
"epoch": 0.59,
"learning_rate": 1.7675054702435774e-05,
"loss": 0.7188,
"step": 1380
},
{
"epoch": 0.6,
"learning_rate": 1.764267982626901e-05,
"loss": 0.775,
"step": 1381
},
{
"epoch": 0.6,
"learning_rate": 1.7610318452715782e-05,
"loss": 0.7102,
"step": 1382
},
{
"epoch": 0.6,
"learning_rate": 1.7577970641167706e-05,
"loss": 0.7563,
"step": 1383
},
{
"epoch": 0.6,
"learning_rate": 1.7545636450991507e-05,
"loss": 0.7591,
"step": 1384
},
{
"epoch": 0.6,
"learning_rate": 1.751331594152889e-05,
"loss": 0.841,
"step": 1385
},
{
"epoch": 0.6,
"learning_rate": 1.74810091720965e-05,
"loss": 0.8448,
"step": 1386
},
{
"epoch": 0.6,
"learning_rate": 1.7448716201985728e-05,
"loss": 0.6866,
"step": 1387
},
{
"epoch": 0.6,
"learning_rate": 1.7416437090462644e-05,
"loss": 0.7427,
"step": 1388
},
{
"epoch": 0.6,
"learning_rate": 1.7384171896767896e-05,
"loss": 0.7489,
"step": 1389
},
{
"epoch": 0.6,
"learning_rate": 1.735192068011658e-05,
"loss": 0.5422,
"step": 1390
},
{
"epoch": 0.6,
"learning_rate": 1.7319683499698152e-05,
"loss": 0.8508,
"step": 1391
},
{
"epoch": 0.6,
"learning_rate": 1.7287460414676283e-05,
"loss": 0.8014,
"step": 1392
},
{
"epoch": 0.6,
"learning_rate": 1.72552514841888e-05,
"loss": 0.7901,
"step": 1393
},
{
"epoch": 0.6,
"learning_rate": 1.7223056767347548e-05,
"loss": 0.6313,
"step": 1394
},
{
"epoch": 0.6,
"learning_rate": 1.719087632323827e-05,
"loss": 0.7272,
"step": 1395
},
{
"epoch": 0.6,
"learning_rate": 1.715871021092054e-05,
"loss": 0.7253,
"step": 1396
},
{
"epoch": 0.6,
"learning_rate": 1.7126558489427622e-05,
"loss": 0.6508,
"step": 1397
},
{
"epoch": 0.6,
"learning_rate": 1.7094421217766344e-05,
"loss": 0.8422,
"step": 1398
},
{
"epoch": 0.6,
"learning_rate": 1.706229845491704e-05,
"loss": 0.801,
"step": 1399
},
{
"epoch": 0.6,
"learning_rate": 1.703019025983342e-05,
"loss": 0.7436,
"step": 1400
},
{
"epoch": 0.6,
"eval_loss": 0.7689833641052246,
"eval_runtime": 10.1586,
"eval_samples_per_second": 3.642,
"eval_steps_per_second": 0.492,
"step": 1400
},
{
"epoch": 0.6,
"learning_rate": 1.699809669144244e-05,
"loss": 0.8388,
"step": 1401
},
{
"epoch": 0.6,
"learning_rate": 1.6966017808644225e-05,
"loss": 0.7115,
"step": 1402
},
{
"epoch": 0.6,
"learning_rate": 1.693395367031195e-05,
"loss": 0.8026,
"step": 1403
},
{
"epoch": 0.61,
"learning_rate": 1.6901904335291717e-05,
"loss": 0.6485,
"step": 1404
},
{
"epoch": 0.61,
"learning_rate": 1.6869869862402466e-05,
"loss": 0.6811,
"step": 1405
},
{
"epoch": 0.61,
"learning_rate": 1.683785031043586e-05,
"loss": 0.6476,
"step": 1406
},
{
"epoch": 0.61,
"learning_rate": 1.6805845738156194e-05,
"loss": 0.7066,
"step": 1407
},
{
"epoch": 0.61,
"learning_rate": 1.677385620430026e-05,
"loss": 0.8859,
"step": 1408
},
{
"epoch": 0.61,
"learning_rate": 1.6741881767577226e-05,
"loss": 0.7926,
"step": 1409
},
{
"epoch": 0.61,
"learning_rate": 1.6709922486668588e-05,
"loss": 0.7507,
"step": 1410
},
{
"epoch": 0.61,
"learning_rate": 1.667797842022802e-05,
"loss": 0.8807,
"step": 1411
},
{
"epoch": 0.61,
"learning_rate": 1.6646049626881265e-05,
"loss": 0.7969,
"step": 1412
},
{
"epoch": 0.61,
"learning_rate": 1.661413616522603e-05,
"loss": 0.7937,
"step": 1413
},
{
"epoch": 0.61,
"learning_rate": 1.6582238093831903e-05,
"loss": 0.845,
"step": 1414
},
{
"epoch": 0.61,
"learning_rate": 1.6550355471240224e-05,
"loss": 0.6704,
"step": 1415
},
{
"epoch": 0.61,
"learning_rate": 1.6518488355963958e-05,
"loss": 0.7718,
"step": 1416
},
{
"epoch": 0.61,
"learning_rate": 1.6486636806487626e-05,
"loss": 0.6275,
"step": 1417
},
{
"epoch": 0.61,
"learning_rate": 1.645480088126719e-05,
"loss": 0.8281,
"step": 1418
},
{
"epoch": 0.61,
"learning_rate": 1.642298063872993e-05,
"loss": 0.8532,
"step": 1419
},
{
"epoch": 0.61,
"learning_rate": 1.6391176137274324e-05,
"loss": 0.6332,
"step": 1420
},
{
"epoch": 0.61,
"learning_rate": 1.6359387435270007e-05,
"loss": 0.7622,
"step": 1421
},
{
"epoch": 0.61,
"learning_rate": 1.6327614591057576e-05,
"loss": 0.7457,
"step": 1422
},
{
"epoch": 0.61,
"learning_rate": 1.6295857662948534e-05,
"loss": 0.7234,
"step": 1423
},
{
"epoch": 0.61,
"learning_rate": 1.6264116709225186e-05,
"loss": 0.7721,
"step": 1424
},
{
"epoch": 0.61,
"learning_rate": 1.623239178814051e-05,
"loss": 0.7289,
"step": 1425
},
{
"epoch": 0.61,
"learning_rate": 1.6200682957918074e-05,
"loss": 0.7806,
"step": 1426
},
{
"epoch": 0.61,
"learning_rate": 1.6168990276751884e-05,
"loss": 0.6087,
"step": 1427
},
{
"epoch": 0.62,
"learning_rate": 1.6137313802806343e-05,
"loss": 0.6868,
"step": 1428
},
{
"epoch": 0.62,
"learning_rate": 1.610565359421609e-05,
"loss": 0.771,
"step": 1429
},
{
"epoch": 0.62,
"learning_rate": 1.6074009709085924e-05,
"loss": 0.7682,
"step": 1430
},
{
"epoch": 0.62,
"learning_rate": 1.604238220549066e-05,
"loss": 0.6348,
"step": 1431
},
{
"epoch": 0.62,
"learning_rate": 1.601077114147509e-05,
"loss": 0.794,
"step": 1432
},
{
"epoch": 0.62,
"learning_rate": 1.59791765750538e-05,
"loss": 0.8844,
"step": 1433
},
{
"epoch": 0.62,
"learning_rate": 1.5947598564211114e-05,
"loss": 0.8491,
"step": 1434
},
{
"epoch": 0.62,
"learning_rate": 1.5916037166900977e-05,
"loss": 0.8331,
"step": 1435
},
{
"epoch": 0.62,
"learning_rate": 1.5884492441046832e-05,
"loss": 0.6503,
"step": 1436
},
{
"epoch": 0.62,
"learning_rate": 1.5852964444541533e-05,
"loss": 0.7658,
"step": 1437
},
{
"epoch": 0.62,
"learning_rate": 1.5821453235247215e-05,
"loss": 0.8397,
"step": 1438
},
{
"epoch": 0.62,
"learning_rate": 1.5789958870995236e-05,
"loss": 0.8869,
"step": 1439
},
{
"epoch": 0.62,
"learning_rate": 1.575848140958601e-05,
"loss": 0.7899,
"step": 1440
},
{
"epoch": 0.62,
"learning_rate": 1.572702090878895e-05,
"loss": 0.7458,
"step": 1441
},
{
"epoch": 0.62,
"learning_rate": 1.5695577426342307e-05,
"loss": 0.6571,
"step": 1442
},
{
"epoch": 0.62,
"learning_rate": 1.5664151019953156e-05,
"loss": 0.8676,
"step": 1443
},
{
"epoch": 0.62,
"learning_rate": 1.5632741747297182e-05,
"loss": 0.8095,
"step": 1444
},
{
"epoch": 0.62,
"learning_rate": 1.560134966601865e-05,
"loss": 0.8415,
"step": 1445
},
{
"epoch": 0.62,
"learning_rate": 1.5569974833730265e-05,
"loss": 0.7099,
"step": 1446
},
{
"epoch": 0.62,
"learning_rate": 1.5538617308013083e-05,
"loss": 0.6655,
"step": 1447
},
{
"epoch": 0.62,
"learning_rate": 1.5507277146416398e-05,
"loss": 0.8824,
"step": 1448
},
{
"epoch": 0.62,
"learning_rate": 1.547595440645761e-05,
"loss": 0.8194,
"step": 1449
},
{
"epoch": 0.62,
"learning_rate": 1.54446491456222e-05,
"loss": 0.6864,
"step": 1450
},
{
"epoch": 0.63,
"learning_rate": 1.5413361421363516e-05,
"loss": 0.7061,
"step": 1451
},
{
"epoch": 0.63,
"learning_rate": 1.5382091291102745e-05,
"loss": 0.7927,
"step": 1452
},
{
"epoch": 0.63,
"learning_rate": 1.5350838812228797e-05,
"loss": 0.7616,
"step": 1453
},
{
"epoch": 0.63,
"learning_rate": 1.5319604042098167e-05,
"loss": 0.7953,
"step": 1454
},
{
"epoch": 0.63,
"learning_rate": 1.5288387038034848e-05,
"loss": 0.7014,
"step": 1455
},
{
"epoch": 0.63,
"learning_rate": 1.5257187857330235e-05,
"loss": 0.6782,
"step": 1456
},
{
"epoch": 0.63,
"learning_rate": 1.5226006557243024e-05,
"loss": 0.761,
"step": 1457
},
{
"epoch": 0.63,
"learning_rate": 1.5194843194999082e-05,
"loss": 0.9225,
"step": 1458
},
{
"epoch": 0.63,
"learning_rate": 1.516369782779135e-05,
"loss": 0.8315,
"step": 1459
},
{
"epoch": 0.63,
"learning_rate": 1.5132570512779764e-05,
"loss": 0.7894,
"step": 1460
},
{
"epoch": 0.63,
"learning_rate": 1.5101461307091113e-05,
"loss": 0.65,
"step": 1461
},
{
"epoch": 0.63,
"learning_rate": 1.507037026781895e-05,
"loss": 0.7349,
"step": 1462
},
{
"epoch": 0.63,
"learning_rate": 1.5039297452023485e-05,
"loss": 0.7232,
"step": 1463
},
{
"epoch": 0.63,
"learning_rate": 1.5008242916731505e-05,
"loss": 0.7363,
"step": 1464
},
{
"epoch": 0.63,
"learning_rate": 1.4977206718936232e-05,
"loss": 0.8518,
"step": 1465
},
{
"epoch": 0.63,
"learning_rate": 1.4946188915597215e-05,
"loss": 0.7308,
"step": 1466
},
{
"epoch": 0.63,
"learning_rate": 1.4915189563640291e-05,
"loss": 0.7339,
"step": 1467
},
{
"epoch": 0.63,
"learning_rate": 1.4884208719957396e-05,
"loss": 0.7194,
"step": 1468
},
{
"epoch": 0.63,
"learning_rate": 1.4853246441406504e-05,
"loss": 0.7901,
"step": 1469
},
{
"epoch": 0.63,
"learning_rate": 1.4822302784811524e-05,
"loss": 0.7507,
"step": 1470
},
{
"epoch": 0.63,
"learning_rate": 1.4791377806962193e-05,
"loss": 0.7285,
"step": 1471
},
{
"epoch": 0.63,
"learning_rate": 1.4760471564613964e-05,
"loss": 0.882,
"step": 1472
},
{
"epoch": 0.63,
"learning_rate": 1.4729584114487895e-05,
"loss": 0.688,
"step": 1473
},
{
"epoch": 0.64,
"learning_rate": 1.4698715513270565e-05,
"loss": 0.7817,
"step": 1474
},
{
"epoch": 0.64,
"learning_rate": 1.4667865817613963e-05,
"loss": 0.7811,
"step": 1475
},
{
"epoch": 0.64,
"learning_rate": 1.4637035084135381e-05,
"loss": 0.7831,
"step": 1476
},
{
"epoch": 0.64,
"learning_rate": 1.4606223369417288e-05,
"loss": 0.776,
"step": 1477
},
{
"epoch": 0.64,
"learning_rate": 1.4575430730007288e-05,
"loss": 0.8137,
"step": 1478
},
{
"epoch": 0.64,
"learning_rate": 1.4544657222417954e-05,
"loss": 0.7777,
"step": 1479
},
{
"epoch": 0.64,
"learning_rate": 1.4513902903126753e-05,
"loss": 0.8117,
"step": 1480
},
{
"epoch": 0.64,
"learning_rate": 1.4483167828575903e-05,
"loss": 0.8106,
"step": 1481
},
{
"epoch": 0.64,
"learning_rate": 1.4452452055172366e-05,
"loss": 0.7196,
"step": 1482
},
{
"epoch": 0.64,
"learning_rate": 1.4421755639287632e-05,
"loss": 0.8217,
"step": 1483
},
{
"epoch": 0.64,
"learning_rate": 1.4391078637257687e-05,
"loss": 0.4995,
"step": 1484
},
{
"epoch": 0.64,
"learning_rate": 1.4360421105382882e-05,
"loss": 0.6736,
"step": 1485
},
{
"epoch": 0.64,
"learning_rate": 1.4329783099927826e-05,
"loss": 0.7658,
"step": 1486
},
{
"epoch": 0.64,
"learning_rate": 1.4299164677121308e-05,
"loss": 0.8036,
"step": 1487
},
{
"epoch": 0.64,
"learning_rate": 1.4268565893156167e-05,
"loss": 0.6824,
"step": 1488
},
{
"epoch": 0.64,
"learning_rate": 1.4237986804189202e-05,
"loss": 0.63,
"step": 1489
},
{
"epoch": 0.64,
"learning_rate": 1.4207427466341073e-05,
"loss": 0.8837,
"step": 1490
},
{
"epoch": 0.64,
"learning_rate": 1.4176887935696164e-05,
"loss": 0.6741,
"step": 1491
},
{
"epoch": 0.64,
"learning_rate": 1.4146368268302573e-05,
"loss": 0.8185,
"step": 1492
},
{
"epoch": 0.64,
"learning_rate": 1.411586852017186e-05,
"loss": 0.723,
"step": 1493
},
{
"epoch": 0.64,
"learning_rate": 1.408538874727908e-05,
"loss": 0.7325,
"step": 1494
},
{
"epoch": 0.64,
"learning_rate": 1.4054929005562605e-05,
"loss": 0.6877,
"step": 1495
},
{
"epoch": 0.64,
"learning_rate": 1.402448935092408e-05,
"loss": 0.5648,
"step": 1496
},
{
"epoch": 0.65,
"learning_rate": 1.3994069839228261e-05,
"loss": 0.7489,
"step": 1497
},
{
"epoch": 0.65,
"learning_rate": 1.39636705263029e-05,
"loss": 0.7981,
"step": 1498
},
{
"epoch": 0.65,
"learning_rate": 1.3933291467938747e-05,
"loss": 0.8061,
"step": 1499
},
{
"epoch": 0.65,
"learning_rate": 1.3902932719889339e-05,
"loss": 0.6837,
"step": 1500
},
{
"epoch": 0.65,
"eval_loss": 0.7673142552375793,
"eval_runtime": 10.1791,
"eval_samples_per_second": 3.635,
"eval_steps_per_second": 0.491,
"step": 1500
},
{
"epoch": 0.65,
"learning_rate": 1.3872594337870945e-05,
"loss": 0.7132,
"step": 1501
},
{
"epoch": 0.65,
"learning_rate": 1.384227637756246e-05,
"loss": 0.8449,
"step": 1502
},
{
"epoch": 0.65,
"learning_rate": 1.3811978894605298e-05,
"loss": 0.7883,
"step": 1503
},
{
"epoch": 0.65,
"learning_rate": 1.3781701944603292e-05,
"loss": 0.7814,
"step": 1504
},
{
"epoch": 0.65,
"learning_rate": 1.3751445583122594e-05,
"loss": 0.7865,
"step": 1505
},
{
"epoch": 0.65,
"learning_rate": 1.3721209865691565e-05,
"loss": 0.7895,
"step": 1506
},
{
"epoch": 0.65,
"learning_rate": 1.3690994847800681e-05,
"loss": 0.8147,
"step": 1507
},
{
"epoch": 0.65,
"learning_rate": 1.3660800584902437e-05,
"loss": 0.833,
"step": 1508
},
{
"epoch": 0.65,
"learning_rate": 1.3630627132411214e-05,
"loss": 0.5843,
"step": 1509
},
{
"epoch": 0.65,
"learning_rate": 1.3600474545703248e-05,
"loss": 0.6556,
"step": 1510
},
{
"epoch": 0.65,
"learning_rate": 1.357034288011642e-05,
"loss": 0.5857,
"step": 1511
},
{
"epoch": 0.65,
"learning_rate": 1.3540232190950258e-05,
"loss": 0.7229,
"step": 1512
},
{
"epoch": 0.65,
"learning_rate": 1.3510142533465764e-05,
"loss": 0.7408,
"step": 1513
},
{
"epoch": 0.65,
"learning_rate": 1.3480073962885382e-05,
"loss": 0.7245,
"step": 1514
},
{
"epoch": 0.65,
"learning_rate": 1.3450026534392823e-05,
"loss": 0.7624,
"step": 1515
},
{
"epoch": 0.65,
"learning_rate": 1.3420000303132985e-05,
"loss": 0.7316,
"step": 1516
},
{
"epoch": 0.65,
"learning_rate": 1.3389995324211907e-05,
"loss": 0.8585,
"step": 1517
},
{
"epoch": 0.65,
"learning_rate": 1.3360011652696588e-05,
"loss": 0.7258,
"step": 1518
},
{
"epoch": 0.65,
"learning_rate": 1.3330049343614951e-05,
"loss": 0.7685,
"step": 1519
},
{
"epoch": 0.66,
"learning_rate": 1.3300108451955663e-05,
"loss": 0.7551,
"step": 1520
},
{
"epoch": 0.66,
"learning_rate": 1.327018903266815e-05,
"loss": 0.7523,
"step": 1521
},
{
"epoch": 0.66,
"learning_rate": 1.3240291140662392e-05,
"loss": 0.6846,
"step": 1522
},
{
"epoch": 0.66,
"learning_rate": 1.321041483080886e-05,
"loss": 0.7882,
"step": 1523
},
{
"epoch": 0.66,
"learning_rate": 1.3180560157938432e-05,
"loss": 0.8501,
"step": 1524
},
{
"epoch": 0.66,
"learning_rate": 1.3150727176842265e-05,
"loss": 0.7051,
"step": 1525
},
{
"epoch": 0.66,
"learning_rate": 1.3120915942271708e-05,
"loss": 0.6396,
"step": 1526
},
{
"epoch": 0.66,
"learning_rate": 1.3091126508938203e-05,
"loss": 0.6782,
"step": 1527
},
{
"epoch": 0.66,
"learning_rate": 1.306135893151318e-05,
"loss": 0.7321,
"step": 1528
},
{
"epoch": 0.66,
"learning_rate": 1.3031613264627951e-05,
"loss": 0.7165,
"step": 1529
},
{
"epoch": 0.66,
"learning_rate": 1.300188956287362e-05,
"loss": 0.7861,
"step": 1530
},
{
"epoch": 0.66,
"learning_rate": 1.2972187880800993e-05,
"loss": 0.7434,
"step": 1531
},
{
"epoch": 0.66,
"learning_rate": 1.2942508272920457e-05,
"loss": 0.7279,
"step": 1532
},
{
"epoch": 0.66,
"learning_rate": 1.2912850793701858e-05,
"loss": 0.8702,
"step": 1533
},
{
"epoch": 0.66,
"learning_rate": 1.2883215497574458e-05,
"loss": 0.7523,
"step": 1534
},
{
"epoch": 0.66,
"learning_rate": 1.2853602438926826e-05,
"loss": 0.8204,
"step": 1535
},
{
"epoch": 0.66,
"learning_rate": 1.2824011672106679e-05,
"loss": 0.7554,
"step": 1536
},
{
"epoch": 0.66,
"learning_rate": 1.2794443251420846e-05,
"loss": 0.8102,
"step": 1537
},
{
"epoch": 0.66,
"learning_rate": 1.2764897231135143e-05,
"loss": 0.672,
"step": 1538
},
{
"epoch": 0.66,
"learning_rate": 1.2735373665474262e-05,
"loss": 0.6154,
"step": 1539
},
{
"epoch": 0.66,
"learning_rate": 1.2705872608621706e-05,
"loss": 0.7744,
"step": 1540
},
{
"epoch": 0.66,
"learning_rate": 1.2676394114719647e-05,
"loss": 0.6868,
"step": 1541
},
{
"epoch": 0.66,
"learning_rate": 1.264693823786886e-05,
"loss": 0.7004,
"step": 1542
},
{
"epoch": 0.66,
"learning_rate": 1.2617505032128618e-05,
"loss": 0.7484,
"step": 1543
},
{
"epoch": 0.67,
"learning_rate": 1.2588094551516575e-05,
"loss": 0.9453,
"step": 1544
},
{
"epoch": 0.67,
"learning_rate": 1.2558706850008678e-05,
"loss": 0.6682,
"step": 1545
},
{
"epoch": 0.67,
"learning_rate": 1.2529341981539071e-05,
"loss": 0.7015,
"step": 1546
},
{
"epoch": 0.67,
"learning_rate": 1.2500000000000006e-05,
"loss": 0.7675,
"step": 1547
},
{
"epoch": 0.67,
"learning_rate": 1.2470680959241702e-05,
"loss": 0.6856,
"step": 1548
},
{
"epoch": 0.67,
"learning_rate": 1.244138491307231e-05,
"loss": 0.7482,
"step": 1549
},
{
"epoch": 0.67,
"learning_rate": 1.2412111915257771e-05,
"loss": 0.7449,
"step": 1550
},
{
"epoch": 0.67,
"learning_rate": 1.23828620195217e-05,
"loss": 0.8686,
"step": 1551
},
{
"epoch": 0.67,
"learning_rate": 1.2353635279545325e-05,
"loss": 0.8173,
"step": 1552
},
{
"epoch": 0.67,
"learning_rate": 1.2324431748967413e-05,
"loss": 0.8125,
"step": 1553
},
{
"epoch": 0.67,
"learning_rate": 1.2295251481384101e-05,
"loss": 0.788,
"step": 1554
},
{
"epoch": 0.67,
"learning_rate": 1.2266094530348812e-05,
"loss": 0.7156,
"step": 1555
},
{
"epoch": 0.67,
"learning_rate": 1.2236960949372234e-05,
"loss": 0.6775,
"step": 1556
},
{
"epoch": 0.67,
"learning_rate": 1.2207850791922123e-05,
"loss": 0.9304,
"step": 1557
},
{
"epoch": 0.67,
"learning_rate": 1.2178764111423258e-05,
"loss": 0.8204,
"step": 1558
},
{
"epoch": 0.67,
"learning_rate": 1.2149700961257334e-05,
"loss": 0.7543,
"step": 1559
},
{
"epoch": 0.67,
"learning_rate": 1.2120661394762858e-05,
"loss": 0.787,
"step": 1560
},
{
"epoch": 0.67,
"learning_rate": 1.2091645465235058e-05,
"loss": 0.7496,
"step": 1561
},
{
"epoch": 0.67,
"learning_rate": 1.2062653225925776e-05,
"loss": 0.7551,
"step": 1562
},
{
"epoch": 0.67,
"learning_rate": 1.2033684730043385e-05,
"loss": 0.7674,
"step": 1563
},
{
"epoch": 0.67,
"learning_rate": 1.2004740030752676e-05,
"loss": 0.7669,
"step": 1564
},
{
"epoch": 0.67,
"learning_rate": 1.1975819181174769e-05,
"loss": 0.7482,
"step": 1565
},
{
"epoch": 0.67,
"learning_rate": 1.1946922234387006e-05,
"loss": 0.7732,
"step": 1566
},
{
"epoch": 0.68,
"learning_rate": 1.1918049243422894e-05,
"loss": 0.7507,
"step": 1567
},
{
"epoch": 0.68,
"learning_rate": 1.1889200261271923e-05,
"loss": 0.7801,
"step": 1568
},
{
"epoch": 0.68,
"learning_rate": 1.1860375340879549e-05,
"loss": 0.7344,
"step": 1569
},
{
"epoch": 0.68,
"learning_rate": 1.183157453514708e-05,
"loss": 0.7422,
"step": 1570
},
{
"epoch": 0.68,
"learning_rate": 1.1802797896931548e-05,
"loss": 0.8728,
"step": 1571
},
{
"epoch": 0.68,
"learning_rate": 1.1774045479045646e-05,
"loss": 0.7828,
"step": 1572
},
{
"epoch": 0.68,
"learning_rate": 1.1745317334257577e-05,
"loss": 0.6144,
"step": 1573
},
{
"epoch": 0.68,
"learning_rate": 1.1716613515291058e-05,
"loss": 0.7564,
"step": 1574
},
{
"epoch": 0.68,
"learning_rate": 1.1687934074825122e-05,
"loss": 0.6821,
"step": 1575
},
{
"epoch": 0.68,
"learning_rate": 1.1659279065494064e-05,
"loss": 0.8643,
"step": 1576
},
{
"epoch": 0.68,
"learning_rate": 1.1630648539887352e-05,
"loss": 0.6761,
"step": 1577
},
{
"epoch": 0.68,
"learning_rate": 1.1602042550549508e-05,
"loss": 0.847,
"step": 1578
},
{
"epoch": 0.68,
"learning_rate": 1.1573461149980038e-05,
"loss": 0.6638,
"step": 1579
},
{
"epoch": 0.68,
"learning_rate": 1.1544904390633312e-05,
"loss": 0.8545,
"step": 1580
},
{
"epoch": 0.68,
"learning_rate": 1.1516372324918478e-05,
"loss": 0.7427,
"step": 1581
},
{
"epoch": 0.68,
"learning_rate": 1.1487865005199364e-05,
"loss": 0.8448,
"step": 1582
},
{
"epoch": 0.68,
"learning_rate": 1.1459382483794382e-05,
"loss": 0.8973,
"step": 1583
},
{
"epoch": 0.68,
"learning_rate": 1.1430924812976429e-05,
"loss": 0.6742,
"step": 1584
},
{
"epoch": 0.68,
"learning_rate": 1.1402492044972826e-05,
"loss": 0.6891,
"step": 1585
},
{
"epoch": 0.68,
"learning_rate": 1.1374084231965137e-05,
"loss": 0.7441,
"step": 1586
},
{
"epoch": 0.68,
"learning_rate": 1.1345701426089157e-05,
"loss": 0.8483,
"step": 1587
},
{
"epoch": 0.68,
"learning_rate": 1.1317343679434797e-05,
"loss": 0.7646,
"step": 1588
},
{
"epoch": 0.68,
"learning_rate": 1.1289011044045972e-05,
"loss": 0.8738,
"step": 1589
},
{
"epoch": 0.69,
"learning_rate": 1.126070357192048e-05,
"loss": 0.7981,
"step": 1590
},
{
"epoch": 0.69,
"learning_rate": 1.1232421315009966e-05,
"loss": 0.7359,
"step": 1591
},
{
"epoch": 0.69,
"learning_rate": 1.120416432521981e-05,
"loss": 0.7633,
"step": 1592
},
{
"epoch": 0.69,
"learning_rate": 1.1175932654408997e-05,
"loss": 0.6446,
"step": 1593
},
{
"epoch": 0.69,
"learning_rate": 1.114772635439005e-05,
"loss": 0.6809,
"step": 1594
},
{
"epoch": 0.69,
"learning_rate": 1.1119545476928939e-05,
"loss": 0.7938,
"step": 1595
},
{
"epoch": 0.69,
"learning_rate": 1.1091390073744965e-05,
"loss": 0.7758,
"step": 1596
},
{
"epoch": 0.69,
"learning_rate": 1.1063260196510683e-05,
"loss": 0.6296,
"step": 1597
},
{
"epoch": 0.69,
"learning_rate": 1.1035155896851804e-05,
"loss": 0.8407,
"step": 1598
},
{
"epoch": 0.69,
"learning_rate": 1.1007077226347087e-05,
"loss": 0.756,
"step": 1599
},
{
"epoch": 0.69,
"learning_rate": 1.097902423652827e-05,
"loss": 0.7576,
"step": 1600
},
{
"epoch": 0.69,
"eval_loss": 0.7660685777664185,
"eval_runtime": 10.1738,
"eval_samples_per_second": 3.637,
"eval_steps_per_second": 0.491,
"step": 1600
},
{
"epoch": 0.69,
"learning_rate": 1.095099697887994e-05,
"loss": 0.857,
"step": 1601
},
{
"epoch": 0.69,
"learning_rate": 1.09229955048395e-05,
"loss": 0.6867,
"step": 1602
},
{
"epoch": 0.69,
"learning_rate": 1.0895019865796974e-05,
"loss": 0.8768,
"step": 1603
},
{
"epoch": 0.69,
"learning_rate": 1.086707011309501e-05,
"loss": 0.7735,
"step": 1604
},
{
"epoch": 0.69,
"learning_rate": 1.0839146298028736e-05,
"loss": 0.7836,
"step": 1605
},
{
"epoch": 0.69,
"learning_rate": 1.0811248471845694e-05,
"loss": 0.9065,
"step": 1606
},
{
"epoch": 0.69,
"learning_rate": 1.0783376685745723e-05,
"loss": 0.6707,
"step": 1607
},
{
"epoch": 0.69,
"learning_rate": 1.0755530990880834e-05,
"loss": 0.6427,
"step": 1608
},
{
"epoch": 0.69,
"learning_rate": 1.0727711438355213e-05,
"loss": 0.8193,
"step": 1609
},
{
"epoch": 0.69,
"learning_rate": 1.0699918079225033e-05,
"loss": 0.8126,
"step": 1610
},
{
"epoch": 0.69,
"learning_rate": 1.0672150964498403e-05,
"loss": 0.8801,
"step": 1611
},
{
"epoch": 0.69,
"learning_rate": 1.0644410145135267e-05,
"loss": 0.7672,
"step": 1612
},
{
"epoch": 0.7,
"learning_rate": 1.0616695672047305e-05,
"loss": 0.681,
"step": 1613
},
{
"epoch": 0.7,
"learning_rate": 1.0589007596097855e-05,
"loss": 0.6755,
"step": 1614
},
{
"epoch": 0.7,
"learning_rate": 1.0561345968101799e-05,
"loss": 0.7523,
"step": 1615
},
{
"epoch": 0.7,
"learning_rate": 1.0533710838825489e-05,
"loss": 0.624,
"step": 1616
},
{
"epoch": 0.7,
"learning_rate": 1.0506102258986634e-05,
"loss": 0.7993,
"step": 1617
},
{
"epoch": 0.7,
"learning_rate": 1.0478520279254231e-05,
"loss": 0.8003,
"step": 1618
},
{
"epoch": 0.7,
"learning_rate": 1.0450964950248437e-05,
"loss": 0.69,
"step": 1619
},
{
"epoch": 0.7,
"learning_rate": 1.0423436322540545e-05,
"loss": 0.7027,
"step": 1620
},
{
"epoch": 0.7,
"learning_rate": 1.039593444665278e-05,
"loss": 0.8992,
"step": 1621
},
{
"epoch": 0.7,
"learning_rate": 1.0368459373058318e-05,
"loss": 0.7951,
"step": 1622
},
{
"epoch": 0.7,
"learning_rate": 1.0341011152181112e-05,
"loss": 0.6282,
"step": 1623
},
{
"epoch": 0.7,
"learning_rate": 1.031358983439589e-05,
"loss": 0.9396,
"step": 1624
},
{
"epoch": 0.7,
"learning_rate": 1.0286195470027928e-05,
"loss": 0.7106,
"step": 1625
},
{
"epoch": 0.7,
"learning_rate": 1.0258828109353086e-05,
"loss": 0.8467,
"step": 1626
},
{
"epoch": 0.7,
"learning_rate": 1.0231487802597665e-05,
"loss": 0.7889,
"step": 1627
},
{
"epoch": 0.7,
"learning_rate": 1.0204174599938293e-05,
"loss": 0.931,
"step": 1628
},
{
"epoch": 0.7,
"learning_rate": 1.0176888551501881e-05,
"loss": 0.7254,
"step": 1629
},
{
"epoch": 0.7,
"learning_rate": 1.0149629707365457e-05,
"loss": 0.7016,
"step": 1630
},
{
"epoch": 0.7,
"learning_rate": 1.0122398117556184e-05,
"loss": 0.7054,
"step": 1631
},
{
"epoch": 0.7,
"learning_rate": 1.0095193832051168e-05,
"loss": 0.6572,
"step": 1632
},
{
"epoch": 0.7,
"learning_rate": 1.0068016900777411e-05,
"loss": 0.8565,
"step": 1633
},
{
"epoch": 0.7,
"learning_rate": 1.0040867373611715e-05,
"loss": 0.7847,
"step": 1634
},
{
"epoch": 0.7,
"learning_rate": 1.0013745300380589e-05,
"loss": 0.7425,
"step": 1635
},
{
"epoch": 0.71,
"learning_rate": 9.986650730860156e-06,
"loss": 0.8952,
"step": 1636
},
{
"epoch": 0.71,
"learning_rate": 9.959583714776063e-06,
"loss": 0.835,
"step": 1637
},
{
"epoch": 0.71,
"learning_rate": 9.932544301803388e-06,
"loss": 0.8751,
"step": 1638
},
{
"epoch": 0.71,
"learning_rate": 9.905532541566549e-06,
"loss": 0.6323,
"step": 1639
},
{
"epoch": 0.71,
"learning_rate": 9.878548483639204e-06,
"loss": 0.7731,
"step": 1640
},
{
"epoch": 0.71,
"learning_rate": 9.851592177544208e-06,
"loss": 0.7715,
"step": 1641
},
{
"epoch": 0.71,
"learning_rate": 9.824663672753453e-06,
"loss": 0.6615,
"step": 1642
},
{
"epoch": 0.71,
"learning_rate": 9.797763018687795e-06,
"loss": 0.8719,
"step": 1643
},
{
"epoch": 0.71,
"learning_rate": 9.770890264716992e-06,
"loss": 0.7178,
"step": 1644
},
{
"epoch": 0.71,
"learning_rate": 9.744045460159624e-06,
"loss": 0.7073,
"step": 1645
},
{
"epoch": 0.71,
"learning_rate": 9.717228654282942e-06,
"loss": 0.7443,
"step": 1646
},
{
"epoch": 0.71,
"learning_rate": 9.690439896302827e-06,
"loss": 0.7285,
"step": 1647
},
{
"epoch": 0.71,
"learning_rate": 9.663679235383662e-06,
"loss": 0.731,
"step": 1648
},
{
"epoch": 0.71,
"learning_rate": 9.636946720638306e-06,
"loss": 0.8314,
"step": 1649
},
{
"epoch": 0.71,
"learning_rate": 9.610242401127934e-06,
"loss": 0.7368,
"step": 1650
},
{
"epoch": 0.71,
"learning_rate": 9.583566325861975e-06,
"loss": 0.7324,
"step": 1651
},
{
"epoch": 0.71,
"learning_rate": 9.556918543798035e-06,
"loss": 0.7315,
"step": 1652
},
{
"epoch": 0.71,
"learning_rate": 9.530299103841785e-06,
"loss": 0.736,
"step": 1653
},
{
"epoch": 0.71,
"learning_rate": 9.503708054846885e-06,
"loss": 0.8074,
"step": 1654
},
{
"epoch": 0.71,
"learning_rate": 9.477145445614888e-06,
"loss": 0.701,
"step": 1655
},
{
"epoch": 0.71,
"learning_rate": 9.450611324895156e-06,
"loss": 0.7082,
"step": 1656
},
{
"epoch": 0.71,
"learning_rate": 9.42410574138476e-06,
"loss": 0.6774,
"step": 1657
},
{
"epoch": 0.71,
"learning_rate": 9.397628743728399e-06,
"loss": 0.7554,
"step": 1658
},
{
"epoch": 0.71,
"learning_rate": 9.371180380518333e-06,
"loss": 0.7239,
"step": 1659
},
{
"epoch": 0.72,
"learning_rate": 9.344760700294231e-06,
"loss": 0.8155,
"step": 1660
},
{
"epoch": 0.72,
"learning_rate": 9.318369751543143e-06,
"loss": 0.6981,
"step": 1661
},
{
"epoch": 0.72,
"learning_rate": 9.292007582699383e-06,
"loss": 0.7622,
"step": 1662
},
{
"epoch": 0.72,
"learning_rate": 9.265674242144462e-06,
"loss": 0.6517,
"step": 1663
},
{
"epoch": 0.72,
"learning_rate": 9.239369778206977e-06,
"loss": 0.6636,
"step": 1664
},
{
"epoch": 0.72,
"learning_rate": 9.213094239162496e-06,
"loss": 0.6402,
"step": 1665
},
{
"epoch": 0.72,
"learning_rate": 9.18684767323355e-06,
"loss": 0.8188,
"step": 1666
},
{
"epoch": 0.72,
"learning_rate": 9.160630128589473e-06,
"loss": 0.6963,
"step": 1667
},
{
"epoch": 0.72,
"learning_rate": 9.134441653346337e-06,
"loss": 0.8683,
"step": 1668
},
{
"epoch": 0.72,
"learning_rate": 9.108282295566873e-06,
"loss": 0.7867,
"step": 1669
},
{
"epoch": 0.72,
"learning_rate": 9.082152103260363e-06,
"loss": 0.8273,
"step": 1670
},
{
"epoch": 0.72,
"learning_rate": 9.056051124382572e-06,
"loss": 0.8001,
"step": 1671
},
{
"epoch": 0.72,
"learning_rate": 9.029979406835643e-06,
"loss": 0.8615,
"step": 1672
},
{
"epoch": 0.72,
"learning_rate": 9.00393699846802e-06,
"loss": 0.7993,
"step": 1673
},
{
"epoch": 0.72,
"learning_rate": 8.977923947074366e-06,
"loss": 0.6579,
"step": 1674
},
{
"epoch": 0.72,
"learning_rate": 8.951940300395445e-06,
"loss": 0.8094,
"step": 1675
},
{
"epoch": 0.72,
"learning_rate": 8.92598610611807e-06,
"loss": 0.7659,
"step": 1676
},
{
"epoch": 0.72,
"learning_rate": 8.900061411875022e-06,
"loss": 0.6938,
"step": 1677
},
{
"epoch": 0.72,
"learning_rate": 8.87416626524489e-06,
"loss": 0.8218,
"step": 1678
},
{
"epoch": 0.72,
"learning_rate": 8.84830071375207e-06,
"loss": 0.7098,
"step": 1679
},
{
"epoch": 0.72,
"learning_rate": 8.822464804866652e-06,
"loss": 0.8403,
"step": 1680
},
{
"epoch": 0.72,
"learning_rate": 8.7966585860043e-06,
"loss": 0.7814,
"step": 1681
},
{
"epoch": 0.72,
"learning_rate": 8.770882104526207e-06,
"loss": 0.7178,
"step": 1682
},
{
"epoch": 0.73,
"learning_rate": 8.745135407738961e-06,
"loss": 0.7318,
"step": 1683
},
{
"epoch": 0.73,
"learning_rate": 8.71941854289453e-06,
"loss": 0.6442,
"step": 1684
},
{
"epoch": 0.73,
"learning_rate": 8.693731557190105e-06,
"loss": 0.6537,
"step": 1685
},
{
"epoch": 0.73,
"learning_rate": 8.66807449776805e-06,
"loss": 0.754,
"step": 1686
},
{
"epoch": 0.73,
"learning_rate": 8.642447411715784e-06,
"loss": 0.7928,
"step": 1687
},
{
"epoch": 0.73,
"learning_rate": 8.616850346065755e-06,
"loss": 0.6301,
"step": 1688
},
{
"epoch": 0.73,
"learning_rate": 8.591283347795293e-06,
"loss": 0.7883,
"step": 1689
},
{
"epoch": 0.73,
"learning_rate": 8.565746463826549e-06,
"loss": 0.7844,
"step": 1690
},
{
"epoch": 0.73,
"learning_rate": 8.540239741026405e-06,
"loss": 0.651,
"step": 1691
},
{
"epoch": 0.73,
"learning_rate": 8.514763226206395e-06,
"loss": 0.6677,
"step": 1692
},
{
"epoch": 0.73,
"learning_rate": 8.489316966122609e-06,
"loss": 0.802,
"step": 1693
},
{
"epoch": 0.73,
"learning_rate": 8.463901007475614e-06,
"loss": 0.8137,
"step": 1694
},
{
"epoch": 0.73,
"learning_rate": 8.438515396910362e-06,
"loss": 0.7098,
"step": 1695
},
{
"epoch": 0.73,
"learning_rate": 8.413160181016116e-06,
"loss": 0.728,
"step": 1696
},
{
"epoch": 0.73,
"learning_rate": 8.387835406326338e-06,
"loss": 0.5884,
"step": 1697
},
{
"epoch": 0.73,
"learning_rate": 8.362541119318659e-06,
"loss": 0.8875,
"step": 1698
},
{
"epoch": 0.73,
"learning_rate": 8.337277366414728e-06,
"loss": 0.772,
"step": 1699
},
{
"epoch": 0.73,
"learning_rate": 8.31204419398015e-06,
"loss": 0.938,
"step": 1700
},
{
"epoch": 0.73,
"eval_loss": 0.7655941247940063,
"eval_runtime": 10.1757,
"eval_samples_per_second": 3.636,
"eval_steps_per_second": 0.491,
"step": 1700
},
{
"epoch": 0.73,
"learning_rate": 8.28684164832442e-06,
"loss": 0.7552,
"step": 1701
},
{
"epoch": 0.73,
"learning_rate": 8.261669775700836e-06,
"loss": 0.6922,
"step": 1702
},
{
"epoch": 0.73,
"learning_rate": 8.236528622306388e-06,
"loss": 0.7711,
"step": 1703
},
{
"epoch": 0.73,
"learning_rate": 8.211418234281686e-06,
"loss": 0.7662,
"step": 1704
},
{
"epoch": 0.73,
"learning_rate": 8.186338657710884e-06,
"loss": 0.8212,
"step": 1705
},
{
"epoch": 0.74,
"learning_rate": 8.16128993862159e-06,
"loss": 0.7022,
"step": 1706
},
{
"epoch": 0.74,
"learning_rate": 8.136272122984778e-06,
"loss": 0.7451,
"step": 1707
},
{
"epoch": 0.74,
"learning_rate": 8.1112852567147e-06,
"loss": 0.7028,
"step": 1708
},
{
"epoch": 0.74,
"learning_rate": 8.086329385668822e-06,
"loss": 0.6798,
"step": 1709
},
{
"epoch": 0.74,
"learning_rate": 8.061404555647711e-06,
"loss": 0.8119,
"step": 1710
},
{
"epoch": 0.74,
"learning_rate": 8.036510812394967e-06,
"loss": 0.9166,
"step": 1711
},
{
"epoch": 0.74,
"learning_rate": 8.011648201597166e-06,
"loss": 0.5711,
"step": 1712
},
{
"epoch": 0.74,
"learning_rate": 7.986816768883699e-06,
"loss": 0.6761,
"step": 1713
},
{
"epoch": 0.74,
"learning_rate": 7.962016559826772e-06,
"loss": 0.723,
"step": 1714
},
{
"epoch": 0.74,
"learning_rate": 7.937247619941265e-06,
"loss": 0.7273,
"step": 1715
},
{
"epoch": 0.74,
"learning_rate": 7.912509994684709e-06,
"loss": 0.805,
"step": 1716
},
{
"epoch": 0.74,
"learning_rate": 7.88780372945713e-06,
"loss": 0.8572,
"step": 1717
},
{
"epoch": 0.74,
"learning_rate": 7.863128869600998e-06,
"loss": 0.8044,
"step": 1718
},
{
"epoch": 0.74,
"learning_rate": 7.838485460401154e-06,
"loss": 0.7109,
"step": 1719
},
{
"epoch": 0.74,
"learning_rate": 7.813873547084741e-06,
"loss": 0.7738,
"step": 1720
},
{
"epoch": 0.74,
"learning_rate": 7.789293174821077e-06,
"loss": 0.6835,
"step": 1721
},
{
"epoch": 0.74,
"learning_rate": 7.764744388721576e-06,
"loss": 0.8413,
"step": 1722
},
{
"epoch": 0.74,
"learning_rate": 7.740227233839725e-06,
"loss": 0.8831,
"step": 1723
},
{
"epoch": 0.74,
"learning_rate": 7.71574175517093e-06,
"loss": 0.7507,
"step": 1724
},
{
"epoch": 0.74,
"learning_rate": 7.69128799765247e-06,
"loss": 0.8442,
"step": 1725
},
{
"epoch": 0.74,
"learning_rate": 7.66686600616341e-06,
"loss": 0.8111,
"step": 1726
},
{
"epoch": 0.74,
"learning_rate": 7.642475825524514e-06,
"loss": 0.9816,
"step": 1727
},
{
"epoch": 0.74,
"learning_rate": 7.618117500498165e-06,
"loss": 0.7127,
"step": 1728
},
{
"epoch": 0.75,
"learning_rate": 7.593791075788281e-06,
"loss": 0.7663,
"step": 1729
},
{
"epoch": 0.75,
"learning_rate": 7.5694965960402375e-06,
"loss": 0.6685,
"step": 1730
},
{
"epoch": 0.75,
"learning_rate": 7.5452341058407816e-06,
"loss": 0.8091,
"step": 1731
},
{
"epoch": 0.75,
"learning_rate": 7.521003649717947e-06,
"loss": 0.6338,
"step": 1732
},
{
"epoch": 0.75,
"learning_rate": 7.496805272140981e-06,
"loss": 0.8271,
"step": 1733
},
{
"epoch": 0.75,
"learning_rate": 7.472639017520272e-06,
"loss": 0.856,
"step": 1734
},
{
"epoch": 0.75,
"learning_rate": 7.4485049302072205e-06,
"loss": 0.7441,
"step": 1735
},
{
"epoch": 0.75,
"learning_rate": 7.424403054494211e-06,
"loss": 0.6767,
"step": 1736
},
{
"epoch": 0.75,
"learning_rate": 7.400333434614528e-06,
"loss": 0.6676,
"step": 1737
},
{
"epoch": 0.75,
"learning_rate": 7.376296114742231e-06,
"loss": 0.8644,
"step": 1738
},
{
"epoch": 0.75,
"learning_rate": 7.3522911389921176e-06,
"loss": 0.7289,
"step": 1739
},
{
"epoch": 0.75,
"learning_rate": 7.3283185514196e-06,
"loss": 0.6892,
"step": 1740
},
{
"epoch": 0.75,
"learning_rate": 7.304378396020689e-06,
"loss": 0.7766,
"step": 1741
},
{
"epoch": 0.75,
"learning_rate": 7.280470716731843e-06,
"loss": 0.6815,
"step": 1742
},
{
"epoch": 0.75,
"learning_rate": 7.256595557429935e-06,
"loss": 0.7115,
"step": 1743
},
{
"epoch": 0.75,
"learning_rate": 7.232752961932141e-06,
"loss": 0.7781,
"step": 1744
},
{
"epoch": 0.75,
"learning_rate": 7.208942973995886e-06,
"loss": 0.7154,
"step": 1745
},
{
"epoch": 0.75,
"learning_rate": 7.185165637318747e-06,
"loss": 0.8243,
"step": 1746
},
{
"epoch": 0.75,
"learning_rate": 7.161420995538373e-06,
"loss": 0.7906,
"step": 1747
},
{
"epoch": 0.75,
"learning_rate": 7.13770909223242e-06,
"loss": 0.6554,
"step": 1748
},
{
"epoch": 0.75,
"learning_rate": 7.114029970918451e-06,
"loss": 0.7639,
"step": 1749
},
{
"epoch": 0.75,
"learning_rate": 7.090383675053866e-06,
"loss": 0.6918,
"step": 1750
},
{
"epoch": 0.75,
"learning_rate": 7.066770248035823e-06,
"loss": 0.8866,
"step": 1751
},
{
"epoch": 0.76,
"learning_rate": 7.043189733201175e-06,
"loss": 0.6485,
"step": 1752
},
{
"epoch": 0.76,
"learning_rate": 7.019642173826335e-06,
"loss": 0.7619,
"step": 1753
},
{
"epoch": 0.76,
"learning_rate": 6.9961276131272565e-06,
"loss": 0.8776,
"step": 1754
},
{
"epoch": 0.76,
"learning_rate": 6.972646094259342e-06,
"loss": 0.7731,
"step": 1755
},
{
"epoch": 0.76,
"learning_rate": 6.949197660317344e-06,
"loss": 0.6611,
"step": 1756
},
{
"epoch": 0.76,
"learning_rate": 6.925782354335278e-06,
"loss": 0.8898,
"step": 1757
},
{
"epoch": 0.76,
"learning_rate": 6.902400219286373e-06,
"loss": 0.8171,
"step": 1758
},
{
"epoch": 0.76,
"learning_rate": 6.879051298083003e-06,
"loss": 0.8203,
"step": 1759
},
{
"epoch": 0.76,
"learning_rate": 6.855735633576557e-06,
"loss": 0.7698,
"step": 1760
},
{
"epoch": 0.76,
"learning_rate": 6.832453268557398e-06,
"loss": 0.8323,
"step": 1761
},
{
"epoch": 0.76,
"learning_rate": 6.8092042457547775e-06,
"loss": 0.7024,
"step": 1762
},
{
"epoch": 0.76,
"learning_rate": 6.785988607836755e-06,
"loss": 0.7692,
"step": 1763
},
{
"epoch": 0.76,
"learning_rate": 6.762806397410118e-06,
"loss": 0.6685,
"step": 1764
},
{
"epoch": 0.76,
"learning_rate": 6.739657657020309e-06,
"loss": 0.8867,
"step": 1765
},
{
"epoch": 0.76,
"learning_rate": 6.716542429151343e-06,
"loss": 0.795,
"step": 1766
},
{
"epoch": 0.76,
"learning_rate": 6.693460756225725e-06,
"loss": 0.6574,
"step": 1767
},
{
"epoch": 0.76,
"learning_rate": 6.670412680604379e-06,
"loss": 0.7285,
"step": 1768
},
{
"epoch": 0.76,
"learning_rate": 6.647398244586595e-06,
"loss": 0.7736,
"step": 1769
},
{
"epoch": 0.76,
"learning_rate": 6.624417490409882e-06,
"loss": 0.7106,
"step": 1770
},
{
"epoch": 0.76,
"learning_rate": 6.601470460249959e-06,
"loss": 0.8025,
"step": 1771
},
{
"epoch": 0.76,
"learning_rate": 6.578557196220647e-06,
"loss": 0.8226,
"step": 1772
},
{
"epoch": 0.76,
"learning_rate": 6.555677740373812e-06,
"loss": 0.8037,
"step": 1773
},
{
"epoch": 0.76,
"learning_rate": 6.53283213469926e-06,
"loss": 0.8083,
"step": 1774
},
{
"epoch": 0.76,
"learning_rate": 6.510020421124649e-06,
"loss": 0.8245,
"step": 1775
},
{
"epoch": 0.77,
"learning_rate": 6.487242641515489e-06,
"loss": 0.8152,
"step": 1776
},
{
"epoch": 0.77,
"learning_rate": 6.4644988376749705e-06,
"loss": 0.9058,
"step": 1777
},
{
"epoch": 0.77,
"learning_rate": 6.441789051343955e-06,
"loss": 0.7909,
"step": 1778
},
{
"epoch": 0.77,
"learning_rate": 6.419113324200835e-06,
"loss": 0.7437,
"step": 1779
},
{
"epoch": 0.77,
"learning_rate": 6.3964716978615416e-06,
"loss": 0.7873,
"step": 1780
},
{
"epoch": 0.77,
"learning_rate": 6.373864213879402e-06,
"loss": 0.7341,
"step": 1781
},
{
"epoch": 0.77,
"learning_rate": 6.351290913745076e-06,
"loss": 0.8479,
"step": 1782
},
{
"epoch": 0.77,
"learning_rate": 6.328751838886496e-06,
"loss": 0.7774,
"step": 1783
},
{
"epoch": 0.77,
"learning_rate": 6.306247030668783e-06,
"loss": 0.7135,
"step": 1784
},
{
"epoch": 0.77,
"learning_rate": 6.283776530394162e-06,
"loss": 0.6989,
"step": 1785
},
{
"epoch": 0.77,
"learning_rate": 6.261340379301894e-06,
"loss": 0.7747,
"step": 1786
},
{
"epoch": 0.77,
"learning_rate": 6.238938618568227e-06,
"loss": 0.7521,
"step": 1787
},
{
"epoch": 0.77,
"learning_rate": 6.2165712893062476e-06,
"loss": 0.7549,
"step": 1788
},
{
"epoch": 0.77,
"learning_rate": 6.194238432565885e-06,
"loss": 0.8281,
"step": 1789
},
{
"epoch": 0.77,
"learning_rate": 6.171940089333786e-06,
"loss": 0.7668,
"step": 1790
},
{
"epoch": 0.77,
"learning_rate": 6.14967630053328e-06,
"loss": 0.7116,
"step": 1791
},
{
"epoch": 0.77,
"learning_rate": 6.127447107024248e-06,
"loss": 0.5955,
"step": 1792
},
{
"epoch": 0.77,
"learning_rate": 6.105252549603088e-06,
"loss": 0.7986,
"step": 1793
},
{
"epoch": 0.77,
"learning_rate": 6.083092669002658e-06,
"loss": 0.6424,
"step": 1794
},
{
"epoch": 0.77,
"learning_rate": 6.060967505892143e-06,
"loss": 0.6974,
"step": 1795
},
{
"epoch": 0.77,
"learning_rate": 6.038877100877036e-06,
"loss": 0.7831,
"step": 1796
},
{
"epoch": 0.77,
"learning_rate": 6.016821494499003e-06,
"loss": 0.8403,
"step": 1797
},
{
"epoch": 0.77,
"learning_rate": 5.994800727235894e-06,
"loss": 0.6605,
"step": 1798
},
{
"epoch": 0.78,
"learning_rate": 5.972814839501589e-06,
"loss": 0.8305,
"step": 1799
},
{
"epoch": 0.78,
"learning_rate": 5.950863871645962e-06,
"loss": 0.8028,
"step": 1800
},
{
"epoch": 0.78,
"eval_loss": 0.7649816274642944,
"eval_runtime": 10.1798,
"eval_samples_per_second": 3.635,
"eval_steps_per_second": 0.491,
"step": 1800
},
{
"epoch": 0.78,
"learning_rate": 5.928947863954798e-06,
"loss": 0.8021,
"step": 1801
},
{
"epoch": 0.78,
"learning_rate": 5.907066856649729e-06,
"loss": 0.687,
"step": 1802
},
{
"epoch": 0.78,
"learning_rate": 5.88522088988814e-06,
"loss": 0.8061,
"step": 1803
},
{
"epoch": 0.78,
"learning_rate": 5.86341000376312e-06,
"loss": 0.6787,
"step": 1804
},
{
"epoch": 0.78,
"learning_rate": 5.841634238303365e-06,
"loss": 0.6511,
"step": 1805
},
{
"epoch": 0.78,
"learning_rate": 5.81989363347312e-06,
"loss": 0.846,
"step": 1806
},
{
"epoch": 0.78,
"learning_rate": 5.7981882291720944e-06,
"loss": 0.6861,
"step": 1807
},
{
"epoch": 0.78,
"learning_rate": 5.77651806523542e-06,
"loss": 0.5914,
"step": 1808
},
{
"epoch": 0.78,
"learning_rate": 5.754883181433529e-06,
"loss": 0.7258,
"step": 1809
},
{
"epoch": 0.78,
"learning_rate": 5.7332836174721014e-06,
"loss": 0.8925,
"step": 1810
},
{
"epoch": 0.78,
"learning_rate": 5.711719412992006e-06,
"loss": 0.8861,
"step": 1811
},
{
"epoch": 0.78,
"learning_rate": 5.690190607569229e-06,
"loss": 0.7975,
"step": 1812
},
{
"epoch": 0.78,
"learning_rate": 5.668697240714782e-06,
"loss": 0.7744,
"step": 1813
},
{
"epoch": 0.78,
"learning_rate": 5.647239351874614e-06,
"loss": 0.6823,
"step": 1814
},
{
"epoch": 0.78,
"learning_rate": 5.625816980429599e-06,
"loss": 0.7567,
"step": 1815
},
{
"epoch": 0.78,
"learning_rate": 5.60443016569541e-06,
"loss": 0.796,
"step": 1816
},
{
"epoch": 0.78,
"learning_rate": 5.58307894692246e-06,
"loss": 0.8224,
"step": 1817
},
{
"epoch": 0.78,
"learning_rate": 5.561763363295841e-06,
"loss": 0.798,
"step": 1818
},
{
"epoch": 0.78,
"learning_rate": 5.540483453935247e-06,
"loss": 0.7715,
"step": 1819
},
{
"epoch": 0.78,
"learning_rate": 5.519239257894893e-06,
"loss": 0.7577,
"step": 1820
},
{
"epoch": 0.78,
"learning_rate": 5.498030814163454e-06,
"loss": 1.0168,
"step": 1821
},
{
"epoch": 0.79,
"learning_rate": 5.476858161663994e-06,
"loss": 0.8213,
"step": 1822
},
{
"epoch": 0.79,
"learning_rate": 5.455721339253886e-06,
"loss": 0.8093,
"step": 1823
},
{
"epoch": 0.79,
"learning_rate": 5.434620385724745e-06,
"loss": 0.8216,
"step": 1824
},
{
"epoch": 0.79,
"learning_rate": 5.413555339802356e-06,
"loss": 0.7165,
"step": 1825
},
{
"epoch": 0.79,
"learning_rate": 5.392526240146622e-06,
"loss": 0.8417,
"step": 1826
},
{
"epoch": 0.79,
"learning_rate": 5.371533125351444e-06,
"loss": 0.7238,
"step": 1827
},
{
"epoch": 0.79,
"learning_rate": 5.350576033944704e-06,
"loss": 0.7569,
"step": 1828
},
{
"epoch": 0.79,
"learning_rate": 5.329655004388157e-06,
"loss": 0.8378,
"step": 1829
},
{
"epoch": 0.79,
"learning_rate": 5.308770075077396e-06,
"loss": 0.7627,
"step": 1830
},
{
"epoch": 0.79,
"learning_rate": 5.2879212843417485e-06,
"loss": 0.8093,
"step": 1831
},
{
"epoch": 0.79,
"learning_rate": 5.267108670444199e-06,
"loss": 0.7012,
"step": 1832
},
{
"epoch": 0.79,
"learning_rate": 5.246332271581375e-06,
"loss": 0.6478,
"step": 1833
},
{
"epoch": 0.79,
"learning_rate": 5.225592125883416e-06,
"loss": 0.869,
"step": 1834
},
{
"epoch": 0.79,
"learning_rate": 5.20488827141393e-06,
"loss": 0.6629,
"step": 1835
},
{
"epoch": 0.79,
"learning_rate": 5.184220746169927e-06,
"loss": 0.8603,
"step": 1836
},
{
"epoch": 0.79,
"learning_rate": 5.163589588081741e-06,
"loss": 0.7278,
"step": 1837
},
{
"epoch": 0.79,
"learning_rate": 5.142994835012957e-06,
"loss": 0.7023,
"step": 1838
},
{
"epoch": 0.79,
"learning_rate": 5.122436524760357e-06,
"loss": 0.9379,
"step": 1839
},
{
"epoch": 0.79,
"learning_rate": 5.101914695053836e-06,
"loss": 0.7581,
"step": 1840
},
{
"epoch": 0.79,
"learning_rate": 5.081429383556333e-06,
"loss": 0.7618,
"step": 1841
},
{
"epoch": 0.79,
"learning_rate": 5.060980627863776e-06,
"loss": 0.7541,
"step": 1842
},
{
"epoch": 0.79,
"learning_rate": 5.040568465504991e-06,
"loss": 0.8867,
"step": 1843
},
{
"epoch": 0.79,
"learning_rate": 5.020192933941672e-06,
"loss": 0.8828,
"step": 1844
},
{
"epoch": 0.8,
"learning_rate": 4.99985407056825e-06,
"loss": 0.7256,
"step": 1845
},
{
"epoch": 0.8,
"learning_rate": 4.979551912711872e-06,
"loss": 0.8781,
"step": 1846
},
{
"epoch": 0.8,
"learning_rate": 4.959286497632345e-06,
"loss": 0.8466,
"step": 1847
},
{
"epoch": 0.8,
"learning_rate": 4.939057862522023e-06,
"loss": 0.7232,
"step": 1848
},
{
"epoch": 0.8,
"learning_rate": 4.918866044505746e-06,
"loss": 0.6573,
"step": 1849
},
{
"epoch": 0.8,
"learning_rate": 4.8987110806408e-06,
"loss": 0.77,
"step": 1850
},
{
"epoch": 0.8,
"learning_rate": 4.878593007916849e-06,
"loss": 0.6516,
"step": 1851
},
{
"epoch": 0.8,
"learning_rate": 4.8585118632558265e-06,
"loss": 0.7282,
"step": 1852
},
{
"epoch": 0.8,
"learning_rate": 4.838467683511913e-06,
"loss": 0.8151,
"step": 1853
},
{
"epoch": 0.8,
"learning_rate": 4.81846050547142e-06,
"loss": 0.6609,
"step": 1854
},
{
"epoch": 0.8,
"learning_rate": 4.798490365852787e-06,
"loss": 0.802,
"step": 1855
},
{
"epoch": 0.8,
"learning_rate": 4.778557301306455e-06,
"loss": 0.7106,
"step": 1856
},
{
"epoch": 0.8,
"learning_rate": 4.758661348414833e-06,
"loss": 0.7277,
"step": 1857
},
{
"epoch": 0.8,
"learning_rate": 4.738802543692214e-06,
"loss": 0.7965,
"step": 1858
},
{
"epoch": 0.8,
"learning_rate": 4.718980923584715e-06,
"loss": 0.8508,
"step": 1859
},
{
"epoch": 0.8,
"learning_rate": 4.699196524470215e-06,
"loss": 0.8593,
"step": 1860
},
{
"epoch": 0.8,
"learning_rate": 4.679449382658277e-06,
"loss": 0.7577,
"step": 1861
},
{
"epoch": 0.8,
"learning_rate": 4.659739534390089e-06,
"loss": 0.7016,
"step": 1862
},
{
"epoch": 0.8,
"learning_rate": 4.640067015838398e-06,
"loss": 0.7406,
"step": 1863
},
{
"epoch": 0.8,
"learning_rate": 4.62043186310743e-06,
"loss": 0.7154,
"step": 1864
},
{
"epoch": 0.8,
"learning_rate": 4.600834112232855e-06,
"loss": 0.8136,
"step": 1865
},
{
"epoch": 0.8,
"learning_rate": 4.581273799181693e-06,
"loss": 0.7784,
"step": 1866
},
{
"epoch": 0.8,
"learning_rate": 4.561750959852238e-06,
"loss": 0.7978,
"step": 1867
},
{
"epoch": 0.8,
"learning_rate": 4.542265630074022e-06,
"loss": 0.7299,
"step": 1868
},
{
"epoch": 0.81,
"learning_rate": 4.5228178456077575e-06,
"loss": 0.5917,
"step": 1869
},
{
"epoch": 0.81,
"learning_rate": 4.503407642145222e-06,
"loss": 0.6906,
"step": 1870
},
{
"epoch": 0.81,
"learning_rate": 4.484035055309235e-06,
"loss": 0.7447,
"step": 1871
},
{
"epoch": 0.81,
"learning_rate": 4.4647001206535785e-06,
"loss": 0.7769,
"step": 1872
},
{
"epoch": 0.81,
"learning_rate": 4.4454028736629324e-06,
"loss": 0.7974,
"step": 1873
},
{
"epoch": 0.81,
"learning_rate": 4.426143349752807e-06,
"loss": 0.7232,
"step": 1874
},
{
"epoch": 0.81,
"learning_rate": 4.406921584269486e-06,
"loss": 0.7412,
"step": 1875
},
{
"epoch": 0.81,
"learning_rate": 4.3877376124899535e-06,
"loss": 0.7451,
"step": 1876
},
{
"epoch": 0.81,
"learning_rate": 4.3685914696218326e-06,
"loss": 0.7912,
"step": 1877
},
{
"epoch": 0.81,
"learning_rate": 4.3494831908033166e-06,
"loss": 0.8545,
"step": 1878
},
{
"epoch": 0.81,
"learning_rate": 4.330412811103126e-06,
"loss": 0.6788,
"step": 1879
},
{
"epoch": 0.81,
"learning_rate": 4.311380365520401e-06,
"loss": 0.7453,
"step": 1880
},
{
"epoch": 0.81,
"learning_rate": 4.2923858889846735e-06,
"loss": 0.778,
"step": 1881
},
{
"epoch": 0.81,
"learning_rate": 4.273429416355792e-06,
"loss": 0.8369,
"step": 1882
},
{
"epoch": 0.81,
"learning_rate": 4.254510982423876e-06,
"loss": 0.6613,
"step": 1883
},
{
"epoch": 0.81,
"learning_rate": 4.235630621909201e-06,
"loss": 0.851,
"step": 1884
},
{
"epoch": 0.81,
"learning_rate": 4.216788369462176e-06,
"loss": 0.9104,
"step": 1885
},
{
"epoch": 0.81,
"learning_rate": 4.197984259663304e-06,
"loss": 0.8859,
"step": 1886
},
{
"epoch": 0.81,
"learning_rate": 4.179218327023046e-06,
"loss": 0.7183,
"step": 1887
},
{
"epoch": 0.81,
"learning_rate": 4.160490605981823e-06,
"loss": 0.8198,
"step": 1888
},
{
"epoch": 0.81,
"learning_rate": 4.1418011309099e-06,
"loss": 0.927,
"step": 1889
},
{
"epoch": 0.81,
"learning_rate": 4.12314993610739e-06,
"loss": 0.7529,
"step": 1890
},
{
"epoch": 0.81,
"learning_rate": 4.104537055804119e-06,
"loss": 0.5568,
"step": 1891
},
{
"epoch": 0.82,
"learning_rate": 4.085962524159612e-06,
"loss": 0.6771,
"step": 1892
},
{
"epoch": 0.82,
"learning_rate": 4.067426375263009e-06,
"loss": 0.7252,
"step": 1893
},
{
"epoch": 0.82,
"learning_rate": 4.048928643133007e-06,
"loss": 0.6822,
"step": 1894
},
{
"epoch": 0.82,
"learning_rate": 4.030469361717801e-06,
"loss": 0.8225,
"step": 1895
},
{
"epoch": 0.82,
"learning_rate": 4.012048564895016e-06,
"loss": 0.902,
"step": 1896
},
{
"epoch": 0.82,
"learning_rate": 3.993666286471648e-06,
"loss": 0.631,
"step": 1897
},
{
"epoch": 0.82,
"learning_rate": 3.975322560184006e-06,
"loss": 0.7094,
"step": 1898
},
{
"epoch": 0.82,
"learning_rate": 3.957017419697637e-06,
"loss": 0.8937,
"step": 1899
},
{
"epoch": 0.82,
"learning_rate": 3.938750898607277e-06,
"loss": 0.7126,
"step": 1900
},
{
"epoch": 0.82,
"eval_loss": 0.764629065990448,
"eval_runtime": 10.2072,
"eval_samples_per_second": 3.625,
"eval_steps_per_second": 0.49,
"step": 1900
}
],
"logging_steps": 1,
"max_steps": 2320,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 100,
"total_flos": 2.671892094989107e+18,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}