Upload checkpoint-350 fair SFT checkpoint
Browse files- trainer_state.json +811 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,811 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.0574669187145558,
|
| 6 |
+
"eval_steps": 50,
|
| 7 |
+
"global_step": 350,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 0.442822265625,
|
| 14 |
+
"epoch": 0.015122873345935728,
|
| 15 |
+
"grad_norm": 2.669071628071661,
|
| 16 |
+
"learning_rate": 2.666666666666667e-06,
|
| 17 |
+
"loss": 0.5928983211517334,
|
| 18 |
+
"mean_token_accuracy": 0.8548422813415527,
|
| 19 |
+
"num_tokens": 325432.0,
|
| 20 |
+
"step": 5
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 0.3935546875,
|
| 24 |
+
"epoch": 0.030245746691871456,
|
| 25 |
+
"grad_norm": 1.634140660130009,
|
| 26 |
+
"learning_rate": 6e-06,
|
| 27 |
+
"loss": 0.3988723993301392,
|
| 28 |
+
"mean_token_accuracy": 0.8959574371576309,
|
| 29 |
+
"num_tokens": 653193.0,
|
| 30 |
+
"step": 10
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 0.2187255859375,
|
| 34 |
+
"epoch": 0.045368620037807186,
|
| 35 |
+
"grad_norm": 0.9021532274129692,
|
| 36 |
+
"learning_rate": 9.333333333333334e-06,
|
| 37 |
+
"loss": 0.19831626415252684,
|
| 38 |
+
"mean_token_accuracy": 0.9389900401234627,
|
| 39 |
+
"num_tokens": 964453.0,
|
| 40 |
+
"step": 15
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 0.10172119140625,
|
| 44 |
+
"epoch": 0.06049149338374291,
|
| 45 |
+
"grad_norm": 0.4697352375821627,
|
| 46 |
+
"learning_rate": 1.2666666666666667e-05,
|
| 47 |
+
"loss": 0.10886554718017578,
|
| 48 |
+
"mean_token_accuracy": 0.9637670248746872,
|
| 49 |
+
"num_tokens": 1307238.0,
|
| 50 |
+
"step": 20
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 0.078204345703125,
|
| 54 |
+
"epoch": 0.07561436672967864,
|
| 55 |
+
"grad_norm": 0.39345906256866536,
|
| 56 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 57 |
+
"loss": 0.0776131808757782,
|
| 58 |
+
"mean_token_accuracy": 0.9740302637219429,
|
| 59 |
+
"num_tokens": 1615672.0,
|
| 60 |
+
"step": 25
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"entropy": 0.05927734375,
|
| 64 |
+
"epoch": 0.09073724007561437,
|
| 65 |
+
"grad_norm": 0.35679792282054484,
|
| 66 |
+
"learning_rate": 1.9333333333333333e-05,
|
| 67 |
+
"loss": 0.06322811841964722,
|
| 68 |
+
"mean_token_accuracy": 0.9787174344062806,
|
| 69 |
+
"num_tokens": 1935122.0,
|
| 70 |
+
"step": 30
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"entropy": 0.055340576171875,
|
| 74 |
+
"epoch": 0.10586011342155009,
|
| 75 |
+
"grad_norm": 0.28127188289574473,
|
| 76 |
+
"learning_rate": 1.99991486051959e-05,
|
| 77 |
+
"loss": 0.055406326055526735,
|
| 78 |
+
"mean_token_accuracy": 0.9811019197106361,
|
| 79 |
+
"num_tokens": 2246001.0,
|
| 80 |
+
"step": 35
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"entropy": 0.050018310546875,
|
| 84 |
+
"epoch": 0.12098298676748583,
|
| 85 |
+
"grad_norm": 0.23485165839069644,
|
| 86 |
+
"learning_rate": 1.9995690062269985e-05,
|
| 87 |
+
"loss": 0.05159962177276611,
|
| 88 |
+
"mean_token_accuracy": 0.9823019146919251,
|
| 89 |
+
"num_tokens": 2572728.0,
|
| 90 |
+
"step": 40
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"entropy": 0.05009765625,
|
| 94 |
+
"epoch": 0.13610586011342155,
|
| 95 |
+
"grad_norm": 0.3286661753766386,
|
| 96 |
+
"learning_rate": 1.9989572078514722e-05,
|
| 97 |
+
"loss": 0.04924889206886292,
|
| 98 |
+
"mean_token_accuracy": 0.983111384510994,
|
| 99 |
+
"num_tokens": 2899222.0,
|
| 100 |
+
"step": 45
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"entropy": 0.03935394287109375,
|
| 104 |
+
"epoch": 0.15122873345935728,
|
| 105 |
+
"grad_norm": 0.250685968179852,
|
| 106 |
+
"learning_rate": 1.998079628167323e-05,
|
| 107 |
+
"loss": 0.041659212112426756,
|
| 108 |
+
"mean_token_accuracy": 0.9857470989227295,
|
| 109 |
+
"num_tokens": 3199913.0,
|
| 110 |
+
"step": 50
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.15122873345935728,
|
| 114 |
+
"eval_entropy": 0.040476572764170424,
|
| 115 |
+
"eval_loss": 0.0391058586537838,
|
| 116 |
+
"eval_mean_token_accuracy": 0.986595441092242,
|
| 117 |
+
"eval_num_tokens": 3199913.0,
|
| 118 |
+
"eval_runtime": 94.4655,
|
| 119 |
+
"eval_samples_per_second": 14.069,
|
| 120 |
+
"eval_steps_per_second": 3.525,
|
| 121 |
+
"step": 50
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"entropy": 0.038671875,
|
| 125 |
+
"epoch": 0.166351606805293,
|
| 126 |
+
"grad_norm": 0.3054203398912923,
|
| 127 |
+
"learning_rate": 1.9969365006623072e-05,
|
| 128 |
+
"loss": 0.03826351165771484,
|
| 129 |
+
"mean_token_accuracy": 0.9868962660431861,
|
| 130 |
+
"num_tokens": 3519539.0,
|
| 131 |
+
"step": 55
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"entropy": 0.035321044921875,
|
| 135 |
+
"epoch": 0.18147448015122875,
|
| 136 |
+
"grad_norm": 0.3848999837215887,
|
| 137 |
+
"learning_rate": 1.9955281294755034e-05,
|
| 138 |
+
"loss": 0.03624363541603089,
|
| 139 |
+
"mean_token_accuracy": 0.9876649409532547,
|
| 140 |
+
"num_tokens": 3846310.0,
|
| 141 |
+
"step": 60
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"entropy": 0.035552978515625,
|
| 145 |
+
"epoch": 0.19659735349716445,
|
| 146 |
+
"grad_norm": 0.2606280391924725,
|
| 147 |
+
"learning_rate": 1.9938548893163924e-05,
|
| 148 |
+
"loss": 0.03628517687320709,
|
| 149 |
+
"mean_token_accuracy": 0.9875744208693504,
|
| 150 |
+
"num_tokens": 4173916.0,
|
| 151 |
+
"step": 65
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"entropy": 0.0336944580078125,
|
| 155 |
+
"epoch": 0.21172022684310018,
|
| 156 |
+
"grad_norm": 0.24422090914230693,
|
| 157 |
+
"learning_rate": 1.9919172253651637e-05,
|
| 158 |
+
"loss": 0.033189314603805545,
|
| 159 |
+
"mean_token_accuracy": 0.9885250389575958,
|
| 160 |
+
"num_tokens": 4494452.0,
|
| 161 |
+
"step": 70
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"entropy": 0.02899322509765625,
|
| 165 |
+
"epoch": 0.22684310018903592,
|
| 166 |
+
"grad_norm": 0.19378251727544818,
|
| 167 |
+
"learning_rate": 1.9897156531542708e-05,
|
| 168 |
+
"loss": 0.029927724599838258,
|
| 169 |
+
"mean_token_accuracy": 0.9899556040763855,
|
| 170 |
+
"num_tokens": 4813996.0,
|
| 171 |
+
"step": 75
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"entropy": 0.0284515380859375,
|
| 175 |
+
"epoch": 0.24196597353497165,
|
| 176 |
+
"grad_norm": 0.2664077757245733,
|
| 177 |
+
"learning_rate": 1.98725075843127e-05,
|
| 178 |
+
"loss": 0.029463762044906618,
|
| 179 |
+
"mean_token_accuracy": 0.9898741751909256,
|
| 180 |
+
"num_tokens": 5154020.0,
|
| 181 |
+
"step": 80
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"entropy": 0.0311767578125,
|
| 185 |
+
"epoch": 0.2570888468809074,
|
| 186 |
+
"grad_norm": 0.22537314543102535,
|
| 187 |
+
"learning_rate": 1.9845231970029774e-05,
|
| 188 |
+
"loss": 0.03209342658519745,
|
| 189 |
+
"mean_token_accuracy": 0.9891638234257698,
|
| 190 |
+
"num_tokens": 5484973.0,
|
| 191 |
+
"step": 85
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"entropy": 0.0276885986328125,
|
| 195 |
+
"epoch": 0.2722117202268431,
|
| 196 |
+
"grad_norm": 0.2077524527319594,
|
| 197 |
+
"learning_rate": 1.9815336945609834e-05,
|
| 198 |
+
"loss": 0.029284730553627014,
|
| 199 |
+
"mean_token_accuracy": 0.9901994511485099,
|
| 200 |
+
"num_tokens": 5811817.0,
|
| 201 |
+
"step": 90
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"entropy": 0.0260345458984375,
|
| 205 |
+
"epoch": 0.28733459357277885,
|
| 206 |
+
"grad_norm": 0.31080264218901027,
|
| 207 |
+
"learning_rate": 1.9782830464885788e-05,
|
| 208 |
+
"loss": 0.02565096616744995,
|
| 209 |
+
"mean_token_accuracy": 0.991170009970665,
|
| 210 |
+
"num_tokens": 6118979.0,
|
| 211 |
+
"step": 95
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"entropy": 0.024970245361328126,
|
| 215 |
+
"epoch": 0.30245746691871456,
|
| 216 |
+
"grad_norm": 0.21180271207392207,
|
| 217 |
+
"learning_rate": 1.974772117649135e-05,
|
| 218 |
+
"loss": 0.024624763429164885,
|
| 219 |
+
"mean_token_accuracy": 0.9913630470633507,
|
| 220 |
+
"num_tokens": 6457003.0,
|
| 221 |
+
"step": 100
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.30245746691871456,
|
| 225 |
+
"eval_entropy": 0.02398535009618994,
|
| 226 |
+
"eval_loss": 0.025832440704107285,
|
| 227 |
+
"eval_mean_token_accuracy": 0.9910635029947436,
|
| 228 |
+
"eval_num_tokens": 6457003.0,
|
| 229 |
+
"eval_runtime": 92.826,
|
| 230 |
+
"eval_samples_per_second": 14.317,
|
| 231 |
+
"eval_steps_per_second": 3.587,
|
| 232 |
+
"step": 100
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"entropy": 0.0255767822265625,
|
| 236 |
+
"epoch": 0.31758034026465026,
|
| 237 |
+
"grad_norm": 0.19039718090172789,
|
| 238 |
+
"learning_rate": 1.9710018421559995e-05,
|
| 239 |
+
"loss": 0.026742279529571533,
|
| 240 |
+
"mean_token_accuracy": 0.9909975379705429,
|
| 241 |
+
"num_tokens": 6768352.0,
|
| 242 |
+
"step": 105
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"entropy": 0.02334747314453125,
|
| 246 |
+
"epoch": 0.332703213610586,
|
| 247 |
+
"grad_norm": 0.15632220147486883,
|
| 248 |
+
"learning_rate": 1.966973223123967e-05,
|
| 249 |
+
"loss": 0.021733593940734864,
|
| 250 |
+
"mean_token_accuracy": 0.9925331950187684,
|
| 251 |
+
"num_tokens": 7086098.0,
|
| 252 |
+
"step": 110
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"entropy": 0.02210235595703125,
|
| 256 |
+
"epoch": 0.34782608695652173,
|
| 257 |
+
"grad_norm": 0.2458450644732953,
|
| 258 |
+
"learning_rate": 1.9626873324023915e-05,
|
| 259 |
+
"loss": 0.02539382576942444,
|
| 260 |
+
"mean_token_accuracy": 0.9917543068528175,
|
| 261 |
+
"num_tokens": 7397835.0,
|
| 262 |
+
"step": 115
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"entropy": 0.022796630859375,
|
| 266 |
+
"epoch": 0.3629489603024575,
|
| 267 |
+
"grad_norm": 0.511944627378666,
|
| 268 |
+
"learning_rate": 1.9581453102900127e-05,
|
| 269 |
+
"loss": 0.022371816635131835,
|
| 270 |
+
"mean_token_accuracy": 0.9924163356423378,
|
| 271 |
+
"num_tokens": 7720248.0,
|
| 272 |
+
"step": 120
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"entropy": 0.02347259521484375,
|
| 276 |
+
"epoch": 0.3780718336483932,
|
| 277 |
+
"grad_norm": 0.22819135667809418,
|
| 278 |
+
"learning_rate": 1.953348365231568e-05,
|
| 279 |
+
"loss": 0.02383442670106888,
|
| 280 |
+
"mean_token_accuracy": 0.9916202604770661,
|
| 281 |
+
"num_tokens": 8046850.0,
|
| 282 |
+
"step": 125
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"entropy": 0.02361907958984375,
|
| 286 |
+
"epoch": 0.3931947069943289,
|
| 287 |
+
"grad_norm": 0.23351254559117343,
|
| 288 |
+
"learning_rate": 1.9482977734962753e-05,
|
| 289 |
+
"loss": 0.0238445445895195,
|
| 290 |
+
"mean_token_accuracy": 0.9918097272515297,
|
| 291 |
+
"num_tokens": 8390721.0,
|
| 292 |
+
"step": 130
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"entropy": 0.0227020263671875,
|
| 296 |
+
"epoch": 0.40831758034026466,
|
| 297 |
+
"grad_norm": 0.18760418206446414,
|
| 298 |
+
"learning_rate": 1.9429948788382726e-05,
|
| 299 |
+
"loss": 0.022852468490600585,
|
| 300 |
+
"mean_token_accuracy": 0.9926092982292175,
|
| 301 |
+
"num_tokens": 8741140.0,
|
| 302 |
+
"step": 135
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"entropy": 0.02464141845703125,
|
| 306 |
+
"epoch": 0.42344045368620037,
|
| 307 |
+
"grad_norm": 0.17532160128342728,
|
| 308 |
+
"learning_rate": 1.937441092139097e-05,
|
| 309 |
+
"loss": 0.025189298391342162,
|
| 310 |
+
"mean_token_accuracy": 0.9915472865104675,
|
| 311 |
+
"num_tokens": 9066217.0,
|
| 312 |
+
"step": 140
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"entropy": 0.022135162353515626,
|
| 316 |
+
"epoch": 0.43856332703213613,
|
| 317 |
+
"grad_norm": 0.2898150379760738,
|
| 318 |
+
"learning_rate": 1.93163789103231e-05,
|
| 319 |
+
"loss": 0.02323928475379944,
|
| 320 |
+
"mean_token_accuracy": 0.9923686638474465,
|
| 321 |
+
"num_tokens": 9398518.0,
|
| 322 |
+
"step": 145
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"entropy": 0.022913360595703126,
|
| 326 |
+
"epoch": 0.45368620037807184,
|
| 327 |
+
"grad_norm": 0.20882908714005557,
|
| 328 |
+
"learning_rate": 1.9255868195103596e-05,
|
| 329 |
+
"loss": 0.023580947518348695,
|
| 330 |
+
"mean_token_accuracy": 0.9922584399580956,
|
| 331 |
+
"num_tokens": 9722146.0,
|
| 332 |
+
"step": 150
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.45368620037807184,
|
| 336 |
+
"eval_entropy": 0.019961612002627627,
|
| 337 |
+
"eval_loss": 0.021232564002275467,
|
| 338 |
+
"eval_mean_token_accuracy": 0.9926312886200868,
|
| 339 |
+
"eval_num_tokens": 9722146.0,
|
| 340 |
+
"eval_runtime": 92.911,
|
| 341 |
+
"eval_samples_per_second": 14.304,
|
| 342 |
+
"eval_steps_per_second": 3.584,
|
| 343 |
+
"step": 150
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"entropy": 0.02125396728515625,
|
| 347 |
+
"epoch": 0.46880907372400754,
|
| 348 |
+
"grad_norm": 0.16611123987930837,
|
| 349 |
+
"learning_rate": 1.9192894875137864e-05,
|
| 350 |
+
"loss": 0.022398817539215087,
|
| 351 |
+
"mean_token_accuracy": 0.9927845776081086,
|
| 352 |
+
"num_tokens": 10051130.0,
|
| 353 |
+
"step": 155
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"entropy": 0.01784820556640625,
|
| 357 |
+
"epoch": 0.4839319470699433,
|
| 358 |
+
"grad_norm": 0.1705996270647505,
|
| 359 |
+
"learning_rate": 1.9127475705028864e-05,
|
| 360 |
+
"loss": 0.020223259925842285,
|
| 361 |
+
"mean_token_accuracy": 0.9935442849993705,
|
| 362 |
+
"num_tokens": 10364936.0,
|
| 363 |
+
"step": 160
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"entropy": 0.02181854248046875,
|
| 367 |
+
"epoch": 0.499054820415879,
|
| 368 |
+
"grad_norm": 0.18993585672995894,
|
| 369 |
+
"learning_rate": 1.9059628090119392e-05,
|
| 370 |
+
"loss": 0.02131263315677643,
|
| 371 |
+
"mean_token_accuracy": 0.9929479479789733,
|
| 372 |
+
"num_tokens": 10684934.0,
|
| 373 |
+
"step": 165
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"entropy": 0.01970977783203125,
|
| 377 |
+
"epoch": 0.5141776937618148,
|
| 378 |
+
"grad_norm": 0.23862681974692412,
|
| 379 |
+
"learning_rate": 1.898937008186123e-05,
|
| 380 |
+
"loss": 0.021003079414367676,
|
| 381 |
+
"mean_token_accuracy": 0.992812205851078,
|
| 382 |
+
"num_tokens": 11016640.0,
|
| 383 |
+
"step": 170
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"entropy": 0.02065277099609375,
|
| 387 |
+
"epoch": 0.5293005671077504,
|
| 388 |
+
"grad_norm": 0.22580792158388535,
|
| 389 |
+
"learning_rate": 1.8916720373012425e-05,
|
| 390 |
+
"loss": 0.020279140770435335,
|
| 391 |
+
"mean_token_accuracy": 0.9934262841939926,
|
| 392 |
+
"num_tokens": 11335375.0,
|
| 393 |
+
"step": 175
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"entropy": 0.02077484130859375,
|
| 397 |
+
"epoch": 0.5444234404536862,
|
| 398 |
+
"grad_norm": 0.18177690286836462,
|
| 399 |
+
"learning_rate": 1.8841698292663887e-05,
|
| 400 |
+
"loss": 0.021297402679920197,
|
| 401 |
+
"mean_token_accuracy": 0.9927030056715012,
|
| 402 |
+
"num_tokens": 11669483.0,
|
| 403 |
+
"step": 180
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"entropy": 0.020357513427734376,
|
| 407 |
+
"epoch": 0.5595463137996219,
|
| 408 |
+
"grad_norm": 0.18951806144385833,
|
| 409 |
+
"learning_rate": 1.876432380109673e-05,
|
| 410 |
+
"loss": 0.02147246301174164,
|
| 411 |
+
"mean_token_accuracy": 0.9929309383034706,
|
| 412 |
+
"num_tokens": 11987399.0,
|
| 413 |
+
"step": 185
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"entropy": 0.021063995361328126,
|
| 417 |
+
"epoch": 0.5746691871455577,
|
| 418 |
+
"grad_norm": 0.28328991051134833,
|
| 419 |
+
"learning_rate": 1.8684617484471662e-05,
|
| 420 |
+
"loss": 0.021729469299316406,
|
| 421 |
+
"mean_token_accuracy": 0.9926604524254798,
|
| 422 |
+
"num_tokens": 12321769.0,
|
| 423 |
+
"step": 190
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"entropy": 0.016484832763671874,
|
| 427 |
+
"epoch": 0.5897920604914934,
|
| 428 |
+
"grad_norm": 0.2604043117050105,
|
| 429 |
+
"learning_rate": 1.860260054935186e-05,
|
| 430 |
+
"loss": 0.016648142039775847,
|
| 431 |
+
"mean_token_accuracy": 0.9943771168589592,
|
| 432 |
+
"num_tokens": 12637459.0,
|
| 433 |
+
"step": 195
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"entropy": 0.016230010986328126,
|
| 437 |
+
"epoch": 0.6049149338374291,
|
| 438 |
+
"grad_norm": 0.1518215594544113,
|
| 439 |
+
"learning_rate": 1.8518294817060777e-05,
|
| 440 |
+
"loss": 0.016636769473552703,
|
| 441 |
+
"mean_token_accuracy": 0.9944247990846634,
|
| 442 |
+
"num_tokens": 12976497.0,
|
| 443 |
+
"step": 200
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.6049149338374291,
|
| 447 |
+
"eval_entropy": 0.01848311324019332,
|
| 448 |
+
"eval_loss": 0.018460111692547798,
|
| 449 |
+
"eval_mean_token_accuracy": 0.9936752308596362,
|
| 450 |
+
"eval_num_tokens": 12976497.0,
|
| 451 |
+
"eval_runtime": 93.7884,
|
| 452 |
+
"eval_samples_per_second": 14.17,
|
| 453 |
+
"eval_steps_per_second": 3.551,
|
| 454 |
+
"step": 200
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"entropy": 0.01815185546875,
|
| 458 |
+
"epoch": 0.6200378071833649,
|
| 459 |
+
"grad_norm": 0.2097083339362261,
|
| 460 |
+
"learning_rate": 1.8431722717876383e-05,
|
| 461 |
+
"loss": 0.01749354600906372,
|
| 462 |
+
"mean_token_accuracy": 0.993867652118206,
|
| 463 |
+
"num_tokens": 13302976.0,
|
| 464 |
+
"step": 205
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"entropy": 0.01755523681640625,
|
| 468 |
+
"epoch": 0.6351606805293005,
|
| 469 |
+
"grad_norm": 0.17477493752672238,
|
| 470 |
+
"learning_rate": 1.834290728506339e-05,
|
| 471 |
+
"loss": 0.018500179052352905,
|
| 472 |
+
"mean_token_accuracy": 0.9936169162392616,
|
| 473 |
+
"num_tokens": 13627997.0,
|
| 474 |
+
"step": 210
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"entropy": 0.018619537353515625,
|
| 478 |
+
"epoch": 0.6502835538752363,
|
| 479 |
+
"grad_norm": 0.1969863482217428,
|
| 480 |
+
"learning_rate": 1.8251872148745075e-05,
|
| 481 |
+
"loss": 0.017858733236789704,
|
| 482 |
+
"mean_token_accuracy": 0.9939072072505951,
|
| 483 |
+
"num_tokens": 13945104.0,
|
| 484 |
+
"step": 215
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"entropy": 0.019687652587890625,
|
| 488 |
+
"epoch": 0.665406427221172,
|
| 489 |
+
"grad_norm": 0.2554011428793856,
|
| 490 |
+
"learning_rate": 1.815864152961624e-05,
|
| 491 |
+
"loss": 0.02146708369255066,
|
| 492 |
+
"mean_token_accuracy": 0.9928086891770362,
|
| 493 |
+
"num_tokens": 14279010.0,
|
| 494 |
+
"step": 220
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"entropy": 0.0164215087890625,
|
| 498 |
+
"epoch": 0.6805293005671077,
|
| 499 |
+
"grad_norm": 0.22053987762264646,
|
| 500 |
+
"learning_rate": 1.806324023249912e-05,
|
| 501 |
+
"loss": 0.017443592846393585,
|
| 502 |
+
"mean_token_accuracy": 0.9943347543478012,
|
| 503 |
+
"num_tokens": 14608053.0,
|
| 504 |
+
"step": 225
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"entropy": 0.01520233154296875,
|
| 508 |
+
"epoch": 0.6956521739130435,
|
| 509 |
+
"grad_norm": 0.15495690088956657,
|
| 510 |
+
"learning_rate": 1.7965693639743816e-05,
|
| 511 |
+
"loss": 0.015465393662452698,
|
| 512 |
+
"mean_token_accuracy": 0.9950109854340553,
|
| 513 |
+
"num_tokens": 14925691.0,
|
| 514 |
+
"step": 230
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"entropy": 0.0169342041015625,
|
| 518 |
+
"epoch": 0.7107750472589792,
|
| 519 |
+
"grad_norm": 0.23516476733142996,
|
| 520 |
+
"learning_rate": 1.786602770447513e-05,
|
| 521 |
+
"loss": 0.017490310966968535,
|
| 522 |
+
"mean_token_accuracy": 0.9939287096261978,
|
| 523 |
+
"num_tokens": 15247510.0,
|
| 524 |
+
"step": 235
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"entropy": 0.015985107421875,
|
| 528 |
+
"epoch": 0.725897920604915,
|
| 529 |
+
"grad_norm": 0.13394070957421397,
|
| 530 |
+
"learning_rate": 1.776426894368746e-05,
|
| 531 |
+
"loss": 0.015802830457687378,
|
| 532 |
+
"mean_token_accuracy": 0.9947777166962624,
|
| 533 |
+
"num_tokens": 15560305.0,
|
| 534 |
+
"step": 240
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"entropy": 0.014200592041015625,
|
| 538 |
+
"epoch": 0.7410207939508506,
|
| 539 |
+
"grad_norm": 0.19432987534155574,
|
| 540 |
+
"learning_rate": 1.766044443118978e-05,
|
| 541 |
+
"loss": 0.014611494541168214,
|
| 542 |
+
"mean_token_accuracy": 0.9948656380176544,
|
| 543 |
+
"num_tokens": 15898787.0,
|
| 544 |
+
"step": 245
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"entropy": 0.0140716552734375,
|
| 548 |
+
"epoch": 0.7561436672967864,
|
| 549 |
+
"grad_norm": 0.16144920711925453,
|
| 550 |
+
"learning_rate": 1.7554581790402372e-05,
|
| 551 |
+
"loss": 0.015314009785652161,
|
| 552 |
+
"mean_token_accuracy": 0.9950292810797692,
|
| 553 |
+
"num_tokens": 16236000.0,
|
| 554 |
+
"step": 250
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.7561436672967864,
|
| 558 |
+
"eval_entropy": 0.01619310350389452,
|
| 559 |
+
"eval_loss": 0.016467994078993797,
|
| 560 |
+
"eval_mean_token_accuracy": 0.9943410491084194,
|
| 561 |
+
"eval_num_tokens": 16236000.0,
|
| 562 |
+
"eval_runtime": 92.8452,
|
| 563 |
+
"eval_samples_per_second": 14.314,
|
| 564 |
+
"eval_steps_per_second": 3.587,
|
| 565 |
+
"step": 250
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"entropy": 0.017685699462890624,
|
| 569 |
+
"epoch": 0.7712665406427222,
|
| 570 |
+
"grad_norm": 0.17754228327893673,
|
| 571 |
+
"learning_rate": 1.7446709187007383e-05,
|
| 572 |
+
"loss": 0.018420988321304323,
|
| 573 |
+
"mean_token_accuracy": 0.9938467770814896,
|
| 574 |
+
"num_tokens": 16568742.0,
|
| 575 |
+
"step": 255
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"entropy": 0.01598701477050781,
|
| 579 |
+
"epoch": 0.7863894139886578,
|
| 580 |
+
"grad_norm": 0.1753647354705704,
|
| 581 |
+
"learning_rate": 1.7336855321455126e-05,
|
| 582 |
+
"loss": 0.016123867034912108,
|
| 583 |
+
"mean_token_accuracy": 0.9946757212281228,
|
| 584 |
+
"num_tokens": 16886171.0,
|
| 585 |
+
"step": 260
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"entropy": 0.01281585693359375,
|
| 589 |
+
"epoch": 0.8015122873345936,
|
| 590 |
+
"grad_norm": 0.1694346707825758,
|
| 591 |
+
"learning_rate": 1.7225049421328024e-05,
|
| 592 |
+
"loss": 0.01325436681509018,
|
| 593 |
+
"mean_token_accuracy": 0.9954081103205681,
|
| 594 |
+
"num_tokens": 17213518.0,
|
| 595 |
+
"step": 265
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"entropy": 0.015825653076171876,
|
| 599 |
+
"epoch": 0.8166351606805293,
|
| 600 |
+
"grad_norm": 0.12320756121826888,
|
| 601 |
+
"learning_rate": 1.7111321233564406e-05,
|
| 602 |
+
"loss": 0.016265182197093962,
|
| 603 |
+
"mean_token_accuracy": 0.9945767849683762,
|
| 604 |
+
"num_tokens": 17526983.0,
|
| 605 |
+
"step": 270
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"entropy": 0.01460418701171875,
|
| 609 |
+
"epoch": 0.831758034026465,
|
| 610 |
+
"grad_norm": 0.15708115348818796,
|
| 611 |
+
"learning_rate": 1.699570101654404e-05,
|
| 612 |
+
"loss": 0.013886567950248719,
|
| 613 |
+
"mean_token_accuracy": 0.9951759412884712,
|
| 614 |
+
"num_tokens": 17844226.0,
|
| 615 |
+
"step": 275
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"entropy": 0.012836456298828125,
|
| 619 |
+
"epoch": 0.8468809073724007,
|
| 620 |
+
"grad_norm": 0.143701427710118,
|
| 621 |
+
"learning_rate": 1.687821953203765e-05,
|
| 622 |
+
"loss": 0.01477896273136139,
|
| 623 |
+
"mean_token_accuracy": 0.9952674239873887,
|
| 624 |
+
"num_tokens": 18154251.0,
|
| 625 |
+
"step": 280
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"entropy": 0.013329315185546874,
|
| 629 |
+
"epoch": 0.8620037807183365,
|
| 630 |
+
"grad_norm": 0.14730660194978837,
|
| 631 |
+
"learning_rate": 1.6758908037022446e-05,
|
| 632 |
+
"loss": 0.01359080672264099,
|
| 633 |
+
"mean_token_accuracy": 0.9955309063196183,
|
| 634 |
+
"num_tokens": 18468667.0,
|
| 635 |
+
"step": 285
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"entropy": 0.01543426513671875,
|
| 639 |
+
"epoch": 0.8771266540642723,
|
| 640 |
+
"grad_norm": 0.13900308540287734,
|
| 641 |
+
"learning_rate": 1.6637798275365976e-05,
|
| 642 |
+
"loss": 0.016150069236755372,
|
| 643 |
+
"mean_token_accuracy": 0.9946491599082947,
|
| 644 |
+
"num_tokens": 18806646.0,
|
| 645 |
+
"step": 290
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"entropy": 0.012978363037109374,
|
| 649 |
+
"epoch": 0.8922495274102079,
|
| 650 |
+
"grad_norm": 0.13732124684095226,
|
| 651 |
+
"learning_rate": 1.651492246938034e-05,
|
| 652 |
+
"loss": 0.012651418149471284,
|
| 653 |
+
"mean_token_accuracy": 0.995641553401947,
|
| 654 |
+
"num_tokens": 19132756.0,
|
| 655 |
+
"step": 295
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"entropy": 0.015012359619140625,
|
| 659 |
+
"epoch": 0.9073724007561437,
|
| 660 |
+
"grad_norm": 0.13977045253544546,
|
| 661 |
+
"learning_rate": 1.6390313311249212e-05,
|
| 662 |
+
"loss": 0.01540960967540741,
|
| 663 |
+
"mean_token_accuracy": 0.9949710816144943,
|
| 664 |
+
"num_tokens": 19465713.0,
|
| 665 |
+
"step": 300
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 0.9073724007561437,
|
| 669 |
+
"eval_entropy": 0.014344957139756944,
|
| 670 |
+
"eval_loss": 0.015235255472362041,
|
| 671 |
+
"eval_mean_token_accuracy": 0.9947822013416806,
|
| 672 |
+
"eval_num_tokens": 19465713.0,
|
| 673 |
+
"eval_runtime": 92.3526,
|
| 674 |
+
"eval_samples_per_second": 14.391,
|
| 675 |
+
"eval_steps_per_second": 3.606,
|
| 676 |
+
"step": 300
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"entropy": 0.013709640502929688,
|
| 680 |
+
"epoch": 0.9224952741020794,
|
| 681 |
+
"grad_norm": 0.19119314364034928,
|
| 682 |
+
"learning_rate": 1.6264003954329762e-05,
|
| 683 |
+
"loss": 0.01464797556400299,
|
| 684 |
+
"mean_token_accuracy": 0.9951745182275772,
|
| 685 |
+
"num_tokens": 19791857.0,
|
| 686 |
+
"step": 305
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"entropy": 0.0113555908203125,
|
| 690 |
+
"epoch": 0.9376181474480151,
|
| 691 |
+
"grad_norm": 0.13382137381631254,
|
| 692 |
+
"learning_rate": 1.613602800433194e-05,
|
| 693 |
+
"loss": 0.012380270659923554,
|
| 694 |
+
"mean_token_accuracy": 0.9959253877401352,
|
| 695 |
+
"num_tokens": 20121726.0,
|
| 696 |
+
"step": 310
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"entropy": 0.013879013061523438,
|
| 700 |
+
"epoch": 0.9527410207939508,
|
| 701 |
+
"grad_norm": 0.1786900745985722,
|
| 702 |
+
"learning_rate": 1.600641951037736e-05,
|
| 703 |
+
"loss": 0.013191039860248565,
|
| 704 |
+
"mean_token_accuracy": 0.9955014079809189,
|
| 705 |
+
"num_tokens": 20434281.0,
|
| 706 |
+
"step": 315
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"entropy": 0.014342498779296876,
|
| 710 |
+
"epoch": 0.9678638941398866,
|
| 711 |
+
"grad_norm": 0.20383676268015227,
|
| 712 |
+
"learning_rate": 1.5875212955940255e-05,
|
| 713 |
+
"loss": 0.01611756980419159,
|
| 714 |
+
"mean_token_accuracy": 0.9947072401642799,
|
| 715 |
+
"num_tokens": 20753074.0,
|
| 716 |
+
"step": 320
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"entropy": 0.013257980346679688,
|
| 720 |
+
"epoch": 0.9829867674858223,
|
| 721 |
+
"grad_norm": 0.1364426748258384,
|
| 722 |
+
"learning_rate": 1.574244324967283e-05,
|
| 723 |
+
"loss": 0.013307541608810425,
|
| 724 |
+
"mean_token_accuracy": 0.9954897597432136,
|
| 725 |
+
"num_tokens": 21086031.0,
|
| 726 |
+
"step": 325
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"entropy": 0.0136016845703125,
|
| 730 |
+
"epoch": 0.998109640831758,
|
| 731 |
+
"grad_norm": 0.11139044118229276,
|
| 732 |
+
"learning_rate": 1.5608145716117507e-05,
|
| 733 |
+
"loss": 0.013101503252983093,
|
| 734 |
+
"mean_token_accuracy": 0.9954297497868538,
|
| 735 |
+
"num_tokens": 21385002.0,
|
| 736 |
+
"step": 330
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"entropy": 0.010610631994298986,
|
| 740 |
+
"epoch": 1.0120982986767486,
|
| 741 |
+
"grad_norm": 0.11937254210200125,
|
| 742 |
+
"learning_rate": 1.547235608630856e-05,
|
| 743 |
+
"loss": 0.009768091142177582,
|
| 744 |
+
"mean_token_accuracy": 0.9965261527009912,
|
| 745 |
+
"num_tokens": 21651447.0,
|
| 746 |
+
"step": 335
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"entropy": 0.008671951293945313,
|
| 750 |
+
"epoch": 1.0272211720226843,
|
| 751 |
+
"grad_norm": 0.12948932598689103,
|
| 752 |
+
"learning_rate": 1.5335110488265497e-05,
|
| 753 |
+
"loss": 0.00888235867023468,
|
| 754 |
+
"mean_token_accuracy": 0.996855866909027,
|
| 755 |
+
"num_tokens": 21981761.0,
|
| 756 |
+
"step": 340
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"entropy": 0.009595108032226563,
|
| 760 |
+
"epoch": 1.0423440453686201,
|
| 761 |
+
"grad_norm": 0.21649437613862998,
|
| 762 |
+
"learning_rate": 1.5196445437380941e-05,
|
| 763 |
+
"loss": 0.010289055854082107,
|
| 764 |
+
"mean_token_accuracy": 0.9964772433042526,
|
| 765 |
+
"num_tokens": 22320903.0,
|
| 766 |
+
"step": 345
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"entropy": 0.00907135009765625,
|
| 770 |
+
"epoch": 1.0574669187145558,
|
| 771 |
+
"grad_norm": 0.12728450692524118,
|
| 772 |
+
"learning_rate": 1.5056397826705356e-05,
|
| 773 |
+
"loss": 0.008513091504573822,
|
| 774 |
+
"mean_token_accuracy": 0.9971326634287834,
|
| 775 |
+
"num_tokens": 22652825.0,
|
| 776 |
+
"step": 350
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.0574669187145558,
|
| 780 |
+
"eval_entropy": 0.011664244505736205,
|
| 781 |
+
"eval_loss": 0.013880621641874313,
|
| 782 |
+
"eval_mean_token_accuracy": 0.9952635595032403,
|
| 783 |
+
"eval_num_tokens": 22652825.0,
|
| 784 |
+
"eval_runtime": 91.6429,
|
| 785 |
+
"eval_samples_per_second": 14.502,
|
| 786 |
+
"eval_steps_per_second": 3.634,
|
| 787 |
+
"step": 350
|
| 788 |
+
}
|
| 789 |
+
],
|
| 790 |
+
"logging_steps": 5,
|
| 791 |
+
"max_steps": 993,
|
| 792 |
+
"num_input_tokens_seen": 0,
|
| 793 |
+
"num_train_epochs": 3,
|
| 794 |
+
"save_steps": 50,
|
| 795 |
+
"stateful_callbacks": {
|
| 796 |
+
"TrainerControl": {
|
| 797 |
+
"args": {
|
| 798 |
+
"should_epoch_stop": false,
|
| 799 |
+
"should_evaluate": false,
|
| 800 |
+
"should_log": false,
|
| 801 |
+
"should_save": true,
|
| 802 |
+
"should_training_stop": false
|
| 803 |
+
},
|
| 804 |
+
"attributes": {}
|
| 805 |
+
}
|
| 806 |
+
},
|
| 807 |
+
"total_flos": 45311893700608.0,
|
| 808 |
+
"train_batch_size": 1,
|
| 809 |
+
"trial_name": null,
|
| 810 |
+
"trial_params": null
|
| 811 |
+
}
|