Training in progress, step 400, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 671149168
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8198e9317f349ff3fd1e2f22f5f0e02e0d464b3b2d14b422bd2542cc62793a64
|
| 3 |
size 671149168
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 341314644
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c6174d7213eebb72e1fe3ca63d3d72d9fd39633ebe8d894d7532140049e8df32
|
| 3 |
size 341314644
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9fb82bc159a82b428f124a5ab73b5266e9f4847bee821a1e25a1056b8c500ae4
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7e99fbb72294539e29c605dada5b1e203613200e6c1e835a4f2f4baed9ec2604
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 100,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2139,6 +2139,714 @@
|
|
| 2139 |
"eval_samples_per_second": 12.722,
|
| 2140 |
"eval_steps_per_second": 3.18,
|
| 2141 |
"step": 300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2142 |
}
|
| 2143 |
],
|
| 2144 |
"logging_steps": 1,
|
|
@@ -2162,12 +2870,12 @@
|
|
| 2162 |
"should_evaluate": false,
|
| 2163 |
"should_log": false,
|
| 2164 |
"should_save": true,
|
| 2165 |
-
"should_training_stop":
|
| 2166 |
},
|
| 2167 |
"attributes": {}
|
| 2168 |
}
|
| 2169 |
},
|
| 2170 |
-
"total_flos":
|
| 2171 |
"train_batch_size": 8,
|
| 2172 |
"trial_name": null,
|
| 2173 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.3466954827308655,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-400",
|
| 4 |
+
"epoch": 0.3137254901960784,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 400,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2139 |
"eval_samples_per_second": 12.722,
|
| 2140 |
"eval_steps_per_second": 3.18,
|
| 2141 |
"step": 300
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"epoch": 0.23607843137254902,
|
| 2145 |
+
"grad_norm": 2.4137773513793945,
|
| 2146 |
+
"learning_rate": 1.0552102216287209e-05,
|
| 2147 |
+
"loss": 3.5898,
|
| 2148 |
+
"step": 301
|
| 2149 |
+
},
|
| 2150 |
+
{
|
| 2151 |
+
"epoch": 0.23686274509803923,
|
| 2152 |
+
"grad_norm": 2.2105584144592285,
|
| 2153 |
+
"learning_rate": 1.0351143194462377e-05,
|
| 2154 |
+
"loss": 1.8317,
|
| 2155 |
+
"step": 302
|
| 2156 |
+
},
|
| 2157 |
+
{
|
| 2158 |
+
"epoch": 0.2376470588235294,
|
| 2159 |
+
"grad_norm": 2.3090407848358154,
|
| 2160 |
+
"learning_rate": 1.0151783601632461e-05,
|
| 2161 |
+
"loss": 1.5182,
|
| 2162 |
+
"step": 303
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 0.2384313725490196,
|
| 2166 |
+
"grad_norm": 2.195463180541992,
|
| 2167 |
+
"learning_rate": 9.954036373955764e-06,
|
| 2168 |
+
"loss": 1.2783,
|
| 2169 |
+
"step": 304
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 0.23921568627450981,
|
| 2173 |
+
"grad_norm": 2.6486854553222656,
|
| 2174 |
+
"learning_rate": 9.757914342966493e-06,
|
| 2175 |
+
"loss": 1.4084,
|
| 2176 |
+
"step": 305
|
| 2177 |
+
},
|
| 2178 |
+
{
|
| 2179 |
+
"epoch": 0.24,
|
| 2180 |
+
"grad_norm": 3.396476984024048,
|
| 2181 |
+
"learning_rate": 9.563430234742194e-06,
|
| 2182 |
+
"loss": 1.3099,
|
| 2183 |
+
"step": 306
|
| 2184 |
+
},
|
| 2185 |
+
{
|
| 2186 |
+
"epoch": 0.2407843137254902,
|
| 2187 |
+
"grad_norm": 2.8271055221557617,
|
| 2188 |
+
"learning_rate": 9.370596669077936e-06,
|
| 2189 |
+
"loss": 1.3866,
|
| 2190 |
+
"step": 307
|
| 2191 |
+
},
|
| 2192 |
+
{
|
| 2193 |
+
"epoch": 0.2415686274509804,
|
| 2194 |
+
"grad_norm": 2.6021790504455566,
|
| 2195 |
+
"learning_rate": 9.179426158667445e-06,
|
| 2196 |
+
"loss": 1.1333,
|
| 2197 |
+
"step": 308
|
| 2198 |
+
},
|
| 2199 |
+
{
|
| 2200 |
+
"epoch": 0.24235294117647058,
|
| 2201 |
+
"grad_norm": 2.9245004653930664,
|
| 2202 |
+
"learning_rate": 8.989931108291197e-06,
|
| 2203 |
+
"loss": 1.0958,
|
| 2204 |
+
"step": 309
|
| 2205 |
+
},
|
| 2206 |
+
{
|
| 2207 |
+
"epoch": 0.24313725490196078,
|
| 2208 |
+
"grad_norm": 2.7905077934265137,
|
| 2209 |
+
"learning_rate": 8.802123814011457e-06,
|
| 2210 |
+
"loss": 1.0016,
|
| 2211 |
+
"step": 310
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 0.24392156862745099,
|
| 2215 |
+
"grad_norm": 3.4618003368377686,
|
| 2216 |
+
"learning_rate": 8.616016462374414e-06,
|
| 2217 |
+
"loss": 1.2777,
|
| 2218 |
+
"step": 311
|
| 2219 |
+
},
|
| 2220 |
+
{
|
| 2221 |
+
"epoch": 0.2447058823529412,
|
| 2222 |
+
"grad_norm": 4.318086624145508,
|
| 2223 |
+
"learning_rate": 8.431621129619399e-06,
|
| 2224 |
+
"loss": 1.139,
|
| 2225 |
+
"step": 312
|
| 2226 |
+
},
|
| 2227 |
+
{
|
| 2228 |
+
"epoch": 0.24549019607843137,
|
| 2229 |
+
"grad_norm": 4.268404483795166,
|
| 2230 |
+
"learning_rate": 8.248949780895306e-06,
|
| 2231 |
+
"loss": 1.6534,
|
| 2232 |
+
"step": 313
|
| 2233 |
+
},
|
| 2234 |
+
{
|
| 2235 |
+
"epoch": 0.24627450980392157,
|
| 2236 |
+
"grad_norm": 3.4030003547668457,
|
| 2237 |
+
"learning_rate": 8.068014269484189e-06,
|
| 2238 |
+
"loss": 1.146,
|
| 2239 |
+
"step": 314
|
| 2240 |
+
},
|
| 2241 |
+
{
|
| 2242 |
+
"epoch": 0.24705882352941178,
|
| 2243 |
+
"grad_norm": 10.152856826782227,
|
| 2244 |
+
"learning_rate": 7.888826336032093e-06,
|
| 2245 |
+
"loss": 1.53,
|
| 2246 |
+
"step": 315
|
| 2247 |
+
},
|
| 2248 |
+
{
|
| 2249 |
+
"epoch": 0.24784313725490195,
|
| 2250 |
+
"grad_norm": 3.9469947814941406,
|
| 2251 |
+
"learning_rate": 7.71139760778722e-06,
|
| 2252 |
+
"loss": 1.47,
|
| 2253 |
+
"step": 316
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 0.24862745098039216,
|
| 2257 |
+
"grad_norm": 4.744470596313477,
|
| 2258 |
+
"learning_rate": 7.535739597845489e-06,
|
| 2259 |
+
"loss": 1.4464,
|
| 2260 |
+
"step": 317
|
| 2261 |
+
},
|
| 2262 |
+
{
|
| 2263 |
+
"epoch": 0.24941176470588236,
|
| 2264 |
+
"grad_norm": 5.53031063079834,
|
| 2265 |
+
"learning_rate": 7.361863704403427e-06,
|
| 2266 |
+
"loss": 1.1041,
|
| 2267 |
+
"step": 318
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 0.25019607843137254,
|
| 2271 |
+
"grad_norm": 4.436324596405029,
|
| 2272 |
+
"learning_rate": 7.189781210018604e-06,
|
| 2273 |
+
"loss": 1.3396,
|
| 2274 |
+
"step": 319
|
| 2275 |
+
},
|
| 2276 |
+
{
|
| 2277 |
+
"epoch": 0.25098039215686274,
|
| 2278 |
+
"grad_norm": 5.226994037628174,
|
| 2279 |
+
"learning_rate": 7.019503280877466e-06,
|
| 2280 |
+
"loss": 1.4027,
|
| 2281 |
+
"step": 320
|
| 2282 |
+
},
|
| 2283 |
+
{
|
| 2284 |
+
"epoch": 0.25176470588235295,
|
| 2285 |
+
"grad_norm": 4.095127582550049,
|
| 2286 |
+
"learning_rate": 6.851040966070809e-06,
|
| 2287 |
+
"loss": 1.0458,
|
| 2288 |
+
"step": 321
|
| 2289 |
+
},
|
| 2290 |
+
{
|
| 2291 |
+
"epoch": 0.25254901960784315,
|
| 2292 |
+
"grad_norm": 3.7459771633148193,
|
| 2293 |
+
"learning_rate": 6.684405196876842e-06,
|
| 2294 |
+
"loss": 0.7628,
|
| 2295 |
+
"step": 322
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 0.25333333333333335,
|
| 2299 |
+
"grad_norm": 4.650934219360352,
|
| 2300 |
+
"learning_rate": 6.519606786051812e-06,
|
| 2301 |
+
"loss": 0.9564,
|
| 2302 |
+
"step": 323
|
| 2303 |
+
},
|
| 2304 |
+
{
|
| 2305 |
+
"epoch": 0.2541176470588235,
|
| 2306 |
+
"grad_norm": 5.133500576019287,
|
| 2307 |
+
"learning_rate": 6.356656427128452e-06,
|
| 2308 |
+
"loss": 1.198,
|
| 2309 |
+
"step": 324
|
| 2310 |
+
},
|
| 2311 |
+
{
|
| 2312 |
+
"epoch": 0.2549019607843137,
|
| 2313 |
+
"grad_norm": 4.67864990234375,
|
| 2314 |
+
"learning_rate": 6.195564693722027e-06,
|
| 2315 |
+
"loss": 1.1566,
|
| 2316 |
+
"step": 325
|
| 2317 |
+
},
|
| 2318 |
+
{
|
| 2319 |
+
"epoch": 0.2556862745098039,
|
| 2320 |
+
"grad_norm": 6.105125427246094,
|
| 2321 |
+
"learning_rate": 6.0363420388442584e-06,
|
| 2322 |
+
"loss": 1.2762,
|
| 2323 |
+
"step": 326
|
| 2324 |
+
},
|
| 2325 |
+
{
|
| 2326 |
+
"epoch": 0.2564705882352941,
|
| 2327 |
+
"grad_norm": 5.678111553192139,
|
| 2328 |
+
"learning_rate": 5.878998794225053e-06,
|
| 2329 |
+
"loss": 0.9718,
|
| 2330 |
+
"step": 327
|
| 2331 |
+
},
|
| 2332 |
+
{
|
| 2333 |
+
"epoch": 0.2572549019607843,
|
| 2334 |
+
"grad_norm": 7.232873916625977,
|
| 2335 |
+
"learning_rate": 5.723545169642064e-06,
|
| 2336 |
+
"loss": 1.1458,
|
| 2337 |
+
"step": 328
|
| 2338 |
+
},
|
| 2339 |
+
{
|
| 2340 |
+
"epoch": 0.2580392156862745,
|
| 2341 |
+
"grad_norm": 4.871210098266602,
|
| 2342 |
+
"learning_rate": 5.569991252258227e-06,
|
| 2343 |
+
"loss": 0.7986,
|
| 2344 |
+
"step": 329
|
| 2345 |
+
},
|
| 2346 |
+
{
|
| 2347 |
+
"epoch": 0.25882352941176473,
|
| 2348 |
+
"grad_norm": 4.941523551940918,
|
| 2349 |
+
"learning_rate": 5.418347005967188e-06,
|
| 2350 |
+
"loss": 0.9478,
|
| 2351 |
+
"step": 330
|
| 2352 |
+
},
|
| 2353 |
+
{
|
| 2354 |
+
"epoch": 0.2596078431372549,
|
| 2355 |
+
"grad_norm": 5.843031406402588,
|
| 2356 |
+
"learning_rate": 5.268622270746769e-06,
|
| 2357 |
+
"loss": 1.3708,
|
| 2358 |
+
"step": 331
|
| 2359 |
+
},
|
| 2360 |
+
{
|
| 2361 |
+
"epoch": 0.2603921568627451,
|
| 2362 |
+
"grad_norm": 5.538339614868164,
|
| 2363 |
+
"learning_rate": 5.120826762020479e-06,
|
| 2364 |
+
"loss": 1.1192,
|
| 2365 |
+
"step": 332
|
| 2366 |
+
},
|
| 2367 |
+
{
|
| 2368 |
+
"epoch": 0.2611764705882353,
|
| 2369 |
+
"grad_norm": 7.33754825592041,
|
| 2370 |
+
"learning_rate": 4.974970070027069e-06,
|
| 2371 |
+
"loss": 1.2714,
|
| 2372 |
+
"step": 333
|
| 2373 |
+
},
|
| 2374 |
+
{
|
| 2375 |
+
"epoch": 0.2619607843137255,
|
| 2376 |
+
"grad_norm": 5.050201892852783,
|
| 2377 |
+
"learning_rate": 4.8310616591982745e-06,
|
| 2378 |
+
"loss": 0.9628,
|
| 2379 |
+
"step": 334
|
| 2380 |
+
},
|
| 2381 |
+
{
|
| 2382 |
+
"epoch": 0.2627450980392157,
|
| 2383 |
+
"grad_norm": 5.330129623413086,
|
| 2384 |
+
"learning_rate": 4.689110867544645e-06,
|
| 2385 |
+
"loss": 0.9378,
|
| 2386 |
+
"step": 335
|
| 2387 |
+
},
|
| 2388 |
+
{
|
| 2389 |
+
"epoch": 0.2635294117647059,
|
| 2390 |
+
"grad_norm": 5.494550704956055,
|
| 2391 |
+
"learning_rate": 4.549126906049629e-06,
|
| 2392 |
+
"loss": 1.3844,
|
| 2393 |
+
"step": 336
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 0.2643137254901961,
|
| 2397 |
+
"grad_norm": 7.918943405151367,
|
| 2398 |
+
"learning_rate": 4.411118858071873e-06,
|
| 2399 |
+
"loss": 1.6752,
|
| 2400 |
+
"step": 337
|
| 2401 |
+
},
|
| 2402 |
+
{
|
| 2403 |
+
"epoch": 0.26509803921568625,
|
| 2404 |
+
"grad_norm": 8.494551658630371,
|
| 2405 |
+
"learning_rate": 4.275095678755838e-06,
|
| 2406 |
+
"loss": 1.3365,
|
| 2407 |
+
"step": 338
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 0.26588235294117646,
|
| 2411 |
+
"grad_norm": 6.204376697540283,
|
| 2412 |
+
"learning_rate": 4.141066194450712e-06,
|
| 2413 |
+
"loss": 1.2237,
|
| 2414 |
+
"step": 339
|
| 2415 |
+
},
|
| 2416 |
+
{
|
| 2417 |
+
"epoch": 0.26666666666666666,
|
| 2418 |
+
"grad_norm": 10.700553894042969,
|
| 2419 |
+
"learning_rate": 4.009039102137657e-06,
|
| 2420 |
+
"loss": 2.0144,
|
| 2421 |
+
"step": 340
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 0.26745098039215687,
|
| 2425 |
+
"grad_norm": 8.845378875732422,
|
| 2426 |
+
"learning_rate": 3.879022968865484e-06,
|
| 2427 |
+
"loss": 2.1293,
|
| 2428 |
+
"step": 341
|
| 2429 |
+
},
|
| 2430 |
+
{
|
| 2431 |
+
"epoch": 0.26823529411764707,
|
| 2432 |
+
"grad_norm": 8.239388465881348,
|
| 2433 |
+
"learning_rate": 3.751026231194761e-06,
|
| 2434 |
+
"loss": 1.779,
|
| 2435 |
+
"step": 342
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 0.2690196078431373,
|
| 2439 |
+
"grad_norm": 9.209283828735352,
|
| 2440 |
+
"learning_rate": 3.6250571946503677e-06,
|
| 2441 |
+
"loss": 1.6962,
|
| 2442 |
+
"step": 343
|
| 2443 |
+
},
|
| 2444 |
+
{
|
| 2445 |
+
"epoch": 0.2698039215686274,
|
| 2446 |
+
"grad_norm": 9.209612846374512,
|
| 2447 |
+
"learning_rate": 3.5011240331825784e-06,
|
| 2448 |
+
"loss": 1.7011,
|
| 2449 |
+
"step": 344
|
| 2450 |
+
},
|
| 2451 |
+
{
|
| 2452 |
+
"epoch": 0.27058823529411763,
|
| 2453 |
+
"grad_norm": 10.101373672485352,
|
| 2454 |
+
"learning_rate": 3.379234788636626e-06,
|
| 2455 |
+
"loss": 1.532,
|
| 2456 |
+
"step": 345
|
| 2457 |
+
},
|
| 2458 |
+
{
|
| 2459 |
+
"epoch": 0.27137254901960783,
|
| 2460 |
+
"grad_norm": 11.359880447387695,
|
| 2461 |
+
"learning_rate": 3.2593973702309106e-06,
|
| 2462 |
+
"loss": 2.0318,
|
| 2463 |
+
"step": 346
|
| 2464 |
+
},
|
| 2465 |
+
{
|
| 2466 |
+
"epoch": 0.27215686274509804,
|
| 2467 |
+
"grad_norm": 10.647089958190918,
|
| 2468 |
+
"learning_rate": 3.1416195540437896e-06,
|
| 2469 |
+
"loss": 1.9065,
|
| 2470 |
+
"step": 347
|
| 2471 |
+
},
|
| 2472 |
+
{
|
| 2473 |
+
"epoch": 0.27294117647058824,
|
| 2474 |
+
"grad_norm": 15.720868110656738,
|
| 2475 |
+
"learning_rate": 3.0259089825089655e-06,
|
| 2476 |
+
"loss": 1.8493,
|
| 2477 |
+
"step": 348
|
| 2478 |
+
},
|
| 2479 |
+
{
|
| 2480 |
+
"epoch": 0.27372549019607845,
|
| 2481 |
+
"grad_norm": 13.961613655090332,
|
| 2482 |
+
"learning_rate": 2.9122731639196154e-06,
|
| 2483 |
+
"loss": 2.564,
|
| 2484 |
+
"step": 349
|
| 2485 |
+
},
|
| 2486 |
+
{
|
| 2487 |
+
"epoch": 0.27450980392156865,
|
| 2488 |
+
"grad_norm": 12.990523338317871,
|
| 2489 |
+
"learning_rate": 2.800719471941152e-06,
|
| 2490 |
+
"loss": 1.8791,
|
| 2491 |
+
"step": 350
|
| 2492 |
+
},
|
| 2493 |
+
{
|
| 2494 |
+
"epoch": 0.2752941176470588,
|
| 2495 |
+
"grad_norm": 2.013960123062134,
|
| 2496 |
+
"learning_rate": 2.691255145132793e-06,
|
| 2497 |
+
"loss": 2.7899,
|
| 2498 |
+
"step": 351
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 0.276078431372549,
|
| 2502 |
+
"grad_norm": 1.9573743343353271,
|
| 2503 |
+
"learning_rate": 2.5838872864778276e-06,
|
| 2504 |
+
"loss": 1.9973,
|
| 2505 |
+
"step": 352
|
| 2506 |
+
},
|
| 2507 |
+
{
|
| 2508 |
+
"epoch": 0.2768627450980392,
|
| 2509 |
+
"grad_norm": 2.0770833492279053,
|
| 2510 |
+
"learning_rate": 2.478622862922738e-06,
|
| 2511 |
+
"loss": 1.5119,
|
| 2512 |
+
"step": 353
|
| 2513 |
+
},
|
| 2514 |
+
{
|
| 2515 |
+
"epoch": 0.2776470588235294,
|
| 2516 |
+
"grad_norm": 2.2673606872558594,
|
| 2517 |
+
"learning_rate": 2.375468704925131e-06,
|
| 2518 |
+
"loss": 1.473,
|
| 2519 |
+
"step": 354
|
| 2520 |
+
},
|
| 2521 |
+
{
|
| 2522 |
+
"epoch": 0.2784313725490196,
|
| 2523 |
+
"grad_norm": 2.035940408706665,
|
| 2524 |
+
"learning_rate": 2.2744315060104845e-06,
|
| 2525 |
+
"loss": 1.1183,
|
| 2526 |
+
"step": 355
|
| 2527 |
+
},
|
| 2528 |
+
{
|
| 2529 |
+
"epoch": 0.2792156862745098,
|
| 2530 |
+
"grad_norm": 2.1854915618896484,
|
| 2531 |
+
"learning_rate": 2.175517822337849e-06,
|
| 2532 |
+
"loss": 1.1662,
|
| 2533 |
+
"step": 356
|
| 2534 |
+
},
|
| 2535 |
+
{
|
| 2536 |
+
"epoch": 0.28,
|
| 2537 |
+
"grad_norm": 2.2467715740203857,
|
| 2538 |
+
"learning_rate": 2.0787340722744e-06,
|
| 2539 |
+
"loss": 0.9999,
|
| 2540 |
+
"step": 357
|
| 2541 |
+
},
|
| 2542 |
+
{
|
| 2543 |
+
"epoch": 0.2807843137254902,
|
| 2544 |
+
"grad_norm": 2.530546188354492,
|
| 2545 |
+
"learning_rate": 1.9840865359789894e-06,
|
| 2546 |
+
"loss": 1.1862,
|
| 2547 |
+
"step": 358
|
| 2548 |
+
},
|
| 2549 |
+
{
|
| 2550 |
+
"epoch": 0.2815686274509804,
|
| 2551 |
+
"grad_norm": 3.1195712089538574,
|
| 2552 |
+
"learning_rate": 1.891581354994618e-06,
|
| 2553 |
+
"loss": 1.2027,
|
| 2554 |
+
"step": 359
|
| 2555 |
+
},
|
| 2556 |
+
{
|
| 2557 |
+
"epoch": 0.2823529411764706,
|
| 2558 |
+
"grad_norm": 3.0070858001708984,
|
| 2559 |
+
"learning_rate": 1.801224531849908e-06,
|
| 2560 |
+
"loss": 1.0958,
|
| 2561 |
+
"step": 360
|
| 2562 |
+
},
|
| 2563 |
+
{
|
| 2564 |
+
"epoch": 0.2831372549019608,
|
| 2565 |
+
"grad_norm": 3.387432813644409,
|
| 2566 |
+
"learning_rate": 1.7130219296696263e-06,
|
| 2567 |
+
"loss": 1.3352,
|
| 2568 |
+
"step": 361
|
| 2569 |
+
},
|
| 2570 |
+
{
|
| 2571 |
+
"epoch": 0.283921568627451,
|
| 2572 |
+
"grad_norm": 3.0927343368530273,
|
| 2573 |
+
"learning_rate": 1.6269792717942316e-06,
|
| 2574 |
+
"loss": 1.222,
|
| 2575 |
+
"step": 362
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"epoch": 0.2847058823529412,
|
| 2579 |
+
"grad_norm": 3.4728779792785645,
|
| 2580 |
+
"learning_rate": 1.5431021414084833e-06,
|
| 2581 |
+
"loss": 1.4163,
|
| 2582 |
+
"step": 363
|
| 2583 |
+
},
|
| 2584 |
+
{
|
| 2585 |
+
"epoch": 0.28549019607843135,
|
| 2586 |
+
"grad_norm": 3.3711791038513184,
|
| 2587 |
+
"learning_rate": 1.461395981179182e-06,
|
| 2588 |
+
"loss": 1.1301,
|
| 2589 |
+
"step": 364
|
| 2590 |
+
},
|
| 2591 |
+
{
|
| 2592 |
+
"epoch": 0.28627450980392155,
|
| 2593 |
+
"grad_norm": 3.495633602142334,
|
| 2594 |
+
"learning_rate": 1.3818660929019714e-06,
|
| 2595 |
+
"loss": 1.2015,
|
| 2596 |
+
"step": 365
|
| 2597 |
+
},
|
| 2598 |
+
{
|
| 2599 |
+
"epoch": 0.28705882352941176,
|
| 2600 |
+
"grad_norm": 4.38746452331543,
|
| 2601 |
+
"learning_rate": 1.3045176371573263e-06,
|
| 2602 |
+
"loss": 1.3248,
|
| 2603 |
+
"step": 366
|
| 2604 |
+
},
|
| 2605 |
+
{
|
| 2606 |
+
"epoch": 0.28784313725490196,
|
| 2607 |
+
"grad_norm": 4.259886741638184,
|
| 2608 |
+
"learning_rate": 1.2293556329757026e-06,
|
| 2609 |
+
"loss": 1.512,
|
| 2610 |
+
"step": 367
|
| 2611 |
+
},
|
| 2612 |
+
{
|
| 2613 |
+
"epoch": 0.28862745098039216,
|
| 2614 |
+
"grad_norm": 3.8518102169036865,
|
| 2615 |
+
"learning_rate": 1.1563849575118322e-06,
|
| 2616 |
+
"loss": 1.1679,
|
| 2617 |
+
"step": 368
|
| 2618 |
+
},
|
| 2619 |
+
{
|
| 2620 |
+
"epoch": 0.28941176470588237,
|
| 2621 |
+
"grad_norm": 3.8261094093322754,
|
| 2622 |
+
"learning_rate": 1.08561034572828e-06,
|
| 2623 |
+
"loss": 1.1916,
|
| 2624 |
+
"step": 369
|
| 2625 |
+
},
|
| 2626 |
+
{
|
| 2627 |
+
"epoch": 0.2901960784313726,
|
| 2628 |
+
"grad_norm": 5.102038383483887,
|
| 2629 |
+
"learning_rate": 1.0170363900881796e-06,
|
| 2630 |
+
"loss": 1.8872,
|
| 2631 |
+
"step": 370
|
| 2632 |
+
},
|
| 2633 |
+
{
|
| 2634 |
+
"epoch": 0.2909803921568627,
|
| 2635 |
+
"grad_norm": 5.184679985046387,
|
| 2636 |
+
"learning_rate": 9.506675402572445e-07,
|
| 2637 |
+
"loss": 1.4387,
|
| 2638 |
+
"step": 371
|
| 2639 |
+
},
|
| 2640 |
+
{
|
| 2641 |
+
"epoch": 0.2917647058823529,
|
| 2642 |
+
"grad_norm": 4.591920375823975,
|
| 2643 |
+
"learning_rate": 8.865081028150279e-07,
|
| 2644 |
+
"loss": 0.9937,
|
| 2645 |
+
"step": 372
|
| 2646 |
+
},
|
| 2647 |
+
{
|
| 2648 |
+
"epoch": 0.29254901960784313,
|
| 2649 |
+
"grad_norm": 5.405058860778809,
|
| 2650 |
+
"learning_rate": 8.24562240975476e-07,
|
| 2651 |
+
"loss": 1.3768,
|
| 2652 |
+
"step": 373
|
| 2653 |
+
},
|
| 2654 |
+
{
|
| 2655 |
+
"epoch": 0.29333333333333333,
|
| 2656 |
+
"grad_norm": 4.751702308654785,
|
| 2657 |
+
"learning_rate": 7.648339743168009e-07,
|
| 2658 |
+
"loss": 1.1257,
|
| 2659 |
+
"step": 374
|
| 2660 |
+
},
|
| 2661 |
+
{
|
| 2662 |
+
"epoch": 0.29411764705882354,
|
| 2663 |
+
"grad_norm": 5.307697772979736,
|
| 2664 |
+
"learning_rate": 7.073271785206313e-07,
|
| 2665 |
+
"loss": 1.3546,
|
| 2666 |
+
"step": 375
|
| 2667 |
+
},
|
| 2668 |
+
{
|
| 2669 |
+
"epoch": 0.29490196078431374,
|
| 2670 |
+
"grad_norm": 5.108140468597412,
|
| 2671 |
+
"learning_rate": 6.520455851205298e-07,
|
| 2672 |
+
"loss": 0.9153,
|
| 2673 |
+
"step": 376
|
| 2674 |
+
},
|
| 2675 |
+
{
|
| 2676 |
+
"epoch": 0.29568627450980395,
|
| 2677 |
+
"grad_norm": 4.52064847946167,
|
| 2678 |
+
"learning_rate": 5.98992781259876e-07,
|
| 2679 |
+
"loss": 0.8798,
|
| 2680 |
+
"step": 377
|
| 2681 |
+
},
|
| 2682 |
+
{
|
| 2683 |
+
"epoch": 0.2964705882352941,
|
| 2684 |
+
"grad_norm": 3.916222095489502,
|
| 2685 |
+
"learning_rate": 5.48172209459079e-07,
|
| 2686 |
+
"loss": 0.7,
|
| 2687 |
+
"step": 378
|
| 2688 |
+
},
|
| 2689 |
+
{
|
| 2690 |
+
"epoch": 0.2972549019607843,
|
| 2691 |
+
"grad_norm": 4.668052673339844,
|
| 2692 |
+
"learning_rate": 4.995871673922213e-07,
|
| 2693 |
+
"loss": 1.0176,
|
| 2694 |
+
"step": 379
|
| 2695 |
+
},
|
| 2696 |
+
{
|
| 2697 |
+
"epoch": 0.2980392156862745,
|
| 2698 |
+
"grad_norm": 4.989783763885498,
|
| 2699 |
+
"learning_rate": 4.532408076730504e-07,
|
| 2700 |
+
"loss": 1.0162,
|
| 2701 |
+
"step": 380
|
| 2702 |
+
},
|
| 2703 |
+
{
|
| 2704 |
+
"epoch": 0.2988235294117647,
|
| 2705 |
+
"grad_norm": 3.6178388595581055,
|
| 2706 |
+
"learning_rate": 4.091361376504382e-07,
|
| 2707 |
+
"loss": 0.746,
|
| 2708 |
+
"step": 381
|
| 2709 |
+
},
|
| 2710 |
+
{
|
| 2711 |
+
"epoch": 0.2996078431372549,
|
| 2712 |
+
"grad_norm": 4.038928985595703,
|
| 2713 |
+
"learning_rate": 3.672760192132107e-07,
|
| 2714 |
+
"loss": 0.8829,
|
| 2715 |
+
"step": 382
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 0.3003921568627451,
|
| 2719 |
+
"grad_norm": 5.406976699829102,
|
| 2720 |
+
"learning_rate": 3.276631686044734e-07,
|
| 2721 |
+
"loss": 1.3866,
|
| 2722 |
+
"step": 383
|
| 2723 |
+
},
|
| 2724 |
+
{
|
| 2725 |
+
"epoch": 0.30117647058823527,
|
| 2726 |
+
"grad_norm": 4.995997905731201,
|
| 2727 |
+
"learning_rate": 2.903001562453289e-07,
|
| 2728 |
+
"loss": 0.8837,
|
| 2729 |
+
"step": 384
|
| 2730 |
+
},
|
| 2731 |
+
{
|
| 2732 |
+
"epoch": 0.30196078431372547,
|
| 2733 |
+
"grad_norm": 6.531996250152588,
|
| 2734 |
+
"learning_rate": 2.5518940656811094e-07,
|
| 2735 |
+
"loss": 1.3823,
|
| 2736 |
+
"step": 385
|
| 2737 |
+
},
|
| 2738 |
+
{
|
| 2739 |
+
"epoch": 0.3027450980392157,
|
| 2740 |
+
"grad_norm": 6.122218132019043,
|
| 2741 |
+
"learning_rate": 2.2233319785904537e-07,
|
| 2742 |
+
"loss": 1.251,
|
| 2743 |
+
"step": 386
|
| 2744 |
+
},
|
| 2745 |
+
{
|
| 2746 |
+
"epoch": 0.3035294117647059,
|
| 2747 |
+
"grad_norm": 6.061277389526367,
|
| 2748 |
+
"learning_rate": 1.9173366211043484e-07,
|
| 2749 |
+
"loss": 1.4177,
|
| 2750 |
+
"step": 387
|
| 2751 |
+
},
|
| 2752 |
+
{
|
| 2753 |
+
"epoch": 0.3043137254901961,
|
| 2754 |
+
"grad_norm": 5.348453521728516,
|
| 2755 |
+
"learning_rate": 1.6339278488229802e-07,
|
| 2756 |
+
"loss": 1.3636,
|
| 2757 |
+
"step": 388
|
| 2758 |
+
},
|
| 2759 |
+
{
|
| 2760 |
+
"epoch": 0.3050980392156863,
|
| 2761 |
+
"grad_norm": 5.866225719451904,
|
| 2762 |
+
"learning_rate": 1.3731240517354802e-07,
|
| 2763 |
+
"loss": 1.1487,
|
| 2764 |
+
"step": 389
|
| 2765 |
+
},
|
| 2766 |
+
{
|
| 2767 |
+
"epoch": 0.3058823529411765,
|
| 2768 |
+
"grad_norm": 5.9875688552856445,
|
| 2769 |
+
"learning_rate": 1.1349421530265245e-07,
|
| 2770 |
+
"loss": 1.3415,
|
| 2771 |
+
"step": 390
|
| 2772 |
+
},
|
| 2773 |
+
{
|
| 2774 |
+
"epoch": 0.30666666666666664,
|
| 2775 |
+
"grad_norm": 7.915737628936768,
|
| 2776 |
+
"learning_rate": 9.19397607978134e-08,
|
| 2777 |
+
"loss": 1.225,
|
| 2778 |
+
"step": 391
|
| 2779 |
+
},
|
| 2780 |
+
{
|
| 2781 |
+
"epoch": 0.30745098039215685,
|
| 2782 |
+
"grad_norm": 8.340410232543945,
|
| 2783 |
+
"learning_rate": 7.265044029669486e-08,
|
| 2784 |
+
"loss": 1.9622,
|
| 2785 |
+
"step": 392
|
| 2786 |
+
},
|
| 2787 |
+
{
|
| 2788 |
+
"epoch": 0.30823529411764705,
|
| 2789 |
+
"grad_norm": 10.420660018920898,
|
| 2790 |
+
"learning_rate": 5.562750545566652e-08,
|
| 2791 |
+
"loss": 1.9871,
|
| 2792 |
+
"step": 393
|
| 2793 |
+
},
|
| 2794 |
+
{
|
| 2795 |
+
"epoch": 0.30901960784313726,
|
| 2796 |
+
"grad_norm": 9.027931213378906,
|
| 2797 |
+
"learning_rate": 4.087206086856754e-08,
|
| 2798 |
+
"loss": 1.585,
|
| 2799 |
+
"step": 394
|
| 2800 |
+
},
|
| 2801 |
+
{
|
| 2802 |
+
"epoch": 0.30980392156862746,
|
| 2803 |
+
"grad_norm": 10.239818572998047,
|
| 2804 |
+
"learning_rate": 2.838506399506446e-08,
|
| 2805 |
+
"loss": 1.9573,
|
| 2806 |
+
"step": 395
|
| 2807 |
+
},
|
| 2808 |
+
{
|
| 2809 |
+
"epoch": 0.31058823529411766,
|
| 2810 |
+
"grad_norm": 7.999627590179443,
|
| 2811 |
+
"learning_rate": 1.8167325098478735e-08,
|
| 2812 |
+
"loss": 1.6673,
|
| 2813 |
+
"step": 396
|
| 2814 |
+
},
|
| 2815 |
+
{
|
| 2816 |
+
"epoch": 0.31137254901960787,
|
| 2817 |
+
"grad_norm": 9.10187816619873,
|
| 2818 |
+
"learning_rate": 1.0219507193247045e-08,
|
| 2819 |
+
"loss": 1.6721,
|
| 2820 |
+
"step": 397
|
| 2821 |
+
},
|
| 2822 |
+
{
|
| 2823 |
+
"epoch": 0.312156862745098,
|
| 2824 |
+
"grad_norm": 13.277602195739746,
|
| 2825 |
+
"learning_rate": 4.542126001886303e-09,
|
| 2826 |
+
"loss": 3.0342,
|
| 2827 |
+
"step": 398
|
| 2828 |
+
},
|
| 2829 |
+
{
|
| 2830 |
+
"epoch": 0.3129411764705882,
|
| 2831 |
+
"grad_norm": 13.699840545654297,
|
| 2832 |
+
"learning_rate": 1.1355499215215391e-09,
|
| 2833 |
+
"loss": 1.7288,
|
| 2834 |
+
"step": 399
|
| 2835 |
+
},
|
| 2836 |
+
{
|
| 2837 |
+
"epoch": 0.3137254901960784,
|
| 2838 |
+
"grad_norm": 11.232860565185547,
|
| 2839 |
+
"learning_rate": 0.0,
|
| 2840 |
+
"loss": 2.2469,
|
| 2841 |
+
"step": 400
|
| 2842 |
+
},
|
| 2843 |
+
{
|
| 2844 |
+
"epoch": 0.3137254901960784,
|
| 2845 |
+
"eval_loss": 0.3466954827308655,
|
| 2846 |
+
"eval_runtime": 169.3978,
|
| 2847 |
+
"eval_samples_per_second": 12.68,
|
| 2848 |
+
"eval_steps_per_second": 3.17,
|
| 2849 |
+
"step": 400
|
| 2850 |
}
|
| 2851 |
],
|
| 2852 |
"logging_steps": 1,
|
|
|
|
| 2870 |
"should_evaluate": false,
|
| 2871 |
"should_log": false,
|
| 2872 |
"should_save": true,
|
| 2873 |
+
"should_training_stop": true
|
| 2874 |
},
|
| 2875 |
"attributes": {}
|
| 2876 |
}
|
| 2877 |
},
|
| 2878 |
+
"total_flos": 6.053128390909624e+17,
|
| 2879 |
"train_batch_size": 8,
|
| 2880 |
"trial_name": null,
|
| 2881 |
"trial_params": null
|