Training in progress, step 369, 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 981504152
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9b967f5fb314ece18203ae13d12ca313ece97191263b173dbd8f52930ec43491
|
| 3 |
size 981504152
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1963151243
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:537833242174698ee5292a62363f0cb60ab06fd0fc3976cd2808153cf5b05b3d
|
| 3 |
size 1963151243
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14899
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:68a756dab9b6a4b92c54e44af273de52dd9b218a62bc7071b823fa8b4d203e34
|
| 3 |
size 14899
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1337
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0ab40a6cb214210dbfd8deaf5483d8adf2bd57ef057345184005fb87c9e322db
|
| 3 |
size 1337
|
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": 0.
|
| 6 |
"eval_steps": 15,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -2108,6 +2108,489 @@
|
|
| 2108 |
"learning_rate": 2.6296296296296297e-06,
|
| 2109 |
"loss": 0.2181,
|
| 2110 |
"step": 300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2111 |
}
|
| 2112 |
],
|
| 2113 |
"logging_steps": 1,
|
|
@@ -2122,7 +2605,7 @@
|
|
| 2122 |
"should_evaluate": false,
|
| 2123 |
"should_log": false,
|
| 2124 |
"should_save": true,
|
| 2125 |
-
"should_training_stop":
|
| 2126 |
},
|
| 2127 |
"attributes": {}
|
| 2128 |
}
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.017102484553015097,
|
| 6 |
"eval_steps": 15,
|
| 7 |
+
"global_step": 369,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 2108 |
"learning_rate": 2.6296296296296297e-06,
|
| 2109 |
"loss": 0.2181,
|
| 2110 |
"step": 300
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 0.013950807182811772,
|
| 2114 |
+
"grad_norm": 31.496883392333984,
|
| 2115 |
+
"learning_rate": 4.2592592592592596e-06,
|
| 2116 |
+
"loss": 0.2155,
|
| 2117 |
+
"step": 301
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 0.013997155379432408,
|
| 2121 |
+
"grad_norm": 30.944461822509766,
|
| 2122 |
+
"learning_rate": 5.888888888888889e-06,
|
| 2123 |
+
"loss": 0.2296,
|
| 2124 |
+
"step": 302
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 0.014043503576053046,
|
| 2128 |
+
"grad_norm": 12.571844100952148,
|
| 2129 |
+
"learning_rate": 7.5185185185185186e-06,
|
| 2130 |
+
"loss": 0.2175,
|
| 2131 |
+
"step": 303
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 0.014089851772673682,
|
| 2135 |
+
"grad_norm": 26.107559204101562,
|
| 2136 |
+
"learning_rate": 9.148148148148149e-06,
|
| 2137 |
+
"loss": 0.2146,
|
| 2138 |
+
"step": 304
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 0.01413619996929432,
|
| 2142 |
+
"grad_norm": 24.720565795898438,
|
| 2143 |
+
"learning_rate": 1.077777777777778e-05,
|
| 2144 |
+
"loss": 0.2173,
|
| 2145 |
+
"step": 305
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 0.014182548165914957,
|
| 2149 |
+
"grad_norm": 10.030445098876953,
|
| 2150 |
+
"learning_rate": 1.240740740740741e-05,
|
| 2151 |
+
"loss": 0.2199,
|
| 2152 |
+
"step": 306
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 0.014228896362535593,
|
| 2156 |
+
"grad_norm": 9.082711219787598,
|
| 2157 |
+
"learning_rate": 1.4037037037037038e-05,
|
| 2158 |
+
"loss": 0.2273,
|
| 2159 |
+
"step": 307
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 0.014275244559156231,
|
| 2163 |
+
"grad_norm": 24.175121307373047,
|
| 2164 |
+
"learning_rate": 1.5666666666666667e-05,
|
| 2165 |
+
"loss": 0.2208,
|
| 2166 |
+
"step": 308
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 0.014321592755776867,
|
| 2170 |
+
"grad_norm": 15.869462966918945,
|
| 2171 |
+
"learning_rate": 1.7296296296296297e-05,
|
| 2172 |
+
"loss": 0.226,
|
| 2173 |
+
"step": 309
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 0.014367940952397506,
|
| 2177 |
+
"grad_norm": 7.70182466506958,
|
| 2178 |
+
"learning_rate": 1.8925925925925928e-05,
|
| 2179 |
+
"loss": 0.2171,
|
| 2180 |
+
"step": 310
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 0.014414289149018142,
|
| 2184 |
+
"grad_norm": 6.203703880310059,
|
| 2185 |
+
"learning_rate": 2.0555555555555558e-05,
|
| 2186 |
+
"loss": 0.2234,
|
| 2187 |
+
"step": 311
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 0.01446063734563878,
|
| 2191 |
+
"grad_norm": 27.60795783996582,
|
| 2192 |
+
"learning_rate": 2.2185185185185188e-05,
|
| 2193 |
+
"loss": 0.2284,
|
| 2194 |
+
"step": 312
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 0.014506985542259416,
|
| 2198 |
+
"grad_norm": 18.060556411743164,
|
| 2199 |
+
"learning_rate": 2.381481481481482e-05,
|
| 2200 |
+
"loss": 0.2248,
|
| 2201 |
+
"step": 313
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 0.014553333738880054,
|
| 2205 |
+
"grad_norm": 5.732203960418701,
|
| 2206 |
+
"learning_rate": 2.544444444444445e-05,
|
| 2207 |
+
"loss": 0.2168,
|
| 2208 |
+
"step": 314
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 0.01459968193550069,
|
| 2212 |
+
"grad_norm": 15.22141170501709,
|
| 2213 |
+
"learning_rate": 2.7074074074074076e-05,
|
| 2214 |
+
"loss": 0.2298,
|
| 2215 |
+
"step": 315
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 0.014646030132121329,
|
| 2219 |
+
"grad_norm": 33.845497131347656,
|
| 2220 |
+
"learning_rate": 2.870370370370371e-05,
|
| 2221 |
+
"loss": 0.2332,
|
| 2222 |
+
"step": 316
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 0.014692378328741965,
|
| 2226 |
+
"grad_norm": 16.4891357421875,
|
| 2227 |
+
"learning_rate": 3.0333333333333337e-05,
|
| 2228 |
+
"loss": 0.2238,
|
| 2229 |
+
"step": 317
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 0.014738726525362603,
|
| 2233 |
+
"grad_norm": 6.198254585266113,
|
| 2234 |
+
"learning_rate": 3.196296296296297e-05,
|
| 2235 |
+
"loss": 0.2233,
|
| 2236 |
+
"step": 318
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 0.01478507472198324,
|
| 2240 |
+
"grad_norm": 4.290202617645264,
|
| 2241 |
+
"learning_rate": 3.359259259259259e-05,
|
| 2242 |
+
"loss": 0.2249,
|
| 2243 |
+
"step": 319
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 0.014831422918603876,
|
| 2247 |
+
"grad_norm": 2.0879788398742676,
|
| 2248 |
+
"learning_rate": 3.522222222222223e-05,
|
| 2249 |
+
"loss": 0.224,
|
| 2250 |
+
"step": 320
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 0.014877771115224514,
|
| 2254 |
+
"grad_norm": 2.1555585861206055,
|
| 2255 |
+
"learning_rate": 3.685185185185185e-05,
|
| 2256 |
+
"loss": 0.2214,
|
| 2257 |
+
"step": 321
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 0.01492411931184515,
|
| 2261 |
+
"grad_norm": 6.381463050842285,
|
| 2262 |
+
"learning_rate": 3.848148148148148e-05,
|
| 2263 |
+
"loss": 0.2093,
|
| 2264 |
+
"step": 322
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 0.014970467508465788,
|
| 2268 |
+
"grad_norm": 47.52433776855469,
|
| 2269 |
+
"learning_rate": 4.011111111111111e-05,
|
| 2270 |
+
"loss": 0.2229,
|
| 2271 |
+
"step": 323
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 0.015016815705086424,
|
| 2275 |
+
"grad_norm": 8.799651145935059,
|
| 2276 |
+
"learning_rate": 4.174074074074074e-05,
|
| 2277 |
+
"loss": 0.2265,
|
| 2278 |
+
"step": 324
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 0.015063163901707062,
|
| 2282 |
+
"grad_norm": 40.708980560302734,
|
| 2283 |
+
"learning_rate": 4.337037037037037e-05,
|
| 2284 |
+
"loss": 0.2318,
|
| 2285 |
+
"step": 325
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 0.015109512098327699,
|
| 2289 |
+
"grad_norm": 42.69667434692383,
|
| 2290 |
+
"learning_rate": 4.5e-05,
|
| 2291 |
+
"loss": 0.218,
|
| 2292 |
+
"step": 326
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 0.015155860294948337,
|
| 2296 |
+
"grad_norm": 6.28260612487793,
|
| 2297 |
+
"learning_rate": 4.4998161456819626e-05,
|
| 2298 |
+
"loss": 0.2196,
|
| 2299 |
+
"step": 327
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 0.015202208491568973,
|
| 2303 |
+
"grad_norm": 94.07806396484375,
|
| 2304 |
+
"learning_rate": 4.4992646134573125e-05,
|
| 2305 |
+
"loss": 0.3395,
|
| 2306 |
+
"step": 328
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 0.015248556688189611,
|
| 2310 |
+
"grad_norm": 21.78129768371582,
|
| 2311 |
+
"learning_rate": 4.498345495509307e-05,
|
| 2312 |
+
"loss": 0.352,
|
| 2313 |
+
"step": 329
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 0.015294904884810247,
|
| 2317 |
+
"grad_norm": 16.720928192138672,
|
| 2318 |
+
"learning_rate": 4.4970589454595836e-05,
|
| 2319 |
+
"loss": 0.3217,
|
| 2320 |
+
"step": 330
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 0.015341253081430885,
|
| 2324 |
+
"grad_norm": 23.63299560546875,
|
| 2325 |
+
"learning_rate": 4.495405178342491e-05,
|
| 2326 |
+
"loss": 0.2872,
|
| 2327 |
+
"step": 331
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 0.015387601278051522,
|
| 2331 |
+
"grad_norm": 11081.3798828125,
|
| 2332 |
+
"learning_rate": 4.4933844705691426e-05,
|
| 2333 |
+
"loss": 0.6065,
|
| 2334 |
+
"step": 332
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 0.015433949474672158,
|
| 2338 |
+
"grad_norm": 5464.720703125,
|
| 2339 |
+
"learning_rate": 4.490997159881221e-05,
|
| 2340 |
+
"loss": 1.5567,
|
| 2341 |
+
"step": 333
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 0.015480297671292796,
|
| 2345 |
+
"grad_norm": 9930.943359375,
|
| 2346 |
+
"learning_rate": 4.488243645294524e-05,
|
| 2347 |
+
"loss": 1.7496,
|
| 2348 |
+
"step": 334
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 0.015526645867913432,
|
| 2352 |
+
"grad_norm": 2429.20166015625,
|
| 2353 |
+
"learning_rate": 4.485124387032275e-05,
|
| 2354 |
+
"loss": 1.6667,
|
| 2355 |
+
"step": 335
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 0.01557299406453407,
|
| 2359 |
+
"grad_norm": 2331.36865234375,
|
| 2360 |
+
"learning_rate": 4.481639906448203e-05,
|
| 2361 |
+
"loss": 1.3543,
|
| 2362 |
+
"step": 336
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 0.015619342261154707,
|
| 2366 |
+
"grad_norm": 1066.0982666015625,
|
| 2367 |
+
"learning_rate": 4.4777907859394e-05,
|
| 2368 |
+
"loss": 0.7081,
|
| 2369 |
+
"step": 337
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 0.015665690457775345,
|
| 2373 |
+
"grad_norm": 2704.669677734375,
|
| 2374 |
+
"learning_rate": 4.4735776688489807e-05,
|
| 2375 |
+
"loss": 0.4249,
|
| 2376 |
+
"step": 338
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 0.015712038654395983,
|
| 2380 |
+
"grad_norm": 570.82763671875,
|
| 2381 |
+
"learning_rate": 4.4690012593585524e-05,
|
| 2382 |
+
"loss": 0.402,
|
| 2383 |
+
"step": 339
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 0.015758386851016618,
|
| 2387 |
+
"grad_norm": 214.00624084472656,
|
| 2388 |
+
"learning_rate": 4.464062322370523e-05,
|
| 2389 |
+
"loss": 0.3586,
|
| 2390 |
+
"step": 340
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 0.015804735047637256,
|
| 2394 |
+
"grad_norm": 163.50082397460938,
|
| 2395 |
+
"learning_rate": 4.4587616833802466e-05,
|
| 2396 |
+
"loss": 0.3269,
|
| 2397 |
+
"step": 341
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 0.015851083244257894,
|
| 2401 |
+
"grad_norm": 226.10430908203125,
|
| 2402 |
+
"learning_rate": 4.4531002283380575e-05,
|
| 2403 |
+
"loss": 0.3277,
|
| 2404 |
+
"step": 342
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 0.01589743144087853,
|
| 2408 |
+
"grad_norm": 267.2003479003906,
|
| 2409 |
+
"learning_rate": 4.447078903501188e-05,
|
| 2410 |
+
"loss": 0.3227,
|
| 2411 |
+
"step": 343
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 0.015943779637499166,
|
| 2415 |
+
"grad_norm": 20.03980827331543,
|
| 2416 |
+
"learning_rate": 4.4406987152756125e-05,
|
| 2417 |
+
"loss": 0.3159,
|
| 2418 |
+
"step": 344
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 0.015990127834119804,
|
| 2422 |
+
"grad_norm": 30.53392219543457,
|
| 2423 |
+
"learning_rate": 4.4339607300478356e-05,
|
| 2424 |
+
"loss": 0.3176,
|
| 2425 |
+
"step": 345
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 0.016036476030740442,
|
| 2429 |
+
"grad_norm": 25.638111114501953,
|
| 2430 |
+
"learning_rate": 4.4268660740066566e-05,
|
| 2431 |
+
"loss": 0.2905,
|
| 2432 |
+
"step": 346
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 0.016082824227361077,
|
| 2436 |
+
"grad_norm": 93.37738800048828,
|
| 2437 |
+
"learning_rate": 4.419415932954937e-05,
|
| 2438 |
+
"loss": 0.2864,
|
| 2439 |
+
"step": 347
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 0.016129172423981715,
|
| 2443 |
+
"grad_norm": 189.1175994873047,
|
| 2444 |
+
"learning_rate": 4.411611552111406e-05,
|
| 2445 |
+
"loss": 0.2869,
|
| 2446 |
+
"step": 348
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 0.016175520620602353,
|
| 2450 |
+
"grad_norm": 13.896951675415039,
|
| 2451 |
+
"learning_rate": 4.403454235902536e-05,
|
| 2452 |
+
"loss": 0.2696,
|
| 2453 |
+
"step": 349
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 0.01622186881722299,
|
| 2457 |
+
"grad_norm": 14.850593566894531,
|
| 2458 |
+
"learning_rate": 4.394945347744514e-05,
|
| 2459 |
+
"loss": 0.2708,
|
| 2460 |
+
"step": 350
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 0.016268217013843626,
|
| 2464 |
+
"grad_norm": 9.54691219329834,
|
| 2465 |
+
"learning_rate": 4.386086309815368e-05,
|
| 2466 |
+
"loss": 0.2602,
|
| 2467 |
+
"step": 351
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 0.016314565210464264,
|
| 2471 |
+
"grad_norm": 9.796507835388184,
|
| 2472 |
+
"learning_rate": 4.37687860281726e-05,
|
| 2473 |
+
"loss": 0.2612,
|
| 2474 |
+
"step": 352
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 0.016360913407084902,
|
| 2478 |
+
"grad_norm": 111.0169906616211,
|
| 2479 |
+
"learning_rate": 4.3673237657289986e-05,
|
| 2480 |
+
"loss": 0.2505,
|
| 2481 |
+
"step": 353
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 0.01640726160370554,
|
| 2485 |
+
"grad_norm": 29.732372283935547,
|
| 2486 |
+
"learning_rate": 4.3574233955488175e-05,
|
| 2487 |
+
"loss": 0.257,
|
| 2488 |
+
"step": 354
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"epoch": 0.016453609800326174,
|
| 2492 |
+
"grad_norm": 728.8468017578125,
|
| 2493 |
+
"learning_rate": 4.347179147027451e-05,
|
| 2494 |
+
"loss": 0.268,
|
| 2495 |
+
"step": 355
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 0.016499957996946812,
|
| 2499 |
+
"grad_norm": 9.074967384338379,
|
| 2500 |
+
"learning_rate": 4.3365927323915595e-05,
|
| 2501 |
+
"loss": 0.2465,
|
| 2502 |
+
"step": 356
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 0.01654630619356745,
|
| 2506 |
+
"grad_norm": 11.821940422058105,
|
| 2507 |
+
"learning_rate": 4.325665921057545e-05,
|
| 2508 |
+
"loss": 0.25,
|
| 2509 |
+
"step": 357
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 0.016592654390188085,
|
| 2513 |
+
"grad_norm": 37.2886848449707,
|
| 2514 |
+
"learning_rate": 4.3144005393358136e-05,
|
| 2515 |
+
"loss": 0.2583,
|
| 2516 |
+
"step": 358
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 0.016639002586808723,
|
| 2520 |
+
"grad_norm": 15.582650184631348,
|
| 2521 |
+
"learning_rate": 4.302798470125523e-05,
|
| 2522 |
+
"loss": 0.2509,
|
| 2523 |
+
"step": 359
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 0.01668535078342936,
|
| 2527 |
+
"grad_norm": 14.96511173248291,
|
| 2528 |
+
"learning_rate": 4.290861652599876e-05,
|
| 2529 |
+
"loss": 0.2387,
|
| 2530 |
+
"step": 360
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"epoch": 0.01673169898005,
|
| 2534 |
+
"grad_norm": 10.343366622924805,
|
| 2535 |
+
"learning_rate": 4.278592081882005e-05,
|
| 2536 |
+
"loss": 0.2301,
|
| 2537 |
+
"step": 361
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 0.016778047176670634,
|
| 2541 |
+
"grad_norm": 16.448528289794922,
|
| 2542 |
+
"learning_rate": 4.265991808711507e-05,
|
| 2543 |
+
"loss": 0.2357,
|
| 2544 |
+
"step": 362
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 0.016824395373291272,
|
| 2548 |
+
"grad_norm": 15.472856521606445,
|
| 2549 |
+
"learning_rate": 4.253062939101684e-05,
|
| 2550 |
+
"loss": 0.2433,
|
| 2551 |
+
"step": 363
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 0.01687074356991191,
|
| 2555 |
+
"grad_norm": 5.16851806640625,
|
| 2556 |
+
"learning_rate": 4.2398076339875376e-05,
|
| 2557 |
+
"loss": 0.2418,
|
| 2558 |
+
"step": 364
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 0.016917091766532548,
|
| 2562 |
+
"grad_norm": 4.729401588439941,
|
| 2563 |
+
"learning_rate": 4.2262281088646e-05,
|
| 2564 |
+
"loss": 0.2349,
|
| 2565 |
+
"step": 365
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 0.016963439963153183,
|
| 2569 |
+
"grad_norm": 2.712707281112671,
|
| 2570 |
+
"learning_rate": 4.212326633418625e-05,
|
| 2571 |
+
"loss": 0.2313,
|
| 2572 |
+
"step": 366
|
| 2573 |
+
},
|
| 2574 |
+
{
|
| 2575 |
+
"epoch": 0.01700978815977382,
|
| 2576 |
+
"grad_norm": 2.699162721633911,
|
| 2577 |
+
"learning_rate": 4.198105531146239e-05,
|
| 2578 |
+
"loss": 0.2288,
|
| 2579 |
+
"step": 367
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 0.01705613635639446,
|
| 2583 |
+
"grad_norm": 2.176575183868408,
|
| 2584 |
+
"learning_rate": 4.183567178966586e-05,
|
| 2585 |
+
"loss": 0.2324,
|
| 2586 |
+
"step": 368
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 0.017102484553015097,
|
| 2590 |
+
"grad_norm": 2.757868528366089,
|
| 2591 |
+
"learning_rate": 4.1687140068240544e-05,
|
| 2592 |
+
"loss": 0.2307,
|
| 2593 |
+
"step": 369
|
| 2594 |
}
|
| 2595 |
],
|
| 2596 |
"logging_steps": 1,
|
|
|
|
| 2605 |
"should_evaluate": false,
|
| 2606 |
"should_log": false,
|
| 2607 |
"should_save": true,
|
| 2608 |
+
"should_training_stop": true
|
| 2609 |
},
|
| 2610 |
"attributes": {}
|
| 2611 |
}
|