Upload trainer_state.json with huggingface_hub
Browse files- trainer_state.json +408 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,408 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.549186676994578,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 500,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.061967467079783116,
|
| 13 |
+
"grad_norm": 0.31789204478263855,
|
| 14 |
+
"learning_rate": 0.0001,
|
| 15 |
+
"logits/chosen": -2.037724018096924,
|
| 16 |
+
"logits/rejected": -1.4958155155181885,
|
| 17 |
+
"logps/chosen": -366.8481750488281,
|
| 18 |
+
"logps/rejected": -216.5625,
|
| 19 |
+
"loss": 0.7398,
|
| 20 |
+
"rewards/accuracies": 0.596875011920929,
|
| 21 |
+
"rewards/chosen": 0.47608914971351624,
|
| 22 |
+
"rewards/margins": 0.7042624950408936,
|
| 23 |
+
"rewards/rejected": -0.2281733751296997,
|
| 24 |
+
"step": 20
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.12393493415956623,
|
| 28 |
+
"grad_norm": 0.6692605018615723,
|
| 29 |
+
"learning_rate": 0.0002,
|
| 30 |
+
"logits/chosen": -2.0875039100646973,
|
| 31 |
+
"logits/rejected": -1.591808795928955,
|
| 32 |
+
"logps/chosen": -343.7860412597656,
|
| 33 |
+
"logps/rejected": -225.52999877929688,
|
| 34 |
+
"loss": 0.1009,
|
| 35 |
+
"rewards/accuracies": 0.987500011920929,
|
| 36 |
+
"rewards/chosen": 1.8180320262908936,
|
| 37 |
+
"rewards/margins": 4.439352989196777,
|
| 38 |
+
"rewards/rejected": -2.6213204860687256,
|
| 39 |
+
"step": 40
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.18590240123934934,
|
| 43 |
+
"grad_norm": 0.05950424447655678,
|
| 44 |
+
"learning_rate": 0.0003,
|
| 45 |
+
"logits/chosen": -2.024808168411255,
|
| 46 |
+
"logits/rejected": -1.550957441329956,
|
| 47 |
+
"logps/chosen": -337.0711364746094,
|
| 48 |
+
"logps/rejected": -244.19100952148438,
|
| 49 |
+
"loss": 0.0377,
|
| 50 |
+
"rewards/accuracies": 0.9906250238418579,
|
| 51 |
+
"rewards/chosen": 1.4306976795196533,
|
| 52 |
+
"rewards/margins": 6.377615928649902,
|
| 53 |
+
"rewards/rejected": -4.94691801071167,
|
| 54 |
+
"step": 60
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.24786986831913246,
|
| 58 |
+
"grad_norm": 0.23692739009857178,
|
| 59 |
+
"learning_rate": 0.0004,
|
| 60 |
+
"logits/chosen": -1.8400609493255615,
|
| 61 |
+
"logits/rejected": -1.322597861289978,
|
| 62 |
+
"logps/chosen": -346.5595397949219,
|
| 63 |
+
"logps/rejected": -269.07080078125,
|
| 64 |
+
"loss": 0.0183,
|
| 65 |
+
"rewards/accuracies": 0.996874988079071,
|
| 66 |
+
"rewards/chosen": 0.668520987033844,
|
| 67 |
+
"rewards/margins": 8.514095306396484,
|
| 68 |
+
"rewards/rejected": -7.845574378967285,
|
| 69 |
+
"step": 80
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.30983733539891556,
|
| 73 |
+
"grad_norm": 0.14662492275238037,
|
| 74 |
+
"learning_rate": 0.0005,
|
| 75 |
+
"logits/chosen": -1.7353684902191162,
|
| 76 |
+
"logits/rejected": -1.286170244216919,
|
| 77 |
+
"logps/chosen": -343.4979248046875,
|
| 78 |
+
"logps/rejected": -293.38458251953125,
|
| 79 |
+
"loss": 0.0154,
|
| 80 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 81 |
+
"rewards/chosen": -1.1620686054229736,
|
| 82 |
+
"rewards/margins": 9.110270500183105,
|
| 83 |
+
"rewards/rejected": -10.272339820861816,
|
| 84 |
+
"step": 100
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.3718048024786987,
|
| 88 |
+
"grad_norm": 0.061400506645441055,
|
| 89 |
+
"learning_rate": 0.0004969220851487844,
|
| 90 |
+
"logits/chosen": -1.9946537017822266,
|
| 91 |
+
"logits/rejected": -1.5800933837890625,
|
| 92 |
+
"logps/chosen": -379.28070068359375,
|
| 93 |
+
"logps/rejected": -327.15374755859375,
|
| 94 |
+
"loss": 0.0258,
|
| 95 |
+
"rewards/accuracies": 0.9906250238418579,
|
| 96 |
+
"rewards/chosen": -1.6450515985488892,
|
| 97 |
+
"rewards/margins": 9.554536819458008,
|
| 98 |
+
"rewards/rejected": -11.19958782196045,
|
| 99 |
+
"step": 120
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.4337722695584818,
|
| 103 |
+
"grad_norm": 0.017976779490709305,
|
| 104 |
+
"learning_rate": 0.0004877641290737884,
|
| 105 |
+
"logits/chosen": -2.088599443435669,
|
| 106 |
+
"logits/rejected": -1.5488828420639038,
|
| 107 |
+
"logps/chosen": -364.15057373046875,
|
| 108 |
+
"logps/rejected": -305.5300598144531,
|
| 109 |
+
"loss": 0.0109,
|
| 110 |
+
"rewards/accuracies": 1.0,
|
| 111 |
+
"rewards/chosen": -0.5407935380935669,
|
| 112 |
+
"rewards/margins": 9.945794105529785,
|
| 113 |
+
"rewards/rejected": -10.486587524414062,
|
| 114 |
+
"step": 140
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.4957397366382649,
|
| 118 |
+
"grad_norm": 0.0334528349339962,
|
| 119 |
+
"learning_rate": 0.00047275163104709196,
|
| 120 |
+
"logits/chosen": -1.9635769128799438,
|
| 121 |
+
"logits/rejected": -1.3838129043579102,
|
| 122 |
+
"logps/chosen": -364.9557189941406,
|
| 123 |
+
"logps/rejected": -315.8921813964844,
|
| 124 |
+
"loss": 0.01,
|
| 125 |
+
"rewards/accuracies": 0.996874988079071,
|
| 126 |
+
"rewards/chosen": -1.1100990772247314,
|
| 127 |
+
"rewards/margins": 10.210677146911621,
|
| 128 |
+
"rewards/rejected": -11.320775985717773,
|
| 129 |
+
"step": 160
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.557707203718048,
|
| 133 |
+
"grad_norm": 0.18209697306156158,
|
| 134 |
+
"learning_rate": 0.0004522542485937369,
|
| 135 |
+
"logits/chosen": -1.9845554828643799,
|
| 136 |
+
"logits/rejected": -1.381089448928833,
|
| 137 |
+
"logps/chosen": -362.4132080078125,
|
| 138 |
+
"logps/rejected": -319.27923583984375,
|
| 139 |
+
"loss": 0.0136,
|
| 140 |
+
"rewards/accuracies": 0.996874988079071,
|
| 141 |
+
"rewards/chosen": -1.4699629545211792,
|
| 142 |
+
"rewards/margins": 10.683820724487305,
|
| 143 |
+
"rewards/rejected": -12.153783798217773,
|
| 144 |
+
"step": 180
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.6196746707978311,
|
| 148 |
+
"grad_norm": 0.4615195095539093,
|
| 149 |
+
"learning_rate": 0.00042677669529663686,
|
| 150 |
+
"logits/chosen": -1.8291380405426025,
|
| 151 |
+
"logits/rejected": -1.2410638332366943,
|
| 152 |
+
"logps/chosen": -388.328125,
|
| 153 |
+
"logps/rejected": -355.81683349609375,
|
| 154 |
+
"loss": 0.0155,
|
| 155 |
+
"rewards/accuracies": 0.9906250238418579,
|
| 156 |
+
"rewards/chosen": -2.9048476219177246,
|
| 157 |
+
"rewards/margins": 11.032526969909668,
|
| 158 |
+
"rewards/rejected": -13.93737506866455,
|
| 159 |
+
"step": 200
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.6816421378776143,
|
| 163 |
+
"grad_norm": 0.008454305119812489,
|
| 164 |
+
"learning_rate": 0.0003969463130731183,
|
| 165 |
+
"logits/chosen": -1.880886435508728,
|
| 166 |
+
"logits/rejected": -1.2066755294799805,
|
| 167 |
+
"logps/chosen": -373.3675231933594,
|
| 168 |
+
"logps/rejected": -329.75762939453125,
|
| 169 |
+
"loss": 0.0088,
|
| 170 |
+
"rewards/accuracies": 1.0,
|
| 171 |
+
"rewards/chosen": -1.4875344038009644,
|
| 172 |
+
"rewards/margins": 11.527753829956055,
|
| 173 |
+
"rewards/rejected": -13.015289306640625,
|
| 174 |
+
"step": 220
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.7436096049573974,
|
| 178 |
+
"grad_norm": 0.0020042492542415857,
|
| 179 |
+
"learning_rate": 0.00036349762493488667,
|
| 180 |
+
"logits/chosen": -1.827252984046936,
|
| 181 |
+
"logits/rejected": -1.122102975845337,
|
| 182 |
+
"logps/chosen": -353.72039794921875,
|
| 183 |
+
"logps/rejected": -332.6769104003906,
|
| 184 |
+
"loss": 0.0046,
|
| 185 |
+
"rewards/accuracies": 1.0,
|
| 186 |
+
"rewards/chosen": -1.0817521810531616,
|
| 187 |
+
"rewards/margins": 12.253175735473633,
|
| 188 |
+
"rewards/rejected": -13.334927558898926,
|
| 189 |
+
"step": 240
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.8055770720371804,
|
| 193 |
+
"grad_norm": 0.021263889968395233,
|
| 194 |
+
"learning_rate": 0.00032725424859373687,
|
| 195 |
+
"logits/chosen": -1.9819982051849365,
|
| 196 |
+
"logits/rejected": -1.4159529209136963,
|
| 197 |
+
"logps/chosen": -357.13189697265625,
|
| 198 |
+
"logps/rejected": -341.510986328125,
|
| 199 |
+
"loss": 0.0116,
|
| 200 |
+
"rewards/accuracies": 0.984375,
|
| 201 |
+
"rewards/chosen": -0.6232878565788269,
|
| 202 |
+
"rewards/margins": 11.093478202819824,
|
| 203 |
+
"rewards/rejected": -11.716766357421875,
|
| 204 |
+
"step": 260
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.8675445391169636,
|
| 208 |
+
"grad_norm": 0.002243664814159274,
|
| 209 |
+
"learning_rate": 0.00028910861626005774,
|
| 210 |
+
"logits/chosen": -1.9268354177474976,
|
| 211 |
+
"logits/rejected": -1.3426536321640015,
|
| 212 |
+
"logps/chosen": -355.0523681640625,
|
| 213 |
+
"logps/rejected": -325.38604736328125,
|
| 214 |
+
"loss": 0.0083,
|
| 215 |
+
"rewards/accuracies": 0.996874988079071,
|
| 216 |
+
"rewards/chosen": -0.8373309969902039,
|
| 217 |
+
"rewards/margins": 11.25169563293457,
|
| 218 |
+
"rewards/rejected": -12.089027404785156,
|
| 219 |
+
"step": 280
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.9295120061967467,
|
| 223 |
+
"grad_norm": 0.004915986675769091,
|
| 224 |
+
"learning_rate": 0.00025,
|
| 225 |
+
"logits/chosen": -1.7971941232681274,
|
| 226 |
+
"logits/rejected": -1.1745736598968506,
|
| 227 |
+
"logps/chosen": -369.9619140625,
|
| 228 |
+
"logps/rejected": -328.6756286621094,
|
| 229 |
+
"loss": 0.0072,
|
| 230 |
+
"rewards/accuracies": 1.0,
|
| 231 |
+
"rewards/chosen": -1.5830962657928467,
|
| 232 |
+
"rewards/margins": 11.627036094665527,
|
| 233 |
+
"rewards/rejected": -13.210131645202637,
|
| 234 |
+
"step": 300
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.9914794732765299,
|
| 238 |
+
"grad_norm": 0.012169785797595978,
|
| 239 |
+
"learning_rate": 0.00021089138373994224,
|
| 240 |
+
"logits/chosen": -1.7294784784317017,
|
| 241 |
+
"logits/rejected": -1.1395697593688965,
|
| 242 |
+
"logps/chosen": -374.56390380859375,
|
| 243 |
+
"logps/rejected": -331.4615478515625,
|
| 244 |
+
"loss": 0.0052,
|
| 245 |
+
"rewards/accuracies": 0.996874988079071,
|
| 246 |
+
"rewards/chosen": -1.92715322971344,
|
| 247 |
+
"rewards/margins": 11.8142671585083,
|
| 248 |
+
"rewards/rejected": -13.741419792175293,
|
| 249 |
+
"step": 320
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 1.053446940356313,
|
| 253 |
+
"grad_norm": 0.0052313837222754955,
|
| 254 |
+
"learning_rate": 0.00017274575140626317,
|
| 255 |
+
"logits/chosen": -1.749742865562439,
|
| 256 |
+
"logits/rejected": -1.1283290386199951,
|
| 257 |
+
"logps/chosen": -393.4547119140625,
|
| 258 |
+
"logps/rejected": -364.1820983886719,
|
| 259 |
+
"loss": 0.0063,
|
| 260 |
+
"rewards/accuracies": 1.0,
|
| 261 |
+
"rewards/chosen": -1.6951110363006592,
|
| 262 |
+
"rewards/margins": 12.771199226379395,
|
| 263 |
+
"rewards/rejected": -14.466310501098633,
|
| 264 |
+
"step": 340
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 1.115414407436096,
|
| 268 |
+
"grad_norm": 0.0015018833801150322,
|
| 269 |
+
"learning_rate": 0.00013650237506511331,
|
| 270 |
+
"logits/chosen": -1.7724621295928955,
|
| 271 |
+
"logits/rejected": -1.0941975116729736,
|
| 272 |
+
"logps/chosen": -374.0537414550781,
|
| 273 |
+
"logps/rejected": -344.38726806640625,
|
| 274 |
+
"loss": 0.0031,
|
| 275 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 276 |
+
"rewards/chosen": -2.0093319416046143,
|
| 277 |
+
"rewards/margins": 13.258508682250977,
|
| 278 |
+
"rewards/rejected": -15.267840385437012,
|
| 279 |
+
"step": 360
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 1.1773818745158793,
|
| 283 |
+
"grad_norm": 0.0007948160055093467,
|
| 284 |
+
"learning_rate": 0.00010305368692688174,
|
| 285 |
+
"logits/chosen": -1.8663429021835327,
|
| 286 |
+
"logits/rejected": -1.2882705926895142,
|
| 287 |
+
"logps/chosen": -377.31768798828125,
|
| 288 |
+
"logps/rejected": -347.67315673828125,
|
| 289 |
+
"loss": 0.0032,
|
| 290 |
+
"rewards/accuracies": 1.0,
|
| 291 |
+
"rewards/chosen": -1.45194411277771,
|
| 292 |
+
"rewards/margins": 12.697643280029297,
|
| 293 |
+
"rewards/rejected": -14.149587631225586,
|
| 294 |
+
"step": 380
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 1.2393493415956622,
|
| 298 |
+
"grad_norm": 0.0005872417241334915,
|
| 299 |
+
"learning_rate": 7.322330470336314e-05,
|
| 300 |
+
"logits/chosen": -1.8675537109375,
|
| 301 |
+
"logits/rejected": -1.2785203456878662,
|
| 302 |
+
"logps/chosen": -373.61773681640625,
|
| 303 |
+
"logps/rejected": -339.9030456542969,
|
| 304 |
+
"loss": 0.0039,
|
| 305 |
+
"rewards/accuracies": 0.996874988079071,
|
| 306 |
+
"rewards/chosen": -1.034354329109192,
|
| 307 |
+
"rewards/margins": 12.89478874206543,
|
| 308 |
+
"rewards/rejected": -13.929142951965332,
|
| 309 |
+
"step": 400
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 1.3013168086754454,
|
| 313 |
+
"grad_norm": 0.01801152527332306,
|
| 314 |
+
"learning_rate": 4.7745751406263163e-05,
|
| 315 |
+
"logits/chosen": -1.8331152200698853,
|
| 316 |
+
"logits/rejected": -1.235565423965454,
|
| 317 |
+
"logps/chosen": -366.6528625488281,
|
| 318 |
+
"logps/rejected": -348.25146484375,
|
| 319 |
+
"loss": 0.0031,
|
| 320 |
+
"rewards/accuracies": 1.0,
|
| 321 |
+
"rewards/chosen": -1.2253501415252686,
|
| 322 |
+
"rewards/margins": 13.49341869354248,
|
| 323 |
+
"rewards/rejected": -14.718768119812012,
|
| 324 |
+
"step": 420
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 1.3632842757552286,
|
| 328 |
+
"grad_norm": 0.004230986814945936,
|
| 329 |
+
"learning_rate": 2.7248368952908055e-05,
|
| 330 |
+
"logits/chosen": -1.8243134021759033,
|
| 331 |
+
"logits/rejected": -1.223825216293335,
|
| 332 |
+
"logps/chosen": -374.45404052734375,
|
| 333 |
+
"logps/rejected": -363.57244873046875,
|
| 334 |
+
"loss": 0.0009,
|
| 335 |
+
"rewards/accuracies": 1.0,
|
| 336 |
+
"rewards/chosen": -1.9178287982940674,
|
| 337 |
+
"rewards/margins": 13.485855102539062,
|
| 338 |
+
"rewards/rejected": -15.403681755065918,
|
| 339 |
+
"step": 440
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.4252517428350115,
|
| 343 |
+
"grad_norm": 0.0037465775385499,
|
| 344 |
+
"learning_rate": 1.2235870926211617e-05,
|
| 345 |
+
"logits/chosen": -1.8060734272003174,
|
| 346 |
+
"logits/rejected": -1.2161097526550293,
|
| 347 |
+
"logps/chosen": -367.4268798828125,
|
| 348 |
+
"logps/rejected": -350.69482421875,
|
| 349 |
+
"loss": 0.0019,
|
| 350 |
+
"rewards/accuracies": 1.0,
|
| 351 |
+
"rewards/chosen": -1.0722858905792236,
|
| 352 |
+
"rewards/margins": 13.695907592773438,
|
| 353 |
+
"rewards/rejected": -14.768194198608398,
|
| 354 |
+
"step": 460
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 1.4872192099147947,
|
| 358 |
+
"grad_norm": 0.004549082834273577,
|
| 359 |
+
"learning_rate": 3.077914851215585e-06,
|
| 360 |
+
"logits/chosen": -1.79794180393219,
|
| 361 |
+
"logits/rejected": -1.14779794216156,
|
| 362 |
+
"logps/chosen": -384.23504638671875,
|
| 363 |
+
"logps/rejected": -348.41314697265625,
|
| 364 |
+
"loss": 0.0052,
|
| 365 |
+
"rewards/accuracies": 0.9906250238418579,
|
| 366 |
+
"rewards/chosen": -1.522377371788025,
|
| 367 |
+
"rewards/margins": 13.347444534301758,
|
| 368 |
+
"rewards/rejected": -14.869821548461914,
|
| 369 |
+
"step": 480
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.549186676994578,
|
| 373 |
+
"grad_norm": 0.0164234209805727,
|
| 374 |
+
"learning_rate": 0.0,
|
| 375 |
+
"logits/chosen": -1.8267322778701782,
|
| 376 |
+
"logits/rejected": -1.2138524055480957,
|
| 377 |
+
"logps/chosen": -366.42333984375,
|
| 378 |
+
"logps/rejected": -364.5099182128906,
|
| 379 |
+
"loss": 0.0051,
|
| 380 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 381 |
+
"rewards/chosen": -1.8385913372039795,
|
| 382 |
+
"rewards/margins": 13.76579761505127,
|
| 383 |
+
"rewards/rejected": -15.604388236999512,
|
| 384 |
+
"step": 500
|
| 385 |
+
}
|
| 386 |
+
],
|
| 387 |
+
"logging_steps": 20,
|
| 388 |
+
"max_steps": 500,
|
| 389 |
+
"num_input_tokens_seen": 0,
|
| 390 |
+
"num_train_epochs": 2,
|
| 391 |
+
"save_steps": 50,
|
| 392 |
+
"stateful_callbacks": {
|
| 393 |
+
"TrainerControl": {
|
| 394 |
+
"args": {
|
| 395 |
+
"should_epoch_stop": false,
|
| 396 |
+
"should_evaluate": false,
|
| 397 |
+
"should_log": false,
|
| 398 |
+
"should_save": true,
|
| 399 |
+
"should_training_stop": true
|
| 400 |
+
},
|
| 401 |
+
"attributes": {}
|
| 402 |
+
}
|
| 403 |
+
},
|
| 404 |
+
"total_flos": 0.0,
|
| 405 |
+
"train_batch_size": 2,
|
| 406 |
+
"trial_name": null,
|
| 407 |
+
"trial_params": null
|
| 408 |
+
}
|