Commit
·
f05d24e
1
Parent(s):
4ceb749
🍻 cheers
Browse files- all_results.json +9 -9
- config.json +2 -18
- eval_results.json +5 -5
- pytorch_model.bin +2 -2
- runs/Dec14_21-37-11_DESKTOP-MQJAOOG/1702564650.3208292/events.out.tfevents.1702564650.DESKTOP-MQJAOOG.17968.1 +3 -0
- runs/Dec14_21-37-11_DESKTOP-MQJAOOG/events.out.tfevents.1702564650.DESKTOP-MQJAOOG.17968.0 +3 -0
- runs/Dec14_21-39-12_DESKTOP-MQJAOOG/1702564766.38767/events.out.tfevents.1702564766.DESKTOP-MQJAOOG.17968.3 +3 -0
- runs/Dec14_21-39-12_DESKTOP-MQJAOOG/events.out.tfevents.1702564766.DESKTOP-MQJAOOG.17968.2 +3 -0
- runs/Dec14_21-39-12_DESKTOP-MQJAOOG/events.out.tfevents.1702565255.DESKTOP-MQJAOOG.17968.4 +3 -0
- train_results.json +4 -4
- trainer_state.json +13 -1285
- training_args.bin +1 -1
all_results.json
CHANGED
|
@@ -1,12 +1,12 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 4.0,
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second": 2.
|
| 7 |
-
"eval_steps_per_second": 0.
|
| 8 |
-
"train_loss": 0.
|
| 9 |
-
"train_runtime":
|
| 10 |
-
"train_samples_per_second": 0.
|
| 11 |
-
"train_steps_per_second": 0.
|
| 12 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 4.0,
|
| 3 |
+
"eval_accuracy": 0.98,
|
| 4 |
+
"eval_loss": 0.10598108917474747,
|
| 5 |
+
"eval_runtime": 37.1403,
|
| 6 |
+
"eval_samples_per_second": 2.692,
|
| 7 |
+
"eval_steps_per_second": 0.35,
|
| 8 |
+
"train_loss": 0.18094830640724727,
|
| 9 |
+
"train_runtime": 416.3142,
|
| 10 |
+
"train_samples_per_second": 0.961,
|
| 11 |
+
"train_steps_per_second": 0.067
|
| 12 |
}
|
config.json
CHANGED
|
@@ -9,30 +9,14 @@
|
|
| 9 |
"hidden_size": 768,
|
| 10 |
"id2label": {
|
| 11 |
"0": "apple_pie",
|
| 12 |
-
"1": "baby_back_ribs"
|
| 13 |
-
"2": "baklava",
|
| 14 |
-
"3": "beef_carpaccio",
|
| 15 |
-
"4": "beef_tartare",
|
| 16 |
-
"5": "beet_salad",
|
| 17 |
-
"6": "beignets",
|
| 18 |
-
"7": "bibimbap",
|
| 19 |
-
"8": "bread_pudding",
|
| 20 |
-
"9": "breakfast_burrito"
|
| 21 |
},
|
| 22 |
"image_size": 224,
|
| 23 |
"initializer_range": 0.02,
|
| 24 |
"intermediate_size": 3072,
|
| 25 |
"label2id": {
|
| 26 |
"apple_pie": "0",
|
| 27 |
-
"baby_back_ribs": "1"
|
| 28 |
-
"baklava": "2",
|
| 29 |
-
"beef_carpaccio": "3",
|
| 30 |
-
"beef_tartare": "4",
|
| 31 |
-
"beet_salad": "5",
|
| 32 |
-
"beignets": "6",
|
| 33 |
-
"bibimbap": "7",
|
| 34 |
-
"bread_pudding": "8",
|
| 35 |
-
"breakfast_burrito": "9"
|
| 36 |
},
|
| 37 |
"layer_norm_eps": 1e-12,
|
| 38 |
"model_type": "vit",
|
|
|
|
| 9 |
"hidden_size": 768,
|
| 10 |
"id2label": {
|
| 11 |
"0": "apple_pie",
|
| 12 |
+
"1": "baby_back_ribs"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
},
|
| 14 |
"image_size": 224,
|
| 15 |
"initializer_range": 0.02,
|
| 16 |
"intermediate_size": 3072,
|
| 17 |
"label2id": {
|
| 18 |
"apple_pie": "0",
|
| 19 |
+
"baby_back_ribs": "1"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
},
|
| 21 |
"layer_norm_eps": 1e-12,
|
| 22 |
"model_type": "vit",
|
eval_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 4.0,
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second": 2.
|
| 7 |
-
"eval_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 4.0,
|
| 3 |
+
"eval_accuracy": 0.98,
|
| 4 |
+
"eval_loss": 0.10598108917474747,
|
| 5 |
+
"eval_runtime": 37.1403,
|
| 6 |
+
"eval_samples_per_second": 2.692,
|
| 7 |
+
"eval_steps_per_second": 0.35
|
| 8 |
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:83ef21ce3f8713c0bbca09103dc638a2c1618560f3f0d22b7c0b5d752c1604ac
|
| 3 |
+
size 343279210
|
runs/Dec14_21-37-11_DESKTOP-MQJAOOG/1702564650.3208292/events.out.tfevents.1702564650.DESKTOP-MQJAOOG.17968.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0a4b461c47f9745057cae5d70805a5734164ab911b0e5f3a0473b22f8fdfaf7c
|
| 3 |
+
size 4560
|
runs/Dec14_21-37-11_DESKTOP-MQJAOOG/events.out.tfevents.1702564650.DESKTOP-MQJAOOG.17968.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:24fac6b07696d9595866e52ccf2ff6f5e9f9ce37492be59dd7da711a17260dd7
|
| 3 |
+
size 3185
|
runs/Dec14_21-39-12_DESKTOP-MQJAOOG/1702564766.38767/events.out.tfevents.1702564766.DESKTOP-MQJAOOG.17968.3
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eccc9a9061b21158d076db553dd31cf7f48374491575414d04899c925898f050
|
| 3 |
+
size 4560
|
runs/Dec14_21-39-12_DESKTOP-MQJAOOG/events.out.tfevents.1702564766.DESKTOP-MQJAOOG.17968.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b8e89a1a307e0e0509adb17b2fdac0874b66d489099f136a81eea672939faa78
|
| 3 |
+
size 3841
|
runs/Dec14_21-39-12_DESKTOP-MQJAOOG/events.out.tfevents.1702565255.DESKTOP-MQJAOOG.17968.4
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e3d68baf6d227e9c3f805a5968b07a4594e9aeac2096f6386f5fbb45db22232e
|
| 3 |
+
size 405
|
train_results.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 4.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples_per_second": 0.
|
| 6 |
-
"train_steps_per_second": 0.
|
| 7 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 4.0,
|
| 3 |
+
"train_loss": 0.18094830640724727,
|
| 4 |
+
"train_runtime": 416.3142,
|
| 5 |
+
"train_samples_per_second": 0.961,
|
| 6 |
+
"train_steps_per_second": 0.067
|
| 7 |
}
|
trainer_state.json
CHANGED
|
@@ -1,1307 +1,35 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric":
|
| 3 |
-
"best_model_checkpoint":
|
| 4 |
"epoch": 4.0,
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
-
{
|
| 11 |
-
"epoch": 0.02,
|
| 12 |
-
"learning_rate": 0.00019893390191897656,
|
| 13 |
-
"loss": 2.1647,
|
| 14 |
-
"step": 10
|
| 15 |
-
},
|
| 16 |
-
{
|
| 17 |
-
"epoch": 0.04,
|
| 18 |
-
"learning_rate": 0.0001978678038379531,
|
| 19 |
-
"loss": 1.7478,
|
| 20 |
-
"step": 20
|
| 21 |
-
},
|
| 22 |
-
{
|
| 23 |
-
"epoch": 0.06,
|
| 24 |
-
"learning_rate": 0.00019680170575692965,
|
| 25 |
-
"loss": 1.4997,
|
| 26 |
-
"step": 30
|
| 27 |
-
},
|
| 28 |
-
{
|
| 29 |
-
"epoch": 0.09,
|
| 30 |
-
"learning_rate": 0.0001957356076759062,
|
| 31 |
-
"loss": 1.1421,
|
| 32 |
-
"step": 40
|
| 33 |
-
},
|
| 34 |
-
{
|
| 35 |
-
"epoch": 0.11,
|
| 36 |
-
"learning_rate": 0.00019466950959488275,
|
| 37 |
-
"loss": 1.0115,
|
| 38 |
-
"step": 50
|
| 39 |
-
},
|
| 40 |
-
{
|
| 41 |
-
"epoch": 0.13,
|
| 42 |
-
"learning_rate": 0.00019360341151385927,
|
| 43 |
-
"loss": 0.8962,
|
| 44 |
-
"step": 60
|
| 45 |
-
},
|
| 46 |
-
{
|
| 47 |
-
"epoch": 0.15,
|
| 48 |
-
"learning_rate": 0.00019253731343283584,
|
| 49 |
-
"loss": 0.774,
|
| 50 |
-
"step": 70
|
| 51 |
-
},
|
| 52 |
-
{
|
| 53 |
-
"epoch": 0.17,
|
| 54 |
-
"learning_rate": 0.0001914712153518124,
|
| 55 |
-
"loss": 0.8508,
|
| 56 |
-
"step": 80
|
| 57 |
-
},
|
| 58 |
-
{
|
| 59 |
-
"epoch": 0.19,
|
| 60 |
-
"learning_rate": 0.0001904051172707889,
|
| 61 |
-
"loss": 0.8852,
|
| 62 |
-
"step": 90
|
| 63 |
-
},
|
| 64 |
-
{
|
| 65 |
-
"epoch": 0.21,
|
| 66 |
-
"learning_rate": 0.00018933901918976546,
|
| 67 |
-
"loss": 0.6346,
|
| 68 |
-
"step": 100
|
| 69 |
-
},
|
| 70 |
-
{
|
| 71 |
-
"epoch": 0.21,
|
| 72 |
-
"eval_accuracy": 0.8136,
|
| 73 |
-
"eval_loss": 0.6600381135940552,
|
| 74 |
-
"eval_runtime": 910.9195,
|
| 75 |
-
"eval_samples_per_second": 2.744,
|
| 76 |
-
"eval_steps_per_second": 0.344,
|
| 77 |
-
"step": 100
|
| 78 |
-
},
|
| 79 |
-
{
|
| 80 |
-
"epoch": 0.23,
|
| 81 |
-
"learning_rate": 0.00018827292110874203,
|
| 82 |
-
"loss": 0.7183,
|
| 83 |
-
"step": 110
|
| 84 |
-
},
|
| 85 |
-
{
|
| 86 |
-
"epoch": 0.26,
|
| 87 |
-
"learning_rate": 0.00018720682302771858,
|
| 88 |
-
"loss": 0.7806,
|
| 89 |
-
"step": 120
|
| 90 |
-
},
|
| 91 |
-
{
|
| 92 |
-
"epoch": 0.28,
|
| 93 |
-
"learning_rate": 0.0001861407249466951,
|
| 94 |
-
"loss": 0.7702,
|
| 95 |
-
"step": 130
|
| 96 |
-
},
|
| 97 |
-
{
|
| 98 |
-
"epoch": 0.3,
|
| 99 |
-
"learning_rate": 0.00018507462686567165,
|
| 100 |
-
"loss": 0.8338,
|
| 101 |
-
"step": 140
|
| 102 |
-
},
|
| 103 |
-
{
|
| 104 |
-
"epoch": 0.32,
|
| 105 |
-
"learning_rate": 0.0001840085287846482,
|
| 106 |
-
"loss": 0.732,
|
| 107 |
-
"step": 150
|
| 108 |
-
},
|
| 109 |
-
{
|
| 110 |
-
"epoch": 0.34,
|
| 111 |
-
"learning_rate": 0.00018294243070362474,
|
| 112 |
-
"loss": 0.6746,
|
| 113 |
-
"step": 160
|
| 114 |
-
},
|
| 115 |
-
{
|
| 116 |
-
"epoch": 0.36,
|
| 117 |
-
"learning_rate": 0.0001818763326226013,
|
| 118 |
-
"loss": 0.5369,
|
| 119 |
-
"step": 170
|
| 120 |
-
},
|
| 121 |
-
{
|
| 122 |
-
"epoch": 0.38,
|
| 123 |
-
"learning_rate": 0.00018081023454157784,
|
| 124 |
-
"loss": 0.6082,
|
| 125 |
-
"step": 180
|
| 126 |
-
},
|
| 127 |
-
{
|
| 128 |
-
"epoch": 0.41,
|
| 129 |
-
"learning_rate": 0.00017974413646055436,
|
| 130 |
-
"loss": 0.5756,
|
| 131 |
-
"step": 190
|
| 132 |
-
},
|
| 133 |
-
{
|
| 134 |
-
"epoch": 0.43,
|
| 135 |
-
"learning_rate": 0.00017867803837953093,
|
| 136 |
-
"loss": 0.4988,
|
| 137 |
-
"step": 200
|
| 138 |
-
},
|
| 139 |
-
{
|
| 140 |
-
"epoch": 0.43,
|
| 141 |
-
"eval_accuracy": 0.8212,
|
| 142 |
-
"eval_loss": 0.5744591951370239,
|
| 143 |
-
"eval_runtime": 901.9851,
|
| 144 |
-
"eval_samples_per_second": 2.772,
|
| 145 |
-
"eval_steps_per_second": 0.347,
|
| 146 |
-
"step": 200
|
| 147 |
-
},
|
| 148 |
-
{
|
| 149 |
-
"epoch": 0.45,
|
| 150 |
-
"learning_rate": 0.00017761194029850748,
|
| 151 |
-
"loss": 0.672,
|
| 152 |
-
"step": 210
|
| 153 |
-
},
|
| 154 |
-
{
|
| 155 |
-
"epoch": 0.47,
|
| 156 |
-
"learning_rate": 0.00017654584221748402,
|
| 157 |
-
"loss": 0.6575,
|
| 158 |
-
"step": 220
|
| 159 |
-
},
|
| 160 |
-
{
|
| 161 |
-
"epoch": 0.49,
|
| 162 |
-
"learning_rate": 0.00017547974413646055,
|
| 163 |
-
"loss": 0.5028,
|
| 164 |
-
"step": 230
|
| 165 |
-
},
|
| 166 |
-
{
|
| 167 |
-
"epoch": 0.51,
|
| 168 |
-
"learning_rate": 0.00017441364605543712,
|
| 169 |
-
"loss": 0.5451,
|
| 170 |
-
"step": 240
|
| 171 |
-
},
|
| 172 |
-
{
|
| 173 |
-
"epoch": 0.53,
|
| 174 |
-
"learning_rate": 0.00017334754797441367,
|
| 175 |
-
"loss": 0.6499,
|
| 176 |
-
"step": 250
|
| 177 |
-
},
|
| 178 |
-
{
|
| 179 |
-
"epoch": 0.55,
|
| 180 |
-
"learning_rate": 0.0001722814498933902,
|
| 181 |
-
"loss": 0.7127,
|
| 182 |
-
"step": 260
|
| 183 |
-
},
|
| 184 |
-
{
|
| 185 |
-
"epoch": 0.58,
|
| 186 |
-
"learning_rate": 0.00017121535181236673,
|
| 187 |
-
"loss": 0.538,
|
| 188 |
-
"step": 270
|
| 189 |
-
},
|
| 190 |
-
{
|
| 191 |
-
"epoch": 0.6,
|
| 192 |
-
"learning_rate": 0.00017014925373134328,
|
| 193 |
-
"loss": 0.6385,
|
| 194 |
-
"step": 280
|
| 195 |
-
},
|
| 196 |
-
{
|
| 197 |
-
"epoch": 0.62,
|
| 198 |
-
"learning_rate": 0.00016908315565031986,
|
| 199 |
-
"loss": 0.4637,
|
| 200 |
-
"step": 290
|
| 201 |
-
},
|
| 202 |
-
{
|
| 203 |
-
"epoch": 0.64,
|
| 204 |
-
"learning_rate": 0.00016801705756929638,
|
| 205 |
-
"loss": 0.4619,
|
| 206 |
-
"step": 300
|
| 207 |
-
},
|
| 208 |
-
{
|
| 209 |
-
"epoch": 0.64,
|
| 210 |
-
"eval_accuracy": 0.8852,
|
| 211 |
-
"eval_loss": 0.3846997916698456,
|
| 212 |
-
"eval_runtime": 886.5096,
|
| 213 |
-
"eval_samples_per_second": 2.82,
|
| 214 |
-
"eval_steps_per_second": 0.353,
|
| 215 |
-
"step": 300
|
| 216 |
-
},
|
| 217 |
-
{
|
| 218 |
-
"epoch": 0.66,
|
| 219 |
-
"learning_rate": 0.00016695095948827292,
|
| 220 |
-
"loss": 0.6006,
|
| 221 |
-
"step": 310
|
| 222 |
-
},
|
| 223 |
-
{
|
| 224 |
-
"epoch": 0.68,
|
| 225 |
-
"learning_rate": 0.00016588486140724947,
|
| 226 |
-
"loss": 0.4678,
|
| 227 |
-
"step": 320
|
| 228 |
-
},
|
| 229 |
-
{
|
| 230 |
-
"epoch": 0.7,
|
| 231 |
-
"learning_rate": 0.00016481876332622602,
|
| 232 |
-
"loss": 0.5459,
|
| 233 |
-
"step": 330
|
| 234 |
-
},
|
| 235 |
-
{
|
| 236 |
-
"epoch": 0.72,
|
| 237 |
-
"learning_rate": 0.00016375266524520257,
|
| 238 |
-
"loss": 0.4216,
|
| 239 |
-
"step": 340
|
| 240 |
-
},
|
| 241 |
-
{
|
| 242 |
-
"epoch": 0.75,
|
| 243 |
-
"learning_rate": 0.00016268656716417911,
|
| 244 |
-
"loss": 0.6596,
|
| 245 |
-
"step": 350
|
| 246 |
-
},
|
| 247 |
-
{
|
| 248 |
-
"epoch": 0.77,
|
| 249 |
-
"learning_rate": 0.00016162046908315566,
|
| 250 |
-
"loss": 0.6491,
|
| 251 |
-
"step": 360
|
| 252 |
-
},
|
| 253 |
-
{
|
| 254 |
-
"epoch": 0.79,
|
| 255 |
-
"learning_rate": 0.0001605543710021322,
|
| 256 |
-
"loss": 0.5269,
|
| 257 |
-
"step": 370
|
| 258 |
-
},
|
| 259 |
-
{
|
| 260 |
-
"epoch": 0.81,
|
| 261 |
-
"learning_rate": 0.00015948827292110876,
|
| 262 |
-
"loss": 0.441,
|
| 263 |
-
"step": 380
|
| 264 |
-
},
|
| 265 |
-
{
|
| 266 |
-
"epoch": 0.83,
|
| 267 |
-
"learning_rate": 0.0001584221748400853,
|
| 268 |
-
"loss": 0.4428,
|
| 269 |
-
"step": 390
|
| 270 |
-
},
|
| 271 |
-
{
|
| 272 |
-
"epoch": 0.85,
|
| 273 |
-
"learning_rate": 0.00015735607675906182,
|
| 274 |
-
"loss": 0.4205,
|
| 275 |
-
"step": 400
|
| 276 |
-
},
|
| 277 |
-
{
|
| 278 |
-
"epoch": 0.85,
|
| 279 |
-
"eval_accuracy": 0.864,
|
| 280 |
-
"eval_loss": 0.43827134370803833,
|
| 281 |
-
"eval_runtime": 905.4707,
|
| 282 |
-
"eval_samples_per_second": 2.761,
|
| 283 |
-
"eval_steps_per_second": 0.346,
|
| 284 |
-
"step": 400
|
| 285 |
-
},
|
| 286 |
-
{
|
| 287 |
-
"epoch": 0.87,
|
| 288 |
-
"learning_rate": 0.0001562899786780384,
|
| 289 |
-
"loss": 0.6422,
|
| 290 |
-
"step": 410
|
| 291 |
-
},
|
| 292 |
-
{
|
| 293 |
-
"epoch": 0.9,
|
| 294 |
-
"learning_rate": 0.00015522388059701495,
|
| 295 |
-
"loss": 0.4558,
|
| 296 |
-
"step": 420
|
| 297 |
-
},
|
| 298 |
-
{
|
| 299 |
-
"epoch": 0.92,
|
| 300 |
-
"learning_rate": 0.0001541577825159915,
|
| 301 |
-
"loss": 0.4575,
|
| 302 |
-
"step": 430
|
| 303 |
-
},
|
| 304 |
-
{
|
| 305 |
-
"epoch": 0.94,
|
| 306 |
-
"learning_rate": 0.000153091684434968,
|
| 307 |
-
"loss": 0.4454,
|
| 308 |
-
"step": 440
|
| 309 |
-
},
|
| 310 |
-
{
|
| 311 |
-
"epoch": 0.96,
|
| 312 |
-
"learning_rate": 0.00015202558635394456,
|
| 313 |
-
"loss": 0.4917,
|
| 314 |
-
"step": 450
|
| 315 |
-
},
|
| 316 |
-
{
|
| 317 |
-
"epoch": 0.98,
|
| 318 |
-
"learning_rate": 0.00015095948827292113,
|
| 319 |
-
"loss": 0.5301,
|
| 320 |
-
"step": 460
|
| 321 |
-
},
|
| 322 |
-
{
|
| 323 |
-
"epoch": 1.0,
|
| 324 |
-
"learning_rate": 0.00014989339019189766,
|
| 325 |
-
"loss": 0.3294,
|
| 326 |
-
"step": 470
|
| 327 |
-
},
|
| 328 |
-
{
|
| 329 |
-
"epoch": 1.02,
|
| 330 |
-
"learning_rate": 0.0001488272921108742,
|
| 331 |
-
"loss": 0.2967,
|
| 332 |
-
"step": 480
|
| 333 |
-
},
|
| 334 |
-
{
|
| 335 |
-
"epoch": 1.04,
|
| 336 |
-
"learning_rate": 0.00014776119402985075,
|
| 337 |
-
"loss": 0.3406,
|
| 338 |
-
"step": 490
|
| 339 |
-
},
|
| 340 |
-
{
|
| 341 |
-
"epoch": 1.07,
|
| 342 |
-
"learning_rate": 0.0001466950959488273,
|
| 343 |
-
"loss": 0.287,
|
| 344 |
-
"step": 500
|
| 345 |
-
},
|
| 346 |
-
{
|
| 347 |
-
"epoch": 1.07,
|
| 348 |
-
"eval_accuracy": 0.8992,
|
| 349 |
-
"eval_loss": 0.3214336931705475,
|
| 350 |
-
"eval_runtime": 905.5537,
|
| 351 |
-
"eval_samples_per_second": 2.761,
|
| 352 |
-
"eval_steps_per_second": 0.346,
|
| 353 |
-
"step": 500
|
| 354 |
-
},
|
| 355 |
-
{
|
| 356 |
-
"epoch": 1.09,
|
| 357 |
-
"learning_rate": 0.00014562899786780384,
|
| 358 |
-
"loss": 0.3278,
|
| 359 |
-
"step": 510
|
| 360 |
-
},
|
| 361 |
-
{
|
| 362 |
-
"epoch": 1.11,
|
| 363 |
-
"learning_rate": 0.0001445628997867804,
|
| 364 |
-
"loss": 0.2406,
|
| 365 |
-
"step": 520
|
| 366 |
-
},
|
| 367 |
-
{
|
| 368 |
-
"epoch": 1.13,
|
| 369 |
-
"learning_rate": 0.00014349680170575694,
|
| 370 |
-
"loss": 0.2839,
|
| 371 |
-
"step": 530
|
| 372 |
-
},
|
| 373 |
-
{
|
| 374 |
-
"epoch": 1.15,
|
| 375 |
-
"learning_rate": 0.0001424307036247335,
|
| 376 |
-
"loss": 0.2843,
|
| 377 |
-
"step": 540
|
| 378 |
-
},
|
| 379 |
-
{
|
| 380 |
-
"epoch": 1.17,
|
| 381 |
-
"learning_rate": 0.00014136460554371003,
|
| 382 |
-
"loss": 0.2085,
|
| 383 |
-
"step": 550
|
| 384 |
-
},
|
| 385 |
-
{
|
| 386 |
-
"epoch": 1.19,
|
| 387 |
-
"learning_rate": 0.00014029850746268658,
|
| 388 |
-
"loss": 0.2465,
|
| 389 |
-
"step": 560
|
| 390 |
-
},
|
| 391 |
-
{
|
| 392 |
-
"epoch": 1.22,
|
| 393 |
-
"learning_rate": 0.0001392324093816631,
|
| 394 |
-
"loss": 0.2444,
|
| 395 |
-
"step": 570
|
| 396 |
-
},
|
| 397 |
-
{
|
| 398 |
-
"epoch": 1.24,
|
| 399 |
-
"learning_rate": 0.00013816631130063965,
|
| 400 |
-
"loss": 0.3261,
|
| 401 |
-
"step": 580
|
| 402 |
-
},
|
| 403 |
-
{
|
| 404 |
-
"epoch": 1.26,
|
| 405 |
-
"learning_rate": 0.00013710021321961622,
|
| 406 |
-
"loss": 0.1768,
|
| 407 |
-
"step": 590
|
| 408 |
-
},
|
| 409 |
-
{
|
| 410 |
-
"epoch": 1.28,
|
| 411 |
-
"learning_rate": 0.00013603411513859277,
|
| 412 |
-
"loss": 0.2963,
|
| 413 |
-
"step": 600
|
| 414 |
-
},
|
| 415 |
-
{
|
| 416 |
-
"epoch": 1.28,
|
| 417 |
-
"eval_accuracy": 0.8796,
|
| 418 |
-
"eval_loss": 0.35504186153411865,
|
| 419 |
-
"eval_runtime": 901.3195,
|
| 420 |
-
"eval_samples_per_second": 2.774,
|
| 421 |
-
"eval_steps_per_second": 0.347,
|
| 422 |
-
"step": 600
|
| 423 |
-
},
|
| 424 |
-
{
|
| 425 |
-
"epoch": 1.3,
|
| 426 |
-
"learning_rate": 0.0001349680170575693,
|
| 427 |
-
"loss": 0.3221,
|
| 428 |
-
"step": 610
|
| 429 |
-
},
|
| 430 |
-
{
|
| 431 |
-
"epoch": 1.32,
|
| 432 |
-
"learning_rate": 0.00013390191897654584,
|
| 433 |
-
"loss": 0.3559,
|
| 434 |
-
"step": 620
|
| 435 |
-
},
|
| 436 |
-
{
|
| 437 |
-
"epoch": 1.34,
|
| 438 |
-
"learning_rate": 0.0001328358208955224,
|
| 439 |
-
"loss": 0.2589,
|
| 440 |
-
"step": 630
|
| 441 |
-
},
|
| 442 |
-
{
|
| 443 |
-
"epoch": 1.36,
|
| 444 |
-
"learning_rate": 0.00013176972281449893,
|
| 445 |
-
"loss": 0.317,
|
| 446 |
-
"step": 640
|
| 447 |
-
},
|
| 448 |
-
{
|
| 449 |
-
"epoch": 1.39,
|
| 450 |
-
"learning_rate": 0.00013070362473347548,
|
| 451 |
-
"loss": 0.2702,
|
| 452 |
-
"step": 650
|
| 453 |
-
},
|
| 454 |
-
{
|
| 455 |
-
"epoch": 1.41,
|
| 456 |
-
"learning_rate": 0.00012963752665245203,
|
| 457 |
-
"loss": 0.2416,
|
| 458 |
-
"step": 660
|
| 459 |
-
},
|
| 460 |
{
|
| 461 |
"epoch": 1.43,
|
| 462 |
"learning_rate": 0.00012857142857142858,
|
| 463 |
-
"loss": 0.
|
| 464 |
-
"step":
|
| 465 |
-
},
|
| 466 |
-
{
|
| 467 |
-
"epoch": 1.45,
|
| 468 |
-
"learning_rate": 0.00012750533049040512,
|
| 469 |
-
"loss": 0.1777,
|
| 470 |
-
"step": 680
|
| 471 |
-
},
|
| 472 |
-
{
|
| 473 |
-
"epoch": 1.47,
|
| 474 |
-
"learning_rate": 0.00012643923240938167,
|
| 475 |
-
"loss": 0.3028,
|
| 476 |
-
"step": 690
|
| 477 |
-
},
|
| 478 |
-
{
|
| 479 |
-
"epoch": 1.49,
|
| 480 |
-
"learning_rate": 0.00012537313432835822,
|
| 481 |
-
"loss": 0.2908,
|
| 482 |
-
"step": 700
|
| 483 |
-
},
|
| 484 |
-
{
|
| 485 |
-
"epoch": 1.49,
|
| 486 |
-
"eval_accuracy": 0.9064,
|
| 487 |
-
"eval_loss": 0.3182031810283661,
|
| 488 |
-
"eval_runtime": 904.1865,
|
| 489 |
-
"eval_samples_per_second": 2.765,
|
| 490 |
-
"eval_steps_per_second": 0.346,
|
| 491 |
-
"step": 700
|
| 492 |
-
},
|
| 493 |
-
{
|
| 494 |
-
"epoch": 1.51,
|
| 495 |
-
"learning_rate": 0.00012430703624733474,
|
| 496 |
-
"loss": 0.236,
|
| 497 |
-
"step": 710
|
| 498 |
-
},
|
| 499 |
-
{
|
| 500 |
-
"epoch": 1.54,
|
| 501 |
-
"learning_rate": 0.0001232409381663113,
|
| 502 |
-
"loss": 0.1903,
|
| 503 |
-
"step": 720
|
| 504 |
-
},
|
| 505 |
-
{
|
| 506 |
-
"epoch": 1.56,
|
| 507 |
-
"learning_rate": 0.00012217484008528786,
|
| 508 |
-
"loss": 0.2509,
|
| 509 |
-
"step": 730
|
| 510 |
-
},
|
| 511 |
-
{
|
| 512 |
-
"epoch": 1.58,
|
| 513 |
-
"learning_rate": 0.00012110874200426441,
|
| 514 |
-
"loss": 0.3468,
|
| 515 |
-
"step": 740
|
| 516 |
-
},
|
| 517 |
-
{
|
| 518 |
-
"epoch": 1.6,
|
| 519 |
-
"learning_rate": 0.00012004264392324094,
|
| 520 |
-
"loss": 0.4192,
|
| 521 |
-
"step": 750
|
| 522 |
-
},
|
| 523 |
-
{
|
| 524 |
-
"epoch": 1.62,
|
| 525 |
-
"learning_rate": 0.00011897654584221749,
|
| 526 |
-
"loss": 0.2604,
|
| 527 |
-
"step": 760
|
| 528 |
-
},
|
| 529 |
-
{
|
| 530 |
-
"epoch": 1.64,
|
| 531 |
-
"learning_rate": 0.00011791044776119405,
|
| 532 |
-
"loss": 0.2523,
|
| 533 |
-
"step": 770
|
| 534 |
-
},
|
| 535 |
-
{
|
| 536 |
-
"epoch": 1.66,
|
| 537 |
-
"learning_rate": 0.00011684434968017057,
|
| 538 |
-
"loss": 0.169,
|
| 539 |
-
"step": 780
|
| 540 |
-
},
|
| 541 |
-
{
|
| 542 |
-
"epoch": 1.68,
|
| 543 |
-
"learning_rate": 0.00011577825159914713,
|
| 544 |
-
"loss": 0.3698,
|
| 545 |
-
"step": 790
|
| 546 |
-
},
|
| 547 |
-
{
|
| 548 |
-
"epoch": 1.71,
|
| 549 |
-
"learning_rate": 0.00011471215351812368,
|
| 550 |
-
"loss": 0.3791,
|
| 551 |
-
"step": 800
|
| 552 |
-
},
|
| 553 |
-
{
|
| 554 |
-
"epoch": 1.71,
|
| 555 |
-
"eval_accuracy": 0.9012,
|
| 556 |
-
"eval_loss": 0.30747249722480774,
|
| 557 |
-
"eval_runtime": 905.3757,
|
| 558 |
-
"eval_samples_per_second": 2.761,
|
| 559 |
-
"eval_steps_per_second": 0.346,
|
| 560 |
-
"step": 800
|
| 561 |
-
},
|
| 562 |
-
{
|
| 563 |
-
"epoch": 1.73,
|
| 564 |
-
"learning_rate": 0.00011364605543710023,
|
| 565 |
-
"loss": 0.4315,
|
| 566 |
-
"step": 810
|
| 567 |
-
},
|
| 568 |
-
{
|
| 569 |
-
"epoch": 1.75,
|
| 570 |
-
"learning_rate": 0.00011257995735607676,
|
| 571 |
-
"loss": 0.2439,
|
| 572 |
-
"step": 820
|
| 573 |
-
},
|
| 574 |
-
{
|
| 575 |
-
"epoch": 1.77,
|
| 576 |
-
"learning_rate": 0.0001115138592750533,
|
| 577 |
-
"loss": 0.1582,
|
| 578 |
-
"step": 830
|
| 579 |
-
},
|
| 580 |
-
{
|
| 581 |
-
"epoch": 1.79,
|
| 582 |
-
"learning_rate": 0.00011044776119402987,
|
| 583 |
-
"loss": 0.1514,
|
| 584 |
-
"step": 840
|
| 585 |
-
},
|
| 586 |
-
{
|
| 587 |
-
"epoch": 1.81,
|
| 588 |
-
"learning_rate": 0.00010938166311300639,
|
| 589 |
-
"loss": 0.302,
|
| 590 |
-
"step": 850
|
| 591 |
-
},
|
| 592 |
-
{
|
| 593 |
-
"epoch": 1.83,
|
| 594 |
-
"learning_rate": 0.00010831556503198295,
|
| 595 |
-
"loss": 0.1793,
|
| 596 |
-
"step": 860
|
| 597 |
-
},
|
| 598 |
-
{
|
| 599 |
-
"epoch": 1.86,
|
| 600 |
-
"learning_rate": 0.0001072494669509595,
|
| 601 |
-
"loss": 0.2064,
|
| 602 |
-
"step": 870
|
| 603 |
-
},
|
| 604 |
-
{
|
| 605 |
-
"epoch": 1.88,
|
| 606 |
-
"learning_rate": 0.00010618336886993603,
|
| 607 |
-
"loss": 0.1789,
|
| 608 |
-
"step": 880
|
| 609 |
-
},
|
| 610 |
-
{
|
| 611 |
-
"epoch": 1.9,
|
| 612 |
-
"learning_rate": 0.00010511727078891258,
|
| 613 |
-
"loss": 0.2323,
|
| 614 |
-
"step": 890
|
| 615 |
-
},
|
| 616 |
-
{
|
| 617 |
-
"epoch": 1.92,
|
| 618 |
-
"learning_rate": 0.00010405117270788914,
|
| 619 |
-
"loss": 0.1399,
|
| 620 |
-
"step": 900
|
| 621 |
-
},
|
| 622 |
-
{
|
| 623 |
-
"epoch": 1.92,
|
| 624 |
-
"eval_accuracy": 0.9108,
|
| 625 |
-
"eval_loss": 0.31127122044563293,
|
| 626 |
-
"eval_runtime": 903.7995,
|
| 627 |
-
"eval_samples_per_second": 2.766,
|
| 628 |
-
"eval_steps_per_second": 0.346,
|
| 629 |
-
"step": 900
|
| 630 |
-
},
|
| 631 |
-
{
|
| 632 |
-
"epoch": 1.94,
|
| 633 |
-
"learning_rate": 0.00010298507462686569,
|
| 634 |
-
"loss": 0.2957,
|
| 635 |
-
"step": 910
|
| 636 |
-
},
|
| 637 |
-
{
|
| 638 |
-
"epoch": 1.96,
|
| 639 |
-
"learning_rate": 0.00010191897654584222,
|
| 640 |
-
"loss": 0.2673,
|
| 641 |
-
"step": 920
|
| 642 |
-
},
|
| 643 |
-
{
|
| 644 |
-
"epoch": 1.98,
|
| 645 |
-
"learning_rate": 0.00010085287846481877,
|
| 646 |
-
"loss": 0.2396,
|
| 647 |
-
"step": 930
|
| 648 |
-
},
|
| 649 |
-
{
|
| 650 |
-
"epoch": 2.0,
|
| 651 |
-
"learning_rate": 9.978678038379531e-05,
|
| 652 |
-
"loss": 0.1854,
|
| 653 |
-
"step": 940
|
| 654 |
-
},
|
| 655 |
-
{
|
| 656 |
-
"epoch": 2.03,
|
| 657 |
-
"learning_rate": 9.872068230277186e-05,
|
| 658 |
-
"loss": 0.0911,
|
| 659 |
-
"step": 950
|
| 660 |
-
},
|
| 661 |
-
{
|
| 662 |
-
"epoch": 2.05,
|
| 663 |
-
"learning_rate": 9.76545842217484e-05,
|
| 664 |
-
"loss": 0.0954,
|
| 665 |
-
"step": 960
|
| 666 |
-
},
|
| 667 |
-
{
|
| 668 |
-
"epoch": 2.07,
|
| 669 |
-
"learning_rate": 9.658848614072496e-05,
|
| 670 |
-
"loss": 0.0634,
|
| 671 |
-
"step": 970
|
| 672 |
-
},
|
| 673 |
-
{
|
| 674 |
-
"epoch": 2.09,
|
| 675 |
-
"learning_rate": 9.552238805970149e-05,
|
| 676 |
-
"loss": 0.1001,
|
| 677 |
-
"step": 980
|
| 678 |
-
},
|
| 679 |
-
{
|
| 680 |
-
"epoch": 2.11,
|
| 681 |
-
"learning_rate": 9.445628997867805e-05,
|
| 682 |
-
"loss": 0.076,
|
| 683 |
-
"step": 990
|
| 684 |
-
},
|
| 685 |
-
{
|
| 686 |
-
"epoch": 2.13,
|
| 687 |
-
"learning_rate": 9.339019189765458e-05,
|
| 688 |
-
"loss": 0.0501,
|
| 689 |
-
"step": 1000
|
| 690 |
-
},
|
| 691 |
-
{
|
| 692 |
-
"epoch": 2.13,
|
| 693 |
-
"eval_accuracy": 0.9004,
|
| 694 |
-
"eval_loss": 0.33236250281333923,
|
| 695 |
-
"eval_runtime": 901.7876,
|
| 696 |
-
"eval_samples_per_second": 2.772,
|
| 697 |
-
"eval_steps_per_second": 0.347,
|
| 698 |
-
"step": 1000
|
| 699 |
-
},
|
| 700 |
-
{
|
| 701 |
-
"epoch": 2.15,
|
| 702 |
-
"learning_rate": 9.232409381663113e-05,
|
| 703 |
-
"loss": 0.1246,
|
| 704 |
-
"step": 1010
|
| 705 |
-
},
|
| 706 |
-
{
|
| 707 |
-
"epoch": 2.17,
|
| 708 |
-
"learning_rate": 9.125799573560768e-05,
|
| 709 |
-
"loss": 0.1432,
|
| 710 |
-
"step": 1020
|
| 711 |
-
},
|
| 712 |
-
{
|
| 713 |
-
"epoch": 2.2,
|
| 714 |
-
"learning_rate": 9.019189765458423e-05,
|
| 715 |
-
"loss": 0.1998,
|
| 716 |
-
"step": 1030
|
| 717 |
-
},
|
| 718 |
-
{
|
| 719 |
-
"epoch": 2.22,
|
| 720 |
-
"learning_rate": 8.912579957356077e-05,
|
| 721 |
-
"loss": 0.1226,
|
| 722 |
-
"step": 1040
|
| 723 |
-
},
|
| 724 |
-
{
|
| 725 |
-
"epoch": 2.24,
|
| 726 |
-
"learning_rate": 8.805970149253732e-05,
|
| 727 |
-
"loss": 0.0544,
|
| 728 |
-
"step": 1050
|
| 729 |
-
},
|
| 730 |
-
{
|
| 731 |
-
"epoch": 2.26,
|
| 732 |
-
"learning_rate": 8.699360341151387e-05,
|
| 733 |
-
"loss": 0.0842,
|
| 734 |
-
"step": 1060
|
| 735 |
-
},
|
| 736 |
-
{
|
| 737 |
-
"epoch": 2.28,
|
| 738 |
-
"learning_rate": 8.59275053304904e-05,
|
| 739 |
-
"loss": 0.063,
|
| 740 |
-
"step": 1070
|
| 741 |
-
},
|
| 742 |
-
{
|
| 743 |
-
"epoch": 2.3,
|
| 744 |
-
"learning_rate": 8.486140724946695e-05,
|
| 745 |
-
"loss": 0.0682,
|
| 746 |
-
"step": 1080
|
| 747 |
-
},
|
| 748 |
-
{
|
| 749 |
-
"epoch": 2.32,
|
| 750 |
-
"learning_rate": 8.37953091684435e-05,
|
| 751 |
-
"loss": 0.0897,
|
| 752 |
-
"step": 1090
|
| 753 |
-
},
|
| 754 |
-
{
|
| 755 |
-
"epoch": 2.35,
|
| 756 |
-
"learning_rate": 8.272921108742005e-05,
|
| 757 |
-
"loss": 0.0873,
|
| 758 |
-
"step": 1100
|
| 759 |
-
},
|
| 760 |
-
{
|
| 761 |
-
"epoch": 2.35,
|
| 762 |
-
"eval_accuracy": 0.9196,
|
| 763 |
-
"eval_loss": 0.28453874588012695,
|
| 764 |
-
"eval_runtime": 903.3406,
|
| 765 |
-
"eval_samples_per_second": 2.768,
|
| 766 |
-
"eval_steps_per_second": 0.346,
|
| 767 |
-
"step": 1100
|
| 768 |
-
},
|
| 769 |
-
{
|
| 770 |
-
"epoch": 2.37,
|
| 771 |
-
"learning_rate": 8.166311300639659e-05,
|
| 772 |
-
"loss": 0.076,
|
| 773 |
-
"step": 1110
|
| 774 |
-
},
|
| 775 |
-
{
|
| 776 |
-
"epoch": 2.39,
|
| 777 |
-
"learning_rate": 8.059701492537314e-05,
|
| 778 |
-
"loss": 0.0576,
|
| 779 |
-
"step": 1120
|
| 780 |
-
},
|
| 781 |
-
{
|
| 782 |
-
"epoch": 2.41,
|
| 783 |
-
"learning_rate": 7.953091684434967e-05,
|
| 784 |
-
"loss": 0.0622,
|
| 785 |
-
"step": 1130
|
| 786 |
-
},
|
| 787 |
-
{
|
| 788 |
-
"epoch": 2.43,
|
| 789 |
-
"learning_rate": 7.846481876332623e-05,
|
| 790 |
-
"loss": 0.0799,
|
| 791 |
-
"step": 1140
|
| 792 |
-
},
|
| 793 |
-
{
|
| 794 |
-
"epoch": 2.45,
|
| 795 |
-
"learning_rate": 7.739872068230277e-05,
|
| 796 |
-
"loss": 0.1253,
|
| 797 |
-
"step": 1150
|
| 798 |
-
},
|
| 799 |
-
{
|
| 800 |
-
"epoch": 2.47,
|
| 801 |
-
"learning_rate": 7.633262260127933e-05,
|
| 802 |
-
"loss": 0.0698,
|
| 803 |
-
"step": 1160
|
| 804 |
-
},
|
| 805 |
-
{
|
| 806 |
-
"epoch": 2.49,
|
| 807 |
-
"learning_rate": 7.526652452025586e-05,
|
| 808 |
-
"loss": 0.1532,
|
| 809 |
-
"step": 1170
|
| 810 |
-
},
|
| 811 |
-
{
|
| 812 |
-
"epoch": 2.52,
|
| 813 |
-
"learning_rate": 7.420042643923242e-05,
|
| 814 |
-
"loss": 0.1728,
|
| 815 |
-
"step": 1180
|
| 816 |
-
},
|
| 817 |
-
{
|
| 818 |
-
"epoch": 2.54,
|
| 819 |
-
"learning_rate": 7.313432835820896e-05,
|
| 820 |
-
"loss": 0.0871,
|
| 821 |
-
"step": 1190
|
| 822 |
-
},
|
| 823 |
-
{
|
| 824 |
-
"epoch": 2.56,
|
| 825 |
-
"learning_rate": 7.20682302771855e-05,
|
| 826 |
-
"loss": 0.0809,
|
| 827 |
-
"step": 1200
|
| 828 |
-
},
|
| 829 |
-
{
|
| 830 |
-
"epoch": 2.56,
|
| 831 |
-
"eval_accuracy": 0.9104,
|
| 832 |
-
"eval_loss": 0.3186958134174347,
|
| 833 |
-
"eval_runtime": 903.5081,
|
| 834 |
-
"eval_samples_per_second": 2.767,
|
| 835 |
-
"eval_steps_per_second": 0.346,
|
| 836 |
-
"step": 1200
|
| 837 |
-
},
|
| 838 |
-
{
|
| 839 |
-
"epoch": 2.58,
|
| 840 |
-
"learning_rate": 7.100213219616205e-05,
|
| 841 |
-
"loss": 0.1217,
|
| 842 |
-
"step": 1210
|
| 843 |
-
},
|
| 844 |
-
{
|
| 845 |
-
"epoch": 2.6,
|
| 846 |
-
"learning_rate": 6.993603411513859e-05,
|
| 847 |
-
"loss": 0.0469,
|
| 848 |
-
"step": 1220
|
| 849 |
-
},
|
| 850 |
-
{
|
| 851 |
-
"epoch": 2.62,
|
| 852 |
-
"learning_rate": 6.886993603411515e-05,
|
| 853 |
-
"loss": 0.0659,
|
| 854 |
-
"step": 1230
|
| 855 |
-
},
|
| 856 |
-
{
|
| 857 |
-
"epoch": 2.64,
|
| 858 |
-
"learning_rate": 6.780383795309168e-05,
|
| 859 |
-
"loss": 0.1019,
|
| 860 |
-
"step": 1240
|
| 861 |
-
},
|
| 862 |
-
{
|
| 863 |
-
"epoch": 2.67,
|
| 864 |
-
"learning_rate": 6.673773987206824e-05,
|
| 865 |
-
"loss": 0.046,
|
| 866 |
-
"step": 1250
|
| 867 |
-
},
|
| 868 |
-
{
|
| 869 |
-
"epoch": 2.69,
|
| 870 |
-
"learning_rate": 6.567164179104478e-05,
|
| 871 |
-
"loss": 0.0203,
|
| 872 |
-
"step": 1260
|
| 873 |
-
},
|
| 874 |
-
{
|
| 875 |
-
"epoch": 2.71,
|
| 876 |
-
"learning_rate": 6.460554371002132e-05,
|
| 877 |
-
"loss": 0.0483,
|
| 878 |
-
"step": 1270
|
| 879 |
-
},
|
| 880 |
-
{
|
| 881 |
-
"epoch": 2.73,
|
| 882 |
-
"learning_rate": 6.353944562899787e-05,
|
| 883 |
-
"loss": 0.0785,
|
| 884 |
-
"step": 1280
|
| 885 |
-
},
|
| 886 |
-
{
|
| 887 |
-
"epoch": 2.75,
|
| 888 |
-
"learning_rate": 6.247334754797442e-05,
|
| 889 |
-
"loss": 0.0358,
|
| 890 |
-
"step": 1290
|
| 891 |
-
},
|
| 892 |
-
{
|
| 893 |
-
"epoch": 2.77,
|
| 894 |
-
"learning_rate": 6.140724946695097e-05,
|
| 895 |
-
"loss": 0.0696,
|
| 896 |
-
"step": 1300
|
| 897 |
-
},
|
| 898 |
-
{
|
| 899 |
-
"epoch": 2.77,
|
| 900 |
-
"eval_accuracy": 0.9184,
|
| 901 |
-
"eval_loss": 0.3137047588825226,
|
| 902 |
-
"eval_runtime": 902.4586,
|
| 903 |
-
"eval_samples_per_second": 2.77,
|
| 904 |
-
"eval_steps_per_second": 0.347,
|
| 905 |
-
"step": 1300
|
| 906 |
-
},
|
| 907 |
-
{
|
| 908 |
-
"epoch": 2.79,
|
| 909 |
-
"learning_rate": 6.0341151385927506e-05,
|
| 910 |
-
"loss": 0.0716,
|
| 911 |
-
"step": 1310
|
| 912 |
-
},
|
| 913 |
-
{
|
| 914 |
-
"epoch": 2.81,
|
| 915 |
-
"learning_rate": 5.927505330490405e-05,
|
| 916 |
-
"loss": 0.0409,
|
| 917 |
-
"step": 1320
|
| 918 |
-
},
|
| 919 |
-
{
|
| 920 |
-
"epoch": 2.84,
|
| 921 |
-
"learning_rate": 5.82089552238806e-05,
|
| 922 |
-
"loss": 0.0595,
|
| 923 |
-
"step": 1330
|
| 924 |
},
|
| 925 |
{
|
| 926 |
"epoch": 2.86,
|
| 927 |
"learning_rate": 5.714285714285714e-05,
|
| 928 |
-
"loss": 0.
|
| 929 |
-
"step":
|
| 930 |
-
},
|
| 931 |
-
{
|
| 932 |
-
"epoch": 2.88,
|
| 933 |
-
"learning_rate": 5.6076759061833696e-05,
|
| 934 |
-
"loss": 0.1064,
|
| 935 |
-
"step": 1350
|
| 936 |
-
},
|
| 937 |
-
{
|
| 938 |
-
"epoch": 2.9,
|
| 939 |
-
"learning_rate": 5.5010660980810236e-05,
|
| 940 |
-
"loss": 0.0783,
|
| 941 |
-
"step": 1360
|
| 942 |
-
},
|
| 943 |
-
{
|
| 944 |
-
"epoch": 2.92,
|
| 945 |
-
"learning_rate": 5.394456289978679e-05,
|
| 946 |
-
"loss": 0.0889,
|
| 947 |
-
"step": 1370
|
| 948 |
-
},
|
| 949 |
-
{
|
| 950 |
-
"epoch": 2.94,
|
| 951 |
-
"learning_rate": 5.287846481876333e-05,
|
| 952 |
-
"loss": 0.1406,
|
| 953 |
-
"step": 1380
|
| 954 |
-
},
|
| 955 |
-
{
|
| 956 |
-
"epoch": 2.96,
|
| 957 |
-
"learning_rate": 5.181236673773987e-05,
|
| 958 |
-
"loss": 0.0857,
|
| 959 |
-
"step": 1390
|
| 960 |
-
},
|
| 961 |
-
{
|
| 962 |
-
"epoch": 2.99,
|
| 963 |
-
"learning_rate": 5.074626865671642e-05,
|
| 964 |
-
"loss": 0.0684,
|
| 965 |
-
"step": 1400
|
| 966 |
-
},
|
| 967 |
-
{
|
| 968 |
-
"epoch": 2.99,
|
| 969 |
-
"eval_accuracy": 0.9208,
|
| 970 |
-
"eval_loss": 0.28956088423728943,
|
| 971 |
-
"eval_runtime": 904.414,
|
| 972 |
-
"eval_samples_per_second": 2.764,
|
| 973 |
-
"eval_steps_per_second": 0.346,
|
| 974 |
-
"step": 1400
|
| 975 |
-
},
|
| 976 |
-
{
|
| 977 |
-
"epoch": 3.01,
|
| 978 |
-
"learning_rate": 4.9680170575692967e-05,
|
| 979 |
-
"loss": 0.038,
|
| 980 |
-
"step": 1410
|
| 981 |
-
},
|
| 982 |
-
{
|
| 983 |
-
"epoch": 3.03,
|
| 984 |
-
"learning_rate": 4.861407249466951e-05,
|
| 985 |
-
"loss": 0.0321,
|
| 986 |
-
"step": 1420
|
| 987 |
-
},
|
| 988 |
-
{
|
| 989 |
-
"epoch": 3.05,
|
| 990 |
-
"learning_rate": 4.7547974413646055e-05,
|
| 991 |
-
"loss": 0.0125,
|
| 992 |
-
"step": 1430
|
| 993 |
-
},
|
| 994 |
-
{
|
| 995 |
-
"epoch": 3.07,
|
| 996 |
-
"learning_rate": 4.64818763326226e-05,
|
| 997 |
-
"loss": 0.0197,
|
| 998 |
-
"step": 1440
|
| 999 |
-
},
|
| 1000 |
-
{
|
| 1001 |
-
"epoch": 3.09,
|
| 1002 |
-
"learning_rate": 4.541577825159915e-05,
|
| 1003 |
-
"loss": 0.0289,
|
| 1004 |
-
"step": 1450
|
| 1005 |
-
},
|
| 1006 |
-
{
|
| 1007 |
-
"epoch": 3.11,
|
| 1008 |
-
"learning_rate": 4.43496801705757e-05,
|
| 1009 |
-
"loss": 0.0273,
|
| 1010 |
-
"step": 1460
|
| 1011 |
-
},
|
| 1012 |
-
{
|
| 1013 |
-
"epoch": 3.13,
|
| 1014 |
-
"learning_rate": 4.328358208955224e-05,
|
| 1015 |
-
"loss": 0.0118,
|
| 1016 |
-
"step": 1470
|
| 1017 |
-
},
|
| 1018 |
-
{
|
| 1019 |
-
"epoch": 3.16,
|
| 1020 |
-
"learning_rate": 4.2217484008528785e-05,
|
| 1021 |
-
"loss": 0.0105,
|
| 1022 |
-
"step": 1480
|
| 1023 |
-
},
|
| 1024 |
-
{
|
| 1025 |
-
"epoch": 3.18,
|
| 1026 |
-
"learning_rate": 4.115138592750533e-05,
|
| 1027 |
-
"loss": 0.008,
|
| 1028 |
-
"step": 1490
|
| 1029 |
-
},
|
| 1030 |
-
{
|
| 1031 |
-
"epoch": 3.2,
|
| 1032 |
-
"learning_rate": 4.008528784648188e-05,
|
| 1033 |
-
"loss": 0.0084,
|
| 1034 |
-
"step": 1500
|
| 1035 |
-
},
|
| 1036 |
-
{
|
| 1037 |
-
"epoch": 3.2,
|
| 1038 |
-
"eval_accuracy": 0.9308,
|
| 1039 |
-
"eval_loss": 0.26776641607284546,
|
| 1040 |
-
"eval_runtime": 908.0277,
|
| 1041 |
-
"eval_samples_per_second": 2.753,
|
| 1042 |
-
"eval_steps_per_second": 0.345,
|
| 1043 |
-
"step": 1500
|
| 1044 |
-
},
|
| 1045 |
-
{
|
| 1046 |
-
"epoch": 3.22,
|
| 1047 |
-
"learning_rate": 3.901918976545843e-05,
|
| 1048 |
-
"loss": 0.0076,
|
| 1049 |
-
"step": 1510
|
| 1050 |
-
},
|
| 1051 |
-
{
|
| 1052 |
-
"epoch": 3.24,
|
| 1053 |
-
"learning_rate": 3.7953091684434974e-05,
|
| 1054 |
-
"loss": 0.0087,
|
| 1055 |
-
"step": 1520
|
| 1056 |
-
},
|
| 1057 |
-
{
|
| 1058 |
-
"epoch": 3.26,
|
| 1059 |
-
"learning_rate": 3.6886993603411515e-05,
|
| 1060 |
-
"loss": 0.019,
|
| 1061 |
-
"step": 1530
|
| 1062 |
-
},
|
| 1063 |
-
{
|
| 1064 |
-
"epoch": 3.28,
|
| 1065 |
-
"learning_rate": 3.582089552238806e-05,
|
| 1066 |
-
"loss": 0.0182,
|
| 1067 |
-
"step": 1540
|
| 1068 |
-
},
|
| 1069 |
-
{
|
| 1070 |
-
"epoch": 3.3,
|
| 1071 |
-
"learning_rate": 3.47547974413646e-05,
|
| 1072 |
-
"loss": 0.0401,
|
| 1073 |
-
"step": 1550
|
| 1074 |
-
},
|
| 1075 |
-
{
|
| 1076 |
-
"epoch": 3.33,
|
| 1077 |
-
"learning_rate": 3.368869936034115e-05,
|
| 1078 |
-
"loss": 0.0461,
|
| 1079 |
-
"step": 1560
|
| 1080 |
-
},
|
| 1081 |
-
{
|
| 1082 |
-
"epoch": 3.35,
|
| 1083 |
-
"learning_rate": 3.26226012793177e-05,
|
| 1084 |
-
"loss": 0.0272,
|
| 1085 |
-
"step": 1570
|
| 1086 |
-
},
|
| 1087 |
-
{
|
| 1088 |
-
"epoch": 3.37,
|
| 1089 |
-
"learning_rate": 3.1556503198294245e-05,
|
| 1090 |
-
"loss": 0.0333,
|
| 1091 |
-
"step": 1580
|
| 1092 |
-
},
|
| 1093 |
-
{
|
| 1094 |
-
"epoch": 3.39,
|
| 1095 |
-
"learning_rate": 3.0490405117270792e-05,
|
| 1096 |
-
"loss": 0.0699,
|
| 1097 |
-
"step": 1590
|
| 1098 |
-
},
|
| 1099 |
-
{
|
| 1100 |
-
"epoch": 3.41,
|
| 1101 |
-
"learning_rate": 2.9424307036247333e-05,
|
| 1102 |
-
"loss": 0.0072,
|
| 1103 |
-
"step": 1600
|
| 1104 |
-
},
|
| 1105 |
-
{
|
| 1106 |
-
"epoch": 3.41,
|
| 1107 |
-
"eval_accuracy": 0.9292,
|
| 1108 |
-
"eval_loss": 0.2749463617801666,
|
| 1109 |
-
"eval_runtime": 908.1664,
|
| 1110 |
-
"eval_samples_per_second": 2.753,
|
| 1111 |
-
"eval_steps_per_second": 0.345,
|
| 1112 |
-
"step": 1600
|
| 1113 |
-
},
|
| 1114 |
-
{
|
| 1115 |
-
"epoch": 3.43,
|
| 1116 |
-
"learning_rate": 2.835820895522388e-05,
|
| 1117 |
-
"loss": 0.0238,
|
| 1118 |
-
"step": 1610
|
| 1119 |
-
},
|
| 1120 |
-
{
|
| 1121 |
-
"epoch": 3.45,
|
| 1122 |
-
"learning_rate": 2.7292110874200428e-05,
|
| 1123 |
-
"loss": 0.0098,
|
| 1124 |
-
"step": 1620
|
| 1125 |
-
},
|
| 1126 |
-
{
|
| 1127 |
-
"epoch": 3.48,
|
| 1128 |
-
"learning_rate": 2.6226012793176975e-05,
|
| 1129 |
-
"loss": 0.0162,
|
| 1130 |
-
"step": 1630
|
| 1131 |
-
},
|
| 1132 |
-
{
|
| 1133 |
-
"epoch": 3.5,
|
| 1134 |
-
"learning_rate": 2.515991471215352e-05,
|
| 1135 |
-
"loss": 0.0075,
|
| 1136 |
-
"step": 1640
|
| 1137 |
-
},
|
| 1138 |
-
{
|
| 1139 |
-
"epoch": 3.52,
|
| 1140 |
-
"learning_rate": 2.4093816631130063e-05,
|
| 1141 |
-
"loss": 0.0326,
|
| 1142 |
-
"step": 1650
|
| 1143 |
-
},
|
| 1144 |
-
{
|
| 1145 |
-
"epoch": 3.54,
|
| 1146 |
-
"learning_rate": 2.302771855010661e-05,
|
| 1147 |
-
"loss": 0.0141,
|
| 1148 |
-
"step": 1660
|
| 1149 |
-
},
|
| 1150 |
-
{
|
| 1151 |
-
"epoch": 3.56,
|
| 1152 |
-
"learning_rate": 2.1961620469083158e-05,
|
| 1153 |
-
"loss": 0.0083,
|
| 1154 |
-
"step": 1670
|
| 1155 |
-
},
|
| 1156 |
-
{
|
| 1157 |
-
"epoch": 3.58,
|
| 1158 |
-
"learning_rate": 2.0895522388059702e-05,
|
| 1159 |
-
"loss": 0.011,
|
| 1160 |
-
"step": 1680
|
| 1161 |
-
},
|
| 1162 |
-
{
|
| 1163 |
-
"epoch": 3.6,
|
| 1164 |
-
"learning_rate": 1.982942430703625e-05,
|
| 1165 |
-
"loss": 0.0067,
|
| 1166 |
-
"step": 1690
|
| 1167 |
-
},
|
| 1168 |
-
{
|
| 1169 |
-
"epoch": 3.62,
|
| 1170 |
-
"learning_rate": 1.8763326226012797e-05,
|
| 1171 |
-
"loss": 0.0585,
|
| 1172 |
-
"step": 1700
|
| 1173 |
-
},
|
| 1174 |
-
{
|
| 1175 |
-
"epoch": 3.62,
|
| 1176 |
-
"eval_accuracy": 0.9316,
|
| 1177 |
-
"eval_loss": 0.2734415531158447,
|
| 1178 |
-
"eval_runtime": 910.1726,
|
| 1179 |
-
"eval_samples_per_second": 2.747,
|
| 1180 |
-
"eval_steps_per_second": 0.344,
|
| 1181 |
-
"step": 1700
|
| 1182 |
-
},
|
| 1183 |
-
{
|
| 1184 |
-
"epoch": 3.65,
|
| 1185 |
-
"learning_rate": 1.769722814498934e-05,
|
| 1186 |
-
"loss": 0.007,
|
| 1187 |
-
"step": 1710
|
| 1188 |
-
},
|
| 1189 |
-
{
|
| 1190 |
-
"epoch": 3.67,
|
| 1191 |
-
"learning_rate": 1.6631130063965885e-05,
|
| 1192 |
-
"loss": 0.0061,
|
| 1193 |
-
"step": 1720
|
| 1194 |
-
},
|
| 1195 |
-
{
|
| 1196 |
-
"epoch": 3.69,
|
| 1197 |
-
"learning_rate": 1.5565031982942432e-05,
|
| 1198 |
-
"loss": 0.0061,
|
| 1199 |
-
"step": 1730
|
| 1200 |
-
},
|
| 1201 |
-
{
|
| 1202 |
-
"epoch": 3.71,
|
| 1203 |
-
"learning_rate": 1.4498933901918976e-05,
|
| 1204 |
-
"loss": 0.0159,
|
| 1205 |
-
"step": 1740
|
| 1206 |
-
},
|
| 1207 |
-
{
|
| 1208 |
-
"epoch": 3.73,
|
| 1209 |
-
"learning_rate": 1.3432835820895523e-05,
|
| 1210 |
-
"loss": 0.0154,
|
| 1211 |
-
"step": 1750
|
| 1212 |
-
},
|
| 1213 |
-
{
|
| 1214 |
-
"epoch": 3.75,
|
| 1215 |
-
"learning_rate": 1.2366737739872069e-05,
|
| 1216 |
-
"loss": 0.0071,
|
| 1217 |
-
"step": 1760
|
| 1218 |
-
},
|
| 1219 |
-
{
|
| 1220 |
-
"epoch": 3.77,
|
| 1221 |
-
"learning_rate": 1.1300639658848615e-05,
|
| 1222 |
-
"loss": 0.0067,
|
| 1223 |
-
"step": 1770
|
| 1224 |
-
},
|
| 1225 |
-
{
|
| 1226 |
-
"epoch": 3.8,
|
| 1227 |
-
"learning_rate": 1.023454157782516e-05,
|
| 1228 |
-
"loss": 0.0387,
|
| 1229 |
-
"step": 1780
|
| 1230 |
-
},
|
| 1231 |
-
{
|
| 1232 |
-
"epoch": 3.82,
|
| 1233 |
-
"learning_rate": 9.168443496801706e-06,
|
| 1234 |
-
"loss": 0.0492,
|
| 1235 |
-
"step": 1790
|
| 1236 |
-
},
|
| 1237 |
-
{
|
| 1238 |
-
"epoch": 3.84,
|
| 1239 |
-
"learning_rate": 8.102345415778252e-06,
|
| 1240 |
-
"loss": 0.0152,
|
| 1241 |
-
"step": 1800
|
| 1242 |
-
},
|
| 1243 |
-
{
|
| 1244 |
-
"epoch": 3.84,
|
| 1245 |
-
"eval_accuracy": 0.9332,
|
| 1246 |
-
"eval_loss": 0.26903361082077026,
|
| 1247 |
-
"eval_runtime": 906.1705,
|
| 1248 |
-
"eval_samples_per_second": 2.759,
|
| 1249 |
-
"eval_steps_per_second": 0.345,
|
| 1250 |
-
"step": 1800
|
| 1251 |
-
},
|
| 1252 |
-
{
|
| 1253 |
-
"epoch": 3.86,
|
| 1254 |
-
"learning_rate": 7.0362473347547975e-06,
|
| 1255 |
-
"loss": 0.0054,
|
| 1256 |
-
"step": 1810
|
| 1257 |
-
},
|
| 1258 |
-
{
|
| 1259 |
-
"epoch": 3.88,
|
| 1260 |
-
"learning_rate": 5.970149253731343e-06,
|
| 1261 |
-
"loss": 0.006,
|
| 1262 |
-
"step": 1820
|
| 1263 |
-
},
|
| 1264 |
-
{
|
| 1265 |
-
"epoch": 3.9,
|
| 1266 |
-
"learning_rate": 4.90405117270789e-06,
|
| 1267 |
-
"loss": 0.0257,
|
| 1268 |
-
"step": 1830
|
| 1269 |
-
},
|
| 1270 |
-
{
|
| 1271 |
-
"epoch": 3.92,
|
| 1272 |
-
"learning_rate": 3.837953091684435e-06,
|
| 1273 |
-
"loss": 0.0076,
|
| 1274 |
-
"step": 1840
|
| 1275 |
-
},
|
| 1276 |
-
{
|
| 1277 |
-
"epoch": 3.94,
|
| 1278 |
-
"learning_rate": 2.771855010660981e-06,
|
| 1279 |
-
"loss": 0.0055,
|
| 1280 |
-
"step": 1850
|
| 1281 |
-
},
|
| 1282 |
-
{
|
| 1283 |
-
"epoch": 3.97,
|
| 1284 |
-
"learning_rate": 1.7057569296375267e-06,
|
| 1285 |
-
"loss": 0.0064,
|
| 1286 |
-
"step": 1860
|
| 1287 |
-
},
|
| 1288 |
-
{
|
| 1289 |
-
"epoch": 3.99,
|
| 1290 |
-
"learning_rate": 6.396588486140725e-07,
|
| 1291 |
-
"loss": 0.0054,
|
| 1292 |
-
"step": 1870
|
| 1293 |
},
|
| 1294 |
{
|
| 1295 |
"epoch": 4.0,
|
| 1296 |
-
"step":
|
| 1297 |
"total_flos": 0.0,
|
| 1298 |
-
"train_loss": 0.
|
| 1299 |
-
"train_runtime":
|
| 1300 |
-
"train_samples_per_second": 0.
|
| 1301 |
-
"train_steps_per_second": 0.
|
| 1302 |
}
|
| 1303 |
],
|
| 1304 |
-
"max_steps":
|
| 1305 |
"num_train_epochs": 4,
|
| 1306 |
"total_flos": 0.0,
|
| 1307 |
"trial_name": null,
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
"epoch": 4.0,
|
| 5 |
+
"global_step": 28,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
{
|
| 11 |
"epoch": 1.43,
|
| 12 |
"learning_rate": 0.00012857142857142858,
|
| 13 |
+
"loss": 0.3956,
|
| 14 |
+
"step": 10
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 2.86,
|
| 18 |
"learning_rate": 5.714285714285714e-05,
|
| 19 |
+
"loss": 0.0812,
|
| 20 |
+
"step": 20
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 4.0,
|
| 24 |
+
"step": 28,
|
| 25 |
"total_flos": 0.0,
|
| 26 |
+
"train_loss": 0.18094830640724727,
|
| 27 |
+
"train_runtime": 416.3142,
|
| 28 |
+
"train_samples_per_second": 0.961,
|
| 29 |
+
"train_steps_per_second": 0.067
|
| 30 |
}
|
| 31 |
],
|
| 32 |
+
"max_steps": 28,
|
| 33 |
"num_train_epochs": 4,
|
| 34 |
"total_flos": 0.0,
|
| 35 |
"trial_name": null,
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3256
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3eb12cbc835ddd417563643373525c41ded388268170fb56d0f75ed3ddab9f7c
|
| 3 |
size 3256
|