Training in progress, step 7500, checkpoint
Browse files
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 328277848
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7c46250bd52dcffd2137953f30321a4ed3d622b1bca6be15bc5f8f084e4fc31f
|
| 3 |
size 328277848
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 318646859
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:71e649d70bbaea3f3c60f2aa0818a879521dffa0038d58ed1695489f8bca966b
|
| 3 |
size 318646859
|
last-checkpoint/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 sha256:7105421fba4235e8fc90f3dbc4569b85e884f75c3232217a25f8f5042cf8247a
|
| 3 |
size 14645
|
last-checkpoint/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:d1a5b64fb90c999b23793906d64020914f128f72d1523c4f0f8e8ea53ab2425c
|
| 3 |
size 1465
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 1.
|
| 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,
|
|
@@ -5027,6 +5027,364 @@
|
|
| 5027 |
"eval_samples_per_second": 274.439,
|
| 5028 |
"eval_steps_per_second": 5.763,
|
| 5029 |
"step": 7000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5030 |
}
|
| 5031 |
],
|
| 5032 |
"logging_steps": 10,
|
|
@@ -5046,7 +5404,7 @@
|
|
| 5046 |
"attributes": {}
|
| 5047 |
}
|
| 5048 |
},
|
| 5049 |
-
"total_flos": 2.
|
| 5050 |
"train_batch_size": 48,
|
| 5051 |
"trial_name": null,
|
| 5052 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.2671059300557528,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 7500,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 5027 |
"eval_samples_per_second": 274.439,
|
| 5028 |
"eval_steps_per_second": 5.763,
|
| 5029 |
"step": 7000
|
| 5030 |
+
},
|
| 5031 |
+
{
|
| 5032 |
+
"epoch": 1.184321675958777,
|
| 5033 |
+
"grad_norm": 0.4751238524913788,
|
| 5034 |
+
"learning_rate": 0.000145689605353268,
|
| 5035 |
+
"loss": 4.423115158081055,
|
| 5036 |
+
"step": 7010
|
| 5037 |
+
},
|
| 5038 |
+
{
|
| 5039 |
+
"epoch": 1.1860111505321844,
|
| 5040 |
+
"grad_norm": 0.5007295608520508,
|
| 5041 |
+
"learning_rate": 0.00014521082727641152,
|
| 5042 |
+
"loss": 4.394451522827149,
|
| 5043 |
+
"step": 7020
|
| 5044 |
+
},
|
| 5045 |
+
{
|
| 5046 |
+
"epoch": 1.1877006251055922,
|
| 5047 |
+
"grad_norm": 0.48377716541290283,
|
| 5048 |
+
"learning_rate": 0.0001447320980362472,
|
| 5049 |
+
"loss": 4.4401298522949215,
|
| 5050 |
+
"step": 7030
|
| 5051 |
+
},
|
| 5052 |
+
{
|
| 5053 |
+
"epoch": 1.189390099679,
|
| 5054 |
+
"grad_norm": 0.5163672566413879,
|
| 5055 |
+
"learning_rate": 0.00014425342251452679,
|
| 5056 |
+
"loss": 4.39234504699707,
|
| 5057 |
+
"step": 7040
|
| 5058 |
+
},
|
| 5059 |
+
{
|
| 5060 |
+
"epoch": 1.1910795742524076,
|
| 5061 |
+
"grad_norm": 0.4772108793258667,
|
| 5062 |
+
"learning_rate": 0.00014377480559245434,
|
| 5063 |
+
"loss": 4.4317176818847654,
|
| 5064 |
+
"step": 7050
|
| 5065 |
+
},
|
| 5066 |
+
{
|
| 5067 |
+
"epoch": 1.1927690488258151,
|
| 5068 |
+
"grad_norm": 0.5122771263122559,
|
| 5069 |
+
"learning_rate": 0.00014329625215063629,
|
| 5070 |
+
"loss": 4.436024856567383,
|
| 5071 |
+
"step": 7060
|
| 5072 |
+
},
|
| 5073 |
+
{
|
| 5074 |
+
"epoch": 1.1944585233992229,
|
| 5075 |
+
"grad_norm": 0.4967247545719147,
|
| 5076 |
+
"learning_rate": 0.00014281776706903177,
|
| 5077 |
+
"loss": 4.399545288085937,
|
| 5078 |
+
"step": 7070
|
| 5079 |
+
},
|
| 5080 |
+
{
|
| 5081 |
+
"epoch": 1.1961479979726306,
|
| 5082 |
+
"grad_norm": 0.5011969208717346,
|
| 5083 |
+
"learning_rate": 0.0001423393552269028,
|
| 5084 |
+
"loss": 4.415679550170898,
|
| 5085 |
+
"step": 7080
|
| 5086 |
+
},
|
| 5087 |
+
{
|
| 5088 |
+
"epoch": 1.1978374725460381,
|
| 5089 |
+
"grad_norm": 0.5074766874313354,
|
| 5090 |
+
"learning_rate": 0.00014186102150276454,
|
| 5091 |
+
"loss": 4.4365581512451175,
|
| 5092 |
+
"step": 7090
|
| 5093 |
+
},
|
| 5094 |
+
{
|
| 5095 |
+
"epoch": 1.1995269471194459,
|
| 5096 |
+
"grad_norm": 0.48712876439094543,
|
| 5097 |
+
"learning_rate": 0.00014138277077433567,
|
| 5098 |
+
"loss": 4.423890686035156,
|
| 5099 |
+
"step": 7100
|
| 5100 |
+
},
|
| 5101 |
+
{
|
| 5102 |
+
"epoch": 1.2012164216928536,
|
| 5103 |
+
"grad_norm": 0.5180738568305969,
|
| 5104 |
+
"learning_rate": 0.00014090460791848827,
|
| 5105 |
+
"loss": 4.416677474975586,
|
| 5106 |
+
"step": 7110
|
| 5107 |
+
},
|
| 5108 |
+
{
|
| 5109 |
+
"epoch": 1.202905896266261,
|
| 5110 |
+
"grad_norm": 0.45985448360443115,
|
| 5111 |
+
"learning_rate": 0.00014042653781119868,
|
| 5112 |
+
"loss": 4.396572494506836,
|
| 5113 |
+
"step": 7120
|
| 5114 |
+
},
|
| 5115 |
+
{
|
| 5116 |
+
"epoch": 1.2045953708396688,
|
| 5117 |
+
"grad_norm": 0.4647870361804962,
|
| 5118 |
+
"learning_rate": 0.0001399485653274973,
|
| 5119 |
+
"loss": 4.408332824707031,
|
| 5120 |
+
"step": 7130
|
| 5121 |
+
},
|
| 5122 |
+
{
|
| 5123 |
+
"epoch": 1.2062848454130766,
|
| 5124 |
+
"grad_norm": 0.47191551327705383,
|
| 5125 |
+
"learning_rate": 0.00013947069534141904,
|
| 5126 |
+
"loss": 4.4233543395996096,
|
| 5127 |
+
"step": 7140
|
| 5128 |
+
},
|
| 5129 |
+
{
|
| 5130 |
+
"epoch": 1.207974319986484,
|
| 5131 |
+
"grad_norm": 0.481503427028656,
|
| 5132 |
+
"learning_rate": 0.00013899293272595355,
|
| 5133 |
+
"loss": 4.440040969848633,
|
| 5134 |
+
"step": 7150
|
| 5135 |
+
},
|
| 5136 |
+
{
|
| 5137 |
+
"epoch": 1.2096637945598918,
|
| 5138 |
+
"grad_norm": 0.47193533182144165,
|
| 5139 |
+
"learning_rate": 0.0001385152823529957,
|
| 5140 |
+
"loss": 4.411207962036133,
|
| 5141 |
+
"step": 7160
|
| 5142 |
+
},
|
| 5143 |
+
{
|
| 5144 |
+
"epoch": 1.2113532691332995,
|
| 5145 |
+
"grad_norm": 0.4883231520652771,
|
| 5146 |
+
"learning_rate": 0.00013803774909329567,
|
| 5147 |
+
"loss": 4.404615020751953,
|
| 5148 |
+
"step": 7170
|
| 5149 |
+
},
|
| 5150 |
+
{
|
| 5151 |
+
"epoch": 1.2130427437067073,
|
| 5152 |
+
"grad_norm": 0.47556227445602417,
|
| 5153 |
+
"learning_rate": 0.0001375603378164095,
|
| 5154 |
+
"loss": 4.381325149536133,
|
| 5155 |
+
"step": 7180
|
| 5156 |
+
},
|
| 5157 |
+
{
|
| 5158 |
+
"epoch": 1.2147322182801148,
|
| 5159 |
+
"grad_norm": 0.47802919149398804,
|
| 5160 |
+
"learning_rate": 0.00013708305339064933,
|
| 5161 |
+
"loss": 4.414505386352539,
|
| 5162 |
+
"step": 7190
|
| 5163 |
+
},
|
| 5164 |
+
{
|
| 5165 |
+
"epoch": 1.2164216928535225,
|
| 5166 |
+
"grad_norm": 0.4807351529598236,
|
| 5167 |
+
"learning_rate": 0.00013660590068303373,
|
| 5168 |
+
"loss": 4.441463470458984,
|
| 5169 |
+
"step": 7200
|
| 5170 |
+
},
|
| 5171 |
+
{
|
| 5172 |
+
"epoch": 1.2181111674269303,
|
| 5173 |
+
"grad_norm": 0.5178191065788269,
|
| 5174 |
+
"learning_rate": 0.00013612888455923804,
|
| 5175 |
+
"loss": 4.416116714477539,
|
| 5176 |
+
"step": 7210
|
| 5177 |
+
},
|
| 5178 |
+
{
|
| 5179 |
+
"epoch": 1.219800642000338,
|
| 5180 |
+
"grad_norm": 0.48551276326179504,
|
| 5181 |
+
"learning_rate": 0.0001356520098835449,
|
| 5182 |
+
"loss": 4.4417884826660154,
|
| 5183 |
+
"step": 7220
|
| 5184 |
+
},
|
| 5185 |
+
{
|
| 5186 |
+
"epoch": 1.2214901165737455,
|
| 5187 |
+
"grad_norm": 0.4903479218482971,
|
| 5188 |
+
"learning_rate": 0.00013517528151879457,
|
| 5189 |
+
"loss": 4.411055374145508,
|
| 5190 |
+
"step": 7230
|
| 5191 |
+
},
|
| 5192 |
+
{
|
| 5193 |
+
"epoch": 1.2231795911471532,
|
| 5194 |
+
"grad_norm": 0.475599080324173,
|
| 5195 |
+
"learning_rate": 0.0001346987043263352,
|
| 5196 |
+
"loss": 4.41358642578125,
|
| 5197 |
+
"step": 7240
|
| 5198 |
+
},
|
| 5199 |
+
{
|
| 5200 |
+
"epoch": 1.224869065720561,
|
| 5201 |
+
"grad_norm": 0.5013647079467773,
|
| 5202 |
+
"learning_rate": 0.00013422228316597356,
|
| 5203 |
+
"loss": 4.4313819885253904,
|
| 5204 |
+
"step": 7250
|
| 5205 |
+
},
|
| 5206 |
+
{
|
| 5207 |
+
"epoch": 1.2265585402939685,
|
| 5208 |
+
"grad_norm": 0.4796617031097412,
|
| 5209 |
+
"learning_rate": 0.00013374602289592508,
|
| 5210 |
+
"loss": 4.431335830688477,
|
| 5211 |
+
"step": 7260
|
| 5212 |
+
},
|
| 5213 |
+
{
|
| 5214 |
+
"epoch": 1.2282480148673762,
|
| 5215 |
+
"grad_norm": 0.5207120180130005,
|
| 5216 |
+
"learning_rate": 0.00013326992837276494,
|
| 5217 |
+
"loss": 4.4034477233886715,
|
| 5218 |
+
"step": 7270
|
| 5219 |
+
},
|
| 5220 |
+
{
|
| 5221 |
+
"epoch": 1.229937489440784,
|
| 5222 |
+
"grad_norm": 0.48880499601364136,
|
| 5223 |
+
"learning_rate": 0.0001327940044513777,
|
| 5224 |
+
"loss": 4.412507629394531,
|
| 5225 |
+
"step": 7280
|
| 5226 |
+
},
|
| 5227 |
+
{
|
| 5228 |
+
"epoch": 1.2316269640141915,
|
| 5229 |
+
"grad_norm": 0.4934345483779907,
|
| 5230 |
+
"learning_rate": 0.00013231825598490854,
|
| 5231 |
+
"loss": 4.4090087890625,
|
| 5232 |
+
"step": 7290
|
| 5233 |
+
},
|
| 5234 |
+
{
|
| 5235 |
+
"epoch": 1.2333164385875992,
|
| 5236 |
+
"grad_norm": 0.4781767427921295,
|
| 5237 |
+
"learning_rate": 0.0001318426878247133,
|
| 5238 |
+
"loss": 4.418943786621094,
|
| 5239 |
+
"step": 7300
|
| 5240 |
+
},
|
| 5241 |
+
{
|
| 5242 |
+
"epoch": 1.235005913161007,
|
| 5243 |
+
"grad_norm": 0.5096566081047058,
|
| 5244 |
+
"learning_rate": 0.00013136730482030928,
|
| 5245 |
+
"loss": 4.422417449951172,
|
| 5246 |
+
"step": 7310
|
| 5247 |
+
},
|
| 5248 |
+
{
|
| 5249 |
+
"epoch": 1.2366953877344147,
|
| 5250 |
+
"grad_norm": 0.4950323700904846,
|
| 5251 |
+
"learning_rate": 0.0001308921118193257,
|
| 5252 |
+
"loss": 4.429974746704102,
|
| 5253 |
+
"step": 7320
|
| 5254 |
+
},
|
| 5255 |
+
{
|
| 5256 |
+
"epoch": 1.2383848623078222,
|
| 5257 |
+
"grad_norm": 0.513830304145813,
|
| 5258 |
+
"learning_rate": 0.00013041711366745408,
|
| 5259 |
+
"loss": 4.422944641113281,
|
| 5260 |
+
"step": 7330
|
| 5261 |
+
},
|
| 5262 |
+
{
|
| 5263 |
+
"epoch": 1.24007433688123,
|
| 5264 |
+
"grad_norm": 0.48782795667648315,
|
| 5265 |
+
"learning_rate": 0.00012994231520839934,
|
| 5266 |
+
"loss": 4.427762222290039,
|
| 5267 |
+
"step": 7340
|
| 5268 |
+
},
|
| 5269 |
+
{
|
| 5270 |
+
"epoch": 1.2417638114546377,
|
| 5271 |
+
"grad_norm": 0.4940618872642517,
|
| 5272 |
+
"learning_rate": 0.0001294677212838297,
|
| 5273 |
+
"loss": 4.397305297851562,
|
| 5274 |
+
"step": 7350
|
| 5275 |
+
},
|
| 5276 |
+
{
|
| 5277 |
+
"epoch": 1.2434532860280454,
|
| 5278 |
+
"grad_norm": 0.47247400879859924,
|
| 5279 |
+
"learning_rate": 0.00012899333673332795,
|
| 5280 |
+
"loss": 4.438782501220703,
|
| 5281 |
+
"step": 7360
|
| 5282 |
+
},
|
| 5283 |
+
{
|
| 5284 |
+
"epoch": 1.245142760601453,
|
| 5285 |
+
"grad_norm": 0.508912205696106,
|
| 5286 |
+
"learning_rate": 0.00012851916639434164,
|
| 5287 |
+
"loss": 4.381727600097657,
|
| 5288 |
+
"step": 7370
|
| 5289 |
+
},
|
| 5290 |
+
{
|
| 5291 |
+
"epoch": 1.2468322351748606,
|
| 5292 |
+
"grad_norm": 0.47959408164024353,
|
| 5293 |
+
"learning_rate": 0.00012804521510213407,
|
| 5294 |
+
"loss": 4.4016876220703125,
|
| 5295 |
+
"step": 7380
|
| 5296 |
+
},
|
| 5297 |
+
{
|
| 5298 |
+
"epoch": 1.2485217097482684,
|
| 5299 |
+
"grad_norm": 0.4996829330921173,
|
| 5300 |
+
"learning_rate": 0.00012757148768973483,
|
| 5301 |
+
"loss": 4.403145217895508,
|
| 5302 |
+
"step": 7390
|
| 5303 |
+
},
|
| 5304 |
+
{
|
| 5305 |
+
"epoch": 1.2502111843216759,
|
| 5306 |
+
"grad_norm": 0.4780764877796173,
|
| 5307 |
+
"learning_rate": 0.00012709798898789042,
|
| 5308 |
+
"loss": 4.438173294067383,
|
| 5309 |
+
"step": 7400
|
| 5310 |
+
},
|
| 5311 |
+
{
|
| 5312 |
+
"epoch": 1.2519006588950836,
|
| 5313 |
+
"grad_norm": 0.4783228039741516,
|
| 5314 |
+
"learning_rate": 0.00012662472382501524,
|
| 5315 |
+
"loss": 4.408795928955078,
|
| 5316 |
+
"step": 7410
|
| 5317 |
+
},
|
| 5318 |
+
{
|
| 5319 |
+
"epoch": 1.2535901334684914,
|
| 5320 |
+
"grad_norm": 0.5030134320259094,
|
| 5321 |
+
"learning_rate": 0.0001261516970271422,
|
| 5322 |
+
"loss": 4.420645523071289,
|
| 5323 |
+
"step": 7420
|
| 5324 |
+
},
|
| 5325 |
+
{
|
| 5326 |
+
"epoch": 1.2552796080418989,
|
| 5327 |
+
"grad_norm": 0.4808700978755951,
|
| 5328 |
+
"learning_rate": 0.0001256789134178735,
|
| 5329 |
+
"loss": 4.436854553222656,
|
| 5330 |
+
"step": 7430
|
| 5331 |
+
},
|
| 5332 |
+
{
|
| 5333 |
+
"epoch": 1.2569690826153066,
|
| 5334 |
+
"grad_norm": 0.493521511554718,
|
| 5335 |
+
"learning_rate": 0.00012520637781833144,
|
| 5336 |
+
"loss": 4.406418991088867,
|
| 5337 |
+
"step": 7440
|
| 5338 |
+
},
|
| 5339 |
+
{
|
| 5340 |
+
"epoch": 1.2586585571887143,
|
| 5341 |
+
"grad_norm": 0.4652424454689026,
|
| 5342 |
+
"learning_rate": 0.0001247340950471094,
|
| 5343 |
+
"loss": 4.417532348632813,
|
| 5344 |
+
"step": 7450
|
| 5345 |
+
},
|
| 5346 |
+
{
|
| 5347 |
+
"epoch": 1.2603480317621218,
|
| 5348 |
+
"grad_norm": 0.5171105265617371,
|
| 5349 |
+
"learning_rate": 0.0001242620699202224,
|
| 5350 |
+
"loss": 4.399831390380859,
|
| 5351 |
+
"step": 7460
|
| 5352 |
+
},
|
| 5353 |
+
{
|
| 5354 |
+
"epoch": 1.2620375063355296,
|
| 5355 |
+
"grad_norm": 0.49303922057151794,
|
| 5356 |
+
"learning_rate": 0.00012379030725105837,
|
| 5357 |
+
"loss": 4.401795959472656,
|
| 5358 |
+
"step": 7470
|
| 5359 |
+
},
|
| 5360 |
+
{
|
| 5361 |
+
"epoch": 1.2637269809089373,
|
| 5362 |
+
"grad_norm": 0.501957893371582,
|
| 5363 |
+
"learning_rate": 0.00012331881185032872,
|
| 5364 |
+
"loss": 4.388158798217773,
|
| 5365 |
+
"step": 7480
|
| 5366 |
+
},
|
| 5367 |
+
{
|
| 5368 |
+
"epoch": 1.265416455482345,
|
| 5369 |
+
"grad_norm": 0.5417965650558472,
|
| 5370 |
+
"learning_rate": 0.00012284758852601962,
|
| 5371 |
+
"loss": 4.410275650024414,
|
| 5372 |
+
"step": 7490
|
| 5373 |
+
},
|
| 5374 |
+
{
|
| 5375 |
+
"epoch": 1.2671059300557528,
|
| 5376 |
+
"grad_norm": 0.5049046277999878,
|
| 5377 |
+
"learning_rate": 0.00012237664208334263,
|
| 5378 |
+
"loss": 4.402982330322265,
|
| 5379 |
+
"step": 7500
|
| 5380 |
+
},
|
| 5381 |
+
{
|
| 5382 |
+
"epoch": 1.2671059300557528,
|
| 5383 |
+
"eval_loss": 4.3625264167785645,
|
| 5384 |
+
"eval_runtime": 3.7028,
|
| 5385 |
+
"eval_samples_per_second": 270.066,
|
| 5386 |
+
"eval_steps_per_second": 5.671,
|
| 5387 |
+
"step": 7500
|
| 5388 |
}
|
| 5389 |
],
|
| 5390 |
"logging_steps": 10,
|
|
|
|
| 5404 |
"attributes": {}
|
| 5405 |
}
|
| 5406 |
},
|
| 5407 |
+
"total_flos": 2.5084052528902963e+17,
|
| 5408 |
"train_batch_size": 48,
|
| 5409 |
"trial_name": null,
|
| 5410 |
"trial_params": null
|