xiamengzhou
commited on
Commit
•
dfefcd9
1
Parent(s):
43d496b
update
Browse files- all_results.json +15 -15
- eval_results.json +12 -12
- model-00001-of-00004.safetensors +1 -1
- model-00002-of-00004.safetensors +1 -1
- model-00003-of-00004.safetensors +1 -1
- model-00004-of-00004.safetensors +1 -1
- train_results.json +3 -3
- trainer_state.json +912 -912
all_results.json
CHANGED
|
@@ -1,22 +1,22 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 0.9989258861439313,
|
| 3 |
-
"eval_logits/chosen": -13.
|
| 4 |
-
"eval_logits/rejected": -
|
| 5 |
-
"eval_logps/chosen": -1.
|
| 6 |
-
"eval_logps/rejected": -
|
| 7 |
-
"eval_loss": 2.
|
| 8 |
-
"eval_rewards/accuracies": 0.
|
| 9 |
-
"eval_rewards/chosen": -
|
| 10 |
-
"eval_rewards/margins": 4.
|
| 11 |
-
"eval_rewards/rejected": -
|
| 12 |
-
"eval_runtime":
|
| 13 |
"eval_samples": 1941,
|
| 14 |
-
"eval_samples_per_second": 28.
|
| 15 |
-
"eval_steps_per_second": 0.
|
| 16 |
"total_flos": 0.0,
|
| 17 |
-
"train_loss":
|
| 18 |
-
"train_runtime":
|
| 19 |
"train_samples": 59569,
|
| 20 |
-
"train_samples_per_second": 9.
|
| 21 |
"train_steps_per_second": 0.076
|
| 22 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 0.9989258861439313,
|
| 3 |
+
"eval_logits/chosen": -13.382477760314941,
|
| 4 |
+
"eval_logits/rejected": -13.246561050415039,
|
| 5 |
+
"eval_logps/chosen": -1.8618243932724,
|
| 6 |
+
"eval_logps/rejected": -2.3445940017700195,
|
| 7 |
+
"eval_loss": 2.7497549057006836,
|
| 8 |
+
"eval_rewards/accuracies": 0.7581967115402222,
|
| 9 |
+
"eval_rewards/chosen": -18.618242263793945,
|
| 10 |
+
"eval_rewards/margins": 4.82769775390625,
|
| 11 |
+
"eval_rewards/rejected": -23.445940017700195,
|
| 12 |
+
"eval_runtime": 69.1462,
|
| 13 |
"eval_samples": 1941,
|
| 14 |
+
"eval_samples_per_second": 28.071,
|
| 15 |
+
"eval_steps_per_second": 0.882,
|
| 16 |
"total_flos": 0.0,
|
| 17 |
+
"train_loss": 3.5518602760889197,
|
| 18 |
+
"train_runtime": 6148.2375,
|
| 19 |
"train_samples": 59569,
|
| 20 |
+
"train_samples_per_second": 9.689,
|
| 21 |
"train_steps_per_second": 0.076
|
| 22 |
}
|
eval_results.json
CHANGED
|
@@ -1,16 +1,16 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 0.9989258861439313,
|
| 3 |
-
"eval_logits/chosen": -13.
|
| 4 |
-
"eval_logits/rejected": -
|
| 5 |
-
"eval_logps/chosen": -1.
|
| 6 |
-
"eval_logps/rejected": -
|
| 7 |
-
"eval_loss": 2.
|
| 8 |
-
"eval_rewards/accuracies": 0.
|
| 9 |
-
"eval_rewards/chosen": -
|
| 10 |
-
"eval_rewards/margins": 4.
|
| 11 |
-
"eval_rewards/rejected": -
|
| 12 |
-
"eval_runtime":
|
| 13 |
"eval_samples": 1941,
|
| 14 |
-
"eval_samples_per_second": 28.
|
| 15 |
-
"eval_steps_per_second": 0.
|
| 16 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 0.9989258861439313,
|
| 3 |
+
"eval_logits/chosen": -13.382477760314941,
|
| 4 |
+
"eval_logits/rejected": -13.246561050415039,
|
| 5 |
+
"eval_logps/chosen": -1.8618243932724,
|
| 6 |
+
"eval_logps/rejected": -2.3445940017700195,
|
| 7 |
+
"eval_loss": 2.7497549057006836,
|
| 8 |
+
"eval_rewards/accuracies": 0.7581967115402222,
|
| 9 |
+
"eval_rewards/chosen": -18.618242263793945,
|
| 10 |
+
"eval_rewards/margins": 4.82769775390625,
|
| 11 |
+
"eval_rewards/rejected": -23.445940017700195,
|
| 12 |
+
"eval_runtime": 69.1462,
|
| 13 |
"eval_samples": 1941,
|
| 14 |
+
"eval_samples_per_second": 28.071,
|
| 15 |
+
"eval_steps_per_second": 0.882
|
| 16 |
}
|
model-00001-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4903351912
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a013695a0890be87cca5a3adb49547ffa909b14ead512002cd72106b9a6e2622
|
| 3 |
size 4903351912
|
model-00002-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4947570872
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3f340da7e3a6f7bc0d6611d852f795bea0dba334f0204ea9d1df1dc2a4362cf1
|
| 3 |
size 4947570872
|
model-00003-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4962221464
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3ad5697776ba58c1f956d94382dd441055a3268db05bfc491422377cefbc875a
|
| 3 |
size 4962221464
|
model-00004-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3670322200
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:344c4bb51c8b32bb20005c785098f554b076890133c02b05535b373e4108c300
|
| 3 |
size 3670322200
|
train_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 0.9989258861439313,
|
| 3 |
"total_flos": 0.0,
|
| 4 |
-
"train_loss":
|
| 5 |
-
"train_runtime":
|
| 6 |
"train_samples": 59569,
|
| 7 |
-
"train_samples_per_second": 9.
|
| 8 |
"train_steps_per_second": 0.076
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 0.9989258861439313,
|
| 3 |
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 3.5518602760889197,
|
| 5 |
+
"train_runtime": 6148.2375,
|
| 6 |
"train_samples": 59569,
|
| 7 |
+
"train_samples_per_second": 9.689,
|
| 8 |
"train_steps_per_second": 0.076
|
| 9 |
}
|
trainer_state.json
CHANGED
|
@@ -10,1406 +10,1406 @@
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.010741138560687433,
|
| 13 |
-
"grad_norm": 94.
|
| 14 |
"learning_rate": 8.51063829787234e-08,
|
| 15 |
-
"logits/chosen": -10.
|
| 16 |
-
"logits/rejected": -10.
|
| 17 |
-
"logps/chosen": -0.
|
| 18 |
-
"logps/rejected": -0.
|
| 19 |
-
"loss":
|
| 20 |
"rewards/accuracies": 0.574999988079071,
|
| 21 |
-
"rewards/chosen": -9.
|
| 22 |
-
"rewards/margins": -0.
|
| 23 |
-
"rewards/rejected": -8.
|
| 24 |
"step": 5
|
| 25 |
},
|
| 26 |
{
|
| 27 |
"epoch": 0.021482277121374866,
|
| 28 |
-
"grad_norm": 128.
|
| 29 |
"learning_rate": 1.702127659574468e-07,
|
| 30 |
-
"logits/chosen": -10.
|
| 31 |
-
"logits/rejected": -10.
|
| 32 |
-
"logps/chosen": -1.
|
| 33 |
-
"logps/rejected": -0.
|
| 34 |
-
"loss":
|
| 35 |
-
"rewards/accuracies": 0.
|
| 36 |
-
"rewards/chosen": -10.
|
| 37 |
-
"rewards/margins": -1.
|
| 38 |
-
"rewards/rejected": -8.
|
| 39 |
"step": 10
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.0322234156820623,
|
| 43 |
-
"grad_norm":
|
| 44 |
"learning_rate": 2.553191489361702e-07,
|
| 45 |
-
"logits/chosen": -10.
|
| 46 |
-
"logits/rejected": -10.
|
| 47 |
-
"logps/chosen": -1.
|
| 48 |
-
"logps/rejected": -1.
|
| 49 |
-
"loss":
|
| 50 |
-
"rewards/accuracies": 0.
|
| 51 |
-
"rewards/chosen": -10.
|
| 52 |
-
"rewards/margins": 0.
|
| 53 |
-
"rewards/rejected": -11.
|
| 54 |
"step": 15
|
| 55 |
},
|
| 56 |
{
|
| 57 |
"epoch": 0.04296455424274973,
|
| 58 |
-
"grad_norm":
|
| 59 |
"learning_rate": 3.404255319148936e-07,
|
| 60 |
-
"logits/chosen": -9.
|
| 61 |
-
"logits/rejected": -10.
|
| 62 |
-
"logps/chosen": -0.
|
| 63 |
-
"logps/rejected": -1.
|
| 64 |
-
"loss":
|
| 65 |
"rewards/accuracies": 0.44999998807907104,
|
| 66 |
-
"rewards/chosen": -9.
|
| 67 |
-
"rewards/margins": 1.
|
| 68 |
-
"rewards/rejected": -11.
|
| 69 |
"step": 20
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 0.05370569280343716,
|
| 73 |
-
"grad_norm":
|
| 74 |
"learning_rate": 4.25531914893617e-07,
|
| 75 |
-
"logits/chosen": -10.
|
| 76 |
-
"logits/rejected": -10.
|
| 77 |
-
"logps/chosen": -0.
|
| 78 |
-
"logps/rejected": -1.
|
| 79 |
-
"loss":
|
| 80 |
-
"rewards/accuracies": 0.
|
| 81 |
-
"rewards/chosen": -8.
|
| 82 |
-
"rewards/margins": 3.
|
| 83 |
-
"rewards/rejected": -11.
|
| 84 |
"step": 25
|
| 85 |
},
|
| 86 |
{
|
| 87 |
"epoch": 0.0644468313641246,
|
| 88 |
-
"grad_norm":
|
| 89 |
"learning_rate": 5.106382978723404e-07,
|
| 90 |
-
"logits/chosen": -9.
|
| 91 |
-
"logits/rejected": -9.
|
| 92 |
-
"logps/chosen": -0.
|
| 93 |
-
"logps/rejected": -0.
|
| 94 |
-
"loss":
|
| 95 |
"rewards/accuracies": 0.612500011920929,
|
| 96 |
-
"rewards/chosen": -6.
|
| 97 |
-
"rewards/margins": -0.
|
| 98 |
-
"rewards/rejected": -6.
|
| 99 |
"step": 30
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"epoch": 0.07518796992481203,
|
| 103 |
-
"grad_norm":
|
| 104 |
"learning_rate": 5.957446808510638e-07,
|
| 105 |
-
"logits/chosen": -8.
|
| 106 |
-
"logits/rejected": -8.
|
| 107 |
-
"logps/chosen": -0.
|
| 108 |
-
"logps/rejected": -0.
|
| 109 |
-
"loss":
|
| 110 |
-
"rewards/accuracies": 0.
|
| 111 |
-
"rewards/chosen": -7.
|
| 112 |
-
"rewards/margins": -0.
|
| 113 |
-
"rewards/rejected": -6.
|
| 114 |
"step": 35
|
| 115 |
},
|
| 116 |
{
|
| 117 |
"epoch": 0.08592910848549946,
|
| 118 |
-
"grad_norm":
|
| 119 |
"learning_rate": 6.808510638297872e-07,
|
| 120 |
-
"logits/chosen": -7.
|
| 121 |
-
"logits/rejected": -7.
|
| 122 |
-
"logps/chosen": -0.
|
| 123 |
-
"logps/rejected": -0.
|
| 124 |
-
"loss":
|
| 125 |
-
"rewards/accuracies": 0.
|
| 126 |
-
"rewards/chosen": -7.
|
| 127 |
-
"rewards/margins": 0.
|
| 128 |
-
"rewards/rejected": -7.
|
| 129 |
"step": 40
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 0.0966702470461869,
|
| 133 |
-
"grad_norm":
|
| 134 |
"learning_rate": 7.659574468085107e-07,
|
| 135 |
-
"logits/chosen": -8.
|
| 136 |
-
"logits/rejected": -8.
|
| 137 |
-
"logps/chosen": -0.
|
| 138 |
-
"logps/rejected": -0.
|
| 139 |
-
"loss":
|
| 140 |
-
"rewards/accuracies": 0.
|
| 141 |
-
"rewards/chosen": -5.
|
| 142 |
-
"rewards/margins": 0.
|
| 143 |
-
"rewards/rejected": -
|
| 144 |
"step": 45
|
| 145 |
},
|
| 146 |
{
|
| 147 |
"epoch": 0.10741138560687433,
|
| 148 |
-
"grad_norm":
|
| 149 |
"learning_rate": 7.998983280184396e-07,
|
| 150 |
-
"logits/chosen": -8.
|
| 151 |
-
"logits/rejected": -8.
|
| 152 |
-
"logps/chosen": -0.
|
| 153 |
-
"logps/rejected": -0.
|
| 154 |
-
"loss":
|
| 155 |
-
"rewards/accuracies": 0.
|
| 156 |
-
"rewards/chosen": -5.
|
| 157 |
-
"rewards/margins": 0.
|
| 158 |
-
"rewards/rejected": -5.
|
| 159 |
"step": 50
|
| 160 |
},
|
| 161 |
{
|
| 162 |
"epoch": 0.11815252416756176,
|
| 163 |
-
"grad_norm":
|
| 164 |
"learning_rate": 7.992771864078597e-07,
|
| 165 |
-
"logits/chosen": -
|
| 166 |
-
"logits/rejected": -
|
| 167 |
-
"logps/chosen": -0.
|
| 168 |
-
"logps/rejected": -0.
|
| 169 |
-
"loss":
|
| 170 |
-
"rewards/accuracies": 0.
|
| 171 |
-
"rewards/chosen": -5.
|
| 172 |
-
"rewards/margins":
|
| 173 |
-
"rewards/rejected": -
|
| 174 |
"step": 55
|
| 175 |
},
|
| 176 |
{
|
| 177 |
"epoch": 0.1288936627282492,
|
| 178 |
-
"grad_norm":
|
| 179 |
"learning_rate": 7.980922636120897e-07,
|
| 180 |
-
"logits/chosen": -8.
|
| 181 |
-
"logits/rejected": -
|
| 182 |
-
"logps/chosen": -0.
|
| 183 |
-
"logps/rejected": -0.
|
| 184 |
-
"loss":
|
| 185 |
-
"rewards/accuracies": 0.
|
| 186 |
-
"rewards/chosen": -5.
|
| 187 |
-
"rewards/margins":
|
| 188 |
-
"rewards/rejected": -6.
|
| 189 |
"step": 60
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"epoch": 0.13963480128893663,
|
| 193 |
-
"grad_norm":
|
| 194 |
"learning_rate": 7.963452327474534e-07,
|
| 195 |
-
"logits/chosen": -
|
| 196 |
-
"logits/rejected": -
|
| 197 |
-
"logps/chosen": -0.
|
| 198 |
-
"logps/rejected": -0.
|
| 199 |
-
"loss":
|
| 200 |
-
"rewards/accuracies": 0.
|
| 201 |
-
"rewards/chosen": -
|
| 202 |
-
"rewards/margins":
|
| 203 |
-
"rewards/rejected": -
|
| 204 |
"step": 65
|
| 205 |
},
|
| 206 |
{
|
| 207 |
"epoch": 0.15037593984962405,
|
| 208 |
-
"grad_norm":
|
| 209 |
"learning_rate": 7.940385606293987e-07,
|
| 210 |
-
"logits/chosen": -8.
|
| 211 |
-
"logits/rejected": -8.
|
| 212 |
-
"logps/chosen": -0.
|
| 213 |
-
"logps/rejected": -0.
|
| 214 |
-
"loss":
|
| 215 |
-
"rewards/accuracies": 0.
|
| 216 |
-
"rewards/chosen": -6.
|
| 217 |
-
"rewards/margins":
|
| 218 |
-
"rewards/rejected": -
|
| 219 |
"step": 70
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.1611170784103115,
|
| 223 |
-
"grad_norm":
|
| 224 |
"learning_rate": 7.911755042893434e-07,
|
| 225 |
-
"logits/chosen": -
|
| 226 |
-
"logits/rejected": -8.
|
| 227 |
-
"logps/chosen": -0.
|
| 228 |
-
"logps/rejected": -0.
|
| 229 |
-
"loss":
|
| 230 |
-
"rewards/accuracies": 0.
|
| 231 |
-
"rewards/chosen": -6.
|
| 232 |
-
"rewards/margins": 0.
|
| 233 |
-
"rewards/rejected": -
|
| 234 |
"step": 75
|
| 235 |
},
|
| 236 |
{
|
| 237 |
"epoch": 0.17185821697099893,
|
| 238 |
-
"grad_norm":
|
| 239 |
"learning_rate": 7.877601063757321e-07,
|
| 240 |
-
"logits/chosen": -
|
| 241 |
-
"logits/rejected": -
|
| 242 |
-
"logps/chosen": -0.
|
| 243 |
-
"logps/rejected": -0.
|
| 244 |
-
"loss":
|
| 245 |
-
"rewards/accuracies": 0.
|
| 246 |
-
"rewards/chosen": -6.
|
| 247 |
-
"rewards/margins": 1.
|
| 248 |
-
"rewards/rejected": -
|
| 249 |
"step": 80
|
| 250 |
},
|
| 251 |
{
|
| 252 |
"epoch": 0.18259935553168635,
|
| 253 |
-
"grad_norm":
|
| 254 |
"learning_rate": 7.837971894457989e-07,
|
| 255 |
-
"logits/chosen": -
|
| 256 |
-
"logits/rejected": -
|
| 257 |
-
"logps/chosen": -0.
|
| 258 |
-
"logps/rejected": -0.
|
| 259 |
-
"loss":
|
| 260 |
"rewards/accuracies": 0.6000000238418579,
|
| 261 |
-
"rewards/chosen": -6.
|
| 262 |
-
"rewards/margins":
|
| 263 |
-
"rewards/rejected": -
|
| 264 |
"step": 85
|
| 265 |
},
|
| 266 |
{
|
| 267 |
"epoch": 0.1933404940923738,
|
| 268 |
-
"grad_norm":
|
| 269 |
"learning_rate": 7.792923491560942e-07,
|
| 270 |
-
"logits/chosen": -
|
| 271 |
-
"logits/rejected": -
|
| 272 |
-
"logps/chosen": -0.
|
| 273 |
-
"logps/rejected": -0.
|
| 274 |
-
"loss":
|
| 275 |
-
"rewards/accuracies": 0.
|
| 276 |
-
"rewards/chosen": -
|
| 277 |
-
"rewards/margins": 0.
|
| 278 |
-
"rewards/rejected": -
|
| 279 |
"step": 90
|
| 280 |
},
|
| 281 |
{
|
| 282 |
"epoch": 0.20408163265306123,
|
| 283 |
-
"grad_norm":
|
| 284 |
"learning_rate": 7.742519463613926e-07,
|
| 285 |
-
"logits/chosen": -
|
| 286 |
-
"logits/rejected": -
|
| 287 |
-
"logps/chosen": -0.
|
| 288 |
-
"logps/rejected": -0.
|
| 289 |
-
"loss":
|
| 290 |
-
"rewards/accuracies": 0.
|
| 291 |
-
"rewards/chosen": -
|
| 292 |
-
"rewards/margins": 0.
|
| 293 |
-
"rewards/rejected": -
|
| 294 |
"step": 95
|
| 295 |
},
|
| 296 |
{
|
| 297 |
"epoch": 0.21482277121374865,
|
| 298 |
-
"grad_norm":
|
| 299 |
"learning_rate": 7.68683098133138e-07,
|
| 300 |
-
"logits/chosen": -8.
|
| 301 |
-
"logits/rejected": -
|
| 302 |
-
"logps/chosen": -0.
|
| 303 |
-
"logps/rejected": -0.
|
| 304 |
-
"loss":
|
| 305 |
-
"rewards/accuracies": 0.
|
| 306 |
-
"rewards/chosen": -
|
| 307 |
-
"rewards/margins": 1.
|
| 308 |
-
"rewards/rejected": -
|
| 309 |
"step": 100
|
| 310 |
},
|
| 311 |
{
|
| 312 |
"epoch": 0.22556390977443608,
|
| 313 |
-
"grad_norm":
|
| 314 |
"learning_rate": 7.625936677101051e-07,
|
| 315 |
-
"logits/chosen": -
|
| 316 |
-
"logits/rejected": -
|
| 317 |
-
"logps/chosen": -0.
|
| 318 |
-
"logps/rejected": -0.
|
| 319 |
-
"loss":
|
| 320 |
-
"rewards/accuracies": 0.
|
| 321 |
-
"rewards/chosen": -
|
| 322 |
-
"rewards/margins": -0.
|
| 323 |
-
"rewards/rejected": -
|
| 324 |
"step": 105
|
| 325 |
},
|
| 326 |
{
|
| 327 |
"epoch": 0.23630504833512353,
|
| 328 |
-
"grad_norm":
|
| 329 |
"learning_rate": 7.559922533954731e-07,
|
| 330 |
-
"logits/chosen": -
|
| 331 |
-
"logits/rejected": -
|
| 332 |
-
"logps/chosen": -0.
|
| 333 |
-
"logps/rejected": -0.
|
| 334 |
-
"loss":
|
| 335 |
-
"rewards/accuracies": 0.
|
| 336 |
-
"rewards/chosen": -
|
| 337 |
-
"rewards/margins": 1.
|
| 338 |
-
"rewards/rejected": -
|
| 339 |
"step": 110
|
| 340 |
},
|
| 341 |
{
|
| 342 |
"epoch": 0.24704618689581095,
|
| 343 |
-
"grad_norm":
|
| 344 |
"learning_rate": 7.488881764159808e-07,
|
| 345 |
-
"logits/chosen": -
|
| 346 |
-
"logits/rejected": -
|
| 347 |
-
"logps/chosen": -0.
|
| 348 |
-
"logps/rejected": -0.
|
| 349 |
-
"loss":
|
| 350 |
"rewards/accuracies": 0.6499999761581421,
|
| 351 |
-
"rewards/chosen": -
|
| 352 |
-
"rewards/margins": 0.
|
| 353 |
-
"rewards/rejected": -7.
|
| 354 |
"step": 115
|
| 355 |
},
|
| 356 |
{
|
| 357 |
"epoch": 0.2577873254564984,
|
| 358 |
-
"grad_norm":
|
| 359 |
"learning_rate": 7.412914677603135e-07,
|
| 360 |
-
"logits/chosen": -
|
| 361 |
-
"logits/rejected": -
|
| 362 |
-
"logps/chosen": -0.
|
| 363 |
-
"logps/rejected": -0.
|
| 364 |
-
"loss":
|
| 365 |
-
"rewards/accuracies": 0.
|
| 366 |
-
"rewards/chosen": -
|
| 367 |
-
"rewards/margins": 0.
|
| 368 |
-
"rewards/rejected": -
|
| 369 |
"step": 120
|
| 370 |
},
|
| 371 |
{
|
| 372 |
"epoch": 0.26852846401718583,
|
| 373 |
-
"grad_norm":
|
| 374 |
"learning_rate": 7.332128540153017e-07,
|
| 375 |
-
"logits/chosen": -
|
| 376 |
-
"logits/rejected": -
|
| 377 |
-
"logps/chosen": -0.
|
| 378 |
-
"logps/rejected": -0.
|
| 379 |
-
"loss":
|
| 380 |
"rewards/accuracies": 0.6625000238418579,
|
| 381 |
-
"rewards/chosen": -
|
| 382 |
-
"rewards/margins": 1.
|
| 383 |
-
"rewards/rejected": -
|
| 384 |
"step": 125
|
| 385 |
},
|
| 386 |
{
|
| 387 |
"epoch": 0.27926960257787325,
|
| 388 |
-
"grad_norm":
|
| 389 |
"learning_rate": 7.246637422199322e-07,
|
| 390 |
-
"logits/chosen": -
|
| 391 |
-
"logits/rejected": -
|
| 392 |
-
"logps/chosen": -0.
|
| 393 |
-
"logps/rejected": -
|
| 394 |
-
"loss":
|
| 395 |
-
"rewards/accuracies": 0.
|
| 396 |
-
"rewards/chosen": -
|
| 397 |
-
"rewards/margins": 1.
|
| 398 |
-
"rewards/rejected": -
|
| 399 |
"step": 130
|
| 400 |
},
|
| 401 |
{
|
| 402 |
"epoch": 0.2900107411385607,
|
| 403 |
-
"grad_norm":
|
| 404 |
"learning_rate": 7.156562037585574e-07,
|
| 405 |
-
"logits/chosen": -
|
| 406 |
-
"logits/rejected": -
|
| 407 |
-
"logps/chosen": -0.
|
| 408 |
-
"logps/rejected": -
|
| 409 |
-
"loss": 3.
|
| 410 |
-
"rewards/accuracies": 0.
|
| 411 |
-
"rewards/chosen": -
|
| 412 |
-
"rewards/margins":
|
| 413 |
-
"rewards/rejected": -
|
| 414 |
"step": 135
|
| 415 |
},
|
| 416 |
{
|
| 417 |
"epoch": 0.3007518796992481,
|
| 418 |
-
"grad_norm":
|
| 419 |
"learning_rate": 7.062029573160467e-07,
|
| 420 |
-
"logits/chosen": -
|
| 421 |
-
"logits/rejected": -
|
| 422 |
-
"logps/chosen": -0.
|
| 423 |
-
"logps/rejected": -
|
| 424 |
-
"loss": 3.
|
| 425 |
-
"rewards/accuracies": 0.
|
| 426 |
-
"rewards/chosen": -
|
| 427 |
-
"rewards/margins":
|
| 428 |
-
"rewards/rejected": -
|
| 429 |
"step": 140
|
| 430 |
},
|
| 431 |
{
|
| 432 |
"epoch": 0.31149301825993553,
|
| 433 |
-
"grad_norm":
|
| 434 |
"learning_rate": 6.963173509189455e-07,
|
| 435 |
-
"logits/chosen": -
|
| 436 |
-
"logits/rejected": -
|
| 437 |
-
"logps/chosen": -
|
| 438 |
-
"logps/rejected": -
|
| 439 |
-
"loss": 3.
|
| 440 |
"rewards/accuracies": 0.5375000238418579,
|
| 441 |
-
"rewards/chosen": -
|
| 442 |
-
"rewards/margins":
|
| 443 |
-
"rewards/rejected": -
|
| 444 |
"step": 145
|
| 445 |
},
|
| 446 |
{
|
| 447 |
"epoch": 0.322234156820623,
|
| 448 |
-
"grad_norm":
|
| 449 |
"learning_rate": 6.860133430880024e-07,
|
| 450 |
-
"logits/chosen": -
|
| 451 |
-
"logits/rejected": -
|
| 452 |
-
"logps/chosen": -
|
| 453 |
-
"logps/rejected": -1.
|
| 454 |
-
"loss": 3.
|
| 455 |
-
"rewards/accuracies": 0.
|
| 456 |
-
"rewards/chosen": -
|
| 457 |
-
"rewards/margins":
|
| 458 |
-
"rewards/rejected": -
|
| 459 |
"step": 150
|
| 460 |
},
|
| 461 |
{
|
| 462 |
"epoch": 0.33297529538131043,
|
| 463 |
-
"grad_norm":
|
| 464 |
"learning_rate": 6.753054831286747e-07,
|
| 465 |
-
"logits/chosen": -
|
| 466 |
-
"logits/rejected": -
|
| 467 |
-
"logps/chosen": -1.
|
| 468 |
-
"logps/rejected": -1.
|
| 469 |
-
"loss": 3.
|
| 470 |
-
"rewards/accuracies": 0.
|
| 471 |
-
"rewards/chosen": -
|
| 472 |
-
"rewards/margins":
|
| 473 |
-
"rewards/rejected": -
|
| 474 |
"step": 155
|
| 475 |
},
|
| 476 |
{
|
| 477 |
"epoch": 0.34371643394199786,
|
| 478 |
-
"grad_norm":
|
| 479 |
"learning_rate": 6.642088905874433e-07,
|
| 480 |
-
"logits/chosen": -
|
| 481 |
-
"logits/rejected": -
|
| 482 |
-
"logps/chosen": -1.
|
| 483 |
-
"logps/rejected": -1.
|
| 484 |
-
"loss": 3.
|
| 485 |
-
"rewards/accuracies": 0.
|
| 486 |
-
"rewards/chosen": -
|
| 487 |
-
"rewards/margins":
|
| 488 |
-
"rewards/rejected": -
|
| 489 |
"step": 160
|
| 490 |
},
|
| 491 |
{
|
| 492 |
"epoch": 0.3544575725026853,
|
| 493 |
-
"grad_norm":
|
| 494 |
"learning_rate": 6.527392339029455e-07,
|
| 495 |
-
"logits/chosen": -
|
| 496 |
-
"logits/rejected": -
|
| 497 |
-
"logps/chosen": -
|
| 498 |
-
"logps/rejected": -1.
|
| 499 |
-
"loss":
|
| 500 |
-
"rewards/accuracies": 0.
|
| 501 |
-
"rewards/chosen": -
|
| 502 |
-
"rewards/margins":
|
| 503 |
-
"rewards/rejected": -
|
| 504 |
"step": 165
|
| 505 |
},
|
| 506 |
{
|
| 507 |
"epoch": 0.3651987110633727,
|
| 508 |
-
"grad_norm":
|
| 509 |
"learning_rate": 6.409127082820689e-07,
|
| 510 |
-
"logits/chosen": -
|
| 511 |
-
"logits/rejected": -
|
| 512 |
-
"logps/chosen": -1.
|
| 513 |
-
"logps/rejected": -1.
|
| 514 |
-
"loss": 3.
|
| 515 |
-
"rewards/accuracies": 0.
|
| 516 |
-
"rewards/chosen": -
|
| 517 |
-
"rewards/margins":
|
| 518 |
-
"rewards/rejected": -
|
| 519 |
"step": 170
|
| 520 |
},
|
| 521 |
{
|
| 522 |
"epoch": 0.37593984962406013,
|
| 523 |
-
"grad_norm":
|
| 524 |
"learning_rate": 6.287460128322457e-07,
|
| 525 |
-
"logits/chosen": -
|
| 526 |
-
"logits/rejected": -
|
| 527 |
-
"logps/chosen": -1.
|
| 528 |
-
"logps/rejected": -1.
|
| 529 |
-
"loss":
|
| 530 |
-
"rewards/accuracies": 0.
|
| 531 |
-
"rewards/chosen": -
|
| 532 |
-
"rewards/margins":
|
| 533 |
-
"rewards/rejected": -
|
| 534 |
"step": 175
|
| 535 |
},
|
| 536 |
{
|
| 537 |
"epoch": 0.3866809881847476,
|
| 538 |
-
"grad_norm":
|
| 539 |
"learning_rate": 6.16256326982239e-07,
|
| 540 |
-
"logits/chosen": -
|
| 541 |
-
"logits/rejected": -
|
| 542 |
-
"logps/chosen": -1.
|
| 543 |
-
"logps/rejected": -1.
|
| 544 |
-
"loss":
|
| 545 |
-
"rewards/accuracies": 0.
|
| 546 |
-
"rewards/chosen": -
|
| 547 |
-
"rewards/margins":
|
| 548 |
-
"rewards/rejected": -
|
| 549 |
"step": 180
|
| 550 |
},
|
| 551 |
{
|
| 552 |
"epoch": 0.39742212674543503,
|
| 553 |
-
"grad_norm":
|
| 554 |
"learning_rate": 6.034612862247114e-07,
|
| 555 |
-
"logits/chosen": -
|
| 556 |
-
"logits/rejected": -
|
| 557 |
-
"logps/chosen": -1.
|
| 558 |
-
"logps/rejected": -1.
|
| 559 |
-
"loss":
|
| 560 |
-
"rewards/accuracies": 0.
|
| 561 |
-
"rewards/chosen": -
|
| 562 |
-
"rewards/margins": 2.
|
| 563 |
-
"rewards/rejected": -
|
| 564 |
"step": 185
|
| 565 |
},
|
| 566 |
{
|
| 567 |
"epoch": 0.40816326530612246,
|
| 568 |
-
"grad_norm":
|
| 569 |
"learning_rate": 5.903789572148295e-07,
|
| 570 |
-
"logits/chosen": -
|
| 571 |
-
"logits/rejected": -
|
| 572 |
-
"logps/chosen": -1.
|
| 573 |
-
"logps/rejected": -
|
| 574 |
-
"loss":
|
| 575 |
-
"rewards/accuracies": 0.
|
| 576 |
-
"rewards/chosen": -
|
| 577 |
-
"rewards/margins":
|
| 578 |
-
"rewards/rejected": -
|
| 579 |
"step": 190
|
| 580 |
},
|
| 581 |
{
|
| 582 |
"epoch": 0.4189044038668099,
|
| 583 |
-
"grad_norm":
|
| 584 |
"learning_rate": 5.770278122600662e-07,
|
| 585 |
-
"logits/chosen": -
|
| 586 |
-
"logits/rejected": -
|
| 587 |
-
"logps/chosen": -1.
|
| 588 |
-
"logps/rejected": -1.
|
| 589 |
-
"loss":
|
| 590 |
-
"rewards/accuracies": 0.
|
| 591 |
-
"rewards/chosen": -
|
| 592 |
-
"rewards/margins":
|
| 593 |
-
"rewards/rejected": -
|
| 594 |
"step": 195
|
| 595 |
},
|
| 596 |
{
|
| 597 |
"epoch": 0.4296455424274973,
|
| 598 |
-
"grad_norm":
|
| 599 |
"learning_rate": 5.634267032372192e-07,
|
| 600 |
-
"logits/chosen": -
|
| 601 |
-
"logits/rejected": -
|
| 602 |
-
"logps/chosen": -1.
|
| 603 |
-
"logps/rejected": -1.
|
| 604 |
-
"loss":
|
| 605 |
"rewards/accuracies": 0.699999988079071,
|
| 606 |
-
"rewards/chosen": -
|
| 607 |
-
"rewards/margins":
|
| 608 |
-
"rewards/rejected": -
|
| 609 |
"step": 200
|
| 610 |
},
|
| 611 |
{
|
| 612 |
"epoch": 0.44038668098818473,
|
| 613 |
-
"grad_norm":
|
| 614 |
"learning_rate": 5.495948349734758e-07,
|
| 615 |
-
"logits/chosen": -
|
| 616 |
-
"logits/rejected": -
|
| 617 |
-
"logps/chosen": -1.
|
| 618 |
-
"logps/rejected": -
|
| 619 |
-
"loss":
|
| 620 |
-
"rewards/accuracies": 0.
|
| 621 |
-
"rewards/chosen": -
|
| 622 |
-
"rewards/margins":
|
| 623 |
-
"rewards/rejected": -
|
| 624 |
"step": 205
|
| 625 |
},
|
| 626 |
{
|
| 627 |
"epoch": 0.45112781954887216,
|
| 628 |
-
"grad_norm":
|
| 629 |
"learning_rate": 5.355517381291105e-07,
|
| 630 |
-
"logits/chosen": -
|
| 631 |
-
"logits/rejected": -
|
| 632 |
-
"logps/chosen": -1.
|
| 633 |
-
"logps/rejected": -
|
| 634 |
-
"loss":
|
| 635 |
"rewards/accuracies": 0.75,
|
| 636 |
-
"rewards/chosen": -
|
| 637 |
-
"rewards/margins":
|
| 638 |
-
"rewards/rejected": -
|
| 639 |
"step": 210
|
| 640 |
},
|
| 641 |
{
|
| 642 |
"epoch": 0.46186895810955964,
|
| 643 |
-
"grad_norm":
|
| 644 |
"learning_rate": 5.21317241620105e-07,
|
| 645 |
-
"logits/chosen": -
|
| 646 |
-
"logits/rejected": -
|
| 647 |
-
"logps/chosen": -1.
|
| 648 |
-
"logps/rejected": -
|
| 649 |
-
"loss":
|
| 650 |
-
"rewards/accuracies": 0.
|
| 651 |
-
"rewards/chosen": -
|
| 652 |
-
"rewards/margins":
|
| 653 |
-
"rewards/rejected": -
|
| 654 |
"step": 215
|
| 655 |
},
|
| 656 |
{
|
| 657 |
"epoch": 0.47261009667024706,
|
| 658 |
-
"grad_norm":
|
| 659 |
"learning_rate": 5.069114446196291e-07,
|
| 660 |
-
"logits/chosen": -
|
| 661 |
-
"logits/rejected": -
|
| 662 |
-
"logps/chosen": -1.
|
| 663 |
-
"logps/rejected": -
|
| 664 |
-
"loss":
|
| 665 |
"rewards/accuracies": 0.8500000238418579,
|
| 666 |
-
"rewards/chosen": -
|
| 667 |
-
"rewards/margins":
|
| 668 |
-
"rewards/rejected": -
|
| 669 |
"step": 220
|
| 670 |
},
|
| 671 |
{
|
| 672 |
"epoch": 0.4833512352309345,
|
| 673 |
-
"grad_norm":
|
| 674 |
"learning_rate": 4.923546881779183e-07,
|
| 675 |
-
"logits/chosen": -
|
| 676 |
-
"logits/rejected": -
|
| 677 |
-
"logps/chosen": -1.
|
| 678 |
-
"logps/rejected": -1.
|
| 679 |
-
"loss":
|
| 680 |
-
"rewards/accuracies": 0.
|
| 681 |
-
"rewards/chosen": -
|
| 682 |
-
"rewards/margins":
|
| 683 |
-
"rewards/rejected": -
|
| 684 |
"step": 225
|
| 685 |
},
|
| 686 |
{
|
| 687 |
"epoch": 0.4940923737916219,
|
| 688 |
-
"grad_norm":
|
| 689 |
"learning_rate": 4.776675265006186e-07,
|
| 690 |
-
"logits/chosen": -
|
| 691 |
-
"logits/rejected": -
|
| 692 |
-
"logps/chosen": -1.
|
| 693 |
-
"logps/rejected": -
|
| 694 |
-
"loss":
|
| 695 |
-
"rewards/accuracies": 0.
|
| 696 |
-
"rewards/chosen": -
|
| 697 |
-
"rewards/margins":
|
| 698 |
-
"rewards/rejected": -
|
| 699 |
"step": 230
|
| 700 |
},
|
| 701 |
{
|
| 702 |
"epoch": 0.5048335123523093,
|
| 703 |
-
"grad_norm":
|
| 704 |
"learning_rate": 4.62870697926156e-07,
|
| 705 |
-
"logits/chosen": -
|
| 706 |
-
"logits/rejected": -
|
| 707 |
-
"logps/chosen": -1.
|
| 708 |
-
"logps/rejected": -
|
| 709 |
-
"loss":
|
| 710 |
-
"rewards/accuracies": 0.
|
| 711 |
-
"rewards/chosen": -
|
| 712 |
-
"rewards/margins":
|
| 713 |
-
"rewards/rejected": -
|
| 714 |
"step": 235
|
| 715 |
},
|
| 716 |
{
|
| 717 |
"epoch": 0.5155746509129968,
|
| 718 |
-
"grad_norm":
|
| 719 |
"learning_rate": 4.479850956431092e-07,
|
| 720 |
-
"logits/chosen": -
|
| 721 |
-
"logits/rejected": -
|
| 722 |
-
"logps/chosen": -1.
|
| 723 |
-
"logps/rejected": -
|
| 724 |
-
"loss":
|
| 725 |
-
"rewards/accuracies": 0.
|
| 726 |
-
"rewards/chosen": -
|
| 727 |
-
"rewards/margins":
|
| 728 |
-
"rewards/rejected": -
|
| 729 |
"step": 240
|
| 730 |
},
|
| 731 |
{
|
| 732 |
"epoch": 0.5263157894736842,
|
| 733 |
-
"grad_norm":
|
| 734 |
"learning_rate": 4.33031738188933e-07,
|
| 735 |
-
"logits/chosen": -
|
| 736 |
-
"logits/rejected": -
|
| 737 |
-
"logps/chosen": -1.
|
| 738 |
-
"logps/rejected": -
|
| 739 |
-
"loss":
|
| 740 |
-
"rewards/accuracies": 0.
|
| 741 |
-
"rewards/chosen": -
|
| 742 |
-
"rewards/margins": 3.
|
| 743 |
-
"rewards/rejected": -
|
| 744 |
"step": 245
|
| 745 |
},
|
| 746 |
{
|
| 747 |
"epoch": 0.5370569280343717,
|
| 748 |
-
"grad_norm":
|
| 749 |
"learning_rate": 4.180317397716889e-07,
|
| 750 |
-
"logits/chosen": -
|
| 751 |
-
"logits/rejected": -
|
| 752 |
-
"logps/chosen": -1.
|
| 753 |
-
"logps/rejected": -
|
| 754 |
-
"loss":
|
| 755 |
-
"rewards/accuracies": 0.
|
| 756 |
-
"rewards/chosen": -
|
| 757 |
-
"rewards/margins":
|
| 758 |
-
"rewards/rejected": -
|
| 759 |
"step": 250
|
| 760 |
},
|
| 761 |
{
|
| 762 |
"epoch": 0.547798066595059,
|
| 763 |
-
"grad_norm":
|
| 764 |
"learning_rate": 4.030062804566888e-07,
|
| 765 |
-
"logits/chosen": -
|
| 766 |
-
"logits/rejected": -
|
| 767 |
-
"logps/chosen": -1.
|
| 768 |
-
"logps/rejected": -1.
|
| 769 |
-
"loss":
|
| 770 |
"rewards/accuracies": 0.6499999761581421,
|
| 771 |
-
"rewards/chosen": -
|
| 772 |
-
"rewards/margins": 2.
|
| 773 |
-
"rewards/rejected": -
|
| 774 |
"step": 255
|
| 775 |
},
|
| 776 |
{
|
| 777 |
"epoch": 0.5585392051557465,
|
| 778 |
-
"grad_norm":
|
| 779 |
"learning_rate": 3.8797657626014614e-07,
|
| 780 |
-
"logits/chosen": -
|
| 781 |
-
"logits/rejected": -
|
| 782 |
-
"logps/chosen": -1.
|
| 783 |
-
"logps/rejected": -
|
| 784 |
-
"loss":
|
| 785 |
-
"rewards/accuracies": 0.
|
| 786 |
-
"rewards/chosen": -
|
| 787 |
-
"rewards/margins":
|
| 788 |
-
"rewards/rejected": -
|
| 789 |
"step": 260
|
| 790 |
},
|
| 791 |
{
|
| 792 |
"epoch": 0.569280343716434,
|
| 793 |
-
"grad_norm":
|
| 794 |
"learning_rate": 3.729638491920669e-07,
|
| 795 |
-
"logits/chosen": -
|
| 796 |
-
"logits/rejected": -
|
| 797 |
-
"logps/chosen": -1.
|
| 798 |
-
"logps/rejected": -1.
|
| 799 |
-
"loss":
|
| 800 |
-
"rewards/accuracies": 0.
|
| 801 |
-
"rewards/chosen": -
|
| 802 |
-
"rewards/margins":
|
| 803 |
-
"rewards/rejected": -
|
| 804 |
"step": 265
|
| 805 |
},
|
| 806 |
{
|
| 807 |
"epoch": 0.5800214822771214,
|
| 808 |
-
"grad_norm":
|
| 809 |
"learning_rate": 3.5798929729067464e-07,
|
| 810 |
-
"logits/chosen": -
|
| 811 |
-
"logits/rejected": -
|
| 812 |
-
"logps/chosen": -1.
|
| 813 |
-
"logps/rejected": -
|
| 814 |
-
"loss": 2.
|
| 815 |
-
"rewards/accuracies": 0.
|
| 816 |
-
"rewards/chosen": -
|
| 817 |
-
"rewards/margins": 3.
|
| 818 |
-
"rewards/rejected": -
|
| 819 |
"step": 270
|
| 820 |
},
|
| 821 |
{
|
| 822 |
"epoch": 0.5907626208378088,
|
| 823 |
-
"grad_norm":
|
| 824 |
"learning_rate": 3.4307406469068595e-07,
|
| 825 |
-
"logits/chosen": -
|
| 826 |
-
"logits/rejected": -
|
| 827 |
-
"logps/chosen": -1.
|
| 828 |
-
"logps/rejected": -
|
| 829 |
-
"loss": 2.
|
| 830 |
"rewards/accuracies": 0.8500000238418579,
|
| 831 |
-
"rewards/chosen": -
|
| 832 |
-
"rewards/margins":
|
| 833 |
-
"rewards/rejected": -
|
| 834 |
"step": 275
|
| 835 |
},
|
| 836 |
{
|
| 837 |
"epoch": 0.6015037593984962,
|
| 838 |
-
"grad_norm":
|
| 839 |
"learning_rate": 3.282392117676968e-07,
|
| 840 |
-
"logits/chosen": -
|
| 841 |
-
"logits/rejected": -
|
| 842 |
-
"logps/chosen": -1.
|
| 843 |
-
"logps/rejected": -2.
|
| 844 |
-
"loss": 2.
|
| 845 |
-
"rewards/accuracies": 0.
|
| 846 |
-
"rewards/chosen": -
|
| 847 |
-
"rewards/margins":
|
| 848 |
-
"rewards/rejected": -
|
| 849 |
"step": 280
|
| 850 |
},
|
| 851 |
{
|
| 852 |
"epoch": 0.6122448979591837,
|
| 853 |
-
"grad_norm":
|
| 854 |
"learning_rate": 3.135056854008371e-07,
|
| 855 |
-
"logits/chosen": -
|
| 856 |
-
"logits/rejected": -
|
| 857 |
-
"logps/chosen": -1.
|
| 858 |
-
"logps/rejected": -
|
| 859 |
-
"loss": 2.
|
| 860 |
"rewards/accuracies": 0.75,
|
| 861 |
-
"rewards/chosen": -
|
| 862 |
-
"rewards/margins":
|
| 863 |
-
"rewards/rejected": -
|
| 864 |
"step": 285
|
| 865 |
},
|
| 866 |
{
|
| 867 |
"epoch": 0.6229860365198711,
|
| 868 |
-
"grad_norm":
|
| 869 |
"learning_rate": 2.988942893956833e-07,
|
| 870 |
-
"logits/chosen": -
|
| 871 |
-
"logits/rejected": -
|
| 872 |
-
"logps/chosen": -1.
|
| 873 |
-
"logps/rejected": -
|
| 874 |
-
"loss":
|
| 875 |
-
"rewards/accuracies": 0.
|
| 876 |
-
"rewards/chosen": -
|
| 877 |
-
"rewards/margins":
|
| 878 |
-
"rewards/rejected": -
|
| 879 |
"step": 290
|
| 880 |
},
|
| 881 |
{
|
| 882 |
"epoch": 0.6337271750805585,
|
| 883 |
-
"grad_norm":
|
| 884 |
"learning_rate": 2.844256551091911e-07,
|
| 885 |
-
"logits/chosen": -
|
| 886 |
-
"logits/rejected": -
|
| 887 |
-
"logps/chosen": -1.
|
| 888 |
-
"logps/rejected": -2.
|
| 889 |
-
"loss": 2.
|
| 890 |
-
"rewards/accuracies": 0.
|
| 891 |
-
"rewards/chosen": -
|
| 892 |
-
"rewards/margins":
|
| 893 |
-
"rewards/rejected": -
|
| 894 |
"step": 295
|
| 895 |
},
|
| 896 |
{
|
| 897 |
"epoch": 0.644468313641246,
|
| 898 |
-
"grad_norm":
|
| 899 |
"learning_rate": 2.7012021231812664e-07,
|
| 900 |
-
"logits/chosen": -
|
| 901 |
-
"logits/rejected": -
|
| 902 |
-
"logps/chosen": -1.
|
| 903 |
-
"logps/rejected": -
|
| 904 |
-
"loss":
|
| 905 |
-
"rewards/accuracies": 0.
|
| 906 |
-
"rewards/chosen": -
|
| 907 |
-
"rewards/margins":
|
| 908 |
-
"rewards/rejected": -
|
| 909 |
"step": 300
|
| 910 |
},
|
| 911 |
{
|
| 912 |
"epoch": 0.6552094522019334,
|
| 913 |
-
"grad_norm":
|
| 914 |
"learning_rate": 2.5599816037212954e-07,
|
| 915 |
-
"logits/chosen": -
|
| 916 |
-
"logits/rejected": -
|
| 917 |
-
"logps/chosen": -1.
|
| 918 |
-
"logps/rejected": -
|
| 919 |
-
"loss": 2.
|
| 920 |
"rewards/accuracies": 0.824999988079071,
|
| 921 |
-
"rewards/chosen": -
|
| 922 |
-
"rewards/margins":
|
| 923 |
-
"rewards/rejected": -
|
| 924 |
"step": 305
|
| 925 |
},
|
| 926 |
{
|
| 927 |
"epoch": 0.6659505907626209,
|
| 928 |
-
"grad_norm":
|
| 929 |
"learning_rate": 2.4207943967214064e-07,
|
| 930 |
-
"logits/chosen": -
|
| 931 |
-
"logits/rejected": -15.
|
| 932 |
-
"logps/chosen": -
|
| 933 |
-
"logps/rejected": -2.
|
| 934 |
-
"loss":
|
| 935 |
-
"rewards/accuracies": 0.
|
| 936 |
-
"rewards/chosen": -
|
| 937 |
-
"rewards/margins": 4.
|
| 938 |
-
"rewards/rejected": -
|
| 939 |
"step": 310
|
| 940 |
},
|
| 941 |
{
|
| 942 |
"epoch": 0.6766917293233082,
|
| 943 |
-
"grad_norm":
|
| 944 |
"learning_rate": 2.2838370351446547e-07,
|
| 945 |
-
"logits/chosen": -15.
|
| 946 |
-
"logits/rejected": -15.
|
| 947 |
-
"logps/chosen": -1.
|
| 948 |
-
"logps/rejected": -
|
| 949 |
-
"loss": 2.
|
| 950 |
"rewards/accuracies": 0.6875,
|
| 951 |
-
"rewards/chosen": -
|
| 952 |
-
"rewards/margins": 4.
|
| 953 |
-
"rewards/rejected": -
|
| 954 |
"step": 315
|
| 955 |
},
|
| 956 |
{
|
| 957 |
"epoch": 0.6874328678839957,
|
| 958 |
-
"grad_norm":
|
| 959 |
"learning_rate": 2.1493029034023188e-07,
|
| 960 |
-
"logits/chosen": -
|
| 961 |
-
"logits/rejected": -
|
| 962 |
-
"logps/chosen": -1.
|
| 963 |
-
"logps/rejected": -
|
| 964 |
-
"loss": 2.
|
| 965 |
-
"rewards/accuracies": 0.
|
| 966 |
-
"rewards/chosen": -
|
| 967 |
-
"rewards/margins": 4.
|
| 968 |
-
"rewards/rejected": -
|
| 969 |
"step": 320
|
| 970 |
},
|
| 971 |
{
|
| 972 |
"epoch": 0.6981740064446831,
|
| 973 |
-
"grad_norm":
|
| 974 |
"learning_rate": 2.0173819642942376e-07,
|
| 975 |
-
"logits/chosen": -14.
|
| 976 |
-
"logits/rejected": -14.
|
| 977 |
-
"logps/chosen": -1.
|
| 978 |
-
"logps/rejected": -2.
|
| 979 |
-
"loss": 2.
|
| 980 |
"rewards/accuracies": 0.7749999761581421,
|
| 981 |
-
"rewards/chosen": -
|
| 982 |
-
"rewards/margins":
|
| 983 |
-
"rewards/rejected": -
|
| 984 |
"step": 325
|
| 985 |
},
|
| 986 |
{
|
| 987 |
"epoch": 0.7089151450053706,
|
| 988 |
-
"grad_norm":
|
| 989 |
"learning_rate": 1.888260490780485e-07,
|
| 990 |
-
"logits/chosen": -
|
| 991 |
-
"logits/rejected": -14.
|
| 992 |
-
"logps/chosen": -1.
|
| 993 |
-
"logps/rejected": -
|
| 994 |
-
"loss":
|
| 995 |
"rewards/accuracies": 0.800000011920929,
|
| 996 |
-
"rewards/chosen": -
|
| 997 |
-
"rewards/margins": 4.
|
| 998 |
-
"rewards/rejected": -
|
| 999 |
"step": 330
|
| 1000 |
},
|
| 1001 |
{
|
| 1002 |
"epoch": 0.719656283566058,
|
| 1003 |
-
"grad_norm":
|
| 1004 |
"learning_rate": 1.7621208029631078e-07,
|
| 1005 |
-
"logits/chosen": -14.
|
| 1006 |
-
"logits/rejected": -14.
|
| 1007 |
-
"logps/chosen": -1.
|
| 1008 |
-
"logps/rejected": -2.
|
| 1009 |
-
"loss": 2.
|
| 1010 |
-
"rewards/accuracies": 0.
|
| 1011 |
-
"rewards/chosen": -
|
| 1012 |
-
"rewards/margins":
|
| 1013 |
-
"rewards/rejected": -
|
| 1014 |
"step": 335
|
| 1015 |
},
|
| 1016 |
{
|
| 1017 |
"epoch": 0.7303974221267454,
|
| 1018 |
-
"grad_norm":
|
| 1019 |
"learning_rate": 1.6391410106493227e-07,
|
| 1020 |
-
"logits/chosen": -14.
|
| 1021 |
-
"logits/rejected": -14.
|
| 1022 |
-
"logps/chosen": -1.
|
| 1023 |
-
"logps/rejected": -2.
|
| 1024 |
-
"loss": 2.
|
| 1025 |
-
"rewards/accuracies": 0.
|
| 1026 |
-
"rewards/chosen": -
|
| 1027 |
-
"rewards/margins":
|
| 1028 |
-
"rewards/rejected": -
|
| 1029 |
"step": 340
|
| 1030 |
},
|
| 1031 |
{
|
| 1032 |
"epoch": 0.7411385606874329,
|
| 1033 |
-
"grad_norm":
|
| 1034 |
"learning_rate": 1.5194947618596673e-07,
|
| 1035 |
-
"logits/chosen": -
|
| 1036 |
-
"logits/rejected": -14.
|
| 1037 |
-
"logps/chosen": -1.
|
| 1038 |
-
"logps/rejected": -
|
| 1039 |
-
"loss":
|
| 1040 |
"rewards/accuracies": 0.800000011920929,
|
| 1041 |
-
"rewards/chosen": -
|
| 1042 |
-
"rewards/margins": 4.
|
| 1043 |
-
"rewards/rejected": -
|
| 1044 |
"step": 345
|
| 1045 |
},
|
| 1046 |
{
|
| 1047 |
"epoch": 0.7518796992481203,
|
| 1048 |
-
"grad_norm":
|
| 1049 |
"learning_rate": 1.4033509976362083e-07,
|
| 1050 |
-
"logits/chosen": -15.
|
| 1051 |
-
"logits/rejected": -15.
|
| 1052 |
-
"logps/chosen": -1.
|
| 1053 |
-
"logps/rejected": -
|
| 1054 |
-
"loss": 2.
|
| 1055 |
"rewards/accuracies": 0.800000011920929,
|
| 1056 |
-
"rewards/chosen": -
|
| 1057 |
-
"rewards/margins":
|
| 1058 |
-
"rewards/rejected": -
|
| 1059 |
"step": 350
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
"epoch": 0.7626208378088077,
|
| 1063 |
-
"grad_norm":
|
| 1064 |
"learning_rate": 1.2908737134970363e-07,
|
| 1065 |
-
"logits/chosen": -
|
| 1066 |
-
"logits/rejected": -14.
|
| 1067 |
-
"logps/chosen": -1.
|
| 1068 |
-
"logps/rejected": -2.
|
| 1069 |
-
"loss":
|
| 1070 |
"rewards/accuracies": 0.737500011920929,
|
| 1071 |
-
"rewards/chosen": -
|
| 1072 |
-
"rewards/margins":
|
| 1073 |
-
"rewards/rejected": -
|
| 1074 |
"step": 355
|
| 1075 |
},
|
| 1076 |
{
|
| 1077 |
"epoch": 0.7733619763694952,
|
| 1078 |
-
"grad_norm":
|
| 1079 |
"learning_rate": 1.1822217278738515e-07,
|
| 1080 |
-
"logits/chosen": -15.
|
| 1081 |
-
"logits/rejected": -15.
|
| 1082 |
-
"logps/chosen": -1.
|
| 1083 |
-
"logps/rejected": -2.
|
| 1084 |
-
"loss":
|
| 1085 |
-
"rewards/accuracies": 0.
|
| 1086 |
-
"rewards/chosen": -
|
| 1087 |
-
"rewards/margins":
|
| 1088 |
-
"rewards/rejected": -
|
| 1089 |
"step": 360
|
| 1090 |
},
|
| 1091 |
{
|
| 1092 |
"epoch": 0.7841031149301826,
|
| 1093 |
-
"grad_norm":
|
| 1094 |
"learning_rate": 1.0775484578596241e-07,
|
| 1095 |
-
"logits/chosen": -15.
|
| 1096 |
-
"logits/rejected": -15.
|
| 1097 |
-
"logps/chosen": -1.
|
| 1098 |
-
"logps/rejected": -2.
|
| 1099 |
-
"loss": 2.
|
| 1100 |
-
"rewards/accuracies": 0.
|
| 1101 |
-
"rewards/chosen": -
|
| 1102 |
-
"rewards/margins": 5.
|
| 1103 |
-
"rewards/rejected": -
|
| 1104 |
"step": 365
|
| 1105 |
},
|
| 1106 |
{
|
| 1107 |
"epoch": 0.7948442534908701,
|
| 1108 |
-
"grad_norm":
|
| 1109 |
"learning_rate": 9.770017025829673e-08,
|
| 1110 |
-
"logits/chosen": -15.
|
| 1111 |
-
"logits/rejected": -15.
|
| 1112 |
-
"logps/chosen": -
|
| 1113 |
-
"logps/rejected": -2.
|
| 1114 |
-
"loss": 2.
|
| 1115 |
-
"rewards/accuracies": 0.
|
| 1116 |
-
"rewards/chosen": -
|
| 1117 |
-
"rewards/margins":
|
| 1118 |
-
"rewards/rejected": -
|
| 1119 |
"step": 370
|
| 1120 |
},
|
| 1121 |
{
|
| 1122 |
"epoch": 0.8055853920515574,
|
| 1123 |
-
"grad_norm":
|
| 1124 |
"learning_rate": 8.807234345151027e-08,
|
| 1125 |
-
"logits/chosen": -14.
|
| 1126 |
-
"logits/rejected": -14.
|
| 1127 |
-
"logps/chosen": -
|
| 1128 |
-
"logps/rejected": -2.
|
| 1129 |
-
"loss": 2.
|
| 1130 |
-
"rewards/accuracies": 0.
|
| 1131 |
-
"rewards/chosen": -
|
| 1132 |
-
"rewards/margins":
|
| 1133 |
-
"rewards/rejected": -
|
| 1134 |
"step": 375
|
| 1135 |
},
|
| 1136 |
{
|
| 1137 |
"epoch": 0.8163265306122449,
|
| 1138 |
-
"grad_norm":
|
| 1139 |
"learning_rate": 7.888495990040924e-08,
|
| 1140 |
-
"logits/chosen": -13.
|
| 1141 |
-
"logits/rejected": -13.
|
| 1142 |
-
"logps/chosen": -
|
| 1143 |
-
"logps/rejected": -2.
|
| 1144 |
-
"loss": 2.
|
| 1145 |
-
"rewards/accuracies": 0.
|
| 1146 |
-
"rewards/chosen": -
|
| 1147 |
-
"rewards/margins":
|
| 1148 |
-
"rewards/rejected": -
|
| 1149 |
"step": 380
|
| 1150 |
},
|
| 1151 |
{
|
| 1152 |
"epoch": 0.8270676691729323,
|
| 1153 |
-
"grad_norm":
|
| 1154 |
"learning_rate": 7.015099223193943e-08,
|
| 1155 |
-
"logits/chosen": -15.
|
| 1156 |
-
"logits/rejected": -15.
|
| 1157 |
-
"logps/chosen": -1.
|
| 1158 |
-
"logps/rejected": -2.
|
| 1159 |
-
"loss": 2.
|
| 1160 |
-
"rewards/accuracies": 0.
|
| 1161 |
-
"rewards/chosen": -
|
| 1162 |
-
"rewards/margins":
|
| 1163 |
-
"rewards/rejected": -
|
| 1164 |
"step": 385
|
| 1165 |
},
|
| 1166 |
{
|
| 1167 |
"epoch": 0.8378088077336198,
|
| 1168 |
-
"grad_norm":
|
| 1169 |
"learning_rate": 6.188277284777857e-08,
|
| 1170 |
-
"logits/chosen": -14.
|
| 1171 |
-
"logits/rejected": -13.
|
| 1172 |
-
"logps/chosen": -1.
|
| 1173 |
-
"logps/rejected": -2.
|
| 1174 |
-
"loss": 2.
|
| 1175 |
-
"rewards/accuracies": 0.
|
| 1176 |
-
"rewards/chosen": -
|
| 1177 |
-
"rewards/margins": 5.
|
| 1178 |
-
"rewards/rejected": -
|
| 1179 |
"step": 390
|
| 1180 |
},
|
| 1181 |
{
|
| 1182 |
"epoch": 0.8485499462943072,
|
| 1183 |
-
"grad_norm":
|
| 1184 |
"learning_rate": 5.409197651092965e-08,
|
| 1185 |
-
"logits/chosen": -15.
|
| 1186 |
-
"logits/rejected": -15.
|
| 1187 |
-
"logps/chosen": -
|
| 1188 |
-
"logps/rejected": -2.
|
| 1189 |
-
"loss": 2.
|
| 1190 |
-
"rewards/accuracies": 0.
|
| 1191 |
-
"rewards/chosen": -
|
| 1192 |
-
"rewards/margins":
|
| 1193 |
-
"rewards/rejected": -
|
| 1194 |
"step": 395
|
| 1195 |
},
|
| 1196 |
{
|
| 1197 |
"epoch": 0.8592910848549946,
|
| 1198 |
-
"grad_norm":
|
| 1199 |
"learning_rate": 4.678960386090298e-08,
|
| 1200 |
-
"logits/chosen": -
|
| 1201 |
-
"logits/rejected": -
|
| 1202 |
-
"logps/chosen": -1.
|
| 1203 |
-
"logps/rejected": -2.
|
| 1204 |
-
"loss": 2.
|
| 1205 |
"rewards/accuracies": 0.8125,
|
| 1206 |
-
"rewards/chosen": -
|
| 1207 |
-
"rewards/margins":
|
| 1208 |
-
"rewards/rejected": -
|
| 1209 |
"step": 400
|
| 1210 |
},
|
| 1211 |
{
|
| 1212 |
"epoch": 0.8700322234156821,
|
| 1213 |
-
"grad_norm":
|
| 1214 |
"learning_rate": 3.998596588076366e-08,
|
| 1215 |
-
"logits/chosen": -13.
|
| 1216 |
-
"logits/rejected": -13.
|
| 1217 |
-
"logps/chosen": -1.
|
| 1218 |
-
"logps/rejected": -2.
|
| 1219 |
-
"loss":
|
| 1220 |
-
"rewards/accuracies": 0.
|
| 1221 |
-
"rewards/chosen": -
|
| 1222 |
-
"rewards/margins":
|
| 1223 |
-
"rewards/rejected": -
|
| 1224 |
"step": 405
|
| 1225 |
},
|
| 1226 |
{
|
| 1227 |
"epoch": 0.8807733619763695,
|
| 1228 |
-
"grad_norm":
|
| 1229 |
"learning_rate": 3.3690669337976996e-08,
|
| 1230 |
-
"logits/chosen": -
|
| 1231 |
-
"logits/rejected": -14.
|
| 1232 |
-
"logps/chosen": -1.
|
| 1233 |
-
"logps/rejected": -
|
| 1234 |
-
"loss": 2.
|
| 1235 |
-
"rewards/accuracies": 0.
|
| 1236 |
-
"rewards/chosen": -
|
| 1237 |
-
"rewards/margins": 4.
|
| 1238 |
-
"rewards/rejected": -
|
| 1239 |
"step": 410
|
| 1240 |
},
|
| 1241 |
{
|
| 1242 |
"epoch": 0.8915145005370569,
|
| 1243 |
-
"grad_norm":
|
| 1244 |
"learning_rate": 2.7912603219609798e-08,
|
| 1245 |
-
"logits/chosen": -15.
|
| 1246 |
-
"logits/rejected": -15.
|
| 1247 |
-
"logps/chosen": -
|
| 1248 |
-
"logps/rejected": -2.
|
| 1249 |
-
"loss": 2.
|
| 1250 |
-
"rewards/accuracies": 0.
|
| 1251 |
-
"rewards/chosen": -
|
| 1252 |
-
"rewards/margins":
|
| 1253 |
-
"rewards/rejected": -
|
| 1254 |
"step": 415
|
| 1255 |
},
|
| 1256 |
{
|
| 1257 |
"epoch": 0.9022556390977443,
|
| 1258 |
-
"grad_norm":
|
| 1259 |
"learning_rate": 2.265992618104029e-08,
|
| 1260 |
-
"logits/chosen": -15.
|
| 1261 |
-
"logits/rejected": -15.
|
| 1262 |
-
"logps/chosen": -
|
| 1263 |
-
"logps/rejected": -2.
|
| 1264 |
-
"loss": 2.
|
| 1265 |
-
"rewards/accuracies": 0.
|
| 1266 |
-
"rewards/chosen": -
|
| 1267 |
-
"rewards/margins":
|
| 1268 |
-
"rewards/rejected": -
|
| 1269 |
"step": 420
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
"epoch": 0.9129967776584318,
|
| 1273 |
-
"grad_norm":
|
| 1274 |
"learning_rate": 1.7940055025900304e-08,
|
| 1275 |
-
"logits/chosen": -
|
| 1276 |
-
"logits/rejected": -13.
|
| 1277 |
-
"logps/chosen": -
|
| 1278 |
-
"logps/rejected": -2.
|
| 1279 |
-
"loss":
|
| 1280 |
-
"rewards/accuracies": 0.
|
| 1281 |
-
"rewards/chosen": -
|
| 1282 |
-
"rewards/margins": 3.
|
| 1283 |
-
"rewards/rejected": -
|
| 1284 |
"step": 425
|
| 1285 |
},
|
| 1286 |
{
|
| 1287 |
"epoch": 0.9237379162191193,
|
| 1288 |
-
"grad_norm":
|
| 1289 |
"learning_rate": 1.3759654233514817e-08,
|
| 1290 |
-
"logits/chosen": -14.
|
| 1291 |
-
"logits/rejected": -14.
|
| 1292 |
-
"logps/chosen": -1.
|
| 1293 |
-
"logps/rejected": -2.
|
| 1294 |
-
"loss": 2.
|
| 1295 |
-
"rewards/accuracies": 0.
|
| 1296 |
-
"rewards/chosen": -
|
| 1297 |
-
"rewards/margins":
|
| 1298 |
-
"rewards/rejected": -
|
| 1299 |
"step": 430
|
| 1300 |
},
|
| 1301 |
{
|
| 1302 |
"epoch": 0.9344790547798066,
|
| 1303 |
-
"grad_norm":
|
| 1304 |
"learning_rate": 1.0124626548627402e-08,
|
| 1305 |
-
"logits/chosen": -
|
| 1306 |
-
"logits/rejected": -15.
|
| 1307 |
-
"logps/chosen": -
|
| 1308 |
-
"logps/rejected": -2.
|
| 1309 |
-
"loss": 2.
|
| 1310 |
-
"rewards/accuracies": 0.
|
| 1311 |
-
"rewards/chosen": -
|
| 1312 |
-
"rewards/margins": 6.
|
| 1313 |
-
"rewards/rejected": -
|
| 1314 |
"step": 435
|
| 1315 |
},
|
| 1316 |
{
|
| 1317 |
"epoch": 0.9452201933404941,
|
| 1318 |
-
"grad_norm":
|
| 1319 |
"learning_rate": 7.040104646698042e-09,
|
| 1320 |
-
"logits/chosen": -
|
| 1321 |
-
"logits/rejected": -
|
| 1322 |
-
"logps/chosen": -
|
| 1323 |
-
"logps/rejected": -2.
|
| 1324 |
-
"loss": 2.
|
| 1325 |
-
"rewards/accuracies": 0.
|
| 1326 |
-
"rewards/chosen": -
|
| 1327 |
-
"rewards/margins":
|
| 1328 |
-
"rewards/rejected": -
|
| 1329 |
"step": 440
|
| 1330 |
},
|
| 1331 |
{
|
| 1332 |
"epoch": 0.9559613319011815,
|
| 1333 |
-
"grad_norm":
|
| 1334 |
"learning_rate": 4.510443886542114e-09,
|
| 1335 |
-
"logits/chosen": -
|
| 1336 |
-
"logits/rejected": -
|
| 1337 |
-
"logps/chosen": -
|
| 1338 |
-
"logps/rejected": -2.
|
| 1339 |
-
"loss": 2.
|
| 1340 |
-
"rewards/accuracies": 0.
|
| 1341 |
-
"rewards/chosen": -
|
| 1342 |
-
"rewards/margins":
|
| 1343 |
-
"rewards/rejected": -
|
| 1344 |
"step": 445
|
| 1345 |
},
|
| 1346 |
{
|
| 1347 |
"epoch": 0.966702470461869,
|
| 1348 |
-
"grad_norm":
|
| 1349 |
"learning_rate": 2.539216160544333e-09,
|
| 1350 |
-
"logits/chosen": -
|
| 1351 |
-
"logits/rejected": -
|
| 1352 |
-
"logps/chosen": -
|
| 1353 |
-
"logps/rejected": -2.
|
| 1354 |
-
"loss": 2.
|
| 1355 |
-
"rewards/accuracies": 0.
|
| 1356 |
-
"rewards/chosen": -
|
| 1357 |
-
"rewards/margins":
|
| 1358 |
-
"rewards/rejected": -
|
| 1359 |
"step": 450
|
| 1360 |
},
|
| 1361 |
{
|
| 1362 |
"epoch": 0.9774436090225563,
|
| 1363 |
-
"grad_norm":
|
| 1364 |
"learning_rate": 1.1292048511303054e-09,
|
| 1365 |
-
"logits/chosen": -14.
|
| 1366 |
-
"logits/rejected": -
|
| 1367 |
-
"logps/chosen": -1.
|
| 1368 |
-
"logps/rejected": -2.
|
| 1369 |
-
"loss": 2.
|
| 1370 |
-
"rewards/accuracies": 0.
|
| 1371 |
-
"rewards/chosen": -
|
| 1372 |
-
"rewards/margins":
|
| 1373 |
-
"rewards/rejected": -
|
| 1374 |
"step": 455
|
| 1375 |
},
|
| 1376 |
{
|
| 1377 |
"epoch": 0.9881847475832438,
|
| 1378 |
-
"grad_norm":
|
| 1379 |
"learning_rate": 2.82400900618418e-10,
|
| 1380 |
-
"logits/chosen": -
|
| 1381 |
-
"logits/rejected": -
|
| 1382 |
-
"logps/chosen": -1.
|
| 1383 |
-
"logps/rejected": -2.
|
| 1384 |
-
"loss":
|
| 1385 |
-
"rewards/accuracies": 0.
|
| 1386 |
-
"rewards/chosen": -
|
| 1387 |
-
"rewards/margins":
|
| 1388 |
-
"rewards/rejected": -
|
| 1389 |
"step": 460
|
| 1390 |
},
|
| 1391 |
{
|
| 1392 |
"epoch": 0.9989258861439313,
|
| 1393 |
-
"grad_norm":
|
| 1394 |
"learning_rate": 0.0,
|
| 1395 |
-
"logits/chosen": -
|
| 1396 |
-
"logits/rejected": -
|
| 1397 |
-
"logps/chosen": -
|
| 1398 |
-
"logps/rejected": -2.
|
| 1399 |
-
"loss":
|
| 1400 |
-
"rewards/accuracies": 0.
|
| 1401 |
-
"rewards/chosen": -
|
| 1402 |
-
"rewards/margins": 4.
|
| 1403 |
-
"rewards/rejected": -
|
| 1404 |
"step": 465
|
| 1405 |
},
|
| 1406 |
{
|
| 1407 |
"epoch": 0.9989258861439313,
|
| 1408 |
"step": 465,
|
| 1409 |
"total_flos": 0.0,
|
| 1410 |
-
"train_loss":
|
| 1411 |
-
"train_runtime":
|
| 1412 |
-
"train_samples_per_second": 9.
|
| 1413 |
"train_steps_per_second": 0.076
|
| 1414 |
}
|
| 1415 |
],
|
|
|
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.010741138560687433,
|
| 13 |
+
"grad_norm": 94.5147817778946,
|
| 14 |
"learning_rate": 8.51063829787234e-08,
|
| 15 |
+
"logits/chosen": -10.583702087402344,
|
| 16 |
+
"logits/rejected": -10.455877304077148,
|
| 17 |
+
"logps/chosen": -0.9049979448318481,
|
| 18 |
+
"logps/rejected": -0.8784100413322449,
|
| 19 |
+
"loss": 6.1451,
|
| 20 |
"rewards/accuracies": 0.574999988079071,
|
| 21 |
+
"rewards/chosen": -9.04997730255127,
|
| 22 |
+
"rewards/margins": -0.2658771872520447,
|
| 23 |
+
"rewards/rejected": -8.784101486206055,
|
| 24 |
"step": 5
|
| 25 |
},
|
| 26 |
{
|
| 27 |
"epoch": 0.021482277121374866,
|
| 28 |
+
"grad_norm": 128.5515421485228,
|
| 29 |
"learning_rate": 1.702127659574468e-07,
|
| 30 |
+
"logits/chosen": -10.710015296936035,
|
| 31 |
+
"logits/rejected": -10.85377311706543,
|
| 32 |
+
"logps/chosen": -1.0046945810317993,
|
| 33 |
+
"logps/rejected": -0.8850045204162598,
|
| 34 |
+
"loss": 5.8491,
|
| 35 |
+
"rewards/accuracies": 0.512499988079071,
|
| 36 |
+
"rewards/chosen": -10.04694652557373,
|
| 37 |
+
"rewards/margins": -1.196901559829712,
|
| 38 |
+
"rewards/rejected": -8.850046157836914,
|
| 39 |
"step": 10
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.0322234156820623,
|
| 43 |
+
"grad_norm": 58.802331595913145,
|
| 44 |
"learning_rate": 2.553191489361702e-07,
|
| 45 |
+
"logits/chosen": -10.312850952148438,
|
| 46 |
+
"logits/rejected": -10.239133834838867,
|
| 47 |
+
"logps/chosen": -1.0889472961425781,
|
| 48 |
+
"logps/rejected": -1.1543949842453003,
|
| 49 |
+
"loss": 6.2505,
|
| 50 |
+
"rewards/accuracies": 0.550000011920929,
|
| 51 |
+
"rewards/chosen": -10.889472961425781,
|
| 52 |
+
"rewards/margins": 0.6544777154922485,
|
| 53 |
+
"rewards/rejected": -11.543951034545898,
|
| 54 |
"step": 15
|
| 55 |
},
|
| 56 |
{
|
| 57 |
"epoch": 0.04296455424274973,
|
| 58 |
+
"grad_norm": 151.21348799898024,
|
| 59 |
"learning_rate": 3.404255319148936e-07,
|
| 60 |
+
"logits/chosen": -9.954164505004883,
|
| 61 |
+
"logits/rejected": -10.053568840026855,
|
| 62 |
+
"logps/chosen": -0.9611791372299194,
|
| 63 |
+
"logps/rejected": -1.1332345008850098,
|
| 64 |
+
"loss": 5.5619,
|
| 65 |
"rewards/accuracies": 0.44999998807907104,
|
| 66 |
+
"rewards/chosen": -9.611791610717773,
|
| 67 |
+
"rewards/margins": 1.7205528020858765,
|
| 68 |
+
"rewards/rejected": -11.332345008850098,
|
| 69 |
"step": 20
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 0.05370569280343716,
|
| 73 |
+
"grad_norm": 99.92865956439873,
|
| 74 |
"learning_rate": 4.25531914893617e-07,
|
| 75 |
+
"logits/chosen": -10.326103210449219,
|
| 76 |
+
"logits/rejected": -10.055009841918945,
|
| 77 |
+
"logps/chosen": -0.8260948061943054,
|
| 78 |
+
"logps/rejected": -1.1549828052520752,
|
| 79 |
+
"loss": 5.2635,
|
| 80 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 81 |
+
"rewards/chosen": -8.260948181152344,
|
| 82 |
+
"rewards/margins": 3.2888808250427246,
|
| 83 |
+
"rewards/rejected": -11.54982852935791,
|
| 84 |
"step": 25
|
| 85 |
},
|
| 86 |
{
|
| 87 |
"epoch": 0.0644468313641246,
|
| 88 |
+
"grad_norm": 81.6600880763309,
|
| 89 |
"learning_rate": 5.106382978723404e-07,
|
| 90 |
+
"logits/chosen": -9.319940567016602,
|
| 91 |
+
"logits/rejected": -9.13192081451416,
|
| 92 |
+
"logps/chosen": -0.6618553996086121,
|
| 93 |
+
"logps/rejected": -0.6553267240524292,
|
| 94 |
+
"loss": 5.0518,
|
| 95 |
"rewards/accuracies": 0.612500011920929,
|
| 96 |
+
"rewards/chosen": -6.61855411529541,
|
| 97 |
+
"rewards/margins": -0.06528709828853607,
|
| 98 |
+
"rewards/rejected": -6.553267002105713,
|
| 99 |
"step": 30
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"epoch": 0.07518796992481203,
|
| 103 |
+
"grad_norm": 92.79131753018717,
|
| 104 |
"learning_rate": 5.957446808510638e-07,
|
| 105 |
+
"logits/chosen": -8.877812385559082,
|
| 106 |
+
"logits/rejected": -8.929550170898438,
|
| 107 |
+
"logps/chosen": -0.717892050743103,
|
| 108 |
+
"logps/rejected": -0.6935927867889404,
|
| 109 |
+
"loss": 5.224,
|
| 110 |
+
"rewards/accuracies": 0.512499988079071,
|
| 111 |
+
"rewards/chosen": -7.178920745849609,
|
| 112 |
+
"rewards/margins": -0.24299363791942596,
|
| 113 |
+
"rewards/rejected": -6.935927391052246,
|
| 114 |
"step": 35
|
| 115 |
},
|
| 116 |
{
|
| 117 |
"epoch": 0.08592910848549946,
|
| 118 |
+
"grad_norm": 89.40050001716304,
|
| 119 |
"learning_rate": 6.808510638297872e-07,
|
| 120 |
+
"logits/chosen": -7.988096714019775,
|
| 121 |
+
"logits/rejected": -7.907191276550293,
|
| 122 |
+
"logps/chosen": -0.7402302622795105,
|
| 123 |
+
"logps/rejected": -0.7434382438659668,
|
| 124 |
+
"loss": 4.8885,
|
| 125 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 126 |
+
"rewards/chosen": -7.4023027420043945,
|
| 127 |
+
"rewards/margins": 0.032080501317977905,
|
| 128 |
+
"rewards/rejected": -7.434383392333984,
|
| 129 |
"step": 40
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 0.0966702470461869,
|
| 133 |
+
"grad_norm": 59.92644904113339,
|
| 134 |
"learning_rate": 7.659574468085107e-07,
|
| 135 |
+
"logits/chosen": -8.71805477142334,
|
| 136 |
+
"logits/rejected": -8.232014656066895,
|
| 137 |
+
"logps/chosen": -0.5317873954772949,
|
| 138 |
+
"logps/rejected": -0.6050616502761841,
|
| 139 |
+
"loss": 4.5879,
|
| 140 |
+
"rewards/accuracies": 0.675000011920929,
|
| 141 |
+
"rewards/chosen": -5.317873477935791,
|
| 142 |
+
"rewards/margins": 0.7327424883842468,
|
| 143 |
+
"rewards/rejected": -6.0506157875061035,
|
| 144 |
"step": 45
|
| 145 |
},
|
| 146 |
{
|
| 147 |
"epoch": 0.10741138560687433,
|
| 148 |
+
"grad_norm": 60.25452206880678,
|
| 149 |
"learning_rate": 7.998983280184396e-07,
|
| 150 |
+
"logits/chosen": -8.83049488067627,
|
| 151 |
+
"logits/rejected": -8.585375785827637,
|
| 152 |
+
"logps/chosen": -0.5144228339195251,
|
| 153 |
+
"logps/rejected": -0.5809676647186279,
|
| 154 |
+
"loss": 4.6549,
|
| 155 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 156 |
+
"rewards/chosen": -5.144228458404541,
|
| 157 |
+
"rewards/margins": 0.6654484868049622,
|
| 158 |
+
"rewards/rejected": -5.8096771240234375,
|
| 159 |
"step": 50
|
| 160 |
},
|
| 161 |
{
|
| 162 |
"epoch": 0.11815252416756176,
|
| 163 |
+
"grad_norm": 46.997605934809734,
|
| 164 |
"learning_rate": 7.992771864078597e-07,
|
| 165 |
+
"logits/chosen": -8.163946151733398,
|
| 166 |
+
"logits/rejected": -8.180994033813477,
|
| 167 |
+
"logps/chosen": -0.5956984758377075,
|
| 168 |
+
"logps/rejected": -0.7000880837440491,
|
| 169 |
+
"loss": 4.6606,
|
| 170 |
+
"rewards/accuracies": 0.612500011920929,
|
| 171 |
+
"rewards/chosen": -5.956984519958496,
|
| 172 |
+
"rewards/margins": 1.0438958406448364,
|
| 173 |
+
"rewards/rejected": -7.000881195068359,
|
| 174 |
"step": 55
|
| 175 |
},
|
| 176 |
{
|
| 177 |
"epoch": 0.1288936627282492,
|
| 178 |
+
"grad_norm": 59.25357465703395,
|
| 179 |
"learning_rate": 7.980922636120897e-07,
|
| 180 |
+
"logits/chosen": -8.718216896057129,
|
| 181 |
+
"logits/rejected": -8.35698127746582,
|
| 182 |
+
"logps/chosen": -0.5706155896186829,
|
| 183 |
+
"logps/rejected": -0.6969493627548218,
|
| 184 |
+
"loss": 4.4885,
|
| 185 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 186 |
+
"rewards/chosen": -5.706155776977539,
|
| 187 |
+
"rewards/margins": 1.2633379697799683,
|
| 188 |
+
"rewards/rejected": -6.969493865966797,
|
| 189 |
"step": 60
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"epoch": 0.13963480128893663,
|
| 193 |
+
"grad_norm": 50.79780612404402,
|
| 194 |
"learning_rate": 7.963452327474534e-07,
|
| 195 |
+
"logits/chosen": -9.234804153442383,
|
| 196 |
+
"logits/rejected": -9.1095609664917,
|
| 197 |
+
"logps/chosen": -0.6090906858444214,
|
| 198 |
+
"logps/rejected": -0.7208055257797241,
|
| 199 |
+
"loss": 4.666,
|
| 200 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 201 |
+
"rewards/chosen": -6.090908050537109,
|
| 202 |
+
"rewards/margins": 1.1171473264694214,
|
| 203 |
+
"rewards/rejected": -7.208055019378662,
|
| 204 |
"step": 65
|
| 205 |
},
|
| 206 |
{
|
| 207 |
"epoch": 0.15037593984962405,
|
| 208 |
+
"grad_norm": 56.36564641791114,
|
| 209 |
"learning_rate": 7.940385606293987e-07,
|
| 210 |
+
"logits/chosen": -8.946883201599121,
|
| 211 |
+
"logits/rejected": -8.716778755187988,
|
| 212 |
+
"logps/chosen": -0.6818052530288696,
|
| 213 |
+
"logps/rejected": -0.7961267828941345,
|
| 214 |
+
"loss": 4.577,
|
| 215 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 216 |
+
"rewards/chosen": -6.818052768707275,
|
| 217 |
+
"rewards/margins": 1.1432150602340698,
|
| 218 |
+
"rewards/rejected": -7.961267948150635,
|
| 219 |
"step": 70
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.1611170784103115,
|
| 223 |
+
"grad_norm": 61.76930510948969,
|
| 224 |
"learning_rate": 7.911755042893434e-07,
|
| 225 |
+
"logits/chosen": -9.067525863647461,
|
| 226 |
+
"logits/rejected": -8.9346923828125,
|
| 227 |
+
"logps/chosen": -0.6832990646362305,
|
| 228 |
+
"logps/rejected": -0.7763570547103882,
|
| 229 |
+
"loss": 4.4179,
|
| 230 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 231 |
+
"rewards/chosen": -6.8329901695251465,
|
| 232 |
+
"rewards/margins": 0.9305804371833801,
|
| 233 |
+
"rewards/rejected": -7.763571262359619,
|
| 234 |
"step": 75
|
| 235 |
},
|
| 236 |
{
|
| 237 |
"epoch": 0.17185821697099893,
|
| 238 |
+
"grad_norm": 63.602633074853536,
|
| 239 |
"learning_rate": 7.877601063757321e-07,
|
| 240 |
+
"logits/chosen": -9.461370468139648,
|
| 241 |
+
"logits/rejected": -8.981520652770996,
|
| 242 |
+
"logps/chosen": -0.6881433129310608,
|
| 243 |
+
"logps/rejected": -0.8508684039115906,
|
| 244 |
+
"loss": 4.3763,
|
| 245 |
+
"rewards/accuracies": 0.699999988079071,
|
| 246 |
+
"rewards/chosen": -6.881433010101318,
|
| 247 |
+
"rewards/margins": 1.6272509098052979,
|
| 248 |
+
"rewards/rejected": -8.508684158325195,
|
| 249 |
"step": 80
|
| 250 |
},
|
| 251 |
{
|
| 252 |
"epoch": 0.18259935553168635,
|
| 253 |
+
"grad_norm": 50.67613033462041,
|
| 254 |
"learning_rate": 7.837971894457989e-07,
|
| 255 |
+
"logits/chosen": -9.557887077331543,
|
| 256 |
+
"logits/rejected": -9.17081069946289,
|
| 257 |
+
"logps/chosen": -0.6830392479896545,
|
| 258 |
+
"logps/rejected": -0.799291729927063,
|
| 259 |
+
"loss": 4.6499,
|
| 260 |
"rewards/accuracies": 0.6000000238418579,
|
| 261 |
+
"rewards/chosen": -6.830392360687256,
|
| 262 |
+
"rewards/margins": 1.1625245809555054,
|
| 263 |
+
"rewards/rejected": -7.992917060852051,
|
| 264 |
"step": 85
|
| 265 |
},
|
| 266 |
{
|
| 267 |
"epoch": 0.1933404940923738,
|
| 268 |
+
"grad_norm": 57.58777625940833,
|
| 269 |
"learning_rate": 7.792923491560942e-07,
|
| 270 |
+
"logits/chosen": -8.579484939575195,
|
| 271 |
+
"logits/rejected": -8.546136856079102,
|
| 272 |
+
"logps/chosen": -0.6667743921279907,
|
| 273 |
+
"logps/rejected": -0.7407978177070618,
|
| 274 |
+
"loss": 4.4492,
|
| 275 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 276 |
+
"rewards/chosen": -6.667743682861328,
|
| 277 |
+
"rewards/margins": 0.7402342557907104,
|
| 278 |
+
"rewards/rejected": -7.407977104187012,
|
| 279 |
"step": 90
|
| 280 |
},
|
| 281 |
{
|
| 282 |
"epoch": 0.20408163265306123,
|
| 283 |
+
"grad_norm": 68.91179001810359,
|
| 284 |
"learning_rate": 7.742519463613926e-07,
|
| 285 |
+
"logits/chosen": -9.336307525634766,
|
| 286 |
+
"logits/rejected": -9.128133773803711,
|
| 287 |
+
"logps/chosen": -0.706219494342804,
|
| 288 |
+
"logps/rejected": -0.7757526636123657,
|
| 289 |
+
"loss": 4.2763,
|
| 290 |
+
"rewards/accuracies": 0.5,
|
| 291 |
+
"rewards/chosen": -7.062193870544434,
|
| 292 |
+
"rewards/margins": 0.6953321099281311,
|
| 293 |
+
"rewards/rejected": -7.7575273513793945,
|
| 294 |
"step": 95
|
| 295 |
},
|
| 296 |
{
|
| 297 |
"epoch": 0.21482277121374865,
|
| 298 |
+
"grad_norm": 140.43504047927686,
|
| 299 |
"learning_rate": 7.68683098133138e-07,
|
| 300 |
+
"logits/chosen": -8.939419746398926,
|
| 301 |
+
"logits/rejected": -8.681028366088867,
|
| 302 |
+
"logps/chosen": -0.7093919515609741,
|
| 303 |
+
"logps/rejected": -0.8932901620864868,
|
| 304 |
+
"loss": 4.4002,
|
| 305 |
+
"rewards/accuracies": 0.675000011920929,
|
| 306 |
+
"rewards/chosen": -7.093919277191162,
|
| 307 |
+
"rewards/margins": 1.838982343673706,
|
| 308 |
+
"rewards/rejected": -8.932901382446289,
|
| 309 |
"step": 100
|
| 310 |
},
|
| 311 |
{
|
| 312 |
"epoch": 0.22556390977443608,
|
| 313 |
+
"grad_norm": 105.52949175741638,
|
| 314 |
"learning_rate": 7.625936677101051e-07,
|
| 315 |
+
"logits/chosen": -8.601816177368164,
|
| 316 |
+
"logits/rejected": -8.625459671020508,
|
| 317 |
+
"logps/chosen": -0.8767679333686829,
|
| 318 |
+
"logps/rejected": -0.8515819311141968,
|
| 319 |
+
"loss": 4.4644,
|
| 320 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 321 |
+
"rewards/chosen": -8.767679214477539,
|
| 322 |
+
"rewards/margins": -0.25185948610305786,
|
| 323 |
+
"rewards/rejected": -8.51581859588623,
|
| 324 |
"step": 105
|
| 325 |
},
|
| 326 |
{
|
| 327 |
"epoch": 0.23630504833512353,
|
| 328 |
+
"grad_norm": 56.44051818244315,
|
| 329 |
"learning_rate": 7.559922533954731e-07,
|
| 330 |
+
"logits/chosen": -9.58240795135498,
|
| 331 |
+
"logits/rejected": -9.501542091369629,
|
| 332 |
+
"logps/chosen": -0.7655607461929321,
|
| 333 |
+
"logps/rejected": -0.8916142582893372,
|
| 334 |
+
"loss": 4.2797,
|
| 335 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 336 |
+
"rewards/chosen": -7.6556077003479,
|
| 337 |
+
"rewards/margins": 1.2605348825454712,
|
| 338 |
+
"rewards/rejected": -8.916143417358398,
|
| 339 |
"step": 110
|
| 340 |
},
|
| 341 |
{
|
| 342 |
"epoch": 0.24704618689581095,
|
| 343 |
+
"grad_norm": 68.9497539150874,
|
| 344 |
"learning_rate": 7.488881764159808e-07,
|
| 345 |
+
"logits/chosen": -9.756335258483887,
|
| 346 |
+
"logits/rejected": -9.543218612670898,
|
| 347 |
+
"logps/chosen": -0.7038711309432983,
|
| 348 |
+
"logps/rejected": -0.7986757159233093,
|
| 349 |
+
"loss": 4.154,
|
| 350 |
"rewards/accuracies": 0.6499999761581421,
|
| 351 |
+
"rewards/chosen": -7.0387115478515625,
|
| 352 |
+
"rewards/margins": 0.9480463862419128,
|
| 353 |
+
"rewards/rejected": -7.986758232116699,
|
| 354 |
"step": 115
|
| 355 |
},
|
| 356 |
{
|
| 357 |
"epoch": 0.2577873254564984,
|
| 358 |
+
"grad_norm": 90.20504733911939,
|
| 359 |
"learning_rate": 7.412914677603135e-07,
|
| 360 |
+
"logits/chosen": -9.883420944213867,
|
| 361 |
+
"logits/rejected": -9.735390663146973,
|
| 362 |
+
"logps/chosen": -0.9017173647880554,
|
| 363 |
+
"logps/rejected": -0.9996153116226196,
|
| 364 |
+
"loss": 4.2168,
|
| 365 |
+
"rewards/accuracies": 0.612500011920929,
|
| 366 |
+
"rewards/chosen": -9.017173767089844,
|
| 367 |
+
"rewards/margins": 0.9789786338806152,
|
| 368 |
+
"rewards/rejected": -9.996152877807617,
|
| 369 |
"step": 120
|
| 370 |
},
|
| 371 |
{
|
| 372 |
"epoch": 0.26852846401718583,
|
| 373 |
+
"grad_norm": 63.31580116626401,
|
| 374 |
"learning_rate": 7.332128540153017e-07,
|
| 375 |
+
"logits/chosen": -10.71928596496582,
|
| 376 |
+
"logits/rejected": -10.555776596069336,
|
| 377 |
+
"logps/chosen": -0.788918673992157,
|
| 378 |
+
"logps/rejected": -0.9437187910079956,
|
| 379 |
+
"loss": 4.1085,
|
| 380 |
"rewards/accuracies": 0.6625000238418579,
|
| 381 |
+
"rewards/chosen": -7.889185905456543,
|
| 382 |
+
"rewards/margins": 1.5480016469955444,
|
| 383 |
+
"rewards/rejected": -9.437189102172852,
|
| 384 |
"step": 125
|
| 385 |
},
|
| 386 |
{
|
| 387 |
"epoch": 0.27926960257787325,
|
| 388 |
+
"grad_norm": 81.18347953083857,
|
| 389 |
"learning_rate": 7.246637422199322e-07,
|
| 390 |
+
"logits/chosen": -10.676037788391113,
|
| 391 |
+
"logits/rejected": -10.630210876464844,
|
| 392 |
+
"logps/chosen": -0.8381876945495605,
|
| 393 |
+
"logps/rejected": -1.0221493244171143,
|
| 394 |
+
"loss": 4.0857,
|
| 395 |
+
"rewards/accuracies": 0.699999988079071,
|
| 396 |
+
"rewards/chosen": -8.381875991821289,
|
| 397 |
+
"rewards/margins": 1.8396151065826416,
|
| 398 |
+
"rewards/rejected": -10.221491813659668,
|
| 399 |
"step": 130
|
| 400 |
},
|
| 401 |
{
|
| 402 |
"epoch": 0.2900107411385607,
|
| 403 |
+
"grad_norm": 80.90445716094146,
|
| 404 |
"learning_rate": 7.156562037585574e-07,
|
| 405 |
+
"logits/chosen": -11.714326858520508,
|
| 406 |
+
"logits/rejected": -11.1636323928833,
|
| 407 |
+
"logps/chosen": -0.8452903628349304,
|
| 408 |
+
"logps/rejected": -1.1077954769134521,
|
| 409 |
+
"loss": 3.9406,
|
| 410 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 411 |
+
"rewards/chosen": -8.452905654907227,
|
| 412 |
+
"rewards/margins": 2.6250510215759277,
|
| 413 |
+
"rewards/rejected": -11.07795524597168,
|
| 414 |
"step": 135
|
| 415 |
},
|
| 416 |
{
|
| 417 |
"epoch": 0.3007518796992481,
|
| 418 |
+
"grad_norm": 88.8085364984583,
|
| 419 |
"learning_rate": 7.062029573160467e-07,
|
| 420 |
+
"logits/chosen": -11.935297012329102,
|
| 421 |
+
"logits/rejected": -11.792046546936035,
|
| 422 |
+
"logps/chosen": -0.9109989404678345,
|
| 423 |
+
"logps/rejected": -1.1446388959884644,
|
| 424 |
+
"loss": 3.6921,
|
| 425 |
+
"rewards/accuracies": 0.8125,
|
| 426 |
+
"rewards/chosen": -9.109989166259766,
|
| 427 |
+
"rewards/margins": 2.3363993167877197,
|
| 428 |
+
"rewards/rejected": -11.446390151977539,
|
| 429 |
"step": 140
|
| 430 |
},
|
| 431 |
{
|
| 432 |
"epoch": 0.31149301825993553,
|
| 433 |
+
"grad_norm": 110.78828702618637,
|
| 434 |
"learning_rate": 6.963173509189455e-07,
|
| 435 |
+
"logits/chosen": -13.552042007446289,
|
| 436 |
+
"logits/rejected": -13.324705123901367,
|
| 437 |
+
"logps/chosen": -1.0502710342407227,
|
| 438 |
+
"logps/rejected": -1.2693006992340088,
|
| 439 |
+
"loss": 3.8236,
|
| 440 |
"rewards/accuracies": 0.5375000238418579,
|
| 441 |
+
"rewards/chosen": -10.502711296081543,
|
| 442 |
+
"rewards/margins": 2.190295696258545,
|
| 443 |
+
"rewards/rejected": -12.69300651550293,
|
| 444 |
"step": 145
|
| 445 |
},
|
| 446 |
{
|
| 447 |
"epoch": 0.322234156820623,
|
| 448 |
+
"grad_norm": 111.15417892636452,
|
| 449 |
"learning_rate": 6.860133430880024e-07,
|
| 450 |
+
"logits/chosen": -14.586761474609375,
|
| 451 |
+
"logits/rejected": -14.23077392578125,
|
| 452 |
+
"logps/chosen": -1.2444875240325928,
|
| 453 |
+
"logps/rejected": -1.4712624549865723,
|
| 454 |
+
"loss": 3.7527,
|
| 455 |
+
"rewards/accuracies": 0.637499988079071,
|
| 456 |
+
"rewards/chosen": -12.444875717163086,
|
| 457 |
+
"rewards/margins": 2.2677478790283203,
|
| 458 |
+
"rewards/rejected": -14.712623596191406,
|
| 459 |
"step": 150
|
| 460 |
},
|
| 461 |
{
|
| 462 |
"epoch": 0.33297529538131043,
|
| 463 |
+
"grad_norm": 119.28100769496392,
|
| 464 |
"learning_rate": 6.753054831286747e-07,
|
| 465 |
+
"logits/chosen": -14.823234558105469,
|
| 466 |
+
"logits/rejected": -14.770757675170898,
|
| 467 |
+
"logps/chosen": -1.3944904804229736,
|
| 468 |
+
"logps/rejected": -1.6155385971069336,
|
| 469 |
+
"loss": 3.893,
|
| 470 |
+
"rewards/accuracies": 0.6875,
|
| 471 |
+
"rewards/chosen": -13.944903373718262,
|
| 472 |
+
"rewards/margins": 2.2104804515838623,
|
| 473 |
+
"rewards/rejected": -16.155384063720703,
|
| 474 |
"step": 155
|
| 475 |
},
|
| 476 |
{
|
| 477 |
"epoch": 0.34371643394199786,
|
| 478 |
+
"grad_norm": 103.61310969210844,
|
| 479 |
"learning_rate": 6.642088905874433e-07,
|
| 480 |
+
"logits/chosen": -14.195696830749512,
|
| 481 |
+
"logits/rejected": -14.07690143585205,
|
| 482 |
+
"logps/chosen": -1.3757129907608032,
|
| 483 |
+
"logps/rejected": -1.5891997814178467,
|
| 484 |
+
"loss": 3.7363,
|
| 485 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 486 |
+
"rewards/chosen": -13.75713062286377,
|
| 487 |
+
"rewards/margins": 2.134868860244751,
|
| 488 |
+
"rewards/rejected": -15.891998291015625,
|
| 489 |
"step": 160
|
| 490 |
},
|
| 491 |
{
|
| 492 |
"epoch": 0.3544575725026853,
|
| 493 |
+
"grad_norm": 95.46872943421629,
|
| 494 |
"learning_rate": 6.527392339029455e-07,
|
| 495 |
+
"logits/chosen": -14.401777267456055,
|
| 496 |
+
"logits/rejected": -14.322749137878418,
|
| 497 |
+
"logps/chosen": -1.2530758380889893,
|
| 498 |
+
"logps/rejected": -1.5336121320724487,
|
| 499 |
+
"loss": 3.4763,
|
| 500 |
+
"rewards/accuracies": 0.737500011920929,
|
| 501 |
+
"rewards/chosen": -12.530759811401367,
|
| 502 |
+
"rewards/margins": 2.8053627014160156,
|
| 503 |
+
"rewards/rejected": -15.336122512817383,
|
| 504 |
"step": 165
|
| 505 |
},
|
| 506 |
{
|
| 507 |
"epoch": 0.3651987110633727,
|
| 508 |
+
"grad_norm": 119.23452912777815,
|
| 509 |
"learning_rate": 6.409127082820689e-07,
|
| 510 |
+
"logits/chosen": -14.566454887390137,
|
| 511 |
+
"logits/rejected": -14.484842300415039,
|
| 512 |
+
"logps/chosen": -1.531166434288025,
|
| 513 |
+
"logps/rejected": -1.7996248006820679,
|
| 514 |
+
"loss": 3.6554,
|
| 515 |
+
"rewards/accuracies": 0.6875,
|
| 516 |
+
"rewards/chosen": -15.311663627624512,
|
| 517 |
+
"rewards/margins": 2.6845829486846924,
|
| 518 |
+
"rewards/rejected": -17.996248245239258,
|
| 519 |
"step": 170
|
| 520 |
},
|
| 521 |
{
|
| 522 |
"epoch": 0.37593984962406013,
|
| 523 |
+
"grad_norm": 123.93953756068933,
|
| 524 |
"learning_rate": 6.287460128322457e-07,
|
| 525 |
+
"logits/chosen": -14.157377243041992,
|
| 526 |
+
"logits/rejected": -14.0371675491333,
|
| 527 |
+
"logps/chosen": -1.5080561637878418,
|
| 528 |
+
"logps/rejected": -1.9248558282852173,
|
| 529 |
+
"loss": 3.4374,
|
| 530 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 531 |
+
"rewards/chosen": -15.080561637878418,
|
| 532 |
+
"rewards/margins": 4.167994022369385,
|
| 533 |
+
"rewards/rejected": -19.24855613708496,
|
| 534 |
"step": 175
|
| 535 |
},
|
| 536 |
{
|
| 537 |
"epoch": 0.3866809881847476,
|
| 538 |
+
"grad_norm": 223.710665367037,
|
| 539 |
"learning_rate": 6.16256326982239e-07,
|
| 540 |
+
"logits/chosen": -16.03777313232422,
|
| 541 |
+
"logits/rejected": -16.100345611572266,
|
| 542 |
+
"logps/chosen": -1.5300877094268799,
|
| 543 |
+
"logps/rejected": -1.8575336933135986,
|
| 544 |
+
"loss": 3.4606,
|
| 545 |
+
"rewards/accuracies": 0.6875,
|
| 546 |
+
"rewards/chosen": -15.300875663757324,
|
| 547 |
+
"rewards/margins": 3.2744598388671875,
|
| 548 |
+
"rewards/rejected": -18.575336456298828,
|
| 549 |
"step": 180
|
| 550 |
},
|
| 551 |
{
|
| 552 |
"epoch": 0.39742212674543503,
|
| 553 |
+
"grad_norm": 294.4348098673441,
|
| 554 |
"learning_rate": 6.034612862247114e-07,
|
| 555 |
+
"logits/chosen": -14.142799377441406,
|
| 556 |
+
"logits/rejected": -13.796422958374023,
|
| 557 |
+
"logps/chosen": -1.5025275945663452,
|
| 558 |
+
"logps/rejected": -1.7810484170913696,
|
| 559 |
+
"loss": 3.1185,
|
| 560 |
+
"rewards/accuracies": 0.6875,
|
| 561 |
+
"rewards/chosen": -15.025274276733398,
|
| 562 |
+
"rewards/margins": 2.7852089405059814,
|
| 563 |
+
"rewards/rejected": -17.810483932495117,
|
| 564 |
"step": 185
|
| 565 |
},
|
| 566 |
{
|
| 567 |
"epoch": 0.40816326530612246,
|
| 568 |
+
"grad_norm": 112.37773385751002,
|
| 569 |
"learning_rate": 5.903789572148295e-07,
|
| 570 |
+
"logits/chosen": -14.8009614944458,
|
| 571 |
+
"logits/rejected": -14.250249862670898,
|
| 572 |
+
"logps/chosen": -1.5931546688079834,
|
| 573 |
+
"logps/rejected": -2.022378444671631,
|
| 574 |
+
"loss": 3.3847,
|
| 575 |
+
"rewards/accuracies": 0.75,
|
| 576 |
+
"rewards/chosen": -15.931546211242676,
|
| 577 |
+
"rewards/margins": 4.292238712310791,
|
| 578 |
+
"rewards/rejected": -20.223783493041992,
|
| 579 |
"step": 190
|
| 580 |
},
|
| 581 |
{
|
| 582 |
"epoch": 0.4189044038668099,
|
| 583 |
+
"grad_norm": 125.77761635765717,
|
| 584 |
"learning_rate": 5.770278122600662e-07,
|
| 585 |
+
"logits/chosen": -14.832977294921875,
|
| 586 |
+
"logits/rejected": -14.608530044555664,
|
| 587 |
+
"logps/chosen": -1.6177564859390259,
|
| 588 |
+
"logps/rejected": -1.9727256298065186,
|
| 589 |
+
"loss": 3.6009,
|
| 590 |
+
"rewards/accuracies": 0.675000011920929,
|
| 591 |
+
"rewards/chosen": -16.177562713623047,
|
| 592 |
+
"rewards/margins": 3.549692153930664,
|
| 593 |
+
"rewards/rejected": -19.72725486755371,
|
| 594 |
"step": 195
|
| 595 |
},
|
| 596 |
{
|
| 597 |
"epoch": 0.4296455424274973,
|
| 598 |
+
"grad_norm": 101.70153964682511,
|
| 599 |
"learning_rate": 5.634267032372192e-07,
|
| 600 |
+
"logits/chosen": -14.803668022155762,
|
| 601 |
+
"logits/rejected": -14.786203384399414,
|
| 602 |
+
"logps/chosen": -1.6423594951629639,
|
| 603 |
+
"logps/rejected": -1.9881032705307007,
|
| 604 |
+
"loss": 3.3904,
|
| 605 |
"rewards/accuracies": 0.699999988079071,
|
| 606 |
+
"rewards/chosen": -16.423595428466797,
|
| 607 |
+
"rewards/margins": 3.457437515258789,
|
| 608 |
+
"rewards/rejected": -19.881032943725586,
|
| 609 |
"step": 200
|
| 610 |
},
|
| 611 |
{
|
| 612 |
"epoch": 0.44038668098818473,
|
| 613 |
+
"grad_norm": 123.22102599092351,
|
| 614 |
"learning_rate": 5.495948349734758e-07,
|
| 615 |
+
"logits/chosen": -14.582061767578125,
|
| 616 |
+
"logits/rejected": -14.51270580291748,
|
| 617 |
+
"logps/chosen": -1.747982382774353,
|
| 618 |
+
"logps/rejected": -2.051506757736206,
|
| 619 |
+
"loss": 3.1521,
|
| 620 |
+
"rewards/accuracies": 0.699999988079071,
|
| 621 |
+
"rewards/chosen": -17.47982406616211,
|
| 622 |
+
"rewards/margins": 3.035243034362793,
|
| 623 |
+
"rewards/rejected": -20.51506805419922,
|
| 624 |
"step": 205
|
| 625 |
},
|
| 626 |
{
|
| 627 |
"epoch": 0.45112781954887216,
|
| 628 |
+
"grad_norm": 106.53206062736002,
|
| 629 |
"learning_rate": 5.355517381291105e-07,
|
| 630 |
+
"logits/chosen": -15.856142044067383,
|
| 631 |
+
"logits/rejected": -15.048059463500977,
|
| 632 |
+
"logps/chosen": -1.9086406230926514,
|
| 633 |
+
"logps/rejected": -2.3940463066101074,
|
| 634 |
+
"loss": 3.1856,
|
| 635 |
"rewards/accuracies": 0.75,
|
| 636 |
+
"rewards/chosen": -19.086406707763672,
|
| 637 |
+
"rewards/margins": 4.854057312011719,
|
| 638 |
+
"rewards/rejected": -23.940462112426758,
|
| 639 |
"step": 210
|
| 640 |
},
|
| 641 |
{
|
| 642 |
"epoch": 0.46186895810955964,
|
| 643 |
+
"grad_norm": 135.01895835142213,
|
| 644 |
"learning_rate": 5.21317241620105e-07,
|
| 645 |
+
"logits/chosen": -17.223520278930664,
|
| 646 |
+
"logits/rejected": -16.822795867919922,
|
| 647 |
+
"logps/chosen": -1.9180253744125366,
|
| 648 |
+
"logps/rejected": -2.2895708084106445,
|
| 649 |
+
"loss": 3.4171,
|
| 650 |
+
"rewards/accuracies": 0.75,
|
| 651 |
+
"rewards/chosen": -19.180253982543945,
|
| 652 |
+
"rewards/margins": 3.7154533863067627,
|
| 653 |
+
"rewards/rejected": -22.895706176757812,
|
| 654 |
"step": 215
|
| 655 |
},
|
| 656 |
{
|
| 657 |
"epoch": 0.47261009667024706,
|
| 658 |
+
"grad_norm": 116.88786548811646,
|
| 659 |
"learning_rate": 5.069114446196291e-07,
|
| 660 |
+
"logits/chosen": -14.405430793762207,
|
| 661 |
+
"logits/rejected": -14.03125,
|
| 662 |
+
"logps/chosen": -1.8195463418960571,
|
| 663 |
+
"logps/rejected": -2.373373508453369,
|
| 664 |
+
"loss": 3.1087,
|
| 665 |
"rewards/accuracies": 0.8500000238418579,
|
| 666 |
+
"rewards/chosen": -18.195463180541992,
|
| 667 |
+
"rewards/margins": 5.538268566131592,
|
| 668 |
+
"rewards/rejected": -23.73373031616211,
|
| 669 |
"step": 220
|
| 670 |
},
|
| 671 |
{
|
| 672 |
"epoch": 0.4833512352309345,
|
| 673 |
+
"grad_norm": 117.33800643418842,
|
| 674 |
"learning_rate": 4.923546881779183e-07,
|
| 675 |
+
"logits/chosen": -15.34239387512207,
|
| 676 |
+
"logits/rejected": -15.118896484375,
|
| 677 |
+
"logps/chosen": -1.5427398681640625,
|
| 678 |
+
"logps/rejected": -1.9821481704711914,
|
| 679 |
+
"loss": 3.0885,
|
| 680 |
+
"rewards/accuracies": 0.699999988079071,
|
| 681 |
+
"rewards/chosen": -15.427398681640625,
|
| 682 |
+
"rewards/margins": 4.394083023071289,
|
| 683 |
+
"rewards/rejected": -19.821481704711914,
|
| 684 |
"step": 225
|
| 685 |
},
|
| 686 |
{
|
| 687 |
"epoch": 0.4940923737916219,
|
| 688 |
+
"grad_norm": 116.49865117484065,
|
| 689 |
"learning_rate": 4.776675265006186e-07,
|
| 690 |
+
"logits/chosen": -14.630195617675781,
|
| 691 |
+
"logits/rejected": -14.59937858581543,
|
| 692 |
+
"logps/chosen": -1.6563146114349365,
|
| 693 |
+
"logps/rejected": -2.0928232669830322,
|
| 694 |
+
"loss": 3.2032,
|
| 695 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 696 |
+
"rewards/chosen": -16.56314468383789,
|
| 697 |
+
"rewards/margins": 4.365086555480957,
|
| 698 |
+
"rewards/rejected": -20.928232192993164,
|
| 699 |
"step": 230
|
| 700 |
},
|
| 701 |
{
|
| 702 |
"epoch": 0.5048335123523093,
|
| 703 |
+
"grad_norm": 203.29377037587201,
|
| 704 |
"learning_rate": 4.62870697926156e-07,
|
| 705 |
+
"logits/chosen": -14.8600435256958,
|
| 706 |
+
"logits/rejected": -15.338279724121094,
|
| 707 |
+
"logps/chosen": -1.7831714153289795,
|
| 708 |
+
"logps/rejected": -2.145383358001709,
|
| 709 |
+
"loss": 3.2633,
|
| 710 |
+
"rewards/accuracies": 0.762499988079071,
|
| 711 |
+
"rewards/chosen": -17.831714630126953,
|
| 712 |
+
"rewards/margins": 3.622117519378662,
|
| 713 |
+
"rewards/rejected": -21.45383071899414,
|
| 714 |
"step": 235
|
| 715 |
},
|
| 716 |
{
|
| 717 |
"epoch": 0.5155746509129968,
|
| 718 |
+
"grad_norm": 116.16395762640381,
|
| 719 |
"learning_rate": 4.479850956431092e-07,
|
| 720 |
+
"logits/chosen": -14.476922988891602,
|
| 721 |
+
"logits/rejected": -14.742956161499023,
|
| 722 |
+
"logps/chosen": -1.6810489892959595,
|
| 723 |
+
"logps/rejected": -2.0232789516448975,
|
| 724 |
+
"loss": 3.0869,
|
| 725 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 726 |
+
"rewards/chosen": -16.810489654541016,
|
| 727 |
+
"rewards/margins": 3.4223015308380127,
|
| 728 |
+
"rewards/rejected": -20.232791900634766,
|
| 729 |
"step": 240
|
| 730 |
},
|
| 731 |
{
|
| 732 |
"epoch": 0.5263157894736842,
|
| 733 |
+
"grad_norm": 98.59837238595348,
|
| 734 |
"learning_rate": 4.33031738188933e-07,
|
| 735 |
+
"logits/chosen": -15.394210815429688,
|
| 736 |
+
"logits/rejected": -14.864255905151367,
|
| 737 |
+
"logps/chosen": -1.6588356494903564,
|
| 738 |
+
"logps/rejected": -2.0340933799743652,
|
| 739 |
+
"loss": 3.2303,
|
| 740 |
+
"rewards/accuracies": 0.762499988079071,
|
| 741 |
+
"rewards/chosen": -16.58835792541504,
|
| 742 |
+
"rewards/margins": 3.752579927444458,
|
| 743 |
+
"rewards/rejected": -20.3409366607666,
|
| 744 |
"step": 245
|
| 745 |
},
|
| 746 |
{
|
| 747 |
"epoch": 0.5370569280343717,
|
| 748 |
+
"grad_norm": 122.58746074195308,
|
| 749 |
"learning_rate": 4.180317397716889e-07,
|
| 750 |
+
"logits/chosen": -15.588345527648926,
|
| 751 |
+
"logits/rejected": -15.222723007202148,
|
| 752 |
+
"logps/chosen": -1.6328115463256836,
|
| 753 |
+
"logps/rejected": -2.2505877017974854,
|
| 754 |
+
"loss": 3.0874,
|
| 755 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 756 |
+
"rewards/chosen": -16.328113555908203,
|
| 757 |
+
"rewards/margins": 6.177763938903809,
|
| 758 |
+
"rewards/rejected": -22.505878448486328,
|
| 759 |
"step": 250
|
| 760 |
},
|
| 761 |
{
|
| 762 |
"epoch": 0.547798066595059,
|
| 763 |
+
"grad_norm": 120.49234048427142,
|
| 764 |
"learning_rate": 4.030062804566888e-07,
|
| 765 |
+
"logits/chosen": -15.462881088256836,
|
| 766 |
+
"logits/rejected": -15.3878173828125,
|
| 767 |
+
"logps/chosen": -1.6617505550384521,
|
| 768 |
+
"logps/rejected": -1.9518375396728516,
|
| 769 |
+
"loss": 3.0471,
|
| 770 |
"rewards/accuracies": 0.6499999761581421,
|
| 771 |
+
"rewards/chosen": -16.61750602722168,
|
| 772 |
+
"rewards/margins": 2.9008688926696777,
|
| 773 |
+
"rewards/rejected": -19.518375396728516,
|
| 774 |
"step": 255
|
| 775 |
},
|
| 776 |
{
|
| 777 |
"epoch": 0.5585392051557465,
|
| 778 |
+
"grad_norm": 102.36282696415171,
|
| 779 |
"learning_rate": 3.8797657626014614e-07,
|
| 780 |
+
"logits/chosen": -15.693799018859863,
|
| 781 |
+
"logits/rejected": -15.475125312805176,
|
| 782 |
+
"logps/chosen": -1.7303409576416016,
|
| 783 |
+
"logps/rejected": -2.111295223236084,
|
| 784 |
+
"loss": 3.1486,
|
| 785 |
+
"rewards/accuracies": 0.699999988079071,
|
| 786 |
+
"rewards/chosen": -17.303409576416016,
|
| 787 |
+
"rewards/margins": 3.809544801712036,
|
| 788 |
+
"rewards/rejected": -21.112953186035156,
|
| 789 |
"step": 260
|
| 790 |
},
|
| 791 |
{
|
| 792 |
"epoch": 0.569280343716434,
|
| 793 |
+
"grad_norm": 120.30352416855824,
|
| 794 |
"learning_rate": 3.729638491920669e-07,
|
| 795 |
+
"logits/chosen": -14.018827438354492,
|
| 796 |
+
"logits/rejected": -14.128240585327148,
|
| 797 |
+
"logps/chosen": -1.5650581121444702,
|
| 798 |
+
"logps/rejected": -1.8844165802001953,
|
| 799 |
+
"loss": 3.0723,
|
| 800 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 801 |
+
"rewards/chosen": -15.650581359863281,
|
| 802 |
+
"rewards/margins": 3.193586826324463,
|
| 803 |
+
"rewards/rejected": -18.84417152404785,
|
| 804 |
"step": 265
|
| 805 |
},
|
| 806 |
{
|
| 807 |
"epoch": 0.5800214822771214,
|
| 808 |
+
"grad_norm": 91.11946128351651,
|
| 809 |
"learning_rate": 3.5798929729067464e-07,
|
| 810 |
+
"logits/chosen": -15.980966567993164,
|
| 811 |
+
"logits/rejected": -15.59577465057373,
|
| 812 |
+
"logps/chosen": -1.759478211402893,
|
| 813 |
+
"logps/rejected": -2.1487960815429688,
|
| 814 |
+
"loss": 2.7323,
|
| 815 |
+
"rewards/accuracies": 0.824999988079071,
|
| 816 |
+
"rewards/chosen": -17.59478187561035,
|
| 817 |
+
"rewards/margins": 3.8931777477264404,
|
| 818 |
+
"rewards/rejected": -21.487960815429688,
|
| 819 |
"step": 270
|
| 820 |
},
|
| 821 |
{
|
| 822 |
"epoch": 0.5907626208378088,
|
| 823 |
+
"grad_norm": 111.96074414088407,
|
| 824 |
"learning_rate": 3.4307406469068595e-07,
|
| 825 |
+
"logits/chosen": -15.691810607910156,
|
| 826 |
+
"logits/rejected": -15.631698608398438,
|
| 827 |
+
"logps/chosen": -1.7383606433868408,
|
| 828 |
+
"logps/rejected": -2.2635440826416016,
|
| 829 |
+
"loss": 2.9053,
|
| 830 |
"rewards/accuracies": 0.8500000238418579,
|
| 831 |
+
"rewards/chosen": -17.38360595703125,
|
| 832 |
+
"rewards/margins": 5.251835823059082,
|
| 833 |
+
"rewards/rejected": -22.635440826416016,
|
| 834 |
"step": 275
|
| 835 |
},
|
| 836 |
{
|
| 837 |
"epoch": 0.6015037593984962,
|
| 838 |
+
"grad_norm": 122.95732420434838,
|
| 839 |
"learning_rate": 3.282392117676968e-07,
|
| 840 |
+
"logits/chosen": -15.389913558959961,
|
| 841 |
+
"logits/rejected": -15.556841850280762,
|
| 842 |
+
"logps/chosen": -1.9042613506317139,
|
| 843 |
+
"logps/rejected": -2.476677417755127,
|
| 844 |
+
"loss": 2.9112,
|
| 845 |
+
"rewards/accuracies": 0.824999988079071,
|
| 846 |
+
"rewards/chosen": -19.04261589050293,
|
| 847 |
+
"rewards/margins": 5.72415828704834,
|
| 848 |
+
"rewards/rejected": -24.766775131225586,
|
| 849 |
"step": 280
|
| 850 |
},
|
| 851 |
{
|
| 852 |
"epoch": 0.6122448979591837,
|
| 853 |
+
"grad_norm": 104.08928006990811,
|
| 854 |
"learning_rate": 3.135056854008371e-07,
|
| 855 |
+
"logits/chosen": -16.152729034423828,
|
| 856 |
+
"logits/rejected": -16.002233505249023,
|
| 857 |
+
"logps/chosen": -1.8163830041885376,
|
| 858 |
+
"logps/rejected": -2.228738784790039,
|
| 859 |
+
"loss": 2.8591,
|
| 860 |
"rewards/accuracies": 0.75,
|
| 861 |
+
"rewards/chosen": -18.163829803466797,
|
| 862 |
+
"rewards/margins": 4.123559474945068,
|
| 863 |
+
"rewards/rejected": -22.28738784790039,
|
| 864 |
"step": 285
|
| 865 |
},
|
| 866 |
{
|
| 867 |
"epoch": 0.6229860365198711,
|
| 868 |
+
"grad_norm": 113.24003069429473,
|
| 869 |
"learning_rate": 2.988942893956833e-07,
|
| 870 |
+
"logits/chosen": -15.338768005371094,
|
| 871 |
+
"logits/rejected": -15.249606132507324,
|
| 872 |
+
"logps/chosen": -1.8918412923812866,
|
| 873 |
+
"logps/rejected": -2.32609224319458,
|
| 874 |
+
"loss": 3.0229,
|
| 875 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 876 |
+
"rewards/chosen": -18.918415069580078,
|
| 877 |
+
"rewards/margins": 4.342508792877197,
|
| 878 |
+
"rewards/rejected": -23.260921478271484,
|
| 879 |
"step": 290
|
| 880 |
},
|
| 881 |
{
|
| 882 |
"epoch": 0.6337271750805585,
|
| 883 |
+
"grad_norm": 151.8876206935079,
|
| 884 |
"learning_rate": 2.844256551091911e-07,
|
| 885 |
+
"logits/chosen": -16.8232421875,
|
| 886 |
+
"logits/rejected": -16.842761993408203,
|
| 887 |
+
"logps/chosen": -1.9518108367919922,
|
| 888 |
+
"logps/rejected": -2.4939913749694824,
|
| 889 |
+
"loss": 2.833,
|
| 890 |
+
"rewards/accuracies": 0.8125,
|
| 891 |
+
"rewards/chosen": -19.518108367919922,
|
| 892 |
+
"rewards/margins": 5.421802997589111,
|
| 893 |
+
"rewards/rejected": -24.939910888671875,
|
| 894 |
"step": 295
|
| 895 |
},
|
| 896 |
{
|
| 897 |
"epoch": 0.644468313641246,
|
| 898 |
+
"grad_norm": 273.6603281637076,
|
| 899 |
"learning_rate": 2.7012021231812664e-07,
|
| 900 |
+
"logits/chosen": -16.766956329345703,
|
| 901 |
+
"logits/rejected": -16.424999237060547,
|
| 902 |
+
"logps/chosen": -1.969435453414917,
|
| 903 |
+
"logps/rejected": -2.308202028274536,
|
| 904 |
+
"loss": 3.2442,
|
| 905 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 906 |
+
"rewards/chosen": -19.694355010986328,
|
| 907 |
+
"rewards/margins": 3.387664318084717,
|
| 908 |
+
"rewards/rejected": -23.082019805908203,
|
| 909 |
"step": 300
|
| 910 |
},
|
| 911 |
{
|
| 912 |
"epoch": 0.6552094522019334,
|
| 913 |
+
"grad_norm": 187.01720889425104,
|
| 914 |
"learning_rate": 2.5599816037212954e-07,
|
| 915 |
+
"logits/chosen": -14.743069648742676,
|
| 916 |
+
"logits/rejected": -14.623723983764648,
|
| 917 |
+
"logps/chosen": -1.8649146556854248,
|
| 918 |
+
"logps/rejected": -2.4030163288116455,
|
| 919 |
+
"loss": 2.9136,
|
| 920 |
"rewards/accuracies": 0.824999988079071,
|
| 921 |
+
"rewards/chosen": -18.64914894104004,
|
| 922 |
+
"rewards/margins": 5.381015777587891,
|
| 923 |
+
"rewards/rejected": -24.03016471862793,
|
| 924 |
"step": 305
|
| 925 |
},
|
| 926 |
{
|
| 927 |
"epoch": 0.6659505907626209,
|
| 928 |
+
"grad_norm": 142.83830462137144,
|
| 929 |
"learning_rate": 2.4207943967214064e-07,
|
| 930 |
+
"logits/chosen": -16.09463119506836,
|
| 931 |
+
"logits/rejected": -15.806689262390137,
|
| 932 |
+
"logps/chosen": -2.033447742462158,
|
| 933 |
+
"logps/rejected": -2.524364471435547,
|
| 934 |
+
"loss": 3.1207,
|
| 935 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 936 |
+
"rewards/chosen": -20.334476470947266,
|
| 937 |
+
"rewards/margins": 4.909164905548096,
|
| 938 |
+
"rewards/rejected": -25.24364471435547,
|
| 939 |
"step": 310
|
| 940 |
},
|
| 941 |
{
|
| 942 |
"epoch": 0.6766917293233082,
|
| 943 |
+
"grad_norm": 106.6308691579763,
|
| 944 |
"learning_rate": 2.2838370351446547e-07,
|
| 945 |
+
"logits/chosen": -15.870585441589355,
|
| 946 |
+
"logits/rejected": -15.55876350402832,
|
| 947 |
+
"logps/chosen": -1.8023881912231445,
|
| 948 |
+
"logps/rejected": -2.277968168258667,
|
| 949 |
+
"loss": 2.7641,
|
| 950 |
"rewards/accuracies": 0.6875,
|
| 951 |
+
"rewards/chosen": -18.023881912231445,
|
| 952 |
+
"rewards/margins": 4.755801200866699,
|
| 953 |
+
"rewards/rejected": -22.779682159423828,
|
| 954 |
"step": 315
|
| 955 |
},
|
| 956 |
{
|
| 957 |
"epoch": 0.6874328678839957,
|
| 958 |
+
"grad_norm": 108.7814729394269,
|
| 959 |
"learning_rate": 2.1493029034023188e-07,
|
| 960 |
+
"logits/chosen": -15.210580825805664,
|
| 961 |
+
"logits/rejected": -15.057415962219238,
|
| 962 |
+
"logps/chosen": -1.789072036743164,
|
| 963 |
+
"logps/rejected": -2.2775633335113525,
|
| 964 |
+
"loss": 2.8863,
|
| 965 |
+
"rewards/accuracies": 0.75,
|
| 966 |
+
"rewards/chosen": -17.89072036743164,
|
| 967 |
+
"rewards/margins": 4.884912014007568,
|
| 968 |
+
"rewards/rejected": -22.775630950927734,
|
| 969 |
"step": 320
|
| 970 |
},
|
| 971 |
{
|
| 972 |
"epoch": 0.6981740064446831,
|
| 973 |
+
"grad_norm": 180.86259023584043,
|
| 974 |
"learning_rate": 2.0173819642942376e-07,
|
| 975 |
+
"logits/chosen": -14.378689765930176,
|
| 976 |
+
"logits/rejected": -14.201916694641113,
|
| 977 |
+
"logps/chosen": -1.9404065608978271,
|
| 978 |
+
"logps/rejected": -2.5850563049316406,
|
| 979 |
+
"loss": 2.9848,
|
| 980 |
"rewards/accuracies": 0.7749999761581421,
|
| 981 |
+
"rewards/chosen": -19.404064178466797,
|
| 982 |
+
"rewards/margins": 6.446499824523926,
|
| 983 |
+
"rewards/rejected": -25.85056495666504,
|
| 984 |
"step": 325
|
| 985 |
},
|
| 986 |
{
|
| 987 |
"epoch": 0.7089151450053706,
|
| 988 |
+
"grad_norm": 132.13798361660807,
|
| 989 |
"learning_rate": 1.888260490780485e-07,
|
| 990 |
+
"logits/chosen": -14.281087875366211,
|
| 991 |
+
"logits/rejected": -14.154438972473145,
|
| 992 |
+
"logps/chosen": -1.7613548040390015,
|
| 993 |
+
"logps/rejected": -2.238495349884033,
|
| 994 |
+
"loss": 3.0495,
|
| 995 |
"rewards/accuracies": 0.800000011920929,
|
| 996 |
+
"rewards/chosen": -17.61355209350586,
|
| 997 |
+
"rewards/margins": 4.771404266357422,
|
| 998 |
+
"rewards/rejected": -22.384952545166016,
|
| 999 |
"step": 330
|
| 1000 |
},
|
| 1001 |
{
|
| 1002 |
"epoch": 0.719656283566058,
|
| 1003 |
+
"grad_norm": 127.44733826541191,
|
| 1004 |
"learning_rate": 1.7621208029631078e-07,
|
| 1005 |
+
"logits/chosen": -14.766406059265137,
|
| 1006 |
+
"logits/rejected": -14.667470932006836,
|
| 1007 |
+
"logps/chosen": -1.9043552875518799,
|
| 1008 |
+
"logps/rejected": -2.4970054626464844,
|
| 1009 |
+
"loss": 2.9109,
|
| 1010 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1011 |
+
"rewards/chosen": -19.04355239868164,
|
| 1012 |
+
"rewards/margins": 5.9265007972717285,
|
| 1013 |
+
"rewards/rejected": -24.970054626464844,
|
| 1014 |
"step": 335
|
| 1015 |
},
|
| 1016 |
{
|
| 1017 |
"epoch": 0.7303974221267454,
|
| 1018 |
+
"grad_norm": 114.16195914051926,
|
| 1019 |
"learning_rate": 1.6391410106493227e-07,
|
| 1020 |
+
"logits/chosen": -14.881872177124023,
|
| 1021 |
+
"logits/rejected": -14.627456665039062,
|
| 1022 |
+
"logps/chosen": -1.9786325693130493,
|
| 1023 |
+
"logps/rejected": -2.5424978733062744,
|
| 1024 |
+
"loss": 2.8817,
|
| 1025 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1026 |
+
"rewards/chosen": -19.786325454711914,
|
| 1027 |
+
"rewards/margins": 5.63865327835083,
|
| 1028 |
+
"rewards/rejected": -25.424976348876953,
|
| 1029 |
"step": 340
|
| 1030 |
},
|
| 1031 |
{
|
| 1032 |
"epoch": 0.7411385606874329,
|
| 1033 |
+
"grad_norm": 99.76652109445614,
|
| 1034 |
"learning_rate": 1.5194947618596673e-07,
|
| 1035 |
+
"logits/chosen": -15.026782035827637,
|
| 1036 |
+
"logits/rejected": -14.58587646484375,
|
| 1037 |
+
"logps/chosen": -1.9211170673370361,
|
| 1038 |
+
"logps/rejected": -2.394774913787842,
|
| 1039 |
+
"loss": 3.0269,
|
| 1040 |
"rewards/accuracies": 0.800000011920929,
|
| 1041 |
+
"rewards/chosen": -19.211170196533203,
|
| 1042 |
+
"rewards/margins": 4.736577033996582,
|
| 1043 |
+
"rewards/rejected": -23.9477481842041,
|
| 1044 |
"step": 345
|
| 1045 |
},
|
| 1046 |
{
|
| 1047 |
"epoch": 0.7518796992481203,
|
| 1048 |
+
"grad_norm": 107.2380444923777,
|
| 1049 |
"learning_rate": 1.4033509976362083e-07,
|
| 1050 |
+
"logits/chosen": -15.670697212219238,
|
| 1051 |
+
"logits/rejected": -15.579752922058105,
|
| 1052 |
+
"logps/chosen": -1.9338979721069336,
|
| 1053 |
+
"logps/rejected": -2.3520257472991943,
|
| 1054 |
+
"loss": 2.9892,
|
| 1055 |
"rewards/accuracies": 0.800000011920929,
|
| 1056 |
+
"rewards/chosen": -19.338979721069336,
|
| 1057 |
+
"rewards/margins": 4.181277751922607,
|
| 1058 |
+
"rewards/rejected": -23.52025604248047,
|
| 1059 |
"step": 350
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
"epoch": 0.7626208378088077,
|
| 1063 |
+
"grad_norm": 108.07049767560719,
|
| 1064 |
"learning_rate": 1.2908737134970363e-07,
|
| 1065 |
+
"logits/chosen": -14.5513334274292,
|
| 1066 |
+
"logits/rejected": -14.516873359680176,
|
| 1067 |
+
"logps/chosen": -1.8658726215362549,
|
| 1068 |
+
"logps/rejected": -2.476576566696167,
|
| 1069 |
+
"loss": 3.1196,
|
| 1070 |
"rewards/accuracies": 0.737500011920929,
|
| 1071 |
+
"rewards/chosen": -18.65872573852539,
|
| 1072 |
+
"rewards/margins": 6.1070404052734375,
|
| 1073 |
+
"rewards/rejected": -24.765766143798828,
|
| 1074 |
"step": 355
|
| 1075 |
},
|
| 1076 |
{
|
| 1077 |
"epoch": 0.7733619763694952,
|
| 1078 |
+
"grad_norm": 103.96338647941381,
|
| 1079 |
"learning_rate": 1.1822217278738515e-07,
|
| 1080 |
+
"logits/chosen": -15.559527397155762,
|
| 1081 |
+
"logits/rejected": -15.545167922973633,
|
| 1082 |
+
"logps/chosen": -1.9091074466705322,
|
| 1083 |
+
"logps/rejected": -2.4284536838531494,
|
| 1084 |
+
"loss": 3.0461,
|
| 1085 |
+
"rewards/accuracies": 0.75,
|
| 1086 |
+
"rewards/chosen": -19.091075897216797,
|
| 1087 |
+
"rewards/margins": 5.193462371826172,
|
| 1088 |
+
"rewards/rejected": -24.284536361694336,
|
| 1089 |
"step": 360
|
| 1090 |
},
|
| 1091 |
{
|
| 1092 |
"epoch": 0.7841031149301826,
|
| 1093 |
+
"grad_norm": 138.78303459839336,
|
| 1094 |
"learning_rate": 1.0775484578596241e-07,
|
| 1095 |
+
"logits/chosen": -15.669352531433105,
|
| 1096 |
+
"logits/rejected": -15.537897109985352,
|
| 1097 |
+
"logps/chosen": -1.94468092918396,
|
| 1098 |
+
"logps/rejected": -2.5436809062957764,
|
| 1099 |
+
"loss": 2.7974,
|
| 1100 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1101 |
+
"rewards/chosen": -19.44681167602539,
|
| 1102 |
+
"rewards/margins": 5.989997386932373,
|
| 1103 |
+
"rewards/rejected": -25.436809539794922,
|
| 1104 |
"step": 365
|
| 1105 |
},
|
| 1106 |
{
|
| 1107 |
"epoch": 0.7948442534908701,
|
| 1108 |
+
"grad_norm": 122.14734643501443,
|
| 1109 |
"learning_rate": 9.770017025829673e-08,
|
| 1110 |
+
"logits/chosen": -15.961019515991211,
|
| 1111 |
+
"logits/rejected": -15.95417308807373,
|
| 1112 |
+
"logps/chosen": -2.1812241077423096,
|
| 1113 |
+
"logps/rejected": -2.6908135414123535,
|
| 1114 |
+
"loss": 2.6374,
|
| 1115 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1116 |
+
"rewards/chosen": -21.812240600585938,
|
| 1117 |
+
"rewards/margins": 5.095890998840332,
|
| 1118 |
+
"rewards/rejected": -26.908132553100586,
|
| 1119 |
"step": 370
|
| 1120 |
},
|
| 1121 |
{
|
| 1122 |
"epoch": 0.8055853920515574,
|
| 1123 |
+
"grad_norm": 114.19485857094502,
|
| 1124 |
"learning_rate": 8.807234345151027e-08,
|
| 1125 |
+
"logits/chosen": -14.920249938964844,
|
| 1126 |
+
"logits/rejected": -14.890344619750977,
|
| 1127 |
+
"logps/chosen": -2.0413661003112793,
|
| 1128 |
+
"logps/rejected": -2.6543760299682617,
|
| 1129 |
+
"loss": 2.8891,
|
| 1130 |
+
"rewards/accuracies": 0.824999988079071,
|
| 1131 |
+
"rewards/chosen": -20.41366195678711,
|
| 1132 |
+
"rewards/margins": 6.13009786605835,
|
| 1133 |
+
"rewards/rejected": -26.543758392333984,
|
| 1134 |
"step": 375
|
| 1135 |
},
|
| 1136 |
{
|
| 1137 |
"epoch": 0.8163265306122449,
|
| 1138 |
+
"grad_norm": 184.4289644042616,
|
| 1139 |
"learning_rate": 7.888495990040924e-08,
|
| 1140 |
+
"logits/chosen": -13.656982421875,
|
| 1141 |
+
"logits/rejected": -13.701431274414062,
|
| 1142 |
+
"logps/chosen": -2.011772871017456,
|
| 1143 |
+
"logps/rejected": -2.7216084003448486,
|
| 1144 |
+
"loss": 2.9801,
|
| 1145 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 1146 |
+
"rewards/chosen": -20.117727279663086,
|
| 1147 |
+
"rewards/margins": 7.098354339599609,
|
| 1148 |
+
"rewards/rejected": -27.21608543395996,
|
| 1149 |
"step": 380
|
| 1150 |
},
|
| 1151 |
{
|
| 1152 |
"epoch": 0.8270676691729323,
|
| 1153 |
+
"grad_norm": 132.46593111411627,
|
| 1154 |
"learning_rate": 7.015099223193943e-08,
|
| 1155 |
+
"logits/chosen": -15.658330917358398,
|
| 1156 |
+
"logits/rejected": -15.693890571594238,
|
| 1157 |
+
"logps/chosen": -1.9581212997436523,
|
| 1158 |
+
"logps/rejected": -2.5001060962677,
|
| 1159 |
+
"loss": 2.8611,
|
| 1160 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1161 |
+
"rewards/chosen": -19.58121681213379,
|
| 1162 |
+
"rewards/margins": 5.419846534729004,
|
| 1163 |
+
"rewards/rejected": -25.001062393188477,
|
| 1164 |
"step": 385
|
| 1165 |
},
|
| 1166 |
{
|
| 1167 |
"epoch": 0.8378088077336198,
|
| 1168 |
+
"grad_norm": 109.73562851103766,
|
| 1169 |
"learning_rate": 6.188277284777857e-08,
|
| 1170 |
+
"logits/chosen": -14.48884391784668,
|
| 1171 |
+
"logits/rejected": -13.800481796264648,
|
| 1172 |
+
"logps/chosen": -1.9877732992172241,
|
| 1173 |
+
"logps/rejected": -2.5746169090270996,
|
| 1174 |
+
"loss": 2.8295,
|
| 1175 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 1176 |
+
"rewards/chosen": -19.877731323242188,
|
| 1177 |
+
"rewards/margins": 5.868436813354492,
|
| 1178 |
+
"rewards/rejected": -25.746166229248047,
|
| 1179 |
"step": 390
|
| 1180 |
},
|
| 1181 |
{
|
| 1182 |
"epoch": 0.8485499462943072,
|
| 1183 |
+
"grad_norm": 119.04235166242496,
|
| 1184 |
"learning_rate": 5.409197651092965e-08,
|
| 1185 |
+
"logits/chosen": -15.729510307312012,
|
| 1186 |
+
"logits/rejected": -15.620780944824219,
|
| 1187 |
+
"logps/chosen": -2.1660141944885254,
|
| 1188 |
+
"logps/rejected": -2.667978286743164,
|
| 1189 |
+
"loss": 2.7298,
|
| 1190 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1191 |
+
"rewards/chosen": -21.660140991210938,
|
| 1192 |
+
"rewards/margins": 5.019640922546387,
|
| 1193 |
+
"rewards/rejected": -26.679784774780273,
|
| 1194 |
"step": 395
|
| 1195 |
},
|
| 1196 |
{
|
| 1197 |
"epoch": 0.8592910848549946,
|
| 1198 |
+
"grad_norm": 105.88618147362118,
|
| 1199 |
"learning_rate": 4.678960386090298e-08,
|
| 1200 |
+
"logits/chosen": -15.191770553588867,
|
| 1201 |
+
"logits/rejected": -15.158576965332031,
|
| 1202 |
+
"logps/chosen": -1.9015194177627563,
|
| 1203 |
+
"logps/rejected": -2.5145716667175293,
|
| 1204 |
+
"loss": 2.7402,
|
| 1205 |
"rewards/accuracies": 0.8125,
|
| 1206 |
+
"rewards/chosen": -19.015193939208984,
|
| 1207 |
+
"rewards/margins": 6.130521297454834,
|
| 1208 |
+
"rewards/rejected": -25.145715713500977,
|
| 1209 |
"step": 400
|
| 1210 |
},
|
| 1211 |
{
|
| 1212 |
"epoch": 0.8700322234156821,
|
| 1213 |
+
"grad_norm": 114.47737417347113,
|
| 1214 |
"learning_rate": 3.998596588076366e-08,
|
| 1215 |
+
"logits/chosen": -13.7559814453125,
|
| 1216 |
+
"logits/rejected": -13.483953475952148,
|
| 1217 |
+
"logps/chosen": -1.9659799337387085,
|
| 1218 |
+
"logps/rejected": -2.390488862991333,
|
| 1219 |
+
"loss": 3.1028,
|
| 1220 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1221 |
+
"rewards/chosen": -19.659801483154297,
|
| 1222 |
+
"rewards/margins": 4.245090484619141,
|
| 1223 |
+
"rewards/rejected": -23.904891967773438,
|
| 1224 |
"step": 405
|
| 1225 |
},
|
| 1226 |
{
|
| 1227 |
"epoch": 0.8807733619763695,
|
| 1228 |
+
"grad_norm": 124.82158209754382,
|
| 1229 |
"learning_rate": 3.3690669337976996e-08,
|
| 1230 |
+
"logits/chosen": -15.061103820800781,
|
| 1231 |
+
"logits/rejected": -14.891242980957031,
|
| 1232 |
+
"logps/chosen": -1.8727645874023438,
|
| 1233 |
+
"logps/rejected": -2.3343753814697266,
|
| 1234 |
+
"loss": 2.7051,
|
| 1235 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1236 |
+
"rewards/chosen": -18.727645874023438,
|
| 1237 |
+
"rewards/margins": 4.6161088943481445,
|
| 1238 |
+
"rewards/rejected": -23.343753814697266,
|
| 1239 |
"step": 410
|
| 1240 |
},
|
| 1241 |
{
|
| 1242 |
"epoch": 0.8915145005370569,
|
| 1243 |
+
"grad_norm": 127.84134494966216,
|
| 1244 |
"learning_rate": 2.7912603219609798e-08,
|
| 1245 |
+
"logits/chosen": -15.650156021118164,
|
| 1246 |
+
"logits/rejected": -15.531412124633789,
|
| 1247 |
+
"logps/chosen": -2.0558724403381348,
|
| 1248 |
+
"logps/rejected": -2.456268787384033,
|
| 1249 |
+
"loss": 2.7054,
|
| 1250 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1251 |
+
"rewards/chosen": -20.558725357055664,
|
| 1252 |
+
"rewards/margins": 4.003961563110352,
|
| 1253 |
+
"rewards/rejected": -24.562685012817383,
|
| 1254 |
"step": 415
|
| 1255 |
},
|
| 1256 |
{
|
| 1257 |
"epoch": 0.9022556390977443,
|
| 1258 |
+
"grad_norm": 121.68417066288369,
|
| 1259 |
"learning_rate": 2.265992618104029e-08,
|
| 1260 |
+
"logits/chosen": -15.883665084838867,
|
| 1261 |
+
"logits/rejected": -15.859227180480957,
|
| 1262 |
+
"logps/chosen": -2.1571857929229736,
|
| 1263 |
+
"logps/rejected": -2.7032248973846436,
|
| 1264 |
+
"loss": 2.7706,
|
| 1265 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1266 |
+
"rewards/chosen": -21.571857452392578,
|
| 1267 |
+
"rewards/margins": 5.460390090942383,
|
| 1268 |
+
"rewards/rejected": -27.03224754333496,
|
| 1269 |
"step": 420
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
"epoch": 0.9129967776584318,
|
| 1273 |
+
"grad_norm": 205.5451817035243,
|
| 1274 |
"learning_rate": 1.7940055025900304e-08,
|
| 1275 |
+
"logits/chosen": -14.086555480957031,
|
| 1276 |
+
"logits/rejected": -13.882139205932617,
|
| 1277 |
+
"logps/chosen": -2.0481374263763428,
|
| 1278 |
+
"logps/rejected": -2.428682804107666,
|
| 1279 |
+
"loss": 3.0577,
|
| 1280 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1281 |
+
"rewards/chosen": -20.481372833251953,
|
| 1282 |
+
"rewards/margins": 3.8054566383361816,
|
| 1283 |
+
"rewards/rejected": -24.28683090209961,
|
| 1284 |
"step": 425
|
| 1285 |
},
|
| 1286 |
{
|
| 1287 |
"epoch": 0.9237379162191193,
|
| 1288 |
+
"grad_norm": 119.46095967048043,
|
| 1289 |
"learning_rate": 1.3759654233514817e-08,
|
| 1290 |
+
"logits/chosen": -14.88987922668457,
|
| 1291 |
+
"logits/rejected": -14.742823600769043,
|
| 1292 |
+
"logps/chosen": -1.941057562828064,
|
| 1293 |
+
"logps/rejected": -2.4432146549224854,
|
| 1294 |
+
"loss": 2.8075,
|
| 1295 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1296 |
+
"rewards/chosen": -19.41057586669922,
|
| 1297 |
+
"rewards/margins": 5.021571636199951,
|
| 1298 |
+
"rewards/rejected": -24.432147979736328,
|
| 1299 |
"step": 430
|
| 1300 |
},
|
| 1301 |
{
|
| 1302 |
"epoch": 0.9344790547798066,
|
| 1303 |
+
"grad_norm": 112.52416047730715,
|
| 1304 |
"learning_rate": 1.0124626548627402e-08,
|
| 1305 |
+
"logits/chosen": -15.55200481414795,
|
| 1306 |
+
"logits/rejected": -15.57677936553955,
|
| 1307 |
+
"logps/chosen": -2.0663094520568848,
|
| 1308 |
+
"logps/rejected": -2.723595380783081,
|
| 1309 |
+
"loss": 2.7934,
|
| 1310 |
+
"rewards/accuracies": 0.800000011920929,
|
| 1311 |
+
"rewards/chosen": -20.663097381591797,
|
| 1312 |
+
"rewards/margins": 6.5728559494018555,
|
| 1313 |
+
"rewards/rejected": -27.235950469970703,
|
| 1314 |
"step": 435
|
| 1315 |
},
|
| 1316 |
{
|
| 1317 |
"epoch": 0.9452201933404941,
|
| 1318 |
+
"grad_norm": 95.43516446345191,
|
| 1319 |
"learning_rate": 7.040104646698042e-09,
|
| 1320 |
+
"logits/chosen": -14.149500846862793,
|
| 1321 |
+
"logits/rejected": -14.121160507202148,
|
| 1322 |
+
"logps/chosen": -2.1648197174072266,
|
| 1323 |
+
"logps/rejected": -2.803515911102295,
|
| 1324 |
+
"loss": 2.7258,
|
| 1325 |
+
"rewards/accuracies": 0.75,
|
| 1326 |
+
"rewards/chosen": -21.648197174072266,
|
| 1327 |
+
"rewards/margins": 6.386962890625,
|
| 1328 |
+
"rewards/rejected": -28.035160064697266,
|
| 1329 |
"step": 440
|
| 1330 |
},
|
| 1331 |
{
|
| 1332 |
"epoch": 0.9559613319011815,
|
| 1333 |
+
"grad_norm": 105.32519506523218,
|
| 1334 |
"learning_rate": 4.510443886542114e-09,
|
| 1335 |
+
"logits/chosen": -15.509679794311523,
|
| 1336 |
+
"logits/rejected": -15.587133407592773,
|
| 1337 |
+
"logps/chosen": -2.019160509109497,
|
| 1338 |
+
"logps/rejected": -2.5319721698760986,
|
| 1339 |
+
"loss": 2.8015,
|
| 1340 |
+
"rewards/accuracies": 0.800000011920929,
|
| 1341 |
+
"rewards/chosen": -20.191600799560547,
|
| 1342 |
+
"rewards/margins": 5.128118991851807,
|
| 1343 |
+
"rewards/rejected": -25.319721221923828,
|
| 1344 |
"step": 445
|
| 1345 |
},
|
| 1346 |
{
|
| 1347 |
"epoch": 0.966702470461869,
|
| 1348 |
+
"grad_norm": 142.71634790119586,
|
| 1349 |
"learning_rate": 2.539216160544333e-09,
|
| 1350 |
+
"logits/chosen": -15.480878829956055,
|
| 1351 |
+
"logits/rejected": -15.102048873901367,
|
| 1352 |
+
"logps/chosen": -2.1352379322052,
|
| 1353 |
+
"logps/rejected": -2.589895486831665,
|
| 1354 |
+
"loss": 2.889,
|
| 1355 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1356 |
+
"rewards/chosen": -21.35237693786621,
|
| 1357 |
+
"rewards/margins": 4.546576023101807,
|
| 1358 |
+
"rewards/rejected": -25.898956298828125,
|
| 1359 |
"step": 450
|
| 1360 |
},
|
| 1361 |
{
|
| 1362 |
"epoch": 0.9774436090225563,
|
| 1363 |
+
"grad_norm": 123.94313579799326,
|
| 1364 |
"learning_rate": 1.1292048511303054e-09,
|
| 1365 |
+
"logits/chosen": -14.889852523803711,
|
| 1366 |
+
"logits/rejected": -15.104809761047363,
|
| 1367 |
+
"logps/chosen": -1.9709218740463257,
|
| 1368 |
+
"logps/rejected": -2.4727888107299805,
|
| 1369 |
+
"loss": 2.9813,
|
| 1370 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1371 |
+
"rewards/chosen": -19.709218978881836,
|
| 1372 |
+
"rewards/margins": 5.018665313720703,
|
| 1373 |
+
"rewards/rejected": -24.727886199951172,
|
| 1374 |
"step": 455
|
| 1375 |
},
|
| 1376 |
{
|
| 1377 |
"epoch": 0.9881847475832438,
|
| 1378 |
+
"grad_norm": 100.41632245481235,
|
| 1379 |
"learning_rate": 2.82400900618418e-10,
|
| 1380 |
+
"logits/chosen": -15.215599060058594,
|
| 1381 |
+
"logits/rejected": -15.19567584991455,
|
| 1382 |
+
"logps/chosen": -1.937787652015686,
|
| 1383 |
+
"logps/rejected": -2.593761920928955,
|
| 1384 |
+
"loss": 2.5524,
|
| 1385 |
+
"rewards/accuracies": 0.8125,
|
| 1386 |
+
"rewards/chosen": -19.377878189086914,
|
| 1387 |
+
"rewards/margins": 6.559741020202637,
|
| 1388 |
+
"rewards/rejected": -25.937618255615234,
|
| 1389 |
"step": 460
|
| 1390 |
},
|
| 1391 |
{
|
| 1392 |
"epoch": 0.9989258861439313,
|
| 1393 |
+
"grad_norm": 139.5358860215804,
|
| 1394 |
"learning_rate": 0.0,
|
| 1395 |
+
"logits/chosen": -15.462666511535645,
|
| 1396 |
+
"logits/rejected": -15.356382369995117,
|
| 1397 |
+
"logps/chosen": -2.155303716659546,
|
| 1398 |
+
"logps/rejected": -2.6153688430786133,
|
| 1399 |
+
"loss": 2.5621,
|
| 1400 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1401 |
+
"rewards/chosen": -21.55303955078125,
|
| 1402 |
+
"rewards/margins": 4.600649356842041,
|
| 1403 |
+
"rewards/rejected": -26.1536865234375,
|
| 1404 |
"step": 465
|
| 1405 |
},
|
| 1406 |
{
|
| 1407 |
"epoch": 0.9989258861439313,
|
| 1408 |
"step": 465,
|
| 1409 |
"total_flos": 0.0,
|
| 1410 |
+
"train_loss": 3.5518602760889197,
|
| 1411 |
+
"train_runtime": 6148.2375,
|
| 1412 |
+
"train_samples_per_second": 9.689,
|
| 1413 |
"train_steps_per_second": 0.076
|
| 1414 |
}
|
| 1415 |
],
|