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 144805440
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:30bed0be07b2d262a7c9c2ee8a0b78c224c223f1dfe6648616e1f608c414714b
|
3 |
size 144805440
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 74292308
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c2c1169ba7d2c7f4cb773678dd77d250d9458b9f604eb63ec725d7237ee35516
|
3 |
size 74292308
|
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 sha256:557756899814dbd279052b049fe3fe1a6703f1aa93f393d968ec951e02adc041
|
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:dfd113e2bf8a699da59a6073bb1bcb18aa32fb9667d59893641ceb623ba2aa2b
|
3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
-
"best_metric": 1.
|
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": 3.977,
|
2140 |
"eval_steps_per_second": 0.994,
|
2141 |
"step": 300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2142 |
}
|
2143 |
],
|
2144 |
"logging_steps": 1,
|
@@ -2167,7 +2875,7 @@
|
|
2167 |
"attributes": {}
|
2168 |
}
|
2169 |
},
|
2170 |
-
"total_flos":
|
2171 |
"train_batch_size": 4,
|
2172 |
"trial_name": null,
|
2173 |
"trial_params": null
|
|
|
1 |
{
|
2 |
+
"best_metric": 1.2253398895263672,
|
3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-400",
|
4 |
+
"epoch": 0.05085217114933852,
|
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": 3.977,
|
2140 |
"eval_steps_per_second": 0.994,
|
2141 |
"step": 300
|
2142 |
+
},
|
2143 |
+
{
|
2144 |
+
"epoch": 0.03826625878987724,
|
2145 |
+
"grad_norm": 0.19571375846862793,
|
2146 |
+
"learning_rate": 8.843335487049712e-05,
|
2147 |
+
"loss": 1.2259,
|
2148 |
+
"step": 301
|
2149 |
+
},
|
2150 |
+
{
|
2151 |
+
"epoch": 0.03839338921775059,
|
2152 |
+
"grad_norm": 0.19907177984714508,
|
2153 |
+
"learning_rate": 8.785781316514841e-05,
|
2154 |
+
"loss": 1.1946,
|
2155 |
+
"step": 302
|
2156 |
+
},
|
2157 |
+
{
|
2158 |
+
"epoch": 0.03852051964562393,
|
2159 |
+
"grad_norm": 0.19650404155254364,
|
2160 |
+
"learning_rate": 8.728267940034078e-05,
|
2161 |
+
"loss": 1.1977,
|
2162 |
+
"step": 303
|
2163 |
+
},
|
2164 |
+
{
|
2165 |
+
"epoch": 0.038647650073497276,
|
2166 |
+
"grad_norm": 0.1937037855386734,
|
2167 |
+
"learning_rate": 8.670797289881915e-05,
|
2168 |
+
"loss": 1.1719,
|
2169 |
+
"step": 304
|
2170 |
+
},
|
2171 |
+
{
|
2172 |
+
"epoch": 0.03877478050137063,
|
2173 |
+
"grad_norm": 0.2028638869524002,
|
2174 |
+
"learning_rate": 8.61337129689737e-05,
|
2175 |
+
"loss": 1.224,
|
2176 |
+
"step": 305
|
2177 |
+
},
|
2178 |
+
{
|
2179 |
+
"epoch": 0.03890191092924397,
|
2180 |
+
"grad_norm": 0.19181466102600098,
|
2181 |
+
"learning_rate": 8.555991890419117e-05,
|
2182 |
+
"loss": 1.2375,
|
2183 |
+
"step": 306
|
2184 |
+
},
|
2185 |
+
{
|
2186 |
+
"epoch": 0.039029041357117315,
|
2187 |
+
"grad_norm": 0.1887066513299942,
|
2188 |
+
"learning_rate": 8.498660998220669e-05,
|
2189 |
+
"loss": 1.1786,
|
2190 |
+
"step": 307
|
2191 |
+
},
|
2192 |
+
{
|
2193 |
+
"epoch": 0.039156171784990666,
|
2194 |
+
"grad_norm": 0.18820159137248993,
|
2195 |
+
"learning_rate": 8.441380546445603e-05,
|
2196 |
+
"loss": 1.2536,
|
2197 |
+
"step": 308
|
2198 |
+
},
|
2199 |
+
{
|
2200 |
+
"epoch": 0.03928330221286401,
|
2201 |
+
"grad_norm": 0.18667809665203094,
|
2202 |
+
"learning_rate": 8.384152459542848e-05,
|
2203 |
+
"loss": 1.2834,
|
2204 |
+
"step": 309
|
2205 |
+
},
|
2206 |
+
{
|
2207 |
+
"epoch": 0.039410432640737354,
|
2208 |
+
"grad_norm": 0.19369632005691528,
|
2209 |
+
"learning_rate": 8.326978660202034e-05,
|
2210 |
+
"loss": 1.2989,
|
2211 |
+
"step": 310
|
2212 |
+
},
|
2213 |
+
{
|
2214 |
+
"epoch": 0.039537563068610705,
|
2215 |
+
"grad_norm": 0.18864746391773224,
|
2216 |
+
"learning_rate": 8.269861069288903e-05,
|
2217 |
+
"loss": 1.292,
|
2218 |
+
"step": 311
|
2219 |
+
},
|
2220 |
+
{
|
2221 |
+
"epoch": 0.03966469349648405,
|
2222 |
+
"grad_norm": 0.19906029105186462,
|
2223 |
+
"learning_rate": 8.212801605780753e-05,
|
2224 |
+
"loss": 1.2855,
|
2225 |
+
"step": 312
|
2226 |
+
},
|
2227 |
+
{
|
2228 |
+
"epoch": 0.03979182392435739,
|
2229 |
+
"grad_norm": 0.19700987637043,
|
2230 |
+
"learning_rate": 8.155802186701984e-05,
|
2231 |
+
"loss": 1.1771,
|
2232 |
+
"step": 313
|
2233 |
+
},
|
2234 |
+
{
|
2235 |
+
"epoch": 0.039918954352230744,
|
2236 |
+
"grad_norm": 0.20134317874908447,
|
2237 |
+
"learning_rate": 8.098864727059685e-05,
|
2238 |
+
"loss": 1.1995,
|
2239 |
+
"step": 314
|
2240 |
+
},
|
2241 |
+
{
|
2242 |
+
"epoch": 0.04004608478010409,
|
2243 |
+
"grad_norm": 0.1883343607187271,
|
2244 |
+
"learning_rate": 8.04199113977929e-05,
|
2245 |
+
"loss": 1.2433,
|
2246 |
+
"step": 315
|
2247 |
+
},
|
2248 |
+
{
|
2249 |
+
"epoch": 0.04017321520797743,
|
2250 |
+
"grad_norm": 0.19041708111763,
|
2251 |
+
"learning_rate": 7.985183335640331e-05,
|
2252 |
+
"loss": 1.2538,
|
2253 |
+
"step": 316
|
2254 |
+
},
|
2255 |
+
{
|
2256 |
+
"epoch": 0.04030034563585078,
|
2257 |
+
"grad_norm": 0.1838679164648056,
|
2258 |
+
"learning_rate": 7.928443223212215e-05,
|
2259 |
+
"loss": 1.2025,
|
2260 |
+
"step": 317
|
2261 |
+
},
|
2262 |
+
{
|
2263 |
+
"epoch": 0.04042747606372413,
|
2264 |
+
"grad_norm": 0.19493237137794495,
|
2265 |
+
"learning_rate": 7.871772708790114e-05,
|
2266 |
+
"loss": 1.2553,
|
2267 |
+
"step": 318
|
2268 |
+
},
|
2269 |
+
{
|
2270 |
+
"epoch": 0.04055460649159747,
|
2271 |
+
"grad_norm": 0.197859525680542,
|
2272 |
+
"learning_rate": 7.815173696330919e-05,
|
2273 |
+
"loss": 1.2661,
|
2274 |
+
"step": 319
|
2275 |
+
},
|
2276 |
+
{
|
2277 |
+
"epoch": 0.04068173691947082,
|
2278 |
+
"grad_norm": 0.19427183270454407,
|
2279 |
+
"learning_rate": 7.758648087389277e-05,
|
2280 |
+
"loss": 1.2121,
|
2281 |
+
"step": 320
|
2282 |
+
},
|
2283 |
+
{
|
2284 |
+
"epoch": 0.040808867347344166,
|
2285 |
+
"grad_norm": 0.19236573576927185,
|
2286 |
+
"learning_rate": 7.702197781053696e-05,
|
2287 |
+
"loss": 1.2375,
|
2288 |
+
"step": 321
|
2289 |
+
},
|
2290 |
+
{
|
2291 |
+
"epoch": 0.04093599777521751,
|
2292 |
+
"grad_norm": 0.19856838881969452,
|
2293 |
+
"learning_rate": 7.645824673882748e-05,
|
2294 |
+
"loss": 1.2648,
|
2295 |
+
"step": 322
|
2296 |
+
},
|
2297 |
+
{
|
2298 |
+
"epoch": 0.04106312820309086,
|
2299 |
+
"grad_norm": 0.20721471309661865,
|
2300 |
+
"learning_rate": 7.589530659841349e-05,
|
2301 |
+
"loss": 1.2503,
|
2302 |
+
"step": 323
|
2303 |
+
},
|
2304 |
+
{
|
2305 |
+
"epoch": 0.041190258630964205,
|
2306 |
+
"grad_norm": 0.19413287937641144,
|
2307 |
+
"learning_rate": 7.533317630237117e-05,
|
2308 |
+
"loss": 1.265,
|
2309 |
+
"step": 324
|
2310 |
+
},
|
2311 |
+
{
|
2312 |
+
"epoch": 0.04131738905883755,
|
2313 |
+
"grad_norm": 0.1948065459728241,
|
2314 |
+
"learning_rate": 7.477187473656853e-05,
|
2315 |
+
"loss": 1.2581,
|
2316 |
+
"step": 325
|
2317 |
+
},
|
2318 |
+
{
|
2319 |
+
"epoch": 0.0414445194867109,
|
2320 |
+
"grad_norm": 0.19630448520183563,
|
2321 |
+
"learning_rate": 7.421142075903067e-05,
|
2322 |
+
"loss": 1.2013,
|
2323 |
+
"step": 326
|
2324 |
+
},
|
2325 |
+
{
|
2326 |
+
"epoch": 0.041571649914584244,
|
2327 |
+
"grad_norm": 0.18867121636867523,
|
2328 |
+
"learning_rate": 7.365183319930635e-05,
|
2329 |
+
"loss": 1.1628,
|
2330 |
+
"step": 327
|
2331 |
+
},
|
2332 |
+
{
|
2333 |
+
"epoch": 0.04169878034245759,
|
2334 |
+
"grad_norm": 0.2017098367214203,
|
2335 |
+
"learning_rate": 7.309313085783524e-05,
|
2336 |
+
"loss": 1.1882,
|
2337 |
+
"step": 328
|
2338 |
+
},
|
2339 |
+
{
|
2340 |
+
"epoch": 0.04182591077033094,
|
2341 |
+
"grad_norm": 0.19574840366840363,
|
2342 |
+
"learning_rate": 7.253533250531656e-05,
|
2343 |
+
"loss": 1.1917,
|
2344 |
+
"step": 329
|
2345 |
+
},
|
2346 |
+
{
|
2347 |
+
"epoch": 0.04195304119820428,
|
2348 |
+
"grad_norm": 0.2003111094236374,
|
2349 |
+
"learning_rate": 7.197845688207805e-05,
|
2350 |
+
"loss": 1.3069,
|
2351 |
+
"step": 330
|
2352 |
+
},
|
2353 |
+
{
|
2354 |
+
"epoch": 0.04208017162607763,
|
2355 |
+
"grad_norm": 0.19444699585437775,
|
2356 |
+
"learning_rate": 7.142252269744665e-05,
|
2357 |
+
"loss": 1.1623,
|
2358 |
+
"step": 331
|
2359 |
+
},
|
2360 |
+
{
|
2361 |
+
"epoch": 0.04220730205395098,
|
2362 |
+
"grad_norm": 0.19306592643260956,
|
2363 |
+
"learning_rate": 7.086754862911982e-05,
|
2364 |
+
"loss": 1.2512,
|
2365 |
+
"step": 332
|
2366 |
+
},
|
2367 |
+
{
|
2368 |
+
"epoch": 0.04233443248182432,
|
2369 |
+
"grad_norm": 0.1940678060054779,
|
2370 |
+
"learning_rate": 7.031355332253795e-05,
|
2371 |
+
"loss": 1.2404,
|
2372 |
+
"step": 333
|
2373 |
+
},
|
2374 |
+
{
|
2375 |
+
"epoch": 0.042461562909697666,
|
2376 |
+
"grad_norm": 0.20104342699050903,
|
2377 |
+
"learning_rate": 6.976055539025818e-05,
|
2378 |
+
"loss": 1.1826,
|
2379 |
+
"step": 334
|
2380 |
+
},
|
2381 |
+
{
|
2382 |
+
"epoch": 0.04258869333757102,
|
2383 |
+
"grad_norm": 0.20509012043476105,
|
2384 |
+
"learning_rate": 6.92085734113288e-05,
|
2385 |
+
"loss": 1.2247,
|
2386 |
+
"step": 335
|
2387 |
+
},
|
2388 |
+
{
|
2389 |
+
"epoch": 0.04271582376544436,
|
2390 |
+
"grad_norm": 0.2038545161485672,
|
2391 |
+
"learning_rate": 6.865762593066513e-05,
|
2392 |
+
"loss": 1.25,
|
2393 |
+
"step": 336
|
2394 |
+
},
|
2395 |
+
{
|
2396 |
+
"epoch": 0.042842954193317705,
|
2397 |
+
"grad_norm": 0.20035366714000702,
|
2398 |
+
"learning_rate": 6.810773145842653e-05,
|
2399 |
+
"loss": 1.2243,
|
2400 |
+
"step": 337
|
2401 |
+
},
|
2402 |
+
{
|
2403 |
+
"epoch": 0.042970084621191056,
|
2404 |
+
"grad_norm": 0.20092110335826874,
|
2405 |
+
"learning_rate": 6.755890846939454e-05,
|
2406 |
+
"loss": 1.2279,
|
2407 |
+
"step": 338
|
2408 |
+
},
|
2409 |
+
{
|
2410 |
+
"epoch": 0.0430972150490644,
|
2411 |
+
"grad_norm": 0.20261281728744507,
|
2412 |
+
"learning_rate": 6.701117540235204e-05,
|
2413 |
+
"loss": 1.2418,
|
2414 |
+
"step": 339
|
2415 |
+
},
|
2416 |
+
{
|
2417 |
+
"epoch": 0.043224345476937744,
|
2418 |
+
"grad_norm": 0.20916980504989624,
|
2419 |
+
"learning_rate": 6.646455065946386e-05,
|
2420 |
+
"loss": 1.2205,
|
2421 |
+
"step": 340
|
2422 |
+
},
|
2423 |
+
{
|
2424 |
+
"epoch": 0.043351475904811095,
|
2425 |
+
"grad_norm": 0.1868792027235031,
|
2426 |
+
"learning_rate": 6.591905260565852e-05,
|
2427 |
+
"loss": 1.2149,
|
2428 |
+
"step": 341
|
2429 |
+
},
|
2430 |
+
{
|
2431 |
+
"epoch": 0.04347860633268444,
|
2432 |
+
"grad_norm": 0.19856908917427063,
|
2433 |
+
"learning_rate": 6.537469956801128e-05,
|
2434 |
+
"loss": 1.2518,
|
2435 |
+
"step": 342
|
2436 |
+
},
|
2437 |
+
{
|
2438 |
+
"epoch": 0.04360573676055778,
|
2439 |
+
"grad_norm": 0.19585344195365906,
|
2440 |
+
"learning_rate": 6.483150983512823e-05,
|
2441 |
+
"loss": 1.2202,
|
2442 |
+
"step": 343
|
2443 |
+
},
|
2444 |
+
{
|
2445 |
+
"epoch": 0.04373286718843113,
|
2446 |
+
"grad_norm": 0.19705970585346222,
|
2447 |
+
"learning_rate": 6.428950165653204e-05,
|
2448 |
+
"loss": 1.2701,
|
2449 |
+
"step": 344
|
2450 |
+
},
|
2451 |
+
{
|
2452 |
+
"epoch": 0.04385999761630448,
|
2453 |
+
"grad_norm": 0.19830965995788574,
|
2454 |
+
"learning_rate": 6.374869324204869e-05,
|
2455 |
+
"loss": 1.2132,
|
2456 |
+
"step": 345
|
2457 |
+
},
|
2458 |
+
{
|
2459 |
+
"epoch": 0.04398712804417782,
|
2460 |
+
"grad_norm": 0.20360921323299408,
|
2461 |
+
"learning_rate": 6.320910276119576e-05,
|
2462 |
+
"loss": 1.1979,
|
2463 |
+
"step": 346
|
2464 |
+
},
|
2465 |
+
{
|
2466 |
+
"epoch": 0.04411425847205117,
|
2467 |
+
"grad_norm": 0.20261693000793457,
|
2468 |
+
"learning_rate": 6.267074834257199e-05,
|
2469 |
+
"loss": 1.2231,
|
2470 |
+
"step": 347
|
2471 |
+
},
|
2472 |
+
{
|
2473 |
+
"epoch": 0.044241388899924516,
|
2474 |
+
"grad_norm": 0.19419489800930023,
|
2475 |
+
"learning_rate": 6.213364807324818e-05,
|
2476 |
+
"loss": 1.1575,
|
2477 |
+
"step": 348
|
2478 |
+
},
|
2479 |
+
{
|
2480 |
+
"epoch": 0.04436851932779786,
|
2481 |
+
"grad_norm": 0.2025313675403595,
|
2482 |
+
"learning_rate": 6.15978199981595e-05,
|
2483 |
+
"loss": 1.2566,
|
2484 |
+
"step": 349
|
2485 |
+
},
|
2486 |
+
{
|
2487 |
+
"epoch": 0.04449564975567121,
|
2488 |
+
"grad_norm": 0.19754880666732788,
|
2489 |
+
"learning_rate": 6.106328211949928e-05,
|
2490 |
+
"loss": 1.2821,
|
2491 |
+
"step": 350
|
2492 |
+
},
|
2493 |
+
{
|
2494 |
+
"epoch": 0.044622780183544555,
|
2495 |
+
"grad_norm": 0.20343464612960815,
|
2496 |
+
"learning_rate": 6.053005239611418e-05,
|
2497 |
+
"loss": 1.2204,
|
2498 |
+
"step": 351
|
2499 |
+
},
|
2500 |
+
{
|
2501 |
+
"epoch": 0.0447499106114179,
|
2502 |
+
"grad_norm": 0.20527192950248718,
|
2503 |
+
"learning_rate": 5.999814874290084e-05,
|
2504 |
+
"loss": 1.2513,
|
2505 |
+
"step": 352
|
2506 |
+
},
|
2507 |
+
{
|
2508 |
+
"epoch": 0.04487704103929125,
|
2509 |
+
"grad_norm": 0.19321362674236298,
|
2510 |
+
"learning_rate": 5.946758903020393e-05,
|
2511 |
+
"loss": 1.2466,
|
2512 |
+
"step": 353
|
2513 |
+
},
|
2514 |
+
{
|
2515 |
+
"epoch": 0.045004171467164594,
|
2516 |
+
"grad_norm": 0.20470896363258362,
|
2517 |
+
"learning_rate": 5.893839108321584e-05,
|
2518 |
+
"loss": 1.2846,
|
2519 |
+
"step": 354
|
2520 |
+
},
|
2521 |
+
{
|
2522 |
+
"epoch": 0.04513130189503794,
|
2523 |
+
"grad_norm": 0.19084323942661285,
|
2524 |
+
"learning_rate": 5.841057268137771e-05,
|
2525 |
+
"loss": 1.2126,
|
2526 |
+
"step": 355
|
2527 |
+
},
|
2528 |
+
{
|
2529 |
+
"epoch": 0.04525843232291129,
|
2530 |
+
"grad_norm": 0.19587008655071259,
|
2531 |
+
"learning_rate": 5.7884151557782305e-05,
|
2532 |
+
"loss": 1.1983,
|
2533 |
+
"step": 356
|
2534 |
+
},
|
2535 |
+
{
|
2536 |
+
"epoch": 0.04538556275078463,
|
2537 |
+
"grad_norm": 0.20390859246253967,
|
2538 |
+
"learning_rate": 5.735914539857798e-05,
|
2539 |
+
"loss": 1.1981,
|
2540 |
+
"step": 357
|
2541 |
+
},
|
2542 |
+
{
|
2543 |
+
"epoch": 0.04551269317865798,
|
2544 |
+
"grad_norm": 0.19584935903549194,
|
2545 |
+
"learning_rate": 5.68355718423746e-05,
|
2546 |
+
"loss": 1.2039,
|
2547 |
+
"step": 358
|
2548 |
+
},
|
2549 |
+
{
|
2550 |
+
"epoch": 0.04563982360653133,
|
2551 |
+
"grad_norm": 0.19530071318149567,
|
2552 |
+
"learning_rate": 5.6313448479650946e-05,
|
2553 |
+
"loss": 1.236,
|
2554 |
+
"step": 359
|
2555 |
+
},
|
2556 |
+
{
|
2557 |
+
"epoch": 0.04576695403440467,
|
2558 |
+
"grad_norm": 0.19659969210624695,
|
2559 |
+
"learning_rate": 5.579279285216369e-05,
|
2560 |
+
"loss": 1.1936,
|
2561 |
+
"step": 360
|
2562 |
+
},
|
2563 |
+
{
|
2564 |
+
"epoch": 0.045894084462278016,
|
2565 |
+
"grad_norm": 0.1933298110961914,
|
2566 |
+
"learning_rate": 5.527362245235805e-05,
|
2567 |
+
"loss": 1.227,
|
2568 |
+
"step": 361
|
2569 |
+
},
|
2570 |
+
{
|
2571 |
+
"epoch": 0.04602121489015137,
|
2572 |
+
"grad_norm": 0.20280398428440094,
|
2573 |
+
"learning_rate": 5.475595472278024e-05,
|
2574 |
+
"loss": 1.2644,
|
2575 |
+
"step": 362
|
2576 |
+
},
|
2577 |
+
{
|
2578 |
+
"epoch": 0.04614834531802471,
|
2579 |
+
"grad_norm": 0.1918189376592636,
|
2580 |
+
"learning_rate": 5.4239807055491135e-05,
|
2581 |
+
"loss": 1.1495,
|
2582 |
+
"step": 363
|
2583 |
+
},
|
2584 |
+
{
|
2585 |
+
"epoch": 0.046275475745898055,
|
2586 |
+
"grad_norm": 0.2044762223958969,
|
2587 |
+
"learning_rate": 5.372519679148227e-05,
|
2588 |
+
"loss": 1.241,
|
2589 |
+
"step": 364
|
2590 |
+
},
|
2591 |
+
{
|
2592 |
+
"epoch": 0.046402606173771406,
|
2593 |
+
"grad_norm": 0.1972542256116867,
|
2594 |
+
"learning_rate": 5.321214122009306e-05,
|
2595 |
+
"loss": 1.1419,
|
2596 |
+
"step": 365
|
2597 |
+
},
|
2598 |
+
{
|
2599 |
+
"epoch": 0.04652973660164475,
|
2600 |
+
"grad_norm": 0.20039339363574982,
|
2601 |
+
"learning_rate": 5.270065757843e-05,
|
2602 |
+
"loss": 1.2718,
|
2603 |
+
"step": 366
|
2604 |
+
},
|
2605 |
+
{
|
2606 |
+
"epoch": 0.046656867029518094,
|
2607 |
+
"grad_norm": 0.1938110589981079,
|
2608 |
+
"learning_rate": 5.219076305078749e-05,
|
2609 |
+
"loss": 1.1947,
|
2610 |
+
"step": 367
|
2611 |
+
},
|
2612 |
+
{
|
2613 |
+
"epoch": 0.046783997457391445,
|
2614 |
+
"grad_norm": 0.20640990138053894,
|
2615 |
+
"learning_rate": 5.168247476807053e-05,
|
2616 |
+
"loss": 1.1526,
|
2617 |
+
"step": 368
|
2618 |
+
},
|
2619 |
+
{
|
2620 |
+
"epoch": 0.04691112788526479,
|
2621 |
+
"grad_norm": 0.198054239153862,
|
2622 |
+
"learning_rate": 5.11758098072193e-05,
|
2623 |
+
"loss": 1.1965,
|
2624 |
+
"step": 369
|
2625 |
+
},
|
2626 |
+
{
|
2627 |
+
"epoch": 0.04703825831313813,
|
2628 |
+
"grad_norm": 0.19604484736919403,
|
2629 |
+
"learning_rate": 5.067078519063514e-05,
|
2630 |
+
"loss": 1.2568,
|
2631 |
+
"step": 370
|
2632 |
+
},
|
2633 |
+
{
|
2634 |
+
"epoch": 0.047165388741011484,
|
2635 |
+
"grad_norm": 0.2095029056072235,
|
2636 |
+
"learning_rate": 5.016741788560889e-05,
|
2637 |
+
"loss": 1.2822,
|
2638 |
+
"step": 371
|
2639 |
+
},
|
2640 |
+
{
|
2641 |
+
"epoch": 0.04729251916888483,
|
2642 |
+
"grad_norm": 0.20356985926628113,
|
2643 |
+
"learning_rate": 4.9665724803750756e-05,
|
2644 |
+
"loss": 1.1434,
|
2645 |
+
"step": 372
|
2646 |
+
},
|
2647 |
+
{
|
2648 |
+
"epoch": 0.04741964959675817,
|
2649 |
+
"grad_norm": 0.19989654421806335,
|
2650 |
+
"learning_rate": 4.9165722800422096e-05,
|
2651 |
+
"loss": 1.2767,
|
2652 |
+
"step": 373
|
2653 |
+
},
|
2654 |
+
{
|
2655 |
+
"epoch": 0.04754678002463152,
|
2656 |
+
"grad_norm": 0.19582509994506836,
|
2657 |
+
"learning_rate": 4.86674286741693e-05,
|
2658 |
+
"loss": 1.2693,
|
2659 |
+
"step": 374
|
2660 |
+
},
|
2661 |
+
{
|
2662 |
+
"epoch": 0.04767391045250487,
|
2663 |
+
"grad_norm": 0.1962389498949051,
|
2664 |
+
"learning_rate": 4.8170859166159144e-05,
|
2665 |
+
"loss": 1.3266,
|
2666 |
+
"step": 375
|
2667 |
+
},
|
2668 |
+
{
|
2669 |
+
"epoch": 0.04780104088037821,
|
2670 |
+
"grad_norm": 0.2056453377008438,
|
2671 |
+
"learning_rate": 4.7676030959616526e-05,
|
2672 |
+
"loss": 1.3004,
|
2673 |
+
"step": 376
|
2674 |
+
},
|
2675 |
+
{
|
2676 |
+
"epoch": 0.04792817130825156,
|
2677 |
+
"grad_norm": 0.19587452709674835,
|
2678 |
+
"learning_rate": 4.71829606792639e-05,
|
2679 |
+
"loss": 1.2154,
|
2680 |
+
"step": 377
|
2681 |
+
},
|
2682 |
+
{
|
2683 |
+
"epoch": 0.048055301736124906,
|
2684 |
+
"grad_norm": 0.19662117958068848,
|
2685 |
+
"learning_rate": 4.669166489076283e-05,
|
2686 |
+
"loss": 1.2434,
|
2687 |
+
"step": 378
|
2688 |
+
},
|
2689 |
+
{
|
2690 |
+
"epoch": 0.04818243216399825,
|
2691 |
+
"grad_norm": 0.19508899748325348,
|
2692 |
+
"learning_rate": 4.620216010015724e-05,
|
2693 |
+
"loss": 1.2319,
|
2694 |
+
"step": 379
|
2695 |
+
},
|
2696 |
+
{
|
2697 |
+
"epoch": 0.0483095625918716,
|
2698 |
+
"grad_norm": 0.19653861224651337,
|
2699 |
+
"learning_rate": 4.571446275331903e-05,
|
2700 |
+
"loss": 1.2006,
|
2701 |
+
"step": 380
|
2702 |
+
},
|
2703 |
+
{
|
2704 |
+
"epoch": 0.048436693019744945,
|
2705 |
+
"grad_norm": 0.1971856951713562,
|
2706 |
+
"learning_rate": 4.5228589235395436e-05,
|
2707 |
+
"loss": 1.2937,
|
2708 |
+
"step": 381
|
2709 |
+
},
|
2710 |
+
{
|
2711 |
+
"epoch": 0.04856382344761829,
|
2712 |
+
"grad_norm": 0.2165059596300125,
|
2713 |
+
"learning_rate": 4.4744555870258694e-05,
|
2714 |
+
"loss": 1.2722,
|
2715 |
+
"step": 382
|
2716 |
+
},
|
2717 |
+
{
|
2718 |
+
"epoch": 0.04869095387549164,
|
2719 |
+
"grad_norm": 0.20020557940006256,
|
2720 |
+
"learning_rate": 4.4262378919957413e-05,
|
2721 |
+
"loss": 1.1947,
|
2722 |
+
"step": 383
|
2723 |
+
},
|
2724 |
+
{
|
2725 |
+
"epoch": 0.048818084303364984,
|
2726 |
+
"grad_norm": 0.19455446302890778,
|
2727 |
+
"learning_rate": 4.378207458417035e-05,
|
2728 |
+
"loss": 1.1956,
|
2729 |
+
"step": 384
|
2730 |
+
},
|
2731 |
+
{
|
2732 |
+
"epoch": 0.04894521473123833,
|
2733 |
+
"grad_norm": 0.202660471200943,
|
2734 |
+
"learning_rate": 4.3303658999662086e-05,
|
2735 |
+
"loss": 1.2553,
|
2736 |
+
"step": 385
|
2737 |
+
},
|
2738 |
+
{
|
2739 |
+
"epoch": 0.04907234515911168,
|
2740 |
+
"grad_norm": 0.19681531190872192,
|
2741 |
+
"learning_rate": 4.282714823974088e-05,
|
2742 |
+
"loss": 1.2031,
|
2743 |
+
"step": 386
|
2744 |
+
},
|
2745 |
+
{
|
2746 |
+
"epoch": 0.04919947558698502,
|
2747 |
+
"grad_norm": 0.20613734424114227,
|
2748 |
+
"learning_rate": 4.2352558313718795e-05,
|
2749 |
+
"loss": 1.2384,
|
2750 |
+
"step": 387
|
2751 |
+
},
|
2752 |
+
{
|
2753 |
+
"epoch": 0.04932660601485837,
|
2754 |
+
"grad_norm": 0.1990024596452713,
|
2755 |
+
"learning_rate": 4.1879905166373614e-05,
|
2756 |
+
"loss": 1.2184,
|
2757 |
+
"step": 388
|
2758 |
+
},
|
2759 |
+
{
|
2760 |
+
"epoch": 0.04945373644273172,
|
2761 |
+
"grad_norm": 0.21309691667556763,
|
2762 |
+
"learning_rate": 4.140920467741325e-05,
|
2763 |
+
"loss": 1.1853,
|
2764 |
+
"step": 389
|
2765 |
+
},
|
2766 |
+
{
|
2767 |
+
"epoch": 0.04958086687060506,
|
2768 |
+
"grad_norm": 0.19488035142421722,
|
2769 |
+
"learning_rate": 4.094047266094225e-05,
|
2770 |
+
"loss": 1.1804,
|
2771 |
+
"step": 390
|
2772 |
+
},
|
2773 |
+
{
|
2774 |
+
"epoch": 0.049707997298478405,
|
2775 |
+
"grad_norm": 0.19738849997520447,
|
2776 |
+
"learning_rate": 4.047372486493054e-05,
|
2777 |
+
"loss": 1.2534,
|
2778 |
+
"step": 391
|
2779 |
+
},
|
2780 |
+
{
|
2781 |
+
"epoch": 0.049835127726351756,
|
2782 |
+
"grad_norm": 0.20008018612861633,
|
2783 |
+
"learning_rate": 4.0008976970684176e-05,
|
2784 |
+
"loss": 1.2723,
|
2785 |
+
"step": 392
|
2786 |
+
},
|
2787 |
+
{
|
2788 |
+
"epoch": 0.0499622581542251,
|
2789 |
+
"grad_norm": 0.19521461427211761,
|
2790 |
+
"learning_rate": 3.954624459231866e-05,
|
2791 |
+
"loss": 1.1705,
|
2792 |
+
"step": 393
|
2793 |
+
},
|
2794 |
+
{
|
2795 |
+
"epoch": 0.050089388582098444,
|
2796 |
+
"grad_norm": 0.20466111600399017,
|
2797 |
+
"learning_rate": 3.908554327623425e-05,
|
2798 |
+
"loss": 1.154,
|
2799 |
+
"step": 394
|
2800 |
+
},
|
2801 |
+
{
|
2802 |
+
"epoch": 0.050216519009971795,
|
2803 |
+
"grad_norm": 0.2047969251871109,
|
2804 |
+
"learning_rate": 3.8626888500593695e-05,
|
2805 |
+
"loss": 1.2139,
|
2806 |
+
"step": 395
|
2807 |
+
},
|
2808 |
+
{
|
2809 |
+
"epoch": 0.05034364943784514,
|
2810 |
+
"grad_norm": 0.1980600655078888,
|
2811 |
+
"learning_rate": 3.817029567480228e-05,
|
2812 |
+
"loss": 1.279,
|
2813 |
+
"step": 396
|
2814 |
+
},
|
2815 |
+
{
|
2816 |
+
"epoch": 0.05047077986571848,
|
2817 |
+
"grad_norm": 0.20217813551425934,
|
2818 |
+
"learning_rate": 3.771578013898996e-05,
|
2819 |
+
"loss": 1.2561,
|
2820 |
+
"step": 397
|
2821 |
+
},
|
2822 |
+
{
|
2823 |
+
"epoch": 0.050597910293591834,
|
2824 |
+
"grad_norm": 0.1985122561454773,
|
2825 |
+
"learning_rate": 3.726335716349612e-05,
|
2826 |
+
"loss": 1.2778,
|
2827 |
+
"step": 398
|
2828 |
+
},
|
2829 |
+
{
|
2830 |
+
"epoch": 0.05072504072146518,
|
2831 |
+
"grad_norm": 0.19889195263385773,
|
2832 |
+
"learning_rate": 3.681304194835641e-05,
|
2833 |
+
"loss": 1.3225,
|
2834 |
+
"step": 399
|
2835 |
+
},
|
2836 |
+
{
|
2837 |
+
"epoch": 0.05085217114933852,
|
2838 |
+
"grad_norm": 0.19213935732841492,
|
2839 |
+
"learning_rate": 3.6364849622792266e-05,
|
2840 |
+
"loss": 1.2308,
|
2841 |
+
"step": 400
|
2842 |
+
},
|
2843 |
+
{
|
2844 |
+
"epoch": 0.05085217114933852,
|
2845 |
+
"eval_loss": 1.2253398895263672,
|
2846 |
+
"eval_runtime": 1257.7786,
|
2847 |
+
"eval_samples_per_second": 3.975,
|
2848 |
+
"eval_steps_per_second": 0.994,
|
2849 |
+
"step": 400
|
2850 |
}
|
2851 |
],
|
2852 |
"logging_steps": 1,
|
|
|
2875 |
"attributes": {}
|
2876 |
}
|
2877 |
},
|
2878 |
+
"total_flos": 4.145600090918093e+18,
|
2879 |
"train_batch_size": 4,
|
2880 |
"trial_name": null,
|
2881 |
"trial_params": null
|