Training in progress, step 450, 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 1101095848
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e75e76ad495c6f04db1db938b83a8c0a347ed4aae0aa2da55a2e4abed15fc6bd
|
3 |
size 1101095848
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 559894868
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4a6754dc365b79a92d5e2347eb928b390eae344331ae7e043e300718c75c9c02
|
3 |
size 559894868
|
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 sha256:1ada0c212eb58bca56bed03d6bd6d8c0e2a39e3c77d936483f633c2e099bd000
|
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:c9636ae38b683f4b5b714bdf172e563b0c593e0efe94f07eea78547963bfbfae
|
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": 150,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -2131,6 +2131,1064 @@
|
|
2131 |
"eval_samples_per_second": 13.344,
|
2132 |
"eval_steps_per_second": 3.336,
|
2133 |
"step": 300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2134 |
}
|
2135 |
],
|
2136 |
"logging_steps": 1,
|
@@ -2154,12 +3212,12 @@
|
|
2154 |
"should_evaluate": false,
|
2155 |
"should_log": false,
|
2156 |
"should_save": true,
|
2157 |
-
"should_training_stop":
|
2158 |
},
|
2159 |
"attributes": {}
|
2160 |
}
|
2161 |
},
|
2162 |
-
"total_flos":
|
2163 |
"train_batch_size": 8,
|
2164 |
"trial_name": null,
|
2165 |
"trial_params": null
|
|
|
1 |
{
|
2 |
+
"best_metric": 0.9613198041915894,
|
3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-450",
|
4 |
+
"epoch": 0.05906868375283037,
|
5 |
"eval_steps": 150,
|
6 |
+
"global_step": 450,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
2131 |
"eval_samples_per_second": 13.344,
|
2132 |
"eval_steps_per_second": 3.336,
|
2133 |
"step": 300
|
2134 |
+
},
|
2135 |
+
{
|
2136 |
+
"epoch": 0.039510386243559875,
|
2137 |
+
"grad_norm": 0.20209898054599762,
|
2138 |
+
"learning_rate": 2.6813618894527138e-05,
|
2139 |
+
"loss": 1.0962,
|
2140 |
+
"step": 301
|
2141 |
+
},
|
2142 |
+
{
|
2143 |
+
"epoch": 0.03964164998523283,
|
2144 |
+
"grad_norm": 0.2070281207561493,
|
2145 |
+
"learning_rate": 2.6490591592961578e-05,
|
2146 |
+
"loss": 1.0854,
|
2147 |
+
"step": 302
|
2148 |
+
},
|
2149 |
+
{
|
2150 |
+
"epoch": 0.039772913726905786,
|
2151 |
+
"grad_norm": 0.21131707727909088,
|
2152 |
+
"learning_rate": 2.6168819172567392e-05,
|
2153 |
+
"loss": 1.1135,
|
2154 |
+
"step": 303
|
2155 |
+
},
|
2156 |
+
{
|
2157 |
+
"epoch": 0.039904177468578744,
|
2158 |
+
"grad_norm": 0.2207150012254715,
|
2159 |
+
"learning_rate": 2.5848318808857606e-05,
|
2160 |
+
"loss": 1.1269,
|
2161 |
+
"step": 304
|
2162 |
+
},
|
2163 |
+
{
|
2164 |
+
"epoch": 0.040035441210251696,
|
2165 |
+
"grad_norm": 0.24536538124084473,
|
2166 |
+
"learning_rate": 2.5529107609445733e-05,
|
2167 |
+
"loss": 1.1705,
|
2168 |
+
"step": 305
|
2169 |
+
},
|
2170 |
+
{
|
2171 |
+
"epoch": 0.040166704951924655,
|
2172 |
+
"grad_norm": 0.27333205938339233,
|
2173 |
+
"learning_rate": 2.521120261313241e-05,
|
2174 |
+
"loss": 1.2272,
|
2175 |
+
"step": 306
|
2176 |
+
},
|
2177 |
+
{
|
2178 |
+
"epoch": 0.04029796869359761,
|
2179 |
+
"grad_norm": 0.42892393469810486,
|
2180 |
+
"learning_rate": 2.4894620788996037e-05,
|
2181 |
+
"loss": 1.2927,
|
2182 |
+
"step": 307
|
2183 |
+
},
|
2184 |
+
{
|
2185 |
+
"epoch": 0.040429232435270565,
|
2186 |
+
"grad_norm": 0.4479203224182129,
|
2187 |
+
"learning_rate": 2.457937903548695e-05,
|
2188 |
+
"loss": 1.1936,
|
2189 |
+
"step": 308
|
2190 |
+
},
|
2191 |
+
{
|
2192 |
+
"epoch": 0.040560496176943524,
|
2193 |
+
"grad_norm": 0.5271134376525879,
|
2194 |
+
"learning_rate": 2.426549417952542e-05,
|
2195 |
+
"loss": 1.2875,
|
2196 |
+
"step": 309
|
2197 |
+
},
|
2198 |
+
{
|
2199 |
+
"epoch": 0.04069175991861648,
|
2200 |
+
"grad_norm": 0.6161456108093262,
|
2201 |
+
"learning_rate": 2.3952982975603496e-05,
|
2202 |
+
"loss": 1.3302,
|
2203 |
+
"step": 310
|
2204 |
+
},
|
2205 |
+
{
|
2206 |
+
"epoch": 0.040823023660289434,
|
2207 |
+
"grad_norm": 0.6787443161010742,
|
2208 |
+
"learning_rate": 2.3641862104890595e-05,
|
2209 |
+
"loss": 1.2439,
|
2210 |
+
"step": 311
|
2211 |
+
},
|
2212 |
+
{
|
2213 |
+
"epoch": 0.04095428740196239,
|
2214 |
+
"grad_norm": 0.8013172149658203,
|
2215 |
+
"learning_rate": 2.3332148174343254e-05,
|
2216 |
+
"loss": 1.3476,
|
2217 |
+
"step": 312
|
2218 |
+
},
|
2219 |
+
{
|
2220 |
+
"epoch": 0.04108555114363535,
|
2221 |
+
"grad_norm": 0.8128992319107056,
|
2222 |
+
"learning_rate": 2.3023857715818532e-05,
|
2223 |
+
"loss": 1.2517,
|
2224 |
+
"step": 313
|
2225 |
+
},
|
2226 |
+
{
|
2227 |
+
"epoch": 0.0412168148853083,
|
2228 |
+
"grad_norm": 0.7722633481025696,
|
2229 |
+
"learning_rate": 2.2717007185191674e-05,
|
2230 |
+
"loss": 1.2515,
|
2231 |
+
"step": 314
|
2232 |
+
},
|
2233 |
+
{
|
2234 |
+
"epoch": 0.04134807862698126,
|
2235 |
+
"grad_norm": 0.7521525621414185,
|
2236 |
+
"learning_rate": 2.24116129614777e-05,
|
2237 |
+
"loss": 1.1528,
|
2238 |
+
"step": 315
|
2239 |
+
},
|
2240 |
+
{
|
2241 |
+
"epoch": 0.04147934236865422,
|
2242 |
+
"grad_norm": 0.7824714183807373,
|
2243 |
+
"learning_rate": 2.2107691345957133e-05,
|
2244 |
+
"loss": 1.2161,
|
2245 |
+
"step": 316
|
2246 |
+
},
|
2247 |
+
{
|
2248 |
+
"epoch": 0.04161060611032717,
|
2249 |
+
"grad_norm": 0.7399042248725891,
|
2250 |
+
"learning_rate": 2.1805258561305862e-05,
|
2251 |
+
"loss": 1.2162,
|
2252 |
+
"step": 317
|
2253 |
+
},
|
2254 |
+
{
|
2255 |
+
"epoch": 0.04174186985200013,
|
2256 |
+
"grad_norm": 0.7101216912269592,
|
2257 |
+
"learning_rate": 2.1504330750729186e-05,
|
2258 |
+
"loss": 1.1717,
|
2259 |
+
"step": 318
|
2260 |
+
},
|
2261 |
+
{
|
2262 |
+
"epoch": 0.04187313359367309,
|
2263 |
+
"grad_norm": 0.6807271838188171,
|
2264 |
+
"learning_rate": 2.120492397710022e-05,
|
2265 |
+
"loss": 1.0815,
|
2266 |
+
"step": 319
|
2267 |
+
},
|
2268 |
+
{
|
2269 |
+
"epoch": 0.04200439733534604,
|
2270 |
+
"grad_norm": 0.7126163244247437,
|
2271 |
+
"learning_rate": 2.090705422210237e-05,
|
2272 |
+
"loss": 1.1976,
|
2273 |
+
"step": 320
|
2274 |
+
},
|
2275 |
+
{
|
2276 |
+
"epoch": 0.042135661077019,
|
2277 |
+
"grad_norm": 0.6584108471870422,
|
2278 |
+
"learning_rate": 2.061073738537635e-05,
|
2279 |
+
"loss": 1.0962,
|
2280 |
+
"step": 321
|
2281 |
+
},
|
2282 |
+
{
|
2283 |
+
"epoch": 0.04226692481869196,
|
2284 |
+
"grad_norm": 0.6849985718727112,
|
2285 |
+
"learning_rate": 2.0315989283671473e-05,
|
2286 |
+
"loss": 1.169,
|
2287 |
+
"step": 322
|
2288 |
+
},
|
2289 |
+
{
|
2290 |
+
"epoch": 0.04239818856036491,
|
2291 |
+
"grad_norm": 0.673360288143158,
|
2292 |
+
"learning_rate": 2.0022825650001387e-05,
|
2293 |
+
"loss": 1.0906,
|
2294 |
+
"step": 323
|
2295 |
+
},
|
2296 |
+
{
|
2297 |
+
"epoch": 0.04252945230203787,
|
2298 |
+
"grad_norm": 0.6716238856315613,
|
2299 |
+
"learning_rate": 1.9731262132804274e-05,
|
2300 |
+
"loss": 1.0272,
|
2301 |
+
"step": 324
|
2302 |
+
},
|
2303 |
+
{
|
2304 |
+
"epoch": 0.04266071604371083,
|
2305 |
+
"grad_norm": 0.6913533806800842,
|
2306 |
+
"learning_rate": 1.9441314295107537e-05,
|
2307 |
+
"loss": 1.0831,
|
2308 |
+
"step": 325
|
2309 |
+
},
|
2310 |
+
{
|
2311 |
+
"epoch": 0.04279197978538378,
|
2312 |
+
"grad_norm": 0.7192745208740234,
|
2313 |
+
"learning_rate": 1.9152997613697183e-05,
|
2314 |
+
"loss": 0.9957,
|
2315 |
+
"step": 326
|
2316 |
+
},
|
2317 |
+
{
|
2318 |
+
"epoch": 0.04292324352705674,
|
2319 |
+
"grad_norm": 0.6721824407577515,
|
2320 |
+
"learning_rate": 1.8866327478291546e-05,
|
2321 |
+
"loss": 1.1222,
|
2322 |
+
"step": 327
|
2323 |
+
},
|
2324 |
+
{
|
2325 |
+
"epoch": 0.0430545072687297,
|
2326 |
+
"grad_norm": 0.6699678301811218,
|
2327 |
+
"learning_rate": 1.8581319190720035e-05,
|
2328 |
+
"loss": 0.9989,
|
2329 |
+
"step": 328
|
2330 |
+
},
|
2331 |
+
{
|
2332 |
+
"epoch": 0.04318577101040265,
|
2333 |
+
"grad_norm": 0.6625519394874573,
|
2334 |
+
"learning_rate": 1.8297987964106115e-05,
|
2335 |
+
"loss": 0.9872,
|
2336 |
+
"step": 329
|
2337 |
+
},
|
2338 |
+
{
|
2339 |
+
"epoch": 0.04331703475207561,
|
2340 |
+
"grad_norm": 0.7252044081687927,
|
2341 |
+
"learning_rate": 1.801634892205545e-05,
|
2342 |
+
"loss": 1.0245,
|
2343 |
+
"step": 330
|
2344 |
+
},
|
2345 |
+
{
|
2346 |
+
"epoch": 0.043448298493748566,
|
2347 |
+
"grad_norm": 0.7490676641464233,
|
2348 |
+
"learning_rate": 1.7736417097848506e-05,
|
2349 |
+
"loss": 1.0401,
|
2350 |
+
"step": 331
|
2351 |
+
},
|
2352 |
+
{
|
2353 |
+
"epoch": 0.04357956223542152,
|
2354 |
+
"grad_norm": 0.6983168721199036,
|
2355 |
+
"learning_rate": 1.7458207433638223e-05,
|
2356 |
+
"loss": 0.9628,
|
2357 |
+
"step": 332
|
2358 |
+
},
|
2359 |
+
{
|
2360 |
+
"epoch": 0.04371082597709448,
|
2361 |
+
"grad_norm": 0.7303330898284912,
|
2362 |
+
"learning_rate": 1.718173477965236e-05,
|
2363 |
+
"loss": 0.9528,
|
2364 |
+
"step": 333
|
2365 |
+
},
|
2366 |
+
{
|
2367 |
+
"epoch": 0.043842089718767435,
|
2368 |
+
"grad_norm": 0.7304815053939819,
|
2369 |
+
"learning_rate": 1.6907013893400837e-05,
|
2370 |
+
"loss": 0.9777,
|
2371 |
+
"step": 334
|
2372 |
+
},
|
2373 |
+
{
|
2374 |
+
"epoch": 0.04397335346044039,
|
2375 |
+
"grad_norm": 0.6960116028785706,
|
2376 |
+
"learning_rate": 1.6634059438888033e-05,
|
2377 |
+
"loss": 0.9643,
|
2378 |
+
"step": 335
|
2379 |
+
},
|
2380 |
+
{
|
2381 |
+
"epoch": 0.044104617202113346,
|
2382 |
+
"grad_norm": 0.6498481631278992,
|
2383 |
+
"learning_rate": 1.636288598583e-05,
|
2384 |
+
"loss": 0.7516,
|
2385 |
+
"step": 336
|
2386 |
+
},
|
2387 |
+
{
|
2388 |
+
"epoch": 0.044235880943786304,
|
2389 |
+
"grad_norm": 0.7724774479866028,
|
2390 |
+
"learning_rate": 1.6093508008876857e-05,
|
2391 |
+
"loss": 0.9599,
|
2392 |
+
"step": 337
|
2393 |
+
},
|
2394 |
+
{
|
2395 |
+
"epoch": 0.044367144685459256,
|
2396 |
+
"grad_norm": 0.7501699924468994,
|
2397 |
+
"learning_rate": 1.5825939886840037e-05,
|
2398 |
+
"loss": 0.8397,
|
2399 |
+
"step": 338
|
2400 |
+
},
|
2401 |
+
{
|
2402 |
+
"epoch": 0.044498408427132215,
|
2403 |
+
"grad_norm": 0.7359670400619507,
|
2404 |
+
"learning_rate": 1.5560195901924894e-05,
|
2405 |
+
"loss": 0.8147,
|
2406 |
+
"step": 339
|
2407 |
+
},
|
2408 |
+
{
|
2409 |
+
"epoch": 0.04462967216880517,
|
2410 |
+
"grad_norm": 0.7074993252754211,
|
2411 |
+
"learning_rate": 1.5296290238968303e-05,
|
2412 |
+
"loss": 0.7271,
|
2413 |
+
"step": 340
|
2414 |
+
},
|
2415 |
+
{
|
2416 |
+
"epoch": 0.044760935910478125,
|
2417 |
+
"grad_norm": 0.7447213530540466,
|
2418 |
+
"learning_rate": 1.50342369846815e-05,
|
2419 |
+
"loss": 0.7732,
|
2420 |
+
"step": 341
|
2421 |
+
},
|
2422 |
+
{
|
2423 |
+
"epoch": 0.044892199652151084,
|
2424 |
+
"grad_norm": 0.7885531187057495,
|
2425 |
+
"learning_rate": 1.4774050126898164e-05,
|
2426 |
+
"loss": 0.6494,
|
2427 |
+
"step": 342
|
2428 |
+
},
|
2429 |
+
{
|
2430 |
+
"epoch": 0.04502346339382404,
|
2431 |
+
"grad_norm": 0.7818116545677185,
|
2432 |
+
"learning_rate": 1.451574355382776e-05,
|
2433 |
+
"loss": 0.7047,
|
2434 |
+
"step": 343
|
2435 |
+
},
|
2436 |
+
{
|
2437 |
+
"epoch": 0.045154727135496994,
|
2438 |
+
"grad_norm": 0.6587514281272888,
|
2439 |
+
"learning_rate": 1.425933105331429e-05,
|
2440 |
+
"loss": 0.4821,
|
2441 |
+
"step": 344
|
2442 |
+
},
|
2443 |
+
{
|
2444 |
+
"epoch": 0.04528599087716995,
|
2445 |
+
"grad_norm": 0.7465260624885559,
|
2446 |
+
"learning_rate": 1.4004826312100216e-05,
|
2447 |
+
"loss": 0.496,
|
2448 |
+
"step": 345
|
2449 |
+
},
|
2450 |
+
{
|
2451 |
+
"epoch": 0.04541725461884291,
|
2452 |
+
"grad_norm": 0.7443161010742188,
|
2453 |
+
"learning_rate": 1.3752242915095992e-05,
|
2454 |
+
"loss": 0.5114,
|
2455 |
+
"step": 346
|
2456 |
+
},
|
2457 |
+
{
|
2458 |
+
"epoch": 0.04554851836051586,
|
2459 |
+
"grad_norm": 0.7163719534873962,
|
2460 |
+
"learning_rate": 1.3501594344654884e-05,
|
2461 |
+
"loss": 0.5043,
|
2462 |
+
"step": 347
|
2463 |
+
},
|
2464 |
+
{
|
2465 |
+
"epoch": 0.04567978210218882,
|
2466 |
+
"grad_norm": 0.7420056462287903,
|
2467 |
+
"learning_rate": 1.3252893979853304e-05,
|
2468 |
+
"loss": 0.5634,
|
2469 |
+
"step": 348
|
2470 |
+
},
|
2471 |
+
{
|
2472 |
+
"epoch": 0.04581104584386178,
|
2473 |
+
"grad_norm": 0.6838533282279968,
|
2474 |
+
"learning_rate": 1.3006155095776707e-05,
|
2475 |
+
"loss": 0.3495,
|
2476 |
+
"step": 349
|
2477 |
+
},
|
2478 |
+
{
|
2479 |
+
"epoch": 0.04594230958553473,
|
2480 |
+
"grad_norm": 0.8066396117210388,
|
2481 |
+
"learning_rate": 1.2761390862810907e-05,
|
2482 |
+
"loss": 0.424,
|
2483 |
+
"step": 350
|
2484 |
+
},
|
2485 |
+
{
|
2486 |
+
"epoch": 0.04607357332720769,
|
2487 |
+
"grad_norm": 0.20562657713890076,
|
2488 |
+
"learning_rate": 1.2518614345939212e-05,
|
2489 |
+
"loss": 1.0868,
|
2490 |
+
"step": 351
|
2491 |
+
},
|
2492 |
+
{
|
2493 |
+
"epoch": 0.04620483706888065,
|
2494 |
+
"grad_norm": 0.19780656695365906,
|
2495 |
+
"learning_rate": 1.227783850404487e-05,
|
2496 |
+
"loss": 1.1027,
|
2497 |
+
"step": 352
|
2498 |
+
},
|
2499 |
+
{
|
2500 |
+
"epoch": 0.0463361008105536,
|
2501 |
+
"grad_norm": 0.21005481481552124,
|
2502 |
+
"learning_rate": 1.2039076189219517e-05,
|
2503 |
+
"loss": 1.1234,
|
2504 |
+
"step": 353
|
2505 |
+
},
|
2506 |
+
{
|
2507 |
+
"epoch": 0.04646736455222656,
|
2508 |
+
"grad_norm": 0.2190805971622467,
|
2509 |
+
"learning_rate": 1.1802340146077045e-05,
|
2510 |
+
"loss": 1.1,
|
2511 |
+
"step": 354
|
2512 |
+
},
|
2513 |
+
{
|
2514 |
+
"epoch": 0.04659862829389952,
|
2515 |
+
"grad_norm": 0.24044179916381836,
|
2516 |
+
"learning_rate": 1.1567643011073392e-05,
|
2517 |
+
"loss": 1.1298,
|
2518 |
+
"step": 355
|
2519 |
+
},
|
2520 |
+
{
|
2521 |
+
"epoch": 0.04672989203557247,
|
2522 |
+
"grad_norm": 0.2597028613090515,
|
2523 |
+
"learning_rate": 1.1334997311832002e-05,
|
2524 |
+
"loss": 1.134,
|
2525 |
+
"step": 356
|
2526 |
+
},
|
2527 |
+
{
|
2528 |
+
"epoch": 0.04686115577724543,
|
2529 |
+
"grad_norm": 0.3216112554073334,
|
2530 |
+
"learning_rate": 1.1104415466475087e-05,
|
2531 |
+
"loss": 1.1735,
|
2532 |
+
"step": 357
|
2533 |
+
},
|
2534 |
+
{
|
2535 |
+
"epoch": 0.04699241951891839,
|
2536 |
+
"grad_norm": 0.4165898561477661,
|
2537 |
+
"learning_rate": 1.0875909782960886e-05,
|
2538 |
+
"loss": 1.2389,
|
2539 |
+
"step": 358
|
2540 |
+
},
|
2541 |
+
{
|
2542 |
+
"epoch": 0.04712368326059134,
|
2543 |
+
"grad_norm": 0.4193871319293976,
|
2544 |
+
"learning_rate": 1.0649492458426564e-05,
|
2545 |
+
"loss": 1.1649,
|
2546 |
+
"step": 359
|
2547 |
+
},
|
2548 |
+
{
|
2549 |
+
"epoch": 0.0472549470022643,
|
2550 |
+
"grad_norm": 0.4490433931350708,
|
2551 |
+
"learning_rate": 1.0425175578537299e-05,
|
2552 |
+
"loss": 1.2062,
|
2553 |
+
"step": 360
|
2554 |
+
},
|
2555 |
+
{
|
2556 |
+
"epoch": 0.04738621074393726,
|
2557 |
+
"grad_norm": 0.4922597110271454,
|
2558 |
+
"learning_rate": 1.020297111684101e-05,
|
2559 |
+
"loss": 1.2046,
|
2560 |
+
"step": 361
|
2561 |
+
},
|
2562 |
+
{
|
2563 |
+
"epoch": 0.047517474485610216,
|
2564 |
+
"grad_norm": 0.5809623003005981,
|
2565 |
+
"learning_rate": 9.98289093412938e-06,
|
2566 |
+
"loss": 1.2574,
|
2567 |
+
"step": 362
|
2568 |
+
},
|
2569 |
+
{
|
2570 |
+
"epoch": 0.04764873822728317,
|
2571 |
+
"grad_norm": 0.5852945446968079,
|
2572 |
+
"learning_rate": 9.764946777804646e-06,
|
2573 |
+
"loss": 1.2615,
|
2574 |
+
"step": 363
|
2575 |
+
},
|
2576 |
+
{
|
2577 |
+
"epoch": 0.047780001968956126,
|
2578 |
+
"grad_norm": 0.6204507946968079,
|
2579 |
+
"learning_rate": 9.549150281252633e-06,
|
2580 |
+
"loss": 1.2761,
|
2581 |
+
"step": 364
|
2582 |
+
},
|
2583 |
+
{
|
2584 |
+
"epoch": 0.047911265710629085,
|
2585 |
+
"grad_norm": 0.6129657626152039,
|
2586 |
+
"learning_rate": 9.335512963221732e-06,
|
2587 |
+
"loss": 1.2447,
|
2588 |
+
"step": 365
|
2589 |
+
},
|
2590 |
+
{
|
2591 |
+
"epoch": 0.04804252945230204,
|
2592 |
+
"grad_norm": 0.613051176071167,
|
2593 |
+
"learning_rate": 9.124046227208082e-06,
|
2594 |
+
"loss": 1.2447,
|
2595 |
+
"step": 366
|
2596 |
+
},
|
2597 |
+
{
|
2598 |
+
"epoch": 0.048173793193974995,
|
2599 |
+
"grad_norm": 0.6088857650756836,
|
2600 |
+
"learning_rate": 8.914761360846869e-06,
|
2601 |
+
"loss": 1.2113,
|
2602 |
+
"step": 367
|
2603 |
+
},
|
2604 |
+
{
|
2605 |
+
"epoch": 0.048305056935647954,
|
2606 |
+
"grad_norm": 0.6392094492912292,
|
2607 |
+
"learning_rate": 8.707669535309793e-06,
|
2608 |
+
"loss": 1.1427,
|
2609 |
+
"step": 368
|
2610 |
+
},
|
2611 |
+
{
|
2612 |
+
"epoch": 0.048436320677320906,
|
2613 |
+
"grad_norm": 0.6354860663414001,
|
2614 |
+
"learning_rate": 8.502781804708826e-06,
|
2615 |
+
"loss": 1.1504,
|
2616 |
+
"step": 369
|
2617 |
+
},
|
2618 |
+
{
|
2619 |
+
"epoch": 0.048567584418993864,
|
2620 |
+
"grad_norm": 0.6489117741584778,
|
2621 |
+
"learning_rate": 8.30010910550611e-06,
|
2622 |
+
"loss": 1.1651,
|
2623 |
+
"step": 370
|
2624 |
+
},
|
2625 |
+
{
|
2626 |
+
"epoch": 0.04869884816066682,
|
2627 |
+
"grad_norm": 0.6395854949951172,
|
2628 |
+
"learning_rate": 8.09966225593024e-06,
|
2629 |
+
"loss": 1.1448,
|
2630 |
+
"step": 371
|
2631 |
+
},
|
2632 |
+
{
|
2633 |
+
"epoch": 0.048830111902339775,
|
2634 |
+
"grad_norm": 0.6275275349617004,
|
2635 |
+
"learning_rate": 7.901451955398792e-06,
|
2636 |
+
"loss": 1.1262,
|
2637 |
+
"step": 372
|
2638 |
+
},
|
2639 |
+
{
|
2640 |
+
"epoch": 0.04896137564401273,
|
2641 |
+
"grad_norm": 0.6410810351371765,
|
2642 |
+
"learning_rate": 7.705488783947202e-06,
|
2643 |
+
"loss": 1.1578,
|
2644 |
+
"step": 373
|
2645 |
+
},
|
2646 |
+
{
|
2647 |
+
"epoch": 0.04909263938568569,
|
2648 |
+
"grad_norm": 0.623022735118866,
|
2649 |
+
"learning_rate": 7.511783201664052e-06,
|
2650 |
+
"loss": 1.0632,
|
2651 |
+
"step": 374
|
2652 |
+
},
|
2653 |
+
{
|
2654 |
+
"epoch": 0.049223903127358644,
|
2655 |
+
"grad_norm": 0.6711010336875916,
|
2656 |
+
"learning_rate": 7.320345548132679e-06,
|
2657 |
+
"loss": 1.0221,
|
2658 |
+
"step": 375
|
2659 |
+
},
|
2660 |
+
{
|
2661 |
+
"epoch": 0.0493551668690316,
|
2662 |
+
"grad_norm": 0.7072113156318665,
|
2663 |
+
"learning_rate": 7.131186041879357e-06,
|
2664 |
+
"loss": 1.1742,
|
2665 |
+
"step": 376
|
2666 |
+
},
|
2667 |
+
{
|
2668 |
+
"epoch": 0.04948643061070456,
|
2669 |
+
"grad_norm": 0.7174622416496277,
|
2670 |
+
"learning_rate": 6.944314779827749e-06,
|
2671 |
+
"loss": 1.0623,
|
2672 |
+
"step": 377
|
2673 |
+
},
|
2674 |
+
{
|
2675 |
+
"epoch": 0.04961769435237751,
|
2676 |
+
"grad_norm": 0.6921060085296631,
|
2677 |
+
"learning_rate": 6.759741736760061e-06,
|
2678 |
+
"loss": 0.9611,
|
2679 |
+
"step": 378
|
2680 |
+
},
|
2681 |
+
{
|
2682 |
+
"epoch": 0.04974895809405047,
|
2683 |
+
"grad_norm": 0.6697357296943665,
|
2684 |
+
"learning_rate": 6.577476764784546e-06,
|
2685 |
+
"loss": 0.9576,
|
2686 |
+
"step": 379
|
2687 |
+
},
|
2688 |
+
{
|
2689 |
+
"epoch": 0.04988022183572343,
|
2690 |
+
"grad_norm": 0.7066754698753357,
|
2691 |
+
"learning_rate": 6.397529592809614e-06,
|
2692 |
+
"loss": 1.0607,
|
2693 |
+
"step": 380
|
2694 |
+
},
|
2695 |
+
{
|
2696 |
+
"epoch": 0.05001148557739638,
|
2697 |
+
"grad_norm": 0.6927306652069092,
|
2698 |
+
"learning_rate": 6.219909826024589e-06,
|
2699 |
+
"loss": 0.9993,
|
2700 |
+
"step": 381
|
2701 |
+
},
|
2702 |
+
{
|
2703 |
+
"epoch": 0.05014274931906934,
|
2704 |
+
"grad_norm": 0.7213545441627502,
|
2705 |
+
"learning_rate": 6.0446269453868945e-06,
|
2706 |
+
"loss": 1.0223,
|
2707 |
+
"step": 382
|
2708 |
+
},
|
2709 |
+
{
|
2710 |
+
"epoch": 0.0502740130607423,
|
2711 |
+
"grad_norm": 0.7020846009254456,
|
2712 |
+
"learning_rate": 5.871690307116107e-06,
|
2713 |
+
"loss": 0.954,
|
2714 |
+
"step": 383
|
2715 |
+
},
|
2716 |
+
{
|
2717 |
+
"epoch": 0.05040527680241525,
|
2718 |
+
"grad_norm": 0.7054339051246643,
|
2719 |
+
"learning_rate": 5.701109142194422e-06,
|
2720 |
+
"loss": 0.8829,
|
2721 |
+
"step": 384
|
2722 |
+
},
|
2723 |
+
{
|
2724 |
+
"epoch": 0.05053654054408821,
|
2725 |
+
"grad_norm": 0.7075545191764832,
|
2726 |
+
"learning_rate": 5.532892555874059e-06,
|
2727 |
+
"loss": 0.8735,
|
2728 |
+
"step": 385
|
2729 |
+
},
|
2730 |
+
{
|
2731 |
+
"epoch": 0.05066780428576117,
|
2732 |
+
"grad_norm": 0.6708042621612549,
|
2733 |
+
"learning_rate": 5.3670495271910925e-06,
|
2734 |
+
"loss": 0.7844,
|
2735 |
+
"step": 386
|
2736 |
+
},
|
2737 |
+
{
|
2738 |
+
"epoch": 0.05079906802743412,
|
2739 |
+
"grad_norm": 0.6817059516906738,
|
2740 |
+
"learning_rate": 5.203588908486279e-06,
|
2741 |
+
"loss": 0.8793,
|
2742 |
+
"step": 387
|
2743 |
+
},
|
2744 |
+
{
|
2745 |
+
"epoch": 0.05093033176910708,
|
2746 |
+
"grad_norm": 0.7711595296859741,
|
2747 |
+
"learning_rate": 5.042519424932513e-06,
|
2748 |
+
"loss": 0.9413,
|
2749 |
+
"step": 388
|
2750 |
+
},
|
2751 |
+
{
|
2752 |
+
"epoch": 0.05106159551078004,
|
2753 |
+
"grad_norm": 0.8680550456047058,
|
2754 |
+
"learning_rate": 4.883849674069058e-06,
|
2755 |
+
"loss": 0.8458,
|
2756 |
+
"step": 389
|
2757 |
+
},
|
2758 |
+
{
|
2759 |
+
"epoch": 0.05119285925245299,
|
2760 |
+
"grad_norm": 0.6733363270759583,
|
2761 |
+
"learning_rate": 4.727588125342669e-06,
|
2762 |
+
"loss": 0.6715,
|
2763 |
+
"step": 390
|
2764 |
+
},
|
2765 |
+
{
|
2766 |
+
"epoch": 0.05132412299412595,
|
2767 |
+
"grad_norm": 0.6882769465446472,
|
2768 |
+
"learning_rate": 4.573743119655516e-06,
|
2769 |
+
"loss": 0.6921,
|
2770 |
+
"step": 391
|
2771 |
+
},
|
2772 |
+
{
|
2773 |
+
"epoch": 0.05145538673579891,
|
2774 |
+
"grad_norm": 0.7219226956367493,
|
2775 |
+
"learning_rate": 4.422322868919937e-06,
|
2776 |
+
"loss": 0.6872,
|
2777 |
+
"step": 392
|
2778 |
+
},
|
2779 |
+
{
|
2780 |
+
"epoch": 0.05158665047747186,
|
2781 |
+
"grad_norm": 0.6878823637962341,
|
2782 |
+
"learning_rate": 4.273335455620097e-06,
|
2783 |
+
"loss": 0.6209,
|
2784 |
+
"step": 393
|
2785 |
+
},
|
2786 |
+
{
|
2787 |
+
"epoch": 0.05171791421914482,
|
2788 |
+
"grad_norm": 0.7288166284561157,
|
2789 |
+
"learning_rate": 4.126788832380629e-06,
|
2790 |
+
"loss": 0.6846,
|
2791 |
+
"step": 394
|
2792 |
+
},
|
2793 |
+
{
|
2794 |
+
"epoch": 0.051849177960817776,
|
2795 |
+
"grad_norm": 0.6843792796134949,
|
2796 |
+
"learning_rate": 3.982690821542035e-06,
|
2797 |
+
"loss": 0.5668,
|
2798 |
+
"step": 395
|
2799 |
+
},
|
2800 |
+
{
|
2801 |
+
"epoch": 0.05198044170249073,
|
2802 |
+
"grad_norm": 0.6692485809326172,
|
2803 |
+
"learning_rate": 3.8410491147432395e-06,
|
2804 |
+
"loss": 0.4893,
|
2805 |
+
"step": 396
|
2806 |
+
},
|
2807 |
+
{
|
2808 |
+
"epoch": 0.052111705444163686,
|
2809 |
+
"grad_norm": 0.5910897850990295,
|
2810 |
+
"learning_rate": 3.7018712725109926e-06,
|
2811 |
+
"loss": 0.4123,
|
2812 |
+
"step": 397
|
2813 |
+
},
|
2814 |
+
{
|
2815 |
+
"epoch": 0.052242969185836645,
|
2816 |
+
"grad_norm": 0.7011106014251709,
|
2817 |
+
"learning_rate": 3.5651647238562904e-06,
|
2818 |
+
"loss": 0.4728,
|
2819 |
+
"step": 398
|
2820 |
+
},
|
2821 |
+
{
|
2822 |
+
"epoch": 0.0523742329275096,
|
2823 |
+
"grad_norm": 0.6013720035552979,
|
2824 |
+
"learning_rate": 3.430936765877857e-06,
|
2825 |
+
"loss": 0.4168,
|
2826 |
+
"step": 399
|
2827 |
+
},
|
2828 |
+
{
|
2829 |
+
"epoch": 0.052505496669182555,
|
2830 |
+
"grad_norm": 0.6685484051704407,
|
2831 |
+
"learning_rate": 3.299194563372604e-06,
|
2832 |
+
"loss": 0.4504,
|
2833 |
+
"step": 400
|
2834 |
+
},
|
2835 |
+
{
|
2836 |
+
"epoch": 0.052636760410855514,
|
2837 |
+
"grad_norm": 0.1889829933643341,
|
2838 |
+
"learning_rate": 3.1699451484532463e-06,
|
2839 |
+
"loss": 1.0635,
|
2840 |
+
"step": 401
|
2841 |
+
},
|
2842 |
+
{
|
2843 |
+
"epoch": 0.052768024152528466,
|
2844 |
+
"grad_norm": 0.2015289068222046,
|
2845 |
+
"learning_rate": 3.0431954201728784e-06,
|
2846 |
+
"loss": 1.0821,
|
2847 |
+
"step": 402
|
2848 |
+
},
|
2849 |
+
{
|
2850 |
+
"epoch": 0.052899287894201424,
|
2851 |
+
"grad_norm": 0.20403572916984558,
|
2852 |
+
"learning_rate": 2.9189521441567726e-06,
|
2853 |
+
"loss": 1.1017,
|
2854 |
+
"step": 403
|
2855 |
+
},
|
2856 |
+
{
|
2857 |
+
"epoch": 0.05303055163587438,
|
2858 |
+
"grad_norm": 0.2169038951396942,
|
2859 |
+
"learning_rate": 2.797221952241219e-06,
|
2860 |
+
"loss": 1.0893,
|
2861 |
+
"step": 404
|
2862 |
+
},
|
2863 |
+
{
|
2864 |
+
"epoch": 0.053161815377547335,
|
2865 |
+
"grad_norm": 0.22941887378692627,
|
2866 |
+
"learning_rate": 2.6780113421195298e-06,
|
2867 |
+
"loss": 1.1287,
|
2868 |
+
"step": 405
|
2869 |
+
},
|
2870 |
+
{
|
2871 |
+
"epoch": 0.053293079119220294,
|
2872 |
+
"grad_norm": 0.27469444274902344,
|
2873 |
+
"learning_rate": 2.561326676995218e-06,
|
2874 |
+
"loss": 1.2098,
|
2875 |
+
"step": 406
|
2876 |
+
},
|
2877 |
+
{
|
2878 |
+
"epoch": 0.05342434286089325,
|
2879 |
+
"grad_norm": 0.3559771180152893,
|
2880 |
+
"learning_rate": 2.4471741852423237e-06,
|
2881 |
+
"loss": 1.1814,
|
2882 |
+
"step": 407
|
2883 |
+
},
|
2884 |
+
{
|
2885 |
+
"epoch": 0.053555606602566204,
|
2886 |
+
"grad_norm": 0.4387180507183075,
|
2887 |
+
"learning_rate": 2.3355599600729915e-06,
|
2888 |
+
"loss": 1.2586,
|
2889 |
+
"step": 408
|
2890 |
+
},
|
2891 |
+
{
|
2892 |
+
"epoch": 0.05368687034423916,
|
2893 |
+
"grad_norm": 0.43076077103614807,
|
2894 |
+
"learning_rate": 2.2264899592121744e-06,
|
2895 |
+
"loss": 1.1326,
|
2896 |
+
"step": 409
|
2897 |
+
},
|
2898 |
+
{
|
2899 |
+
"epoch": 0.05381813408591212,
|
2900 |
+
"grad_norm": 0.4769658148288727,
|
2901 |
+
"learning_rate": 2.1199700045797077e-06,
|
2902 |
+
"loss": 1.1753,
|
2903 |
+
"step": 410
|
2904 |
+
},
|
2905 |
+
{
|
2906 |
+
"epoch": 0.05394939782758507,
|
2907 |
+
"grad_norm": 0.49549686908721924,
|
2908 |
+
"learning_rate": 2.0160057819794466e-06,
|
2909 |
+
"loss": 1.1529,
|
2910 |
+
"step": 411
|
2911 |
+
},
|
2912 |
+
{
|
2913 |
+
"epoch": 0.05408066156925803,
|
2914 |
+
"grad_norm": 0.5492319464683533,
|
2915 |
+
"learning_rate": 1.9146028407958484e-06,
|
2916 |
+
"loss": 1.2282,
|
2917 |
+
"step": 412
|
2918 |
+
},
|
2919 |
+
{
|
2920 |
+
"epoch": 0.05421192531093099,
|
2921 |
+
"grad_norm": 0.5631099939346313,
|
2922 |
+
"learning_rate": 1.8157665936977263e-06,
|
2923 |
+
"loss": 1.2408,
|
2924 |
+
"step": 413
|
2925 |
+
},
|
2926 |
+
{
|
2927 |
+
"epoch": 0.05434318905260394,
|
2928 |
+
"grad_norm": 0.5960178375244141,
|
2929 |
+
"learning_rate": 1.7195023163493252e-06,
|
2930 |
+
"loss": 1.2945,
|
2931 |
+
"step": 414
|
2932 |
+
},
|
2933 |
+
{
|
2934 |
+
"epoch": 0.0544744527942769,
|
2935 |
+
"grad_norm": 0.5799109935760498,
|
2936 |
+
"learning_rate": 1.6258151471287396e-06,
|
2937 |
+
"loss": 1.1414,
|
2938 |
+
"step": 415
|
2939 |
+
},
|
2940 |
+
{
|
2941 |
+
"epoch": 0.05460571653594986,
|
2942 |
+
"grad_norm": 0.6928263306617737,
|
2943 |
+
"learning_rate": 1.5347100868536246e-06,
|
2944 |
+
"loss": 1.2234,
|
2945 |
+
"step": 416
|
2946 |
+
},
|
2947 |
+
{
|
2948 |
+
"epoch": 0.05473698027762281,
|
2949 |
+
"grad_norm": 0.6215956807136536,
|
2950 |
+
"learning_rate": 1.4461919985142735e-06,
|
2951 |
+
"loss": 1.1876,
|
2952 |
+
"step": 417
|
2953 |
+
},
|
2954 |
+
{
|
2955 |
+
"epoch": 0.05486824401929577,
|
2956 |
+
"grad_norm": 0.6092692613601685,
|
2957 |
+
"learning_rate": 1.3602656070140275e-06,
|
2958 |
+
"loss": 1.1264,
|
2959 |
+
"step": 418
|
2960 |
+
},
|
2961 |
+
{
|
2962 |
+
"epoch": 0.05499950776096873,
|
2963 |
+
"grad_norm": 0.6288320422172546,
|
2964 |
+
"learning_rate": 1.27693549891707e-06,
|
2965 |
+
"loss": 1.1799,
|
2966 |
+
"step": 419
|
2967 |
+
},
|
2968 |
+
{
|
2969 |
+
"epoch": 0.05513077150264168,
|
2970 |
+
"grad_norm": 0.6499227285385132,
|
2971 |
+
"learning_rate": 1.196206122203647e-06,
|
2972 |
+
"loss": 1.14,
|
2973 |
+
"step": 420
|
2974 |
+
},
|
2975 |
+
{
|
2976 |
+
"epoch": 0.05526203524431464,
|
2977 |
+
"grad_norm": 0.6873496174812317,
|
2978 |
+
"learning_rate": 1.1180817860325599e-06,
|
2979 |
+
"loss": 1.2302,
|
2980 |
+
"step": 421
|
2981 |
+
},
|
2982 |
+
{
|
2983 |
+
"epoch": 0.0553932989859876,
|
2984 |
+
"grad_norm": 0.6340437531471252,
|
2985 |
+
"learning_rate": 1.0425666605112517e-06,
|
2986 |
+
"loss": 1.1138,
|
2987 |
+
"step": 422
|
2988 |
+
},
|
2989 |
+
{
|
2990 |
+
"epoch": 0.05552456272766055,
|
2991 |
+
"grad_norm": 0.663411557674408,
|
2992 |
+
"learning_rate": 9.696647764731337e-07,
|
2993 |
+
"loss": 1.1017,
|
2994 |
+
"step": 423
|
2995 |
+
},
|
2996 |
+
{
|
2997 |
+
"epoch": 0.05565582646933351,
|
2998 |
+
"grad_norm": 0.6668544411659241,
|
2999 |
+
"learning_rate": 8.993800252624862e-07,
|
3000 |
+
"loss": 1.0881,
|
3001 |
+
"step": 424
|
3002 |
+
},
|
3003 |
+
{
|
3004 |
+
"epoch": 0.05578709021100647,
|
3005 |
+
"grad_norm": 0.6418658494949341,
|
3006 |
+
"learning_rate": 8.317161585266964e-07,
|
3007 |
+
"loss": 1.0638,
|
3008 |
+
"step": 425
|
3009 |
+
},
|
3010 |
+
{
|
3011 |
+
"epoch": 0.05591835395267942,
|
3012 |
+
"grad_norm": 0.703060507774353,
|
3013 |
+
"learning_rate": 7.666767880160464e-07,
|
3014 |
+
"loss": 1.1153,
|
3015 |
+
"step": 426
|
3016 |
+
},
|
3017 |
+
{
|
3018 |
+
"epoch": 0.05604961769435238,
|
3019 |
+
"grad_norm": 0.6755343079566956,
|
3020 |
+
"learning_rate": 7.042653853909064e-07,
|
3021 |
+
"loss": 0.9271,
|
3022 |
+
"step": 427
|
3023 |
+
},
|
3024 |
+
{
|
3025 |
+
"epoch": 0.056180881436025336,
|
3026 |
+
"grad_norm": 0.7167825698852539,
|
3027 |
+
"learning_rate": 6.444852820364222e-07,
|
3028 |
+
"loss": 1.1272,
|
3029 |
+
"step": 428
|
3030 |
+
},
|
3031 |
+
{
|
3032 |
+
"epoch": 0.05631214517769829,
|
3033 |
+
"grad_norm": 0.7692311406135559,
|
3034 |
+
"learning_rate": 5.87339668884701e-07,
|
3035 |
+
"loss": 1.1391,
|
3036 |
+
"step": 429
|
3037 |
+
},
|
3038 |
+
{
|
3039 |
+
"epoch": 0.056443408919371246,
|
3040 |
+
"grad_norm": 0.7314695119857788,
|
3041 |
+
"learning_rate": 5.328315962444874e-07,
|
3042 |
+
"loss": 1.0775,
|
3043 |
+
"step": 430
|
3044 |
+
},
|
3045 |
+
{
|
3046 |
+
"epoch": 0.056574672661044205,
|
3047 |
+
"grad_norm": 0.7138880491256714,
|
3048 |
+
"learning_rate": 4.809639736383431e-07,
|
3049 |
+
"loss": 0.9616,
|
3050 |
+
"step": 431
|
3051 |
+
},
|
3052 |
+
{
|
3053 |
+
"epoch": 0.05670593640271716,
|
3054 |
+
"grad_norm": 0.7646660208702087,
|
3055 |
+
"learning_rate": 4.317395696473214e-07,
|
3056 |
+
"loss": 1.0419,
|
3057 |
+
"step": 432
|
3058 |
+
},
|
3059 |
+
{
|
3060 |
+
"epoch": 0.056837200144390115,
|
3061 |
+
"grad_norm": 0.7127510905265808,
|
3062 |
+
"learning_rate": 3.851610117632354e-07,
|
3063 |
+
"loss": 0.9304,
|
3064 |
+
"step": 433
|
3065 |
+
},
|
3066 |
+
{
|
3067 |
+
"epoch": 0.056968463886063074,
|
3068 |
+
"grad_norm": 0.6503668427467346,
|
3069 |
+
"learning_rate": 3.4123078624834216e-07,
|
3070 |
+
"loss": 0.8488,
|
3071 |
+
"step": 434
|
3072 |
+
},
|
3073 |
+
{
|
3074 |
+
"epoch": 0.057099727627736026,
|
3075 |
+
"grad_norm": 0.7812151908874512,
|
3076 |
+
"learning_rate": 2.9995123800270476e-07,
|
3077 |
+
"loss": 0.971,
|
3078 |
+
"step": 435
|
3079 |
+
},
|
3080 |
+
{
|
3081 |
+
"epoch": 0.057230991369408984,
|
3082 |
+
"grad_norm": 0.7327147722244263,
|
3083 |
+
"learning_rate": 2.613245704389644e-07,
|
3084 |
+
"loss": 0.9165,
|
3085 |
+
"step": 436
|
3086 |
+
},
|
3087 |
+
{
|
3088 |
+
"epoch": 0.05736225511108194,
|
3089 |
+
"grad_norm": 0.7133559584617615,
|
3090 |
+
"learning_rate": 2.2535284536476242e-07,
|
3091 |
+
"loss": 0.8214,
|
3092 |
+
"step": 437
|
3093 |
+
},
|
3094 |
+
{
|
3095 |
+
"epoch": 0.057493518852754895,
|
3096 |
+
"grad_norm": 0.6528218984603882,
|
3097 |
+
"learning_rate": 1.920379828726726e-07,
|
3098 |
+
"loss": 0.6965,
|
3099 |
+
"step": 438
|
3100 |
+
},
|
3101 |
+
{
|
3102 |
+
"epoch": 0.057624782594427854,
|
3103 |
+
"grad_norm": 0.7258195281028748,
|
3104 |
+
"learning_rate": 1.6138176123770554e-07,
|
3105 |
+
"loss": 0.8052,
|
3106 |
+
"step": 439
|
3107 |
+
},
|
3108 |
+
{
|
3109 |
+
"epoch": 0.05775604633610081,
|
3110 |
+
"grad_norm": 0.7617157697677612,
|
3111 |
+
"learning_rate": 1.333858168224178e-07,
|
3112 |
+
"loss": 0.7333,
|
3113 |
+
"step": 440
|
3114 |
+
},
|
3115 |
+
{
|
3116 |
+
"epoch": 0.057887310077773764,
|
3117 |
+
"grad_norm": 0.7330428957939148,
|
3118 |
+
"learning_rate": 1.0805164398952072e-07,
|
3119 |
+
"loss": 0.8641,
|
3120 |
+
"step": 441
|
3121 |
+
},
|
3122 |
+
{
|
3123 |
+
"epoch": 0.05801857381944672,
|
3124 |
+
"grad_norm": 0.713188648223877,
|
3125 |
+
"learning_rate": 8.53805950221498e-08,
|
3126 |
+
"loss": 0.5969,
|
3127 |
+
"step": 442
|
3128 |
+
},
|
3129 |
+
{
|
3130 |
+
"epoch": 0.05814983756111968,
|
3131 |
+
"grad_norm": 0.6643161773681641,
|
3132 |
+
"learning_rate": 6.537388005167233e-08,
|
3133 |
+
"loss": 0.6351,
|
3134 |
+
"step": 443
|
3135 |
+
},
|
3136 |
+
{
|
3137 |
+
"epoch": 0.05828110130279263,
|
3138 |
+
"grad_norm": 0.7042210102081299,
|
3139 |
+
"learning_rate": 4.8032566993089225e-08,
|
3140 |
+
"loss": 0.6916,
|
3141 |
+
"step": 444
|
3142 |
+
},
|
3143 |
+
{
|
3144 |
+
"epoch": 0.05841236504446559,
|
3145 |
+
"grad_norm": 0.7722613215446472,
|
3146 |
+
"learning_rate": 3.3357581488030475e-08,
|
3147 |
+
"loss": 0.6977,
|
3148 |
+
"step": 445
|
3149 |
+
},
|
3150 |
+
{
|
3151 |
+
"epoch": 0.05854362878613855,
|
3152 |
+
"grad_norm": 0.6429854035377502,
|
3153 |
+
"learning_rate": 2.134970685536697e-08,
|
3154 |
+
"loss": 0.4461,
|
3155 |
+
"step": 446
|
3156 |
+
},
|
3157 |
+
{
|
3158 |
+
"epoch": 0.0586748925278115,
|
3159 |
+
"grad_norm": 0.628583550453186,
|
3160 |
+
"learning_rate": 1.200958404936059e-08,
|
3161 |
+
"loss": 0.4453,
|
3162 |
+
"step": 447
|
3163 |
+
},
|
3164 |
+
{
|
3165 |
+
"epoch": 0.05880615626948446,
|
3166 |
+
"grad_norm": 0.6538881063461304,
|
3167 |
+
"learning_rate": 5.337711625497121e-09,
|
3168 |
+
"loss": 0.462,
|
3169 |
+
"step": 448
|
3170 |
+
},
|
3171 |
+
{
|
3172 |
+
"epoch": 0.05893742001115742,
|
3173 |
+
"grad_norm": 0.6723350286483765,
|
3174 |
+
"learning_rate": 1.3344457138297906e-09,
|
3175 |
+
"loss": 0.4467,
|
3176 |
+
"step": 449
|
3177 |
+
},
|
3178 |
+
{
|
3179 |
+
"epoch": 0.05906868375283037,
|
3180 |
+
"grad_norm": 0.8126013875007629,
|
3181 |
+
"learning_rate": 0.0,
|
3182 |
+
"loss": 0.4989,
|
3183 |
+
"step": 450
|
3184 |
+
},
|
3185 |
+
{
|
3186 |
+
"epoch": 0.05906868375283037,
|
3187 |
+
"eval_loss": 0.9613198041915894,
|
3188 |
+
"eval_runtime": 960.9133,
|
3189 |
+
"eval_samples_per_second": 13.353,
|
3190 |
+
"eval_steps_per_second": 3.338,
|
3191 |
+
"step": 450
|
3192 |
}
|
3193 |
],
|
3194 |
"logging_steps": 1,
|
|
|
3212 |
"should_evaluate": false,
|
3213 |
"should_log": false,
|
3214 |
"should_save": true,
|
3215 |
+
"should_training_stop": true
|
3216 |
},
|
3217 |
"attributes": {}
|
3218 |
}
|
3219 |
},
|
3220 |
+
"total_flos": 6.602152785708319e+17,
|
3221 |
"train_batch_size": 8,
|
3222 |
"trial_name": null,
|
3223 |
"trial_params": null
|