Training in progress, step 361, 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 639691872
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cc0496d758ec86c37d1971c49b2e7b5876067a9ce8635cb3d8c4520d38cd938b
|
| 3 |
size 639691872
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 325340244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:de404a495901928b91622cc6faceee097ce90f7963c9310ed4a92168b38a33d6
|
| 3 |
size 325340244
|
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 sha256:0f761d844d5f040bacf1f99959e3a5fd3da1b16fd7877660f2d80d9193f6afa1
|
| 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:b857be7b95ff8324c4727de3c0f481a268cea8c6e2533b10d776846f18e23993
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 1.556799054145813,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-300",
|
| 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 +2139,433 @@
|
|
| 2139 |
"eval_samples_per_second": 6.356,
|
| 2140 |
"eval_steps_per_second": 1.589,
|
| 2141 |
"step": 300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2142 |
}
|
| 2143 |
],
|
| 2144 |
"logging_steps": 1,
|
|
@@ -2162,12 +2589,12 @@
|
|
| 2162 |
"should_evaluate": false,
|
| 2163 |
"should_log": false,
|
| 2164 |
"should_save": true,
|
| 2165 |
-
"should_training_stop":
|
| 2166 |
},
|
| 2167 |
"attributes": {}
|
| 2168 |
}
|
| 2169 |
},
|
| 2170 |
-
"total_flos":
|
| 2171 |
"train_batch_size": 8,
|
| 2172 |
"trial_name": null,
|
| 2173 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 1.556799054145813,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-300",
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 361,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2139 |
"eval_samples_per_second": 6.356,
|
| 2140 |
"eval_steps_per_second": 1.589,
|
| 2141 |
"step": 300
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"epoch": 0.8337950138504155,
|
| 2145 |
+
"grad_norm": 0.38641664385795593,
|
| 2146 |
+
"learning_rate": 7.446405675168938e-06,
|
| 2147 |
+
"loss": 1.6862,
|
| 2148 |
+
"step": 301
|
| 2149 |
+
},
|
| 2150 |
+
{
|
| 2151 |
+
"epoch": 0.8365650969529086,
|
| 2152 |
+
"grad_norm": 0.386618435382843,
|
| 2153 |
+
"learning_rate": 7.206354198753862e-06,
|
| 2154 |
+
"loss": 1.6566,
|
| 2155 |
+
"step": 302
|
| 2156 |
+
},
|
| 2157 |
+
{
|
| 2158 |
+
"epoch": 0.8393351800554016,
|
| 2159 |
+
"grad_norm": 0.4190298318862915,
|
| 2160 |
+
"learning_rate": 6.969934901005809e-06,
|
| 2161 |
+
"loss": 1.6447,
|
| 2162 |
+
"step": 303
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 0.8421052631578947,
|
| 2166 |
+
"grad_norm": 0.40949639678001404,
|
| 2167 |
+
"learning_rate": 6.7371678483882264e-06,
|
| 2168 |
+
"loss": 1.6331,
|
| 2169 |
+
"step": 304
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 0.8448753462603878,
|
| 2173 |
+
"grad_norm": 0.42585617303848267,
|
| 2174 |
+
"learning_rate": 6.508072797374454e-06,
|
| 2175 |
+
"loss": 1.6439,
|
| 2176 |
+
"step": 305
|
| 2177 |
+
},
|
| 2178 |
+
{
|
| 2179 |
+
"epoch": 0.8476454293628809,
|
| 2180 |
+
"grad_norm": 0.41650789976119995,
|
| 2181 |
+
"learning_rate": 6.282669192770896e-06,
|
| 2182 |
+
"loss": 1.616,
|
| 2183 |
+
"step": 306
|
| 2184 |
+
},
|
| 2185 |
+
{
|
| 2186 |
+
"epoch": 0.850415512465374,
|
| 2187 |
+
"grad_norm": 0.4084133803844452,
|
| 2188 |
+
"learning_rate": 6.060976166066546e-06,
|
| 2189 |
+
"loss": 1.587,
|
| 2190 |
+
"step": 307
|
| 2191 |
+
},
|
| 2192 |
+
{
|
| 2193 |
+
"epoch": 0.853185595567867,
|
| 2194 |
+
"grad_norm": 0.4127242863178253,
|
| 2195 |
+
"learning_rate": 5.8430125338092115e-06,
|
| 2196 |
+
"loss": 1.6102,
|
| 2197 |
+
"step": 308
|
| 2198 |
+
},
|
| 2199 |
+
{
|
| 2200 |
+
"epoch": 0.8559556786703602,
|
| 2201 |
+
"grad_norm": 0.42294085025787354,
|
| 2202 |
+
"learning_rate": 5.628796796008434e-06,
|
| 2203 |
+
"loss": 1.6542,
|
| 2204 |
+
"step": 309
|
| 2205 |
+
},
|
| 2206 |
+
{
|
| 2207 |
+
"epoch": 0.8587257617728532,
|
| 2208 |
+
"grad_norm": 0.41091540455818176,
|
| 2209 |
+
"learning_rate": 5.418347134565249e-06,
|
| 2210 |
+
"loss": 1.5955,
|
| 2211 |
+
"step": 310
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 0.8614958448753463,
|
| 2215 |
+
"grad_norm": 0.41265037655830383,
|
| 2216 |
+
"learning_rate": 5.211681411728969e-06,
|
| 2217 |
+
"loss": 1.6711,
|
| 2218 |
+
"step": 311
|
| 2219 |
+
},
|
| 2220 |
+
{
|
| 2221 |
+
"epoch": 0.8642659279778393,
|
| 2222 |
+
"grad_norm": 0.44204774498939514,
|
| 2223 |
+
"learning_rate": 5.008817168581137e-06,
|
| 2224 |
+
"loss": 1.5871,
|
| 2225 |
+
"step": 312
|
| 2226 |
+
},
|
| 2227 |
+
{
|
| 2228 |
+
"epoch": 0.8670360110803325,
|
| 2229 |
+
"grad_norm": 0.4217774271965027,
|
| 2230 |
+
"learning_rate": 4.809771623546627e-06,
|
| 2231 |
+
"loss": 1.5632,
|
| 2232 |
+
"step": 313
|
| 2233 |
+
},
|
| 2234 |
+
{
|
| 2235 |
+
"epoch": 0.8698060941828255,
|
| 2236 |
+
"grad_norm": 0.43789371848106384,
|
| 2237 |
+
"learning_rate": 4.614561670932288e-06,
|
| 2238 |
+
"loss": 1.5672,
|
| 2239 |
+
"step": 314
|
| 2240 |
+
},
|
| 2241 |
+
{
|
| 2242 |
+
"epoch": 0.8725761772853186,
|
| 2243 |
+
"grad_norm": 0.44473928213119507,
|
| 2244 |
+
"learning_rate": 4.423203879492943e-06,
|
| 2245 |
+
"loss": 1.5175,
|
| 2246 |
+
"step": 315
|
| 2247 |
+
},
|
| 2248 |
+
{
|
| 2249 |
+
"epoch": 0.8753462603878116,
|
| 2250 |
+
"grad_norm": 0.4420805871486664,
|
| 2251 |
+
"learning_rate": 4.2357144910251e-06,
|
| 2252 |
+
"loss": 1.5502,
|
| 2253 |
+
"step": 316
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 0.8781163434903048,
|
| 2257 |
+
"grad_norm": 0.448369562625885,
|
| 2258 |
+
"learning_rate": 4.05210941898847e-06,
|
| 2259 |
+
"loss": 1.5517,
|
| 2260 |
+
"step": 317
|
| 2261 |
+
},
|
| 2262 |
+
{
|
| 2263 |
+
"epoch": 0.8808864265927978,
|
| 2264 |
+
"grad_norm": 0.44154343008995056,
|
| 2265 |
+
"learning_rate": 3.872404247155193e-06,
|
| 2266 |
+
"loss": 1.4863,
|
| 2267 |
+
"step": 318
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 0.8836565096952909,
|
| 2271 |
+
"grad_norm": 0.44638606905937195,
|
| 2272 |
+
"learning_rate": 3.696614228287187e-06,
|
| 2273 |
+
"loss": 1.5786,
|
| 2274 |
+
"step": 319
|
| 2275 |
+
},
|
| 2276 |
+
{
|
| 2277 |
+
"epoch": 0.8864265927977839,
|
| 2278 |
+
"grad_norm": 0.43334829807281494,
|
| 2279 |
+
"learning_rate": 3.5247542828415747e-06,
|
| 2280 |
+
"loss": 1.5491,
|
| 2281 |
+
"step": 320
|
| 2282 |
+
},
|
| 2283 |
+
{
|
| 2284 |
+
"epoch": 0.889196675900277,
|
| 2285 |
+
"grad_norm": 0.4540192484855652,
|
| 2286 |
+
"learning_rate": 3.356838997704226e-06,
|
| 2287 |
+
"loss": 1.5194,
|
| 2288 |
+
"step": 321
|
| 2289 |
+
},
|
| 2290 |
+
{
|
| 2291 |
+
"epoch": 0.8919667590027701,
|
| 2292 |
+
"grad_norm": 0.4565938115119934,
|
| 2293 |
+
"learning_rate": 3.1928826249516987e-06,
|
| 2294 |
+
"loss": 1.5226,
|
| 2295 |
+
"step": 322
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 0.8947368421052632,
|
| 2299 |
+
"grad_norm": 0.4570363461971283,
|
| 2300 |
+
"learning_rate": 3.0328990806415934e-06,
|
| 2301 |
+
"loss": 1.5702,
|
| 2302 |
+
"step": 323
|
| 2303 |
+
},
|
| 2304 |
+
{
|
| 2305 |
+
"epoch": 0.8975069252077562,
|
| 2306 |
+
"grad_norm": 0.4690124988555908,
|
| 2307 |
+
"learning_rate": 2.8769019436313715e-06,
|
| 2308 |
+
"loss": 1.5384,
|
| 2309 |
+
"step": 324
|
| 2310 |
+
},
|
| 2311 |
+
{
|
| 2312 |
+
"epoch": 0.9002770083102493,
|
| 2313 |
+
"grad_norm": 0.47177836298942566,
|
| 2314 |
+
"learning_rate": 2.7249044544258363e-06,
|
| 2315 |
+
"loss": 1.5368,
|
| 2316 |
+
"step": 325
|
| 2317 |
+
},
|
| 2318 |
+
{
|
| 2319 |
+
"epoch": 0.9030470914127424,
|
| 2320 |
+
"grad_norm": 0.4627552628517151,
|
| 2321 |
+
"learning_rate": 2.576919514053355e-06,
|
| 2322 |
+
"loss": 1.5475,
|
| 2323 |
+
"step": 326
|
| 2324 |
+
},
|
| 2325 |
+
{
|
| 2326 |
+
"epoch": 0.9058171745152355,
|
| 2327 |
+
"grad_norm": 0.4692405164241791,
|
| 2328 |
+
"learning_rate": 2.4329596829708144e-06,
|
| 2329 |
+
"loss": 1.4843,
|
| 2330 |
+
"step": 327
|
| 2331 |
+
},
|
| 2332 |
+
{
|
| 2333 |
+
"epoch": 0.9085872576177285,
|
| 2334 |
+
"grad_norm": 0.5105109214782715,
|
| 2335 |
+
"learning_rate": 2.2930371799975594e-06,
|
| 2336 |
+
"loss": 1.5008,
|
| 2337 |
+
"step": 328
|
| 2338 |
+
},
|
| 2339 |
+
{
|
| 2340 |
+
"epoch": 0.9113573407202216,
|
| 2341 |
+
"grad_norm": 0.48938921093940735,
|
| 2342 |
+
"learning_rate": 2.157163881278312e-06,
|
| 2343 |
+
"loss": 1.5492,
|
| 2344 |
+
"step": 329
|
| 2345 |
+
},
|
| 2346 |
+
{
|
| 2347 |
+
"epoch": 0.9141274238227147,
|
| 2348 |
+
"grad_norm": 0.485343337059021,
|
| 2349 |
+
"learning_rate": 2.0253513192751373e-06,
|
| 2350 |
+
"loss": 1.5183,
|
| 2351 |
+
"step": 330
|
| 2352 |
+
},
|
| 2353 |
+
{
|
| 2354 |
+
"epoch": 0.9168975069252078,
|
| 2355 |
+
"grad_norm": 0.4986342489719391,
|
| 2356 |
+
"learning_rate": 1.8976106817886196e-06,
|
| 2357 |
+
"loss": 1.5177,
|
| 2358 |
+
"step": 331
|
| 2359 |
+
},
|
| 2360 |
+
{
|
| 2361 |
+
"epoch": 0.9196675900277008,
|
| 2362 |
+
"grad_norm": 0.49586573243141174,
|
| 2363 |
+
"learning_rate": 1.7739528110083004e-06,
|
| 2364 |
+
"loss": 1.519,
|
| 2365 |
+
"step": 332
|
| 2366 |
+
},
|
| 2367 |
+
{
|
| 2368 |
+
"epoch": 0.9224376731301939,
|
| 2369 |
+
"grad_norm": 0.5050438642501831,
|
| 2370 |
+
"learning_rate": 1.6543882025923886e-06,
|
| 2371 |
+
"loss": 1.4329,
|
| 2372 |
+
"step": 333
|
| 2373 |
+
},
|
| 2374 |
+
{
|
| 2375 |
+
"epoch": 0.925207756232687,
|
| 2376 |
+
"grad_norm": 0.5025500655174255,
|
| 2377 |
+
"learning_rate": 1.5389270047769578e-06,
|
| 2378 |
+
"loss": 1.486,
|
| 2379 |
+
"step": 334
|
| 2380 |
+
},
|
| 2381 |
+
{
|
| 2382 |
+
"epoch": 0.9279778393351801,
|
| 2383 |
+
"grad_norm": 0.50665682554245,
|
| 2384 |
+
"learning_rate": 1.4275790175145908e-06,
|
| 2385 |
+
"loss": 1.4943,
|
| 2386 |
+
"step": 335
|
| 2387 |
+
},
|
| 2388 |
+
{
|
| 2389 |
+
"epoch": 0.9307479224376731,
|
| 2390 |
+
"grad_norm": 0.518688440322876,
|
| 2391 |
+
"learning_rate": 1.3203536916425841e-06,
|
| 2392 |
+
"loss": 1.526,
|
| 2393 |
+
"step": 336
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 0.9335180055401662,
|
| 2397 |
+
"grad_norm": 0.5425696969032288,
|
| 2398 |
+
"learning_rate": 1.217260128080816e-06,
|
| 2399 |
+
"loss": 1.513,
|
| 2400 |
+
"step": 337
|
| 2401 |
+
},
|
| 2402 |
+
{
|
| 2403 |
+
"epoch": 0.9362880886426593,
|
| 2404 |
+
"grad_norm": 0.5442622303962708,
|
| 2405 |
+
"learning_rate": 1.1183070770592441e-06,
|
| 2406 |
+
"loss": 1.5735,
|
| 2407 |
+
"step": 338
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 0.9390581717451524,
|
| 2411 |
+
"grad_norm": 0.5330432057380676,
|
| 2412 |
+
"learning_rate": 1.0235029373752758e-06,
|
| 2413 |
+
"loss": 1.5291,
|
| 2414 |
+
"step": 339
|
| 2415 |
+
},
|
| 2416 |
+
{
|
| 2417 |
+
"epoch": 0.9418282548476454,
|
| 2418 |
+
"grad_norm": 0.5193708539009094,
|
| 2419 |
+
"learning_rate": 9.32855755680867e-07,
|
| 2420 |
+
"loss": 1.418,
|
| 2421 |
+
"step": 340
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 0.9445983379501385,
|
| 2425 |
+
"grad_norm": 0.5180433392524719,
|
| 2426 |
+
"learning_rate": 8.463732257995571e-07,
|
| 2427 |
+
"loss": 1.495,
|
| 2428 |
+
"step": 341
|
| 2429 |
+
},
|
| 2430 |
+
{
|
| 2431 |
+
"epoch": 0.9473684210526315,
|
| 2432 |
+
"grad_norm": 0.5407363772392273,
|
| 2433 |
+
"learning_rate": 7.640626880734581e-07,
|
| 2434 |
+
"loss": 1.5243,
|
| 2435 |
+
"step": 342
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 0.9501385041551247,
|
| 2439 |
+
"grad_norm": 0.5722284913063049,
|
| 2440 |
+
"learning_rate": 6.859311287402081e-07,
|
| 2441 |
+
"loss": 1.4204,
|
| 2442 |
+
"step": 343
|
| 2443 |
+
},
|
| 2444 |
+
{
|
| 2445 |
+
"epoch": 0.9529085872576177,
|
| 2446 |
+
"grad_norm": 0.5504374504089355,
|
| 2447 |
+
"learning_rate": 6.119851793400189e-07,
|
| 2448 |
+
"loss": 1.4763,
|
| 2449 |
+
"step": 344
|
| 2450 |
+
},
|
| 2451 |
+
{
|
| 2452 |
+
"epoch": 0.9556786703601108,
|
| 2453 |
+
"grad_norm": 0.5620555281639099,
|
| 2454 |
+
"learning_rate": 5.422311161528193e-07,
|
| 2455 |
+
"loss": 1.4551,
|
| 2456 |
+
"step": 345
|
| 2457 |
+
},
|
| 2458 |
+
{
|
| 2459 |
+
"epoch": 0.9584487534626038,
|
| 2460 |
+
"grad_norm": 0.5651283860206604,
|
| 2461 |
+
"learning_rate": 4.7667485966552685e-07,
|
| 2462 |
+
"loss": 1.4817,
|
| 2463 |
+
"step": 346
|
| 2464 |
+
},
|
| 2465 |
+
{
|
| 2466 |
+
"epoch": 0.961218836565097,
|
| 2467 |
+
"grad_norm": 0.6057314872741699,
|
| 2468 |
+
"learning_rate": 4.153219740695435e-07,
|
| 2469 |
+
"loss": 1.5466,
|
| 2470 |
+
"step": 347
|
| 2471 |
+
},
|
| 2472 |
+
{
|
| 2473 |
+
"epoch": 0.96398891966759,
|
| 2474 |
+
"grad_norm": 0.6597415208816528,
|
| 2475 |
+
"learning_rate": 3.5817766678850615e-07,
|
| 2476 |
+
"loss": 1.5562,
|
| 2477 |
+
"step": 348
|
| 2478 |
+
},
|
| 2479 |
+
{
|
| 2480 |
+
"epoch": 0.9667590027700831,
|
| 2481 |
+
"grad_norm": 0.7001429796218872,
|
| 2482 |
+
"learning_rate": 3.052467880362675e-07,
|
| 2483 |
+
"loss": 1.899,
|
| 2484 |
+
"step": 349
|
| 2485 |
+
},
|
| 2486 |
+
{
|
| 2487 |
+
"epoch": 0.9695290858725761,
|
| 2488 |
+
"grad_norm": 0.8798688054084778,
|
| 2489 |
+
"learning_rate": 2.5653383040524227e-07,
|
| 2490 |
+
"loss": 1.96,
|
| 2491 |
+
"step": 350
|
| 2492 |
+
},
|
| 2493 |
+
{
|
| 2494 |
+
"epoch": 0.9722991689750693,
|
| 2495 |
+
"grad_norm": 0.3499889671802521,
|
| 2496 |
+
"learning_rate": 2.1204292848509555e-07,
|
| 2497 |
+
"loss": 1.6123,
|
| 2498 |
+
"step": 351
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 0.9750692520775623,
|
| 2502 |
+
"grad_norm": 0.38690975308418274,
|
| 2503 |
+
"learning_rate": 1.717778585118013e-07,
|
| 2504 |
+
"loss": 1.5579,
|
| 2505 |
+
"step": 352
|
| 2506 |
+
},
|
| 2507 |
+
{
|
| 2508 |
+
"epoch": 0.9778393351800554,
|
| 2509 |
+
"grad_norm": 0.41460373997688293,
|
| 2510 |
+
"learning_rate": 1.3574203804713747e-07,
|
| 2511 |
+
"loss": 1.585,
|
| 2512 |
+
"step": 353
|
| 2513 |
+
},
|
| 2514 |
+
{
|
| 2515 |
+
"epoch": 0.9806094182825484,
|
| 2516 |
+
"grad_norm": 0.4353157877922058,
|
| 2517 |
+
"learning_rate": 1.0393852568860719e-07,
|
| 2518 |
+
"loss": 1.5611,
|
| 2519 |
+
"step": 354
|
| 2520 |
+
},
|
| 2521 |
+
{
|
| 2522 |
+
"epoch": 0.9833795013850416,
|
| 2523 |
+
"grad_norm": 0.450083464384079,
|
| 2524 |
+
"learning_rate": 7.637002080985168e-08,
|
| 2525 |
+
"loss": 1.5568,
|
| 2526 |
+
"step": 355
|
| 2527 |
+
},
|
| 2528 |
+
{
|
| 2529 |
+
"epoch": 0.9861495844875346,
|
| 2530 |
+
"grad_norm": 0.47905081510543823,
|
| 2531 |
+
"learning_rate": 5.303886333151153e-08,
|
| 2532 |
+
"loss": 1.5438,
|
| 2533 |
+
"step": 356
|
| 2534 |
+
},
|
| 2535 |
+
{
|
| 2536 |
+
"epoch": 0.9889196675900277,
|
| 2537 |
+
"grad_norm": 0.4843379855155945,
|
| 2538 |
+
"learning_rate": 3.394703352263551e-08,
|
| 2539 |
+
"loss": 1.4646,
|
| 2540 |
+
"step": 357
|
| 2541 |
+
},
|
| 2542 |
+
{
|
| 2543 |
+
"epoch": 0.9916897506925207,
|
| 2544 |
+
"grad_norm": 0.5109671950340271,
|
| 2545 |
+
"learning_rate": 1.9096151832609375e-08,
|
| 2546 |
+
"loss": 1.513,
|
| 2547 |
+
"step": 358
|
| 2548 |
+
},
|
| 2549 |
+
{
|
| 2550 |
+
"epoch": 0.9944598337950139,
|
| 2551 |
+
"grad_norm": 0.5229476690292358,
|
| 2552 |
+
"learning_rate": 8.487478753615997e-09,
|
| 2553 |
+
"loss": 1.4675,
|
| 2554 |
+
"step": 359
|
| 2555 |
+
},
|
| 2556 |
+
{
|
| 2557 |
+
"epoch": 0.997229916897507,
|
| 2558 |
+
"grad_norm": 0.5755655765533447,
|
| 2559 |
+
"learning_rate": 2.1219147136264382e-09,
|
| 2560 |
+
"loss": 1.5564,
|
| 2561 |
+
"step": 360
|
| 2562 |
+
},
|
| 2563 |
+
{
|
| 2564 |
+
"epoch": 1.0,
|
| 2565 |
+
"grad_norm": 0.7581580877304077,
|
| 2566 |
+
"learning_rate": 0.0,
|
| 2567 |
+
"loss": 1.8149,
|
| 2568 |
+
"step": 361
|
| 2569 |
}
|
| 2570 |
],
|
| 2571 |
"logging_steps": 1,
|
|
|
|
| 2589 |
"should_evaluate": false,
|
| 2590 |
"should_log": false,
|
| 2591 |
"should_save": true,
|
| 2592 |
+
"should_training_stop": true
|
| 2593 |
},
|
| 2594 |
"attributes": {}
|
| 2595 |
}
|
| 2596 |
},
|
| 2597 |
+
"total_flos": 5.1502575448070554e+17,
|
| 2598 |
"train_batch_size": 8,
|
| 2599 |
"trial_name": null,
|
| 2600 |
"trial_params": null
|