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 201353800
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1dc391235528041d2fc64c517de0de5a35d6e73291b9575d80be3c327678412e
|
| 3 |
size 201353800
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 102463162
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b759274f0d85a5e93ff596ccde0cd1ab9ef3d3a503db9c3340bcdec7982c30b6
|
| 3 |
size 102463162
|
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:ea1cdf60dba2aa43fbb9730847255fb9348f31329cac03f340a7414262a148e2
|
| 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:ce6af8c2b755ab309a73060b69c0ffea28d5cf03ddedb3dc24064e45dedc183e
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 0.008372417651116848,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
| 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": 6.896,
|
| 2140 |
"eval_steps_per_second": 1.726,
|
| 2141 |
"step": 300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2142 |
}
|
| 2143 |
],
|
| 2144 |
"logging_steps": 1,
|
|
@@ -2153,7 +2861,7 @@
|
|
| 2153 |
"early_stopping_threshold": 0.0
|
| 2154 |
},
|
| 2155 |
"attributes": {
|
| 2156 |
-
"early_stopping_patience_counter":
|
| 2157 |
}
|
| 2158 |
},
|
| 2159 |
"TrainerControl": {
|
|
@@ -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": 1.
|
| 2171 |
"train_batch_size": 4,
|
| 2172 |
"trial_name": null,
|
| 2173 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 0.008372417651116848,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
| 4 |
+
"epoch": 0.3273657289002558,
|
| 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": 6.896,
|
| 2140 |
"eval_steps_per_second": 1.726,
|
| 2141 |
"step": 300
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"epoch": 0.24634271099744245,
|
| 2145 |
+
"grad_norm": 1.1538922786712646,
|
| 2146 |
+
"learning_rate": 0.000193028873651704,
|
| 2147 |
+
"loss": 0.0128,
|
| 2148 |
+
"step": 301
|
| 2149 |
+
},
|
| 2150 |
+
{
|
| 2151 |
+
"epoch": 0.2471611253196931,
|
| 2152 |
+
"grad_norm": 1.4959121942520142,
|
| 2153 |
+
"learning_rate": 0.000192981449234947,
|
| 2154 |
+
"loss": 0.0151,
|
| 2155 |
+
"step": 302
|
| 2156 |
+
},
|
| 2157 |
+
{
|
| 2158 |
+
"epoch": 0.24797953964194375,
|
| 2159 |
+
"grad_norm": 0.7326957583427429,
|
| 2160 |
+
"learning_rate": 0.00019293386991696648,
|
| 2161 |
+
"loss": 0.0074,
|
| 2162 |
+
"step": 303
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 0.24879795396419438,
|
| 2166 |
+
"grad_norm": 5.472940921783447,
|
| 2167 |
+
"learning_rate": 0.00019288613577702655,
|
| 2168 |
+
"loss": 0.1211,
|
| 2169 |
+
"step": 304
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 0.249616368286445,
|
| 2173 |
+
"grad_norm": 2.403130531311035,
|
| 2174 |
+
"learning_rate": 0.00019283824689464926,
|
| 2175 |
+
"loss": 0.059,
|
| 2176 |
+
"step": 305
|
| 2177 |
+
},
|
| 2178 |
+
{
|
| 2179 |
+
"epoch": 0.25043478260869567,
|
| 2180 |
+
"grad_norm": 4.317528247833252,
|
| 2181 |
+
"learning_rate": 0.00019279020334961447,
|
| 2182 |
+
"loss": 0.1937,
|
| 2183 |
+
"step": 306
|
| 2184 |
+
},
|
| 2185 |
+
{
|
| 2186 |
+
"epoch": 0.2512531969309463,
|
| 2187 |
+
"grad_norm": 2.334576368331909,
|
| 2188 |
+
"learning_rate": 0.0001927420052219597,
|
| 2189 |
+
"loss": 0.1942,
|
| 2190 |
+
"step": 307
|
| 2191 |
+
},
|
| 2192 |
+
{
|
| 2193 |
+
"epoch": 0.25207161125319694,
|
| 2194 |
+
"grad_norm": 2.2626686096191406,
|
| 2195 |
+
"learning_rate": 0.00019269365259198,
|
| 2196 |
+
"loss": 0.0893,
|
| 2197 |
+
"step": 308
|
| 2198 |
+
},
|
| 2199 |
+
{
|
| 2200 |
+
"epoch": 0.2528900255754476,
|
| 2201 |
+
"grad_norm": 1.1748579740524292,
|
| 2202 |
+
"learning_rate": 0.0001926451455402277,
|
| 2203 |
+
"loss": 0.011,
|
| 2204 |
+
"step": 309
|
| 2205 |
+
},
|
| 2206 |
+
{
|
| 2207 |
+
"epoch": 0.2537084398976982,
|
| 2208 |
+
"grad_norm": 0.6484130620956421,
|
| 2209 |
+
"learning_rate": 0.00019259648414751265,
|
| 2210 |
+
"loss": 0.0094,
|
| 2211 |
+
"step": 310
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 0.25452685421994886,
|
| 2215 |
+
"grad_norm": 0.47816258668899536,
|
| 2216 |
+
"learning_rate": 0.00019254766849490152,
|
| 2217 |
+
"loss": 0.0062,
|
| 2218 |
+
"step": 311
|
| 2219 |
+
},
|
| 2220 |
+
{
|
| 2221 |
+
"epoch": 0.25534526854219947,
|
| 2222 |
+
"grad_norm": 3.412865161895752,
|
| 2223 |
+
"learning_rate": 0.00019249869866371817,
|
| 2224 |
+
"loss": 0.1235,
|
| 2225 |
+
"step": 312
|
| 2226 |
+
},
|
| 2227 |
+
{
|
| 2228 |
+
"epoch": 0.2561636828644501,
|
| 2229 |
+
"grad_norm": 12.466361999511719,
|
| 2230 |
+
"learning_rate": 0.00019244957473554323,
|
| 2231 |
+
"loss": 0.0131,
|
| 2232 |
+
"step": 313
|
| 2233 |
+
},
|
| 2234 |
+
{
|
| 2235 |
+
"epoch": 0.2569820971867008,
|
| 2236 |
+
"grad_norm": 1.225807547569275,
|
| 2237 |
+
"learning_rate": 0.00019240029679221408,
|
| 2238 |
+
"loss": 0.0215,
|
| 2239 |
+
"step": 314
|
| 2240 |
+
},
|
| 2241 |
+
{
|
| 2242 |
+
"epoch": 0.2578005115089514,
|
| 2243 |
+
"grad_norm": 0.0476732961833477,
|
| 2244 |
+
"learning_rate": 0.00019235086491582463,
|
| 2245 |
+
"loss": 0.0017,
|
| 2246 |
+
"step": 315
|
| 2247 |
+
},
|
| 2248 |
+
{
|
| 2249 |
+
"epoch": 0.25861892583120205,
|
| 2250 |
+
"grad_norm": 4.701657772064209,
|
| 2251 |
+
"learning_rate": 0.0001923012791887253,
|
| 2252 |
+
"loss": 0.0362,
|
| 2253 |
+
"step": 316
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 0.2594373401534527,
|
| 2257 |
+
"grad_norm": 18.314184188842773,
|
| 2258 |
+
"learning_rate": 0.00019225153969352275,
|
| 2259 |
+
"loss": 0.4503,
|
| 2260 |
+
"step": 317
|
| 2261 |
+
},
|
| 2262 |
+
{
|
| 2263 |
+
"epoch": 0.2602557544757033,
|
| 2264 |
+
"grad_norm": 2.99334716796875,
|
| 2265 |
+
"learning_rate": 0.00019220164651307986,
|
| 2266 |
+
"loss": 0.0408,
|
| 2267 |
+
"step": 318
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 0.261074168797954,
|
| 2271 |
+
"grad_norm": 3.7102837562561035,
|
| 2272 |
+
"learning_rate": 0.00019215159973051552,
|
| 2273 |
+
"loss": 0.1288,
|
| 2274 |
+
"step": 319
|
| 2275 |
+
},
|
| 2276 |
+
{
|
| 2277 |
+
"epoch": 0.2618925831202046,
|
| 2278 |
+
"grad_norm": 2.6803882122039795,
|
| 2279 |
+
"learning_rate": 0.0001921013994292045,
|
| 2280 |
+
"loss": 0.0465,
|
| 2281 |
+
"step": 320
|
| 2282 |
+
},
|
| 2283 |
+
{
|
| 2284 |
+
"epoch": 0.26271099744245524,
|
| 2285 |
+
"grad_norm": 8.001237869262695,
|
| 2286 |
+
"learning_rate": 0.00019205104569277733,
|
| 2287 |
+
"loss": 0.0932,
|
| 2288 |
+
"step": 321
|
| 2289 |
+
},
|
| 2290 |
+
{
|
| 2291 |
+
"epoch": 0.2635294117647059,
|
| 2292 |
+
"grad_norm": 0.7996736764907837,
|
| 2293 |
+
"learning_rate": 0.00019200053860512014,
|
| 2294 |
+
"loss": 0.0281,
|
| 2295 |
+
"step": 322
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 0.2643478260869565,
|
| 2299 |
+
"grad_norm": 0.10956323891878128,
|
| 2300 |
+
"learning_rate": 0.0001919498782503746,
|
| 2301 |
+
"loss": 0.0034,
|
| 2302 |
+
"step": 323
|
| 2303 |
+
},
|
| 2304 |
+
{
|
| 2305 |
+
"epoch": 0.26516624040920717,
|
| 2306 |
+
"grad_norm": 10.366631507873535,
|
| 2307 |
+
"learning_rate": 0.0001918990647129376,
|
| 2308 |
+
"loss": 0.054,
|
| 2309 |
+
"step": 324
|
| 2310 |
+
},
|
| 2311 |
+
{
|
| 2312 |
+
"epoch": 0.2659846547314578,
|
| 2313 |
+
"grad_norm": 16.79122543334961,
|
| 2314 |
+
"learning_rate": 0.0001918480980774613,
|
| 2315 |
+
"loss": 0.3747,
|
| 2316 |
+
"step": 325
|
| 2317 |
+
},
|
| 2318 |
+
{
|
| 2319 |
+
"epoch": 0.26680306905370843,
|
| 2320 |
+
"grad_norm": 16.79612159729004,
|
| 2321 |
+
"learning_rate": 0.00019179697842885293,
|
| 2322 |
+
"loss": 0.1583,
|
| 2323 |
+
"step": 326
|
| 2324 |
+
},
|
| 2325 |
+
{
|
| 2326 |
+
"epoch": 0.2676214833759591,
|
| 2327 |
+
"grad_norm": 0.3873645067214966,
|
| 2328 |
+
"learning_rate": 0.0001917457058522746,
|
| 2329 |
+
"loss": 0.0056,
|
| 2330 |
+
"step": 327
|
| 2331 |
+
},
|
| 2332 |
+
{
|
| 2333 |
+
"epoch": 0.2684398976982097,
|
| 2334 |
+
"grad_norm": 1.963756799697876,
|
| 2335 |
+
"learning_rate": 0.00019169428043314314,
|
| 2336 |
+
"loss": 0.073,
|
| 2337 |
+
"step": 328
|
| 2338 |
+
},
|
| 2339 |
+
{
|
| 2340 |
+
"epoch": 0.26925831202046036,
|
| 2341 |
+
"grad_norm": 9.859981536865234,
|
| 2342 |
+
"learning_rate": 0.00019164270225713008,
|
| 2343 |
+
"loss": 0.0644,
|
| 2344 |
+
"step": 329
|
| 2345 |
+
},
|
| 2346 |
+
{
|
| 2347 |
+
"epoch": 0.270076726342711,
|
| 2348 |
+
"grad_norm": 3.775178909301758,
|
| 2349 |
+
"learning_rate": 0.0001915909714101614,
|
| 2350 |
+
"loss": 0.1595,
|
| 2351 |
+
"step": 330
|
| 2352 |
+
},
|
| 2353 |
+
{
|
| 2354 |
+
"epoch": 0.2708951406649616,
|
| 2355 |
+
"grad_norm": 2.9009556770324707,
|
| 2356 |
+
"learning_rate": 0.00019153908797841742,
|
| 2357 |
+
"loss": 0.0377,
|
| 2358 |
+
"step": 331
|
| 2359 |
+
},
|
| 2360 |
+
{
|
| 2361 |
+
"epoch": 0.2717135549872123,
|
| 2362 |
+
"grad_norm": 11.263628005981445,
|
| 2363 |
+
"learning_rate": 0.0001914870520483327,
|
| 2364 |
+
"loss": 0.061,
|
| 2365 |
+
"step": 332
|
| 2366 |
+
},
|
| 2367 |
+
{
|
| 2368 |
+
"epoch": 0.27253196930946294,
|
| 2369 |
+
"grad_norm": 0.01919454149901867,
|
| 2370 |
+
"learning_rate": 0.00019143486370659573,
|
| 2371 |
+
"loss": 0.0008,
|
| 2372 |
+
"step": 333
|
| 2373 |
+
},
|
| 2374 |
+
{
|
| 2375 |
+
"epoch": 0.27335038363171354,
|
| 2376 |
+
"grad_norm": 2.147348642349243,
|
| 2377 |
+
"learning_rate": 0.00019138252304014907,
|
| 2378 |
+
"loss": 0.2487,
|
| 2379 |
+
"step": 334
|
| 2380 |
+
},
|
| 2381 |
+
{
|
| 2382 |
+
"epoch": 0.2741687979539642,
|
| 2383 |
+
"grad_norm": 0.05764702707529068,
|
| 2384 |
+
"learning_rate": 0.0001913300301361889,
|
| 2385 |
+
"loss": 0.0016,
|
| 2386 |
+
"step": 335
|
| 2387 |
+
},
|
| 2388 |
+
{
|
| 2389 |
+
"epoch": 0.2749872122762148,
|
| 2390 |
+
"grad_norm": 0.025458738207817078,
|
| 2391 |
+
"learning_rate": 0.00019127738508216516,
|
| 2392 |
+
"loss": 0.0009,
|
| 2393 |
+
"step": 336
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 0.27580562659846547,
|
| 2397 |
+
"grad_norm": 2.147672653198242,
|
| 2398 |
+
"learning_rate": 0.00019122458796578114,
|
| 2399 |
+
"loss": 0.1389,
|
| 2400 |
+
"step": 337
|
| 2401 |
+
},
|
| 2402 |
+
{
|
| 2403 |
+
"epoch": 0.27662404092071613,
|
| 2404 |
+
"grad_norm": 2.0472524166107178,
|
| 2405 |
+
"learning_rate": 0.0001911716388749935,
|
| 2406 |
+
"loss": 0.0667,
|
| 2407 |
+
"step": 338
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 0.27744245524296673,
|
| 2411 |
+
"grad_norm": 0.15356223285198212,
|
| 2412 |
+
"learning_rate": 0.0001911185378980121,
|
| 2413 |
+
"loss": 0.0025,
|
| 2414 |
+
"step": 339
|
| 2415 |
+
},
|
| 2416 |
+
{
|
| 2417 |
+
"epoch": 0.2782608695652174,
|
| 2418 |
+
"grad_norm": 0.0540032684803009,
|
| 2419 |
+
"learning_rate": 0.00019106528512329978,
|
| 2420 |
+
"loss": 0.002,
|
| 2421 |
+
"step": 340
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 0.27907928388746805,
|
| 2425 |
+
"grad_norm": 3.762300968170166,
|
| 2426 |
+
"learning_rate": 0.00019101188063957235,
|
| 2427 |
+
"loss": 0.1637,
|
| 2428 |
+
"step": 341
|
| 2429 |
+
},
|
| 2430 |
+
{
|
| 2431 |
+
"epoch": 0.27989769820971866,
|
| 2432 |
+
"grad_norm": 1.0720164775848389,
|
| 2433 |
+
"learning_rate": 0.0001909583245357983,
|
| 2434 |
+
"loss": 0.0124,
|
| 2435 |
+
"step": 342
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 0.2807161125319693,
|
| 2439 |
+
"grad_norm": 3.95150089263916,
|
| 2440 |
+
"learning_rate": 0.00019090461690119868,
|
| 2441 |
+
"loss": 0.021,
|
| 2442 |
+
"step": 343
|
| 2443 |
+
},
|
| 2444 |
+
{
|
| 2445 |
+
"epoch": 0.2815345268542199,
|
| 2446 |
+
"grad_norm": 5.743108749389648,
|
| 2447 |
+
"learning_rate": 0.00019085075782524703,
|
| 2448 |
+
"loss": 0.1296,
|
| 2449 |
+
"step": 344
|
| 2450 |
+
},
|
| 2451 |
+
{
|
| 2452 |
+
"epoch": 0.2823529411764706,
|
| 2453 |
+
"grad_norm": 1.5119516849517822,
|
| 2454 |
+
"learning_rate": 0.00019079674739766917,
|
| 2455 |
+
"loss": 0.0188,
|
| 2456 |
+
"step": 345
|
| 2457 |
+
},
|
| 2458 |
+
{
|
| 2459 |
+
"epoch": 0.28317135549872124,
|
| 2460 |
+
"grad_norm": 0.20997177064418793,
|
| 2461 |
+
"learning_rate": 0.0001907425857084431,
|
| 2462 |
+
"loss": 0.0049,
|
| 2463 |
+
"step": 346
|
| 2464 |
+
},
|
| 2465 |
+
{
|
| 2466 |
+
"epoch": 0.28398976982097185,
|
| 2467 |
+
"grad_norm": 3.7278308868408203,
|
| 2468 |
+
"learning_rate": 0.00019068827284779874,
|
| 2469 |
+
"loss": 0.1293,
|
| 2470 |
+
"step": 347
|
| 2471 |
+
},
|
| 2472 |
+
{
|
| 2473 |
+
"epoch": 0.2848081841432225,
|
| 2474 |
+
"grad_norm": 3.974945545196533,
|
| 2475 |
+
"learning_rate": 0.0001906338089062179,
|
| 2476 |
+
"loss": 0.0574,
|
| 2477 |
+
"step": 348
|
| 2478 |
+
},
|
| 2479 |
+
{
|
| 2480 |
+
"epoch": 0.28562659846547317,
|
| 2481 |
+
"grad_norm": 34.30356216430664,
|
| 2482 |
+
"learning_rate": 0.00019057919397443403,
|
| 2483 |
+
"loss": 0.8184,
|
| 2484 |
+
"step": 349
|
| 2485 |
+
},
|
| 2486 |
+
{
|
| 2487 |
+
"epoch": 0.2864450127877238,
|
| 2488 |
+
"grad_norm": 0.13904468715190887,
|
| 2489 |
+
"learning_rate": 0.0001905244281434322,
|
| 2490 |
+
"loss": 0.0013,
|
| 2491 |
+
"step": 350
|
| 2492 |
+
},
|
| 2493 |
+
{
|
| 2494 |
+
"epoch": 0.28726342710997443,
|
| 2495 |
+
"grad_norm": 1.397051215171814,
|
| 2496 |
+
"learning_rate": 0.00019046951150444882,
|
| 2497 |
+
"loss": 0.0348,
|
| 2498 |
+
"step": 351
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 0.28808184143222504,
|
| 2502 |
+
"grad_norm": 0.16247797012329102,
|
| 2503 |
+
"learning_rate": 0.00019041444414897153,
|
| 2504 |
+
"loss": 0.0031,
|
| 2505 |
+
"step": 352
|
| 2506 |
+
},
|
| 2507 |
+
{
|
| 2508 |
+
"epoch": 0.2889002557544757,
|
| 2509 |
+
"grad_norm": 0.17214684188365936,
|
| 2510 |
+
"learning_rate": 0.0001903592261687391,
|
| 2511 |
+
"loss": 0.0038,
|
| 2512 |
+
"step": 353
|
| 2513 |
+
},
|
| 2514 |
+
{
|
| 2515 |
+
"epoch": 0.28971867007672636,
|
| 2516 |
+
"grad_norm": 0.03287775442004204,
|
| 2517 |
+
"learning_rate": 0.00019030385765574114,
|
| 2518 |
+
"loss": 0.0008,
|
| 2519 |
+
"step": 354
|
| 2520 |
+
},
|
| 2521 |
+
{
|
| 2522 |
+
"epoch": 0.29053708439897696,
|
| 2523 |
+
"grad_norm": 1.5617440938949585,
|
| 2524 |
+
"learning_rate": 0.00019024833870221817,
|
| 2525 |
+
"loss": 0.128,
|
| 2526 |
+
"step": 355
|
| 2527 |
+
},
|
| 2528 |
+
{
|
| 2529 |
+
"epoch": 0.2913554987212276,
|
| 2530 |
+
"grad_norm": 0.4113420248031616,
|
| 2531 |
+
"learning_rate": 0.00019019266940066117,
|
| 2532 |
+
"loss": 0.0022,
|
| 2533 |
+
"step": 356
|
| 2534 |
+
},
|
| 2535 |
+
{
|
| 2536 |
+
"epoch": 0.2921739130434783,
|
| 2537 |
+
"grad_norm": 9.879963874816895,
|
| 2538 |
+
"learning_rate": 0.00019013684984381176,
|
| 2539 |
+
"loss": 0.1339,
|
| 2540 |
+
"step": 357
|
| 2541 |
+
},
|
| 2542 |
+
{
|
| 2543 |
+
"epoch": 0.2929923273657289,
|
| 2544 |
+
"grad_norm": 2.0313758850097656,
|
| 2545 |
+
"learning_rate": 0.00019008088012466179,
|
| 2546 |
+
"loss": 0.0137,
|
| 2547 |
+
"step": 358
|
| 2548 |
+
},
|
| 2549 |
+
{
|
| 2550 |
+
"epoch": 0.29381074168797955,
|
| 2551 |
+
"grad_norm": 2.5029163360595703,
|
| 2552 |
+
"learning_rate": 0.00019002476033645326,
|
| 2553 |
+
"loss": 0.0366,
|
| 2554 |
+
"step": 359
|
| 2555 |
+
},
|
| 2556 |
+
{
|
| 2557 |
+
"epoch": 0.29462915601023015,
|
| 2558 |
+
"grad_norm": 2.537306785583496,
|
| 2559 |
+
"learning_rate": 0.00018996849057267815,
|
| 2560 |
+
"loss": 0.0249,
|
| 2561 |
+
"step": 360
|
| 2562 |
+
},
|
| 2563 |
+
{
|
| 2564 |
+
"epoch": 0.2954475703324808,
|
| 2565 |
+
"grad_norm": 0.12636743485927582,
|
| 2566 |
+
"learning_rate": 0.0001899120709270784,
|
| 2567 |
+
"loss": 0.0024,
|
| 2568 |
+
"step": 361
|
| 2569 |
+
},
|
| 2570 |
+
{
|
| 2571 |
+
"epoch": 0.2962659846547315,
|
| 2572 |
+
"grad_norm": 2.858910083770752,
|
| 2573 |
+
"learning_rate": 0.00018985550149364552,
|
| 2574 |
+
"loss": 0.0341,
|
| 2575 |
+
"step": 362
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"epoch": 0.2970843989769821,
|
| 2579 |
+
"grad_norm": 0.029364800080657005,
|
| 2580 |
+
"learning_rate": 0.0001897987823666207,
|
| 2581 |
+
"loss": 0.0006,
|
| 2582 |
+
"step": 363
|
| 2583 |
+
},
|
| 2584 |
+
{
|
| 2585 |
+
"epoch": 0.29790281329923274,
|
| 2586 |
+
"grad_norm": 4.783391952514648,
|
| 2587 |
+
"learning_rate": 0.00018974191364049434,
|
| 2588 |
+
"loss": 0.242,
|
| 2589 |
+
"step": 364
|
| 2590 |
+
},
|
| 2591 |
+
{
|
| 2592 |
+
"epoch": 0.2987212276214834,
|
| 2593 |
+
"grad_norm": 1.4791219234466553,
|
| 2594 |
+
"learning_rate": 0.0001896848954100062,
|
| 2595 |
+
"loss": 0.0586,
|
| 2596 |
+
"step": 365
|
| 2597 |
+
},
|
| 2598 |
+
{
|
| 2599 |
+
"epoch": 0.299539641943734,
|
| 2600 |
+
"grad_norm": 7.496979713439941,
|
| 2601 |
+
"learning_rate": 0.000189627727770145,
|
| 2602 |
+
"loss": 0.0033,
|
| 2603 |
+
"step": 366
|
| 2604 |
+
},
|
| 2605 |
+
{
|
| 2606 |
+
"epoch": 0.30035805626598466,
|
| 2607 |
+
"grad_norm": 2.8373289108276367,
|
| 2608 |
+
"learning_rate": 0.00018957041081614845,
|
| 2609 |
+
"loss": 0.0062,
|
| 2610 |
+
"step": 367
|
| 2611 |
+
},
|
| 2612 |
+
{
|
| 2613 |
+
"epoch": 0.30117647058823527,
|
| 2614 |
+
"grad_norm": 0.04176490381360054,
|
| 2615 |
+
"learning_rate": 0.000189512944643503,
|
| 2616 |
+
"loss": 0.0006,
|
| 2617 |
+
"step": 368
|
| 2618 |
+
},
|
| 2619 |
+
{
|
| 2620 |
+
"epoch": 0.30199488491048593,
|
| 2621 |
+
"grad_norm": 6.637153625488281,
|
| 2622 |
+
"learning_rate": 0.00018945532934794363,
|
| 2623 |
+
"loss": 0.0217,
|
| 2624 |
+
"step": 369
|
| 2625 |
+
},
|
| 2626 |
+
{
|
| 2627 |
+
"epoch": 0.3028132992327366,
|
| 2628 |
+
"grad_norm": 5.947137355804443,
|
| 2629 |
+
"learning_rate": 0.0001893975650254538,
|
| 2630 |
+
"loss": 0.0624,
|
| 2631 |
+
"step": 370
|
| 2632 |
+
},
|
| 2633 |
+
{
|
| 2634 |
+
"epoch": 0.3036317135549872,
|
| 2635 |
+
"grad_norm": 4.605132102966309,
|
| 2636 |
+
"learning_rate": 0.0001893396517722652,
|
| 2637 |
+
"loss": 0.0287,
|
| 2638 |
+
"step": 371
|
| 2639 |
+
},
|
| 2640 |
+
{
|
| 2641 |
+
"epoch": 0.30445012787723785,
|
| 2642 |
+
"grad_norm": 0.5078949928283691,
|
| 2643 |
+
"learning_rate": 0.00018928158968485769,
|
| 2644 |
+
"loss": 0.005,
|
| 2645 |
+
"step": 372
|
| 2646 |
+
},
|
| 2647 |
+
{
|
| 2648 |
+
"epoch": 0.3052685421994885,
|
| 2649 |
+
"grad_norm": 2.0189433097839355,
|
| 2650 |
+
"learning_rate": 0.00018922337885995903,
|
| 2651 |
+
"loss": 0.0471,
|
| 2652 |
+
"step": 373
|
| 2653 |
+
},
|
| 2654 |
+
{
|
| 2655 |
+
"epoch": 0.3060869565217391,
|
| 2656 |
+
"grad_norm": 1.7046236991882324,
|
| 2657 |
+
"learning_rate": 0.00018916501939454476,
|
| 2658 |
+
"loss": 0.1305,
|
| 2659 |
+
"step": 374
|
| 2660 |
+
},
|
| 2661 |
+
{
|
| 2662 |
+
"epoch": 0.3069053708439898,
|
| 2663 |
+
"grad_norm": 23.23792839050293,
|
| 2664 |
+
"learning_rate": 0.00018910651138583808,
|
| 2665 |
+
"loss": 0.1648,
|
| 2666 |
+
"step": 375
|
| 2667 |
+
},
|
| 2668 |
+
{
|
| 2669 |
+
"epoch": 0.3077237851662404,
|
| 2670 |
+
"grad_norm": 10.033699989318848,
|
| 2671 |
+
"learning_rate": 0.00018904785493130963,
|
| 2672 |
+
"loss": 0.1637,
|
| 2673 |
+
"step": 376
|
| 2674 |
+
},
|
| 2675 |
+
{
|
| 2676 |
+
"epoch": 0.30854219948849104,
|
| 2677 |
+
"grad_norm": 1.7353355884552002,
|
| 2678 |
+
"learning_rate": 0.00018898905012867736,
|
| 2679 |
+
"loss": 0.041,
|
| 2680 |
+
"step": 377
|
| 2681 |
+
},
|
| 2682 |
+
{
|
| 2683 |
+
"epoch": 0.3093606138107417,
|
| 2684 |
+
"grad_norm": 3.0973310470581055,
|
| 2685 |
+
"learning_rate": 0.00018893009707590636,
|
| 2686 |
+
"loss": 0.1848,
|
| 2687 |
+
"step": 378
|
| 2688 |
+
},
|
| 2689 |
+
{
|
| 2690 |
+
"epoch": 0.3101790281329923,
|
| 2691 |
+
"grad_norm": 6.835923194885254,
|
| 2692 |
+
"learning_rate": 0.0001888709958712087,
|
| 2693 |
+
"loss": 0.1367,
|
| 2694 |
+
"step": 379
|
| 2695 |
+
},
|
| 2696 |
+
{
|
| 2697 |
+
"epoch": 0.31099744245524297,
|
| 2698 |
+
"grad_norm": 5.92802619934082,
|
| 2699 |
+
"learning_rate": 0.00018881174661304327,
|
| 2700 |
+
"loss": 0.0561,
|
| 2701 |
+
"step": 380
|
| 2702 |
+
},
|
| 2703 |
+
{
|
| 2704 |
+
"epoch": 0.3118158567774936,
|
| 2705 |
+
"grad_norm": 2.9860589504241943,
|
| 2706 |
+
"learning_rate": 0.00018875234940011557,
|
| 2707 |
+
"loss": 0.106,
|
| 2708 |
+
"step": 381
|
| 2709 |
+
},
|
| 2710 |
+
{
|
| 2711 |
+
"epoch": 0.31263427109974423,
|
| 2712 |
+
"grad_norm": 2.7648398876190186,
|
| 2713 |
+
"learning_rate": 0.00018869280433137759,
|
| 2714 |
+
"loss": 0.0348,
|
| 2715 |
+
"step": 382
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 0.3134526854219949,
|
| 2719 |
+
"grad_norm": 1.2207233905792236,
|
| 2720 |
+
"learning_rate": 0.00018863311150602773,
|
| 2721 |
+
"loss": 0.0328,
|
| 2722 |
+
"step": 383
|
| 2723 |
+
},
|
| 2724 |
+
{
|
| 2725 |
+
"epoch": 0.3142710997442455,
|
| 2726 |
+
"grad_norm": 0.7907407879829407,
|
| 2727 |
+
"learning_rate": 0.00018857327102351034,
|
| 2728 |
+
"loss": 0.0191,
|
| 2729 |
+
"step": 384
|
| 2730 |
+
},
|
| 2731 |
+
{
|
| 2732 |
+
"epoch": 0.31508951406649616,
|
| 2733 |
+
"grad_norm": 0.05348999425768852,
|
| 2734 |
+
"learning_rate": 0.000188513282983516,
|
| 2735 |
+
"loss": 0.0023,
|
| 2736 |
+
"step": 385
|
| 2737 |
+
},
|
| 2738 |
+
{
|
| 2739 |
+
"epoch": 0.3159079283887468,
|
| 2740 |
+
"grad_norm": 1.7986161708831787,
|
| 2741 |
+
"learning_rate": 0.00018845314748598094,
|
| 2742 |
+
"loss": 0.0403,
|
| 2743 |
+
"step": 386
|
| 2744 |
+
},
|
| 2745 |
+
{
|
| 2746 |
+
"epoch": 0.3167263427109974,
|
| 2747 |
+
"grad_norm": 0.08972010016441345,
|
| 2748 |
+
"learning_rate": 0.00018839286463108717,
|
| 2749 |
+
"loss": 0.0015,
|
| 2750 |
+
"step": 387
|
| 2751 |
+
},
|
| 2752 |
+
{
|
| 2753 |
+
"epoch": 0.3175447570332481,
|
| 2754 |
+
"grad_norm": 2.7837443351745605,
|
| 2755 |
+
"learning_rate": 0.000188332434519262,
|
| 2756 |
+
"loss": 0.0452,
|
| 2757 |
+
"step": 388
|
| 2758 |
+
},
|
| 2759 |
+
{
|
| 2760 |
+
"epoch": 0.31836317135549874,
|
| 2761 |
+
"grad_norm": 0.03810626640915871,
|
| 2762 |
+
"learning_rate": 0.00018827185725117827,
|
| 2763 |
+
"loss": 0.0005,
|
| 2764 |
+
"step": 389
|
| 2765 |
+
},
|
| 2766 |
+
{
|
| 2767 |
+
"epoch": 0.31918158567774935,
|
| 2768 |
+
"grad_norm": 3.4926528930664062,
|
| 2769 |
+
"learning_rate": 0.00018821113292775388,
|
| 2770 |
+
"loss": 0.1284,
|
| 2771 |
+
"step": 390
|
| 2772 |
+
},
|
| 2773 |
+
{
|
| 2774 |
+
"epoch": 0.32,
|
| 2775 |
+
"grad_norm": 4.0867791175842285,
|
| 2776 |
+
"learning_rate": 0.0001881502616501517,
|
| 2777 |
+
"loss": 0.0975,
|
| 2778 |
+
"step": 391
|
| 2779 |
+
},
|
| 2780 |
+
{
|
| 2781 |
+
"epoch": 0.32081841432225067,
|
| 2782 |
+
"grad_norm": 0.2339545637369156,
|
| 2783 |
+
"learning_rate": 0.00018808924351977944,
|
| 2784 |
+
"loss": 0.0018,
|
| 2785 |
+
"step": 392
|
| 2786 |
+
},
|
| 2787 |
+
{
|
| 2788 |
+
"epoch": 0.32163682864450127,
|
| 2789 |
+
"grad_norm": 9.076642990112305,
|
| 2790 |
+
"learning_rate": 0.00018802807863828945,
|
| 2791 |
+
"loss": 0.1044,
|
| 2792 |
+
"step": 393
|
| 2793 |
+
},
|
| 2794 |
+
{
|
| 2795 |
+
"epoch": 0.32245524296675193,
|
| 2796 |
+
"grad_norm": 11.172904014587402,
|
| 2797 |
+
"learning_rate": 0.00018796676710757854,
|
| 2798 |
+
"loss": 0.0998,
|
| 2799 |
+
"step": 394
|
| 2800 |
+
},
|
| 2801 |
+
{
|
| 2802 |
+
"epoch": 0.32327365728900254,
|
| 2803 |
+
"grad_norm": 5.419793128967285,
|
| 2804 |
+
"learning_rate": 0.00018790530902978788,
|
| 2805 |
+
"loss": 0.0529,
|
| 2806 |
+
"step": 395
|
| 2807 |
+
},
|
| 2808 |
+
{
|
| 2809 |
+
"epoch": 0.3240920716112532,
|
| 2810 |
+
"grad_norm": 0.1783570945262909,
|
| 2811 |
+
"learning_rate": 0.00018784370450730274,
|
| 2812 |
+
"loss": 0.0013,
|
| 2813 |
+
"step": 396
|
| 2814 |
+
},
|
| 2815 |
+
{
|
| 2816 |
+
"epoch": 0.32491048593350386,
|
| 2817 |
+
"grad_norm": 2.391124725341797,
|
| 2818 |
+
"learning_rate": 0.00018778195364275234,
|
| 2819 |
+
"loss": 0.0167,
|
| 2820 |
+
"step": 397
|
| 2821 |
+
},
|
| 2822 |
+
{
|
| 2823 |
+
"epoch": 0.32572890025575446,
|
| 2824 |
+
"grad_norm": 11.476174354553223,
|
| 2825 |
+
"learning_rate": 0.00018772005653900977,
|
| 2826 |
+
"loss": 0.0821,
|
| 2827 |
+
"step": 398
|
| 2828 |
+
},
|
| 2829 |
+
{
|
| 2830 |
+
"epoch": 0.3265473145780051,
|
| 2831 |
+
"grad_norm": 4.880053997039795,
|
| 2832 |
+
"learning_rate": 0.00018765801329919166,
|
| 2833 |
+
"loss": 0.0475,
|
| 2834 |
+
"step": 399
|
| 2835 |
+
},
|
| 2836 |
+
{
|
| 2837 |
+
"epoch": 0.3273657289002558,
|
| 2838 |
+
"grad_norm": 4.687880039215088,
|
| 2839 |
+
"learning_rate": 0.00018759582402665814,
|
| 2840 |
+
"loss": 0.3302,
|
| 2841 |
+
"step": 400
|
| 2842 |
+
},
|
| 2843 |
+
{
|
| 2844 |
+
"epoch": 0.3273657289002558,
|
| 2845 |
+
"eval_loss": 0.01040154229849577,
|
| 2846 |
+
"eval_runtime": 236.3558,
|
| 2847 |
+
"eval_samples_per_second": 6.896,
|
| 2848 |
+
"eval_steps_per_second": 1.726,
|
| 2849 |
+
"step": 400
|
| 2850 |
}
|
| 2851 |
],
|
| 2852 |
"logging_steps": 1,
|
|
|
|
| 2861 |
"early_stopping_threshold": 0.0
|
| 2862 |
},
|
| 2863 |
"attributes": {
|
| 2864 |
+
"early_stopping_patience_counter": 2
|
| 2865 |
}
|
| 2866 |
},
|
| 2867 |
"TrainerControl": {
|
|
|
|
| 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": 1.980132394795008e+17,
|
| 2879 |
"train_batch_size": 4,
|
| 2880 |
"trial_name": null,
|
| 2881 |
"trial_params": null
|