Upload folder using huggingface_hub
Browse files- model-00001-of-00002.safetensors +1 -1
- model-00002-of-00002.safetensors +1 -1
- optimizer.pt +2 -2
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +127 -162
- training_args.bin +1 -1
model-00001-of-00002.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4991031824
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4fdc69c9532d50d8cbcfc7357cd186c5f961216500c340a60e088c6248848814
|
| 3 |
size 4991031824
|
model-00002-of-00002.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1610725592
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5acc5280e967a03411a879cc37d8f4c9b24d1d8a7df4b74a0ab654ca2b2308b4
|
| 3 |
size 1610725592
|
optimizer.pt
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 sha256:2acac140f381a83b89f9d555cc696d4840505d66eeeaf95bf3c42656fe32c150
|
| 3 |
+
size 13203678103
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14645
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8e2011629d8bed3ef560fa11175cac55684c4e12a72634bb24abf767b6c7399
|
| 3 |
size 14645
|
scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:993186ddfb6142501973771452980601dfb89818da88b2abc705869965101f78
|
| 3 |
size 1465
|
trainer_state.json
CHANGED
|
@@ -2,264 +2,229 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch":
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
| 11 |
"log_history": [
|
| 12 |
{
|
| 13 |
-
"epoch": 0.
|
| 14 |
-
"grad_norm":
|
| 15 |
-
"learning_rate":
|
| 16 |
-
"loss": 5.
|
| 17 |
"step": 50
|
| 18 |
},
|
| 19 |
{
|
| 20 |
-
"epoch": 0.
|
| 21 |
-
"grad_norm":
|
| 22 |
-
"learning_rate":
|
| 23 |
-
"loss": 5.
|
| 24 |
"step": 100
|
| 25 |
},
|
| 26 |
{
|
| 27 |
-
"epoch": 0.
|
| 28 |
-
"grad_norm":
|
| 29 |
-
"learning_rate":
|
| 30 |
-
"loss": 5.
|
| 31 |
"step": 150
|
| 32 |
},
|
| 33 |
{
|
| 34 |
-
"epoch": 0.
|
| 35 |
-
"grad_norm":
|
| 36 |
-
"learning_rate":
|
| 37 |
-
"loss": 5.
|
| 38 |
"step": 200
|
| 39 |
},
|
| 40 |
{
|
| 41 |
-
"epoch": 0.
|
| 42 |
-
"grad_norm":
|
| 43 |
-
"learning_rate": 4.
|
| 44 |
-
"loss": 5.
|
| 45 |
"step": 250
|
| 46 |
},
|
| 47 |
{
|
| 48 |
-
"epoch": 0.
|
| 49 |
-
"grad_norm": 6.
|
| 50 |
-
"learning_rate": 4.
|
| 51 |
-
"loss": 5.
|
| 52 |
"step": 300
|
| 53 |
},
|
| 54 |
{
|
| 55 |
-
"epoch": 0.
|
| 56 |
-
"grad_norm":
|
| 57 |
-
"learning_rate": 4.
|
| 58 |
-
"loss": 5.
|
| 59 |
"step": 350
|
| 60 |
},
|
| 61 |
{
|
| 62 |
-
"epoch": 0.
|
| 63 |
-
"grad_norm":
|
| 64 |
-
"learning_rate":
|
| 65 |
-
"loss":
|
| 66 |
"step": 400
|
| 67 |
},
|
| 68 |
{
|
| 69 |
-
"epoch":
|
| 70 |
-
"grad_norm":
|
| 71 |
-
"learning_rate":
|
| 72 |
-
"loss": 4.
|
| 73 |
"step": 450
|
| 74 |
},
|
| 75 |
{
|
| 76 |
-
"epoch":
|
| 77 |
-
"grad_norm":
|
| 78 |
-
"learning_rate":
|
| 79 |
-
"loss":
|
| 80 |
"step": 500
|
| 81 |
},
|
| 82 |
{
|
| 83 |
-
"epoch":
|
| 84 |
-
"grad_norm":
|
| 85 |
-
"learning_rate":
|
| 86 |
-
"loss":
|
| 87 |
"step": 550
|
| 88 |
},
|
| 89 |
{
|
| 90 |
-
"epoch":
|
| 91 |
-
"grad_norm":
|
| 92 |
-
"learning_rate":
|
| 93 |
-
"loss":
|
| 94 |
"step": 600
|
| 95 |
},
|
| 96 |
{
|
| 97 |
-
"epoch":
|
| 98 |
-
"grad_norm": 7.
|
| 99 |
-
"learning_rate":
|
| 100 |
-
"loss": 4.
|
| 101 |
"step": 650
|
| 102 |
},
|
| 103 |
{
|
| 104 |
-
"epoch":
|
| 105 |
-
"grad_norm":
|
| 106 |
-
"learning_rate":
|
| 107 |
-
"loss":
|
| 108 |
"step": 700
|
| 109 |
},
|
| 110 |
{
|
| 111 |
-
"epoch":
|
| 112 |
-
"grad_norm":
|
| 113 |
-
"learning_rate":
|
| 114 |
-
"loss": 4.
|
| 115 |
"step": 750
|
| 116 |
},
|
| 117 |
{
|
| 118 |
-
"epoch":
|
| 119 |
-
"grad_norm":
|
| 120 |
-
"learning_rate":
|
| 121 |
-
"loss": 4.
|
| 122 |
"step": 800
|
| 123 |
},
|
| 124 |
{
|
| 125 |
-
"epoch":
|
| 126 |
-
"grad_norm":
|
| 127 |
-
"learning_rate":
|
| 128 |
-
"loss": 4.
|
| 129 |
"step": 850
|
| 130 |
},
|
| 131 |
{
|
| 132 |
-
"epoch":
|
| 133 |
-
"grad_norm":
|
| 134 |
-
"learning_rate":
|
| 135 |
-
"loss": 4.
|
| 136 |
"step": 900
|
| 137 |
},
|
| 138 |
{
|
| 139 |
-
"epoch":
|
| 140 |
-
"grad_norm":
|
| 141 |
-
"learning_rate":
|
| 142 |
-
"loss": 4.
|
| 143 |
"step": 950
|
| 144 |
},
|
| 145 |
{
|
| 146 |
-
"epoch":
|
| 147 |
-
"grad_norm":
|
| 148 |
-
"learning_rate":
|
| 149 |
-
"loss": 4.
|
| 150 |
"step": 1000
|
| 151 |
},
|
| 152 |
{
|
| 153 |
-
"epoch":
|
| 154 |
-
"grad_norm":
|
| 155 |
-
"learning_rate":
|
| 156 |
-
"loss": 4.
|
| 157 |
"step": 1050
|
| 158 |
},
|
| 159 |
{
|
| 160 |
-
"epoch":
|
| 161 |
-
"grad_norm":
|
| 162 |
-
"learning_rate":
|
| 163 |
-
"loss": 4.
|
| 164 |
"step": 1100
|
| 165 |
},
|
| 166 |
{
|
| 167 |
-
"epoch":
|
| 168 |
-
"grad_norm":
|
| 169 |
-
"learning_rate":
|
| 170 |
-
"loss": 4.
|
| 171 |
"step": 1150
|
| 172 |
},
|
| 173 |
{
|
| 174 |
-
"epoch":
|
| 175 |
-
"grad_norm":
|
| 176 |
-
"learning_rate":
|
| 177 |
-
"loss": 4.
|
| 178 |
"step": 1200
|
| 179 |
},
|
| 180 |
{
|
| 181 |
-
"epoch":
|
| 182 |
-
"grad_norm":
|
| 183 |
-
"learning_rate":
|
| 184 |
-
"loss": 4.
|
| 185 |
"step": 1250
|
| 186 |
},
|
| 187 |
{
|
| 188 |
-
"epoch":
|
| 189 |
-
"grad_norm":
|
| 190 |
-
"learning_rate":
|
| 191 |
-
"loss": 4.
|
| 192 |
"step": 1300
|
| 193 |
},
|
| 194 |
{
|
| 195 |
-
"epoch":
|
| 196 |
-
"grad_norm":
|
| 197 |
-
"learning_rate":
|
| 198 |
-
"loss": 4.
|
| 199 |
"step": 1350
|
| 200 |
},
|
| 201 |
{
|
| 202 |
-
"epoch":
|
| 203 |
-
"grad_norm":
|
| 204 |
-
"learning_rate":
|
| 205 |
-
"loss":
|
| 206 |
"step": 1400
|
| 207 |
},
|
| 208 |
{
|
| 209 |
-
"epoch":
|
| 210 |
-
"grad_norm":
|
| 211 |
-
"learning_rate":
|
| 212 |
-
"loss":
|
| 213 |
"step": 1450
|
| 214 |
},
|
| 215 |
{
|
| 216 |
-
"epoch":
|
| 217 |
-
"grad_norm":
|
| 218 |
-
"learning_rate":
|
| 219 |
-
"loss":
|
| 220 |
"step": 1500
|
| 221 |
-
},
|
| 222 |
-
{
|
| 223 |
-
"epoch": 0.8688340807174888,
|
| 224 |
-
"grad_norm": 7.0,
|
| 225 |
-
"learning_rate": 6.5863228699551565e-06,
|
| 226 |
-
"loss": 4.9036,
|
| 227 |
-
"step": 1550
|
| 228 |
-
},
|
| 229 |
-
{
|
| 230 |
-
"epoch": 0.8968609865470852,
|
| 231 |
-
"grad_norm": 10.0,
|
| 232 |
-
"learning_rate": 5.184977578475336e-06,
|
| 233 |
-
"loss": 4.8653,
|
| 234 |
-
"step": 1600
|
| 235 |
-
},
|
| 236 |
-
{
|
| 237 |
-
"epoch": 0.9248878923766816,
|
| 238 |
-
"grad_norm": 7.46875,
|
| 239 |
-
"learning_rate": 3.783632286995516e-06,
|
| 240 |
-
"loss": 4.8676,
|
| 241 |
-
"step": 1650
|
| 242 |
-
},
|
| 243 |
-
{
|
| 244 |
-
"epoch": 0.952914798206278,
|
| 245 |
-
"grad_norm": 8.25,
|
| 246 |
-
"learning_rate": 2.3822869955156952e-06,
|
| 247 |
-
"loss": 4.8952,
|
| 248 |
-
"step": 1700
|
| 249 |
-
},
|
| 250 |
-
{
|
| 251 |
-
"epoch": 0.9809417040358744,
|
| 252 |
-
"grad_norm": 6.3125,
|
| 253 |
-
"learning_rate": 9.809417040358745e-07,
|
| 254 |
-
"loss": 4.9213,
|
| 255 |
-
"step": 1750
|
| 256 |
}
|
| 257 |
],
|
| 258 |
"logging_steps": 50,
|
| 259 |
-
"max_steps":
|
| 260 |
"num_input_tokens_seen": 0,
|
| 261 |
-
"num_train_epochs":
|
| 262 |
-
"save_steps":
|
| 263 |
"stateful_callbacks": {
|
| 264 |
"TrainerControl": {
|
| 265 |
"args": {
|
|
@@ -267,12 +232,12 @@
|
|
| 267 |
"should_evaluate": false,
|
| 268 |
"should_log": false,
|
| 269 |
"should_save": true,
|
| 270 |
-
"should_training_stop":
|
| 271 |
},
|
| 272 |
"attributes": {}
|
| 273 |
}
|
| 274 |
},
|
| 275 |
-
"total_flos":
|
| 276 |
"train_batch_size": 1,
|
| 277 |
"trial_name": null,
|
| 278 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 3.3632286995515694,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 1500,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
| 11 |
"log_history": [
|
| 12 |
{
|
| 13 |
+
"epoch": 0.11210762331838565,
|
| 14 |
+
"grad_norm": 4.625,
|
| 15 |
+
"learning_rate": 9.14179104477612e-06,
|
| 16 |
+
"loss": 5.492,
|
| 17 |
"step": 50
|
| 18 |
},
|
| 19 |
{
|
| 20 |
+
"epoch": 0.2242152466367713,
|
| 21 |
+
"grad_norm": 4.90625,
|
| 22 |
+
"learning_rate": 1.8470149253731344e-05,
|
| 23 |
+
"loss": 5.439,
|
| 24 |
"step": 100
|
| 25 |
},
|
| 26 |
{
|
| 27 |
+
"epoch": 0.336322869955157,
|
| 28 |
+
"grad_norm": 3.84375,
|
| 29 |
+
"learning_rate": 2.7798507462686568e-05,
|
| 30 |
+
"loss": 5.2502,
|
| 31 |
"step": 150
|
| 32 |
},
|
| 33 |
{
|
| 34 |
+
"epoch": 0.4484304932735426,
|
| 35 |
+
"grad_norm": 6.59375,
|
| 36 |
+
"learning_rate": 3.7126865671641795e-05,
|
| 37 |
+
"loss": 5.161,
|
| 38 |
"step": 200
|
| 39 |
},
|
| 40 |
{
|
| 41 |
+
"epoch": 0.5605381165919282,
|
| 42 |
+
"grad_norm": 6.25,
|
| 43 |
+
"learning_rate": 4.645522388059701e-05,
|
| 44 |
+
"loss": 5.0548,
|
| 45 |
"step": 250
|
| 46 |
},
|
| 47 |
{
|
| 48 |
+
"epoch": 0.672645739910314,
|
| 49 |
+
"grad_norm": 6.71875,
|
| 50 |
+
"learning_rate": 4.999325361589072e-05,
|
| 51 |
+
"loss": 5.0148,
|
| 52 |
"step": 300
|
| 53 |
},
|
| 54 |
{
|
| 55 |
+
"epoch": 0.7847533632286996,
|
| 56 |
+
"grad_norm": 5.6875,
|
| 57 |
+
"learning_rate": 4.9953952730494324e-05,
|
| 58 |
+
"loss": 5.0028,
|
| 59 |
"step": 350
|
| 60 |
},
|
| 61 |
{
|
| 62 |
+
"epoch": 0.8968609865470852,
|
| 63 |
+
"grad_norm": 5.09375,
|
| 64 |
+
"learning_rate": 4.987961816680492e-05,
|
| 65 |
+
"loss": 4.9576,
|
| 66 |
"step": 400
|
| 67 |
},
|
| 68 |
{
|
| 69 |
+
"epoch": 1.0089686098654709,
|
| 70 |
+
"grad_norm": 4.375,
|
| 71 |
+
"learning_rate": 4.977035428557125e-05,
|
| 72 |
+
"loss": 4.929,
|
| 73 |
"step": 450
|
| 74 |
},
|
| 75 |
{
|
| 76 |
+
"epoch": 1.1210762331838564,
|
| 77 |
+
"grad_norm": 4.78125,
|
| 78 |
+
"learning_rate": 4.9626314485964385e-05,
|
| 79 |
+
"loss": 4.7677,
|
| 80 |
"step": 500
|
| 81 |
},
|
| 82 |
{
|
| 83 |
+
"epoch": 1.2331838565022422,
|
| 84 |
+
"grad_norm": 4.875,
|
| 85 |
+
"learning_rate": 4.944770099021562e-05,
|
| 86 |
+
"loss": 4.7483,
|
| 87 |
"step": 550
|
| 88 |
},
|
| 89 |
{
|
| 90 |
+
"epoch": 1.3452914798206277,
|
| 91 |
+
"grad_norm": 5.9375,
|
| 92 |
+
"learning_rate": 4.923476455971e-05,
|
| 93 |
+
"loss": 4.7362,
|
| 94 |
"step": 600
|
| 95 |
},
|
| 96 |
{
|
| 97 |
+
"epoch": 1.4573991031390134,
|
| 98 |
+
"grad_norm": 7.0625,
|
| 99 |
+
"learning_rate": 4.898780414293411e-05,
|
| 100 |
+
"loss": 4.7181,
|
| 101 |
"step": 650
|
| 102 |
},
|
| 103 |
{
|
| 104 |
+
"epoch": 1.5695067264573992,
|
| 105 |
+
"grad_norm": 4.1875,
|
| 106 |
+
"learning_rate": 4.870716645577244e-05,
|
| 107 |
+
"loss": 4.719,
|
| 108 |
"step": 700
|
| 109 |
},
|
| 110 |
{
|
| 111 |
+
"epoch": 1.6816143497757847,
|
| 112 |
+
"grad_norm": 5.15625,
|
| 113 |
+
"learning_rate": 4.839324549474148e-05,
|
| 114 |
+
"loss": 4.727,
|
| 115 |
"step": 750
|
| 116 |
},
|
| 117 |
{
|
| 118 |
+
"epoch": 1.7937219730941703,
|
| 119 |
+
"grad_norm": 4.46875,
|
| 120 |
+
"learning_rate": 4.804648198384507e-05,
|
| 121 |
+
"loss": 4.7355,
|
| 122 |
"step": 800
|
| 123 |
},
|
| 124 |
{
|
| 125 |
+
"epoch": 1.905829596412556,
|
| 126 |
+
"grad_norm": 4.59375,
|
| 127 |
+
"learning_rate": 4.7667362755827306e-05,
|
| 128 |
+
"loss": 4.7111,
|
| 129 |
"step": 850
|
| 130 |
},
|
| 131 |
{
|
| 132 |
+
"epoch": 2.0179372197309418,
|
| 133 |
+
"grad_norm": 4.15625,
|
| 134 |
+
"learning_rate": 4.725642006869207e-05,
|
| 135 |
+
"loss": 4.6229,
|
| 136 |
"step": 900
|
| 137 |
},
|
| 138 |
{
|
| 139 |
+
"epoch": 2.1300448430493275,
|
| 140 |
+
"grad_norm": 5.53125,
|
| 141 |
+
"learning_rate": 4.68142308584484e-05,
|
| 142 |
+
"loss": 4.3583,
|
| 143 |
"step": 950
|
| 144 |
},
|
| 145 |
{
|
| 146 |
+
"epoch": 2.242152466367713,
|
| 147 |
+
"grad_norm": 3.875,
|
| 148 |
+
"learning_rate": 4.634141592913097e-05,
|
| 149 |
+
"loss": 4.3665,
|
| 150 |
"step": 1000
|
| 151 |
},
|
| 152 |
{
|
| 153 |
+
"epoch": 2.3542600896860986,
|
| 154 |
+
"grad_norm": 3.890625,
|
| 155 |
+
"learning_rate": 4.583863908123282e-05,
|
| 156 |
+
"loss": 4.4125,
|
| 157 |
"step": 1050
|
| 158 |
},
|
| 159 |
{
|
| 160 |
+
"epoch": 2.4663677130044843,
|
| 161 |
+
"grad_norm": 4.21875,
|
| 162 |
+
"learning_rate": 4.530660617977393e-05,
|
| 163 |
+
"loss": 4.3592,
|
| 164 |
"step": 1100
|
| 165 |
},
|
| 166 |
{
|
| 167 |
+
"epoch": 2.57847533632287,
|
| 168 |
+
"grad_norm": 4.5,
|
| 169 |
+
"learning_rate": 4.474606416331397e-05,
|
| 170 |
+
"loss": 4.371,
|
| 171 |
"step": 1150
|
| 172 |
},
|
| 173 |
{
|
| 174 |
+
"epoch": 2.6905829596412554,
|
| 175 |
+
"grad_norm": 4.3125,
|
| 176 |
+
"learning_rate": 4.415779999530064e-05,
|
| 177 |
+
"loss": 4.4046,
|
| 178 |
"step": 1200
|
| 179 |
},
|
| 180 |
{
|
| 181 |
+
"epoch": 2.802690582959641,
|
| 182 |
+
"grad_norm": 4.21875,
|
| 183 |
+
"learning_rate": 4.354263955922568e-05,
|
| 184 |
+
"loss": 4.3805,
|
| 185 |
"step": 1250
|
| 186 |
},
|
| 187 |
{
|
| 188 |
+
"epoch": 2.914798206278027,
|
| 189 |
+
"grad_norm": 4.15625,
|
| 190 |
+
"learning_rate": 4.290144649913973e-05,
|
| 191 |
+
"loss": 4.3829,
|
| 192 |
"step": 1300
|
| 193 |
},
|
| 194 |
{
|
| 195 |
+
"epoch": 3.0269058295964126,
|
| 196 |
+
"grad_norm": 4.6875,
|
| 197 |
+
"learning_rate": 4.2235121007153975e-05,
|
| 198 |
+
"loss": 4.2465,
|
| 199 |
"step": 1350
|
| 200 |
},
|
| 201 |
{
|
| 202 |
+
"epoch": 3.1390134529147984,
|
| 203 |
+
"grad_norm": 6.0,
|
| 204 |
+
"learning_rate": 4.1544598559630694e-05,
|
| 205 |
+
"loss": 3.8933,
|
| 206 |
"step": 1400
|
| 207 |
},
|
| 208 |
{
|
| 209 |
+
"epoch": 3.2511210762331837,
|
| 210 |
+
"grad_norm": 4.875,
|
| 211 |
+
"learning_rate": 4.083084860383708e-05,
|
| 212 |
+
"loss": 3.9146,
|
| 213 |
"step": 1450
|
| 214 |
},
|
| 215 |
{
|
| 216 |
+
"epoch": 3.3632286995515694,
|
| 217 |
+
"grad_norm": 4.6875,
|
| 218 |
+
"learning_rate": 4.009487319690626e-05,
|
| 219 |
+
"loss": 3.9086,
|
| 220 |
"step": 1500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 221 |
}
|
| 222 |
],
|
| 223 |
"logging_steps": 50,
|
| 224 |
+
"max_steps": 4460,
|
| 225 |
"num_input_tokens_seen": 0,
|
| 226 |
+
"num_train_epochs": 10,
|
| 227 |
+
"save_steps": 500,
|
| 228 |
"stateful_callbacks": {
|
| 229 |
"TrainerControl": {
|
| 230 |
"args": {
|
|
|
|
| 232 |
"should_evaluate": false,
|
| 233 |
"should_log": false,
|
| 234 |
"should_save": true,
|
| 235 |
+
"should_training_stop": false
|
| 236 |
},
|
| 237 |
"attributes": {}
|
| 238 |
}
|
| 239 |
},
|
| 240 |
+
"total_flos": 5.135152845033677e+16,
|
| 241 |
"train_batch_size": 1,
|
| 242 |
"trial_name": null,
|
| 243 |
"trial_params": null
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5777
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b07c97d6cade78ad89a0fb0e3020fbd76a7fe36e45ae68b0391dd5d87dad3625
|
| 3 |
size 5777
|