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 289452128
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b085157f0473441fb9977d3df11a7a4e193b6bdf47d705257fba06b83141453a
|
| 3 |
size 289452128
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 147360212
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:034873755ff7773119caf8c281745405cc671d8e01312b19148a01f4c30f2e83
|
| 3 |
size 147360212
|
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:c42e7102d7712bedf568055f461ee69fd9419d0183ab66940d3e4fb3204d09fe
|
| 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 sha256:505f9225762b105f8ca5168f44d99b2f8467174f4ade85f1cc95f684fbd828e0
|
| 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": 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": 31.861,
|
| 2140 |
"eval_steps_per_second": 7.967,
|
| 2141 |
"step": 300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2142 |
}
|
| 2143 |
],
|
| 2144 |
"logging_steps": 1,
|
|
@@ -2167,7 +2875,7 @@
|
|
| 2167 |
"attributes": {}
|
| 2168 |
}
|
| 2169 |
},
|
| 2170 |
-
"total_flos":
|
| 2171 |
"train_batch_size": 8,
|
| 2172 |
"trial_name": null,
|
| 2173 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.6768385171890259,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-400",
|
| 4 |
+
"epoch": 0.07969318125217911,
|
| 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": 31.861,
|
| 2140 |
"eval_steps_per_second": 7.967,
|
| 2141 |
"step": 300
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"epoch": 0.05996911889226478,
|
| 2145 |
+
"grad_norm": 0.1829785406589508,
|
| 2146 |
+
"learning_rate": 3.6743228687829595e-05,
|
| 2147 |
+
"loss": 0.4575,
|
| 2148 |
+
"step": 301
|
| 2149 |
+
},
|
| 2150 |
+
{
|
| 2151 |
+
"epoch": 0.06016835184539523,
|
| 2152 |
+
"grad_norm": 0.2460087090730667,
|
| 2153 |
+
"learning_rate": 3.642797750674629e-05,
|
| 2154 |
+
"loss": 0.5853,
|
| 2155 |
+
"step": 302
|
| 2156 |
+
},
|
| 2157 |
+
{
|
| 2158 |
+
"epoch": 0.06036758479852568,
|
| 2159 |
+
"grad_norm": 0.25549712777137756,
|
| 2160 |
+
"learning_rate": 3.6113307705935396e-05,
|
| 2161 |
+
"loss": 0.6608,
|
| 2162 |
+
"step": 303
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 0.06056681775165612,
|
| 2166 |
+
"grad_norm": 0.25257304310798645,
|
| 2167 |
+
"learning_rate": 3.579923276480387e-05,
|
| 2168 |
+
"loss": 0.6838,
|
| 2169 |
+
"step": 304
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 0.06076605070478657,
|
| 2173 |
+
"grad_norm": 0.27871885895729065,
|
| 2174 |
+
"learning_rate": 3.5485766137276894e-05,
|
| 2175 |
+
"loss": 0.7465,
|
| 2176 |
+
"step": 305
|
| 2177 |
+
},
|
| 2178 |
+
{
|
| 2179 |
+
"epoch": 0.06096528365791702,
|
| 2180 |
+
"grad_norm": 0.21810035407543182,
|
| 2181 |
+
"learning_rate": 3.5172921251221455e-05,
|
| 2182 |
+
"loss": 0.6733,
|
| 2183 |
+
"step": 306
|
| 2184 |
+
},
|
| 2185 |
+
{
|
| 2186 |
+
"epoch": 0.061164516611047465,
|
| 2187 |
+
"grad_norm": 0.19901888072490692,
|
| 2188 |
+
"learning_rate": 3.486071150787128e-05,
|
| 2189 |
+
"loss": 0.6081,
|
| 2190 |
+
"step": 307
|
| 2191 |
+
},
|
| 2192 |
+
{
|
| 2193 |
+
"epoch": 0.061363749564177915,
|
| 2194 |
+
"grad_norm": 0.21920450031757355,
|
| 2195 |
+
"learning_rate": 3.4549150281252636e-05,
|
| 2196 |
+
"loss": 0.6168,
|
| 2197 |
+
"step": 308
|
| 2198 |
+
},
|
| 2199 |
+
{
|
| 2200 |
+
"epoch": 0.061562982517308365,
|
| 2201 |
+
"grad_norm": 0.22061915695667267,
|
| 2202 |
+
"learning_rate": 3.423825091761153e-05,
|
| 2203 |
+
"loss": 0.6026,
|
| 2204 |
+
"step": 309
|
| 2205 |
+
},
|
| 2206 |
+
{
|
| 2207 |
+
"epoch": 0.06176221547043881,
|
| 2208 |
+
"grad_norm": 0.22976446151733398,
|
| 2209 |
+
"learning_rate": 3.392802673484193e-05,
|
| 2210 |
+
"loss": 0.7188,
|
| 2211 |
+
"step": 310
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 0.06196144842356926,
|
| 2215 |
+
"grad_norm": 0.23608386516571045,
|
| 2216 |
+
"learning_rate": 3.361849102191533e-05,
|
| 2217 |
+
"loss": 0.6482,
|
| 2218 |
+
"step": 311
|
| 2219 |
+
},
|
| 2220 |
+
{
|
| 2221 |
+
"epoch": 0.06216068137669971,
|
| 2222 |
+
"grad_norm": 0.24407218396663666,
|
| 2223 |
+
"learning_rate": 3.330965703831146e-05,
|
| 2224 |
+
"loss": 0.6767,
|
| 2225 |
+
"step": 312
|
| 2226 |
+
},
|
| 2227 |
+
{
|
| 2228 |
+
"epoch": 0.06235991432983015,
|
| 2229 |
+
"grad_norm": 0.2779025137424469,
|
| 2230 |
+
"learning_rate": 3.300153801345028e-05,
|
| 2231 |
+
"loss": 0.7787,
|
| 2232 |
+
"step": 313
|
| 2233 |
+
},
|
| 2234 |
+
{
|
| 2235 |
+
"epoch": 0.0625591472829606,
|
| 2236 |
+
"grad_norm": 0.21833088994026184,
|
| 2237 |
+
"learning_rate": 3.2694147146125345e-05,
|
| 2238 |
+
"loss": 0.5775,
|
| 2239 |
+
"step": 314
|
| 2240 |
+
},
|
| 2241 |
+
{
|
| 2242 |
+
"epoch": 0.06275838023609105,
|
| 2243 |
+
"grad_norm": 0.2857174277305603,
|
| 2244 |
+
"learning_rate": 3.2387497603938326e-05,
|
| 2245 |
+
"loss": 0.7358,
|
| 2246 |
+
"step": 315
|
| 2247 |
+
},
|
| 2248 |
+
{
|
| 2249 |
+
"epoch": 0.0629576131892215,
|
| 2250 |
+
"grad_norm": 0.2404899299144745,
|
| 2251 |
+
"learning_rate": 3.2081602522734986e-05,
|
| 2252 |
+
"loss": 0.6752,
|
| 2253 |
+
"step": 316
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 0.06315684614235194,
|
| 2257 |
+
"grad_norm": 0.25618627667427063,
|
| 2258 |
+
"learning_rate": 3.177647500604252e-05,
|
| 2259 |
+
"loss": 0.7334,
|
| 2260 |
+
"step": 317
|
| 2261 |
+
},
|
| 2262 |
+
{
|
| 2263 |
+
"epoch": 0.0633560790954824,
|
| 2264 |
+
"grad_norm": 0.21781964600086212,
|
| 2265 |
+
"learning_rate": 3.147212812450819e-05,
|
| 2266 |
+
"loss": 0.5937,
|
| 2267 |
+
"step": 318
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 0.06355531204861284,
|
| 2271 |
+
"grad_norm": 0.21672700345516205,
|
| 2272 |
+
"learning_rate": 3.116857491533947e-05,
|
| 2273 |
+
"loss": 0.6095,
|
| 2274 |
+
"step": 319
|
| 2275 |
+
},
|
| 2276 |
+
{
|
| 2277 |
+
"epoch": 0.06375454500174328,
|
| 2278 |
+
"grad_norm": 0.2422553449869156,
|
| 2279 |
+
"learning_rate": 3.086582838174551e-05,
|
| 2280 |
+
"loss": 0.6426,
|
| 2281 |
+
"step": 320
|
| 2282 |
+
},
|
| 2283 |
+
{
|
| 2284 |
+
"epoch": 0.06395377795487374,
|
| 2285 |
+
"grad_norm": 0.2380169779062271,
|
| 2286 |
+
"learning_rate": 3.056390149238022e-05,
|
| 2287 |
+
"loss": 0.6648,
|
| 2288 |
+
"step": 321
|
| 2289 |
+
},
|
| 2290 |
+
{
|
| 2291 |
+
"epoch": 0.06415301090800418,
|
| 2292 |
+
"grad_norm": 0.224848210811615,
|
| 2293 |
+
"learning_rate": 3.0262807180786647e-05,
|
| 2294 |
+
"loss": 0.6811,
|
| 2295 |
+
"step": 322
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 0.06435224386113463,
|
| 2299 |
+
"grad_norm": 0.2206374555826187,
|
| 2300 |
+
"learning_rate": 2.996255834484296e-05,
|
| 2301 |
+
"loss": 0.6629,
|
| 2302 |
+
"step": 323
|
| 2303 |
+
},
|
| 2304 |
+
{
|
| 2305 |
+
"epoch": 0.06455147681426508,
|
| 2306 |
+
"grad_norm": 0.22571969032287598,
|
| 2307 |
+
"learning_rate": 2.9663167846209998e-05,
|
| 2308 |
+
"loss": 0.641,
|
| 2309 |
+
"step": 324
|
| 2310 |
+
},
|
| 2311 |
+
{
|
| 2312 |
+
"epoch": 0.06475070976739553,
|
| 2313 |
+
"grad_norm": 0.22498847544193268,
|
| 2314 |
+
"learning_rate": 2.936464850978027e-05,
|
| 2315 |
+
"loss": 0.642,
|
| 2316 |
+
"step": 325
|
| 2317 |
+
},
|
| 2318 |
+
{
|
| 2319 |
+
"epoch": 0.06494994272052597,
|
| 2320 |
+
"grad_norm": 0.26433423161506653,
|
| 2321 |
+
"learning_rate": 2.9067013123128613e-05,
|
| 2322 |
+
"loss": 0.7015,
|
| 2323 |
+
"step": 326
|
| 2324 |
+
},
|
| 2325 |
+
{
|
| 2326 |
+
"epoch": 0.06514917567365643,
|
| 2327 |
+
"grad_norm": 0.2343292236328125,
|
| 2328 |
+
"learning_rate": 2.8770274435964355e-05,
|
| 2329 |
+
"loss": 0.6088,
|
| 2330 |
+
"step": 327
|
| 2331 |
+
},
|
| 2332 |
+
{
|
| 2333 |
+
"epoch": 0.06534840862678687,
|
| 2334 |
+
"grad_norm": 0.2766936123371124,
|
| 2335 |
+
"learning_rate": 2.8474445159585235e-05,
|
| 2336 |
+
"loss": 0.6763,
|
| 2337 |
+
"step": 328
|
| 2338 |
+
},
|
| 2339 |
+
{
|
| 2340 |
+
"epoch": 0.06554764157991731,
|
| 2341 |
+
"grad_norm": 0.2363848239183426,
|
| 2342 |
+
"learning_rate": 2.8179537966332887e-05,
|
| 2343 |
+
"loss": 0.7461,
|
| 2344 |
+
"step": 329
|
| 2345 |
+
},
|
| 2346 |
+
{
|
| 2347 |
+
"epoch": 0.06574687453304777,
|
| 2348 |
+
"grad_norm": 0.2353663295507431,
|
| 2349 |
+
"learning_rate": 2.7885565489049946e-05,
|
| 2350 |
+
"loss": 0.7216,
|
| 2351 |
+
"step": 330
|
| 2352 |
+
},
|
| 2353 |
+
{
|
| 2354 |
+
"epoch": 0.06594610748617821,
|
| 2355 |
+
"grad_norm": 0.2601001262664795,
|
| 2356 |
+
"learning_rate": 2.759254032053888e-05,
|
| 2357 |
+
"loss": 0.7722,
|
| 2358 |
+
"step": 331
|
| 2359 |
+
},
|
| 2360 |
+
{
|
| 2361 |
+
"epoch": 0.06614534043930866,
|
| 2362 |
+
"grad_norm": 0.23950740694999695,
|
| 2363 |
+
"learning_rate": 2.7300475013022663e-05,
|
| 2364 |
+
"loss": 0.5964,
|
| 2365 |
+
"step": 332
|
| 2366 |
+
},
|
| 2367 |
+
{
|
| 2368 |
+
"epoch": 0.06634457339243911,
|
| 2369 |
+
"grad_norm": 0.23393948376178741,
|
| 2370 |
+
"learning_rate": 2.700938207760701e-05,
|
| 2371 |
+
"loss": 0.6738,
|
| 2372 |
+
"step": 333
|
| 2373 |
+
},
|
| 2374 |
+
{
|
| 2375 |
+
"epoch": 0.06654380634556956,
|
| 2376 |
+
"grad_norm": 0.2648164629936218,
|
| 2377 |
+
"learning_rate": 2.671927398374443e-05,
|
| 2378 |
+
"loss": 0.6643,
|
| 2379 |
+
"step": 334
|
| 2380 |
+
},
|
| 2381 |
+
{
|
| 2382 |
+
"epoch": 0.0667430392987,
|
| 2383 |
+
"grad_norm": 0.263930082321167,
|
| 2384 |
+
"learning_rate": 2.6430163158700115e-05,
|
| 2385 |
+
"loss": 0.6688,
|
| 2386 |
+
"step": 335
|
| 2387 |
+
},
|
| 2388 |
+
{
|
| 2389 |
+
"epoch": 0.06694227225183046,
|
| 2390 |
+
"grad_norm": 0.2163466513156891,
|
| 2391 |
+
"learning_rate": 2.6142061987019577e-05,
|
| 2392 |
+
"loss": 0.5839,
|
| 2393 |
+
"step": 336
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 0.0671415052049609,
|
| 2397 |
+
"grad_norm": 0.23970170319080353,
|
| 2398 |
+
"learning_rate": 2.5854982809998153e-05,
|
| 2399 |
+
"loss": 0.6826,
|
| 2400 |
+
"step": 337
|
| 2401 |
+
},
|
| 2402 |
+
{
|
| 2403 |
+
"epoch": 0.06734073815809134,
|
| 2404 |
+
"grad_norm": 0.26360487937927246,
|
| 2405 |
+
"learning_rate": 2.556893792515227e-05,
|
| 2406 |
+
"loss": 0.7177,
|
| 2407 |
+
"step": 338
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 0.0675399711112218,
|
| 2411 |
+
"grad_norm": 0.2591302692890167,
|
| 2412 |
+
"learning_rate": 2.5283939585692783e-05,
|
| 2413 |
+
"loss": 0.6781,
|
| 2414 |
+
"step": 339
|
| 2415 |
+
},
|
| 2416 |
+
{
|
| 2417 |
+
"epoch": 0.06773920406435224,
|
| 2418 |
+
"grad_norm": 0.26808708906173706,
|
| 2419 |
+
"learning_rate": 2.500000000000001e-05,
|
| 2420 |
+
"loss": 0.6558,
|
| 2421 |
+
"step": 340
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 0.06793843701748269,
|
| 2425 |
+
"grad_norm": 0.2927316725254059,
|
| 2426 |
+
"learning_rate": 2.471713133110078e-05,
|
| 2427 |
+
"loss": 0.715,
|
| 2428 |
+
"step": 341
|
| 2429 |
+
},
|
| 2430 |
+
{
|
| 2431 |
+
"epoch": 0.06813766997061314,
|
| 2432 |
+
"grad_norm": 0.25598007440567017,
|
| 2433 |
+
"learning_rate": 2.4435345696147403e-05,
|
| 2434 |
+
"loss": 0.6857,
|
| 2435 |
+
"step": 342
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 0.06833690292374359,
|
| 2439 |
+
"grad_norm": 0.27012765407562256,
|
| 2440 |
+
"learning_rate": 2.4154655165898627e-05,
|
| 2441 |
+
"loss": 0.6704,
|
| 2442 |
+
"step": 343
|
| 2443 |
+
},
|
| 2444 |
+
{
|
| 2445 |
+
"epoch": 0.06853613587687403,
|
| 2446 |
+
"grad_norm": 0.27995193004608154,
|
| 2447 |
+
"learning_rate": 2.3875071764202563e-05,
|
| 2448 |
+
"loss": 0.6795,
|
| 2449 |
+
"step": 344
|
| 2450 |
+
},
|
| 2451 |
+
{
|
| 2452 |
+
"epoch": 0.06873536883000449,
|
| 2453 |
+
"grad_norm": 0.2844606935977936,
|
| 2454 |
+
"learning_rate": 2.3596607467481603e-05,
|
| 2455 |
+
"loss": 0.7176,
|
| 2456 |
+
"step": 345
|
| 2457 |
+
},
|
| 2458 |
+
{
|
| 2459 |
+
"epoch": 0.06893460178313493,
|
| 2460 |
+
"grad_norm": 0.3271996080875397,
|
| 2461 |
+
"learning_rate": 2.3319274204219428e-05,
|
| 2462 |
+
"loss": 0.7862,
|
| 2463 |
+
"step": 346
|
| 2464 |
+
},
|
| 2465 |
+
{
|
| 2466 |
+
"epoch": 0.06913383473626537,
|
| 2467 |
+
"grad_norm": 0.24104076623916626,
|
| 2468 |
+
"learning_rate": 2.3043083854449988e-05,
|
| 2469 |
+
"loss": 0.6574,
|
| 2470 |
+
"step": 347
|
| 2471 |
+
},
|
| 2472 |
+
{
|
| 2473 |
+
"epoch": 0.06933306768939583,
|
| 2474 |
+
"grad_norm": 0.27659985423088074,
|
| 2475 |
+
"learning_rate": 2.2768048249248648e-05,
|
| 2476 |
+
"loss": 0.7289,
|
| 2477 |
+
"step": 348
|
| 2478 |
+
},
|
| 2479 |
+
{
|
| 2480 |
+
"epoch": 0.06953230064252627,
|
| 2481 |
+
"grad_norm": 0.29960310459136963,
|
| 2482 |
+
"learning_rate": 2.2494179170225333e-05,
|
| 2483 |
+
"loss": 0.6818,
|
| 2484 |
+
"step": 349
|
| 2485 |
+
},
|
| 2486 |
+
{
|
| 2487 |
+
"epoch": 0.06973153359565672,
|
| 2488 |
+
"grad_norm": 0.3256206810474396,
|
| 2489 |
+
"learning_rate": 2.2221488349019903e-05,
|
| 2490 |
+
"loss": 0.7484,
|
| 2491 |
+
"step": 350
|
| 2492 |
+
},
|
| 2493 |
+
{
|
| 2494 |
+
"epoch": 0.06993076654878717,
|
| 2495 |
+
"grad_norm": 0.21653087437152863,
|
| 2496 |
+
"learning_rate": 2.194998746679952e-05,
|
| 2497 |
+
"loss": 0.6332,
|
| 2498 |
+
"step": 351
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 0.07012999950191762,
|
| 2502 |
+
"grad_norm": 0.24491187930107117,
|
| 2503 |
+
"learning_rate": 2.167968815375837e-05,
|
| 2504 |
+
"loss": 0.6758,
|
| 2505 |
+
"step": 352
|
| 2506 |
+
},
|
| 2507 |
+
{
|
| 2508 |
+
"epoch": 0.07032923245504806,
|
| 2509 |
+
"grad_norm": 0.2615233361721039,
|
| 2510 |
+
"learning_rate": 2.1410601988619394e-05,
|
| 2511 |
+
"loss": 0.6832,
|
| 2512 |
+
"step": 353
|
| 2513 |
+
},
|
| 2514 |
+
{
|
| 2515 |
+
"epoch": 0.07052846540817852,
|
| 2516 |
+
"grad_norm": 0.3030376136302948,
|
| 2517 |
+
"learning_rate": 2.1142740498138324e-05,
|
| 2518 |
+
"loss": 0.7022,
|
| 2519 |
+
"step": 354
|
| 2520 |
+
},
|
| 2521 |
+
{
|
| 2522 |
+
"epoch": 0.07072769836130896,
|
| 2523 |
+
"grad_norm": 0.24537107348442078,
|
| 2524 |
+
"learning_rate": 2.08761151566099e-05,
|
| 2525 |
+
"loss": 0.6833,
|
| 2526 |
+
"step": 355
|
| 2527 |
+
},
|
| 2528 |
+
{
|
| 2529 |
+
"epoch": 0.0709269313144394,
|
| 2530 |
+
"grad_norm": 0.22852526605129242,
|
| 2531 |
+
"learning_rate": 2.061073738537635e-05,
|
| 2532 |
+
"loss": 0.6045,
|
| 2533 |
+
"step": 356
|
| 2534 |
+
},
|
| 2535 |
+
{
|
| 2536 |
+
"epoch": 0.07112616426756986,
|
| 2537 |
+
"grad_norm": 0.2126518040895462,
|
| 2538 |
+
"learning_rate": 2.034661855233815e-05,
|
| 2539 |
+
"loss": 0.6393,
|
| 2540 |
+
"step": 357
|
| 2541 |
+
},
|
| 2542 |
+
{
|
| 2543 |
+
"epoch": 0.0713253972207003,
|
| 2544 |
+
"grad_norm": 0.2170926332473755,
|
| 2545 |
+
"learning_rate": 2.008376997146705e-05,
|
| 2546 |
+
"loss": 0.636,
|
| 2547 |
+
"step": 358
|
| 2548 |
+
},
|
| 2549 |
+
{
|
| 2550 |
+
"epoch": 0.07152463017383075,
|
| 2551 |
+
"grad_norm": 0.24603933095932007,
|
| 2552 |
+
"learning_rate": 1.982220290232143e-05,
|
| 2553 |
+
"loss": 0.6854,
|
| 2554 |
+
"step": 359
|
| 2555 |
+
},
|
| 2556 |
+
{
|
| 2557 |
+
"epoch": 0.0717238631269612,
|
| 2558 |
+
"grad_norm": 0.25937581062316895,
|
| 2559 |
+
"learning_rate": 1.9561928549563968e-05,
|
| 2560 |
+
"loss": 0.6989,
|
| 2561 |
+
"step": 360
|
| 2562 |
+
},
|
| 2563 |
+
{
|
| 2564 |
+
"epoch": 0.07192309608009165,
|
| 2565 |
+
"grad_norm": 0.2045435905456543,
|
| 2566 |
+
"learning_rate": 1.9302958062481673e-05,
|
| 2567 |
+
"loss": 0.5754,
|
| 2568 |
+
"step": 361
|
| 2569 |
+
},
|
| 2570 |
+
{
|
| 2571 |
+
"epoch": 0.07212232903322209,
|
| 2572 |
+
"grad_norm": 0.24368344247341156,
|
| 2573 |
+
"learning_rate": 1.9045302534508297e-05,
|
| 2574 |
+
"loss": 0.6537,
|
| 2575 |
+
"step": 362
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"epoch": 0.07232156198635255,
|
| 2579 |
+
"grad_norm": 0.2470693588256836,
|
| 2580 |
+
"learning_rate": 1.8788973002749112e-05,
|
| 2581 |
+
"loss": 0.675,
|
| 2582 |
+
"step": 363
|
| 2583 |
+
},
|
| 2584 |
+
{
|
| 2585 |
+
"epoch": 0.07252079493948299,
|
| 2586 |
+
"grad_norm": 0.23468363285064697,
|
| 2587 |
+
"learning_rate": 1.8533980447508137e-05,
|
| 2588 |
+
"loss": 0.675,
|
| 2589 |
+
"step": 364
|
| 2590 |
+
},
|
| 2591 |
+
{
|
| 2592 |
+
"epoch": 0.07272002789261343,
|
| 2593 |
+
"grad_norm": 0.23341692984104156,
|
| 2594 |
+
"learning_rate": 1.8280335791817733e-05,
|
| 2595 |
+
"loss": 0.6899,
|
| 2596 |
+
"step": 365
|
| 2597 |
+
},
|
| 2598 |
+
{
|
| 2599 |
+
"epoch": 0.07291926084574389,
|
| 2600 |
+
"grad_norm": 0.24296964704990387,
|
| 2601 |
+
"learning_rate": 1.8028049900970767e-05,
|
| 2602 |
+
"loss": 0.6898,
|
| 2603 |
+
"step": 366
|
| 2604 |
+
},
|
| 2605 |
+
{
|
| 2606 |
+
"epoch": 0.07311849379887433,
|
| 2607 |
+
"grad_norm": 0.21605587005615234,
|
| 2608 |
+
"learning_rate": 1.777713358205514e-05,
|
| 2609 |
+
"loss": 0.615,
|
| 2610 |
+
"step": 367
|
| 2611 |
+
},
|
| 2612 |
+
{
|
| 2613 |
+
"epoch": 0.07331772675200478,
|
| 2614 |
+
"grad_norm": 0.23035632073879242,
|
| 2615 |
+
"learning_rate": 1.7527597583490822e-05,
|
| 2616 |
+
"loss": 0.7135,
|
| 2617 |
+
"step": 368
|
| 2618 |
+
},
|
| 2619 |
+
{
|
| 2620 |
+
"epoch": 0.07351695970513523,
|
| 2621 |
+
"grad_norm": 0.24987879395484924,
|
| 2622 |
+
"learning_rate": 1.7279452594569483e-05,
|
| 2623 |
+
"loss": 0.6646,
|
| 2624 |
+
"step": 369
|
| 2625 |
+
},
|
| 2626 |
+
{
|
| 2627 |
+
"epoch": 0.07371619265826568,
|
| 2628 |
+
"grad_norm": 0.23275192081928253,
|
| 2629 |
+
"learning_rate": 1.703270924499656e-05,
|
| 2630 |
+
"loss": 0.6405,
|
| 2631 |
+
"step": 370
|
| 2632 |
+
},
|
| 2633 |
+
{
|
| 2634 |
+
"epoch": 0.07391542561139612,
|
| 2635 |
+
"grad_norm": 0.2441992163658142,
|
| 2636 |
+
"learning_rate": 1.678737810443593e-05,
|
| 2637 |
+
"loss": 0.7028,
|
| 2638 |
+
"step": 371
|
| 2639 |
+
},
|
| 2640 |
+
{
|
| 2641 |
+
"epoch": 0.07411465856452658,
|
| 2642 |
+
"grad_norm": 0.21933166682720184,
|
| 2643 |
+
"learning_rate": 1.6543469682057106e-05,
|
| 2644 |
+
"loss": 0.5545,
|
| 2645 |
+
"step": 372
|
| 2646 |
+
},
|
| 2647 |
+
{
|
| 2648 |
+
"epoch": 0.07431389151765702,
|
| 2649 |
+
"grad_norm": 0.24288050830364227,
|
| 2650 |
+
"learning_rate": 1.6300994426085103e-05,
|
| 2651 |
+
"loss": 0.6971,
|
| 2652 |
+
"step": 373
|
| 2653 |
+
},
|
| 2654 |
+
{
|
| 2655 |
+
"epoch": 0.07451312447078746,
|
| 2656 |
+
"grad_norm": 0.2461194097995758,
|
| 2657 |
+
"learning_rate": 1.605996272335291e-05,
|
| 2658 |
+
"loss": 0.6359,
|
| 2659 |
+
"step": 374
|
| 2660 |
+
},
|
| 2661 |
+
{
|
| 2662 |
+
"epoch": 0.07471235742391792,
|
| 2663 |
+
"grad_norm": 0.25669926404953003,
|
| 2664 |
+
"learning_rate": 1.5820384898856434e-05,
|
| 2665 |
+
"loss": 0.6917,
|
| 2666 |
+
"step": 375
|
| 2667 |
+
},
|
| 2668 |
+
{
|
| 2669 |
+
"epoch": 0.07491159037704836,
|
| 2670 |
+
"grad_norm": 0.23437441885471344,
|
| 2671 |
+
"learning_rate": 1.5582271215312294e-05,
|
| 2672 |
+
"loss": 0.6353,
|
| 2673 |
+
"step": 376
|
| 2674 |
+
},
|
| 2675 |
+
{
|
| 2676 |
+
"epoch": 0.07511082333017881,
|
| 2677 |
+
"grad_norm": 0.2659025192260742,
|
| 2678 |
+
"learning_rate": 1.5345631872718214e-05,
|
| 2679 |
+
"loss": 0.7663,
|
| 2680 |
+
"step": 377
|
| 2681 |
+
},
|
| 2682 |
+
{
|
| 2683 |
+
"epoch": 0.07531005628330927,
|
| 2684 |
+
"grad_norm": 0.23349608480930328,
|
| 2685 |
+
"learning_rate": 1.5110477007916001e-05,
|
| 2686 |
+
"loss": 0.6564,
|
| 2687 |
+
"step": 378
|
| 2688 |
+
},
|
| 2689 |
+
{
|
| 2690 |
+
"epoch": 0.07550928923643971,
|
| 2691 |
+
"grad_norm": 0.24196575582027435,
|
| 2692 |
+
"learning_rate": 1.4876816694157419e-05,
|
| 2693 |
+
"loss": 0.6607,
|
| 2694 |
+
"step": 379
|
| 2695 |
+
},
|
| 2696 |
+
{
|
| 2697 |
+
"epoch": 0.07570852218957015,
|
| 2698 |
+
"grad_norm": 0.23814047873020172,
|
| 2699 |
+
"learning_rate": 1.4644660940672627e-05,
|
| 2700 |
+
"loss": 0.6354,
|
| 2701 |
+
"step": 380
|
| 2702 |
+
},
|
| 2703 |
+
{
|
| 2704 |
+
"epoch": 0.07590775514270061,
|
| 2705 |
+
"grad_norm": 0.23850150406360626,
|
| 2706 |
+
"learning_rate": 1.4414019692241437e-05,
|
| 2707 |
+
"loss": 0.6643,
|
| 2708 |
+
"step": 381
|
| 2709 |
+
},
|
| 2710 |
+
{
|
| 2711 |
+
"epoch": 0.07610698809583105,
|
| 2712 |
+
"grad_norm": 0.22468359768390656,
|
| 2713 |
+
"learning_rate": 1.4184902828767287e-05,
|
| 2714 |
+
"loss": 0.6246,
|
| 2715 |
+
"step": 382
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 0.0763062210489615,
|
| 2719 |
+
"grad_norm": 0.23000527918338776,
|
| 2720 |
+
"learning_rate": 1.3957320164854059e-05,
|
| 2721 |
+
"loss": 0.6283,
|
| 2722 |
+
"step": 383
|
| 2723 |
+
},
|
| 2724 |
+
{
|
| 2725 |
+
"epoch": 0.07650545400209195,
|
| 2726 |
+
"grad_norm": 0.2583581507205963,
|
| 2727 |
+
"learning_rate": 1.373128144938563e-05,
|
| 2728 |
+
"loss": 0.6927,
|
| 2729 |
+
"step": 384
|
| 2730 |
+
},
|
| 2731 |
+
{
|
| 2732 |
+
"epoch": 0.0767046869552224,
|
| 2733 |
+
"grad_norm": 0.2393674999475479,
|
| 2734 |
+
"learning_rate": 1.3506796365108232e-05,
|
| 2735 |
+
"loss": 0.668,
|
| 2736 |
+
"step": 385
|
| 2737 |
+
},
|
| 2738 |
+
{
|
| 2739 |
+
"epoch": 0.07690391990835284,
|
| 2740 |
+
"grad_norm": 0.2238411009311676,
|
| 2741 |
+
"learning_rate": 1.3283874528215733e-05,
|
| 2742 |
+
"loss": 0.6291,
|
| 2743 |
+
"step": 386
|
| 2744 |
+
},
|
| 2745 |
+
{
|
| 2746 |
+
"epoch": 0.0771031528614833,
|
| 2747 |
+
"grad_norm": 0.22692270576953888,
|
| 2748 |
+
"learning_rate": 1.3062525487937699e-05,
|
| 2749 |
+
"loss": 0.6488,
|
| 2750 |
+
"step": 387
|
| 2751 |
+
},
|
| 2752 |
+
{
|
| 2753 |
+
"epoch": 0.07730238581461374,
|
| 2754 |
+
"grad_norm": 0.29837989807128906,
|
| 2755 |
+
"learning_rate": 1.2842758726130283e-05,
|
| 2756 |
+
"loss": 0.7459,
|
| 2757 |
+
"step": 388
|
| 2758 |
+
},
|
| 2759 |
+
{
|
| 2760 |
+
"epoch": 0.07750161876774418,
|
| 2761 |
+
"grad_norm": 0.26298198103904724,
|
| 2762 |
+
"learning_rate": 1.2624583656870154e-05,
|
| 2763 |
+
"loss": 0.6305,
|
| 2764 |
+
"step": 389
|
| 2765 |
+
},
|
| 2766 |
+
{
|
| 2767 |
+
"epoch": 0.07770085172087464,
|
| 2768 |
+
"grad_norm": 0.2501063644886017,
|
| 2769 |
+
"learning_rate": 1.2408009626051137e-05,
|
| 2770 |
+
"loss": 0.6475,
|
| 2771 |
+
"step": 390
|
| 2772 |
+
},
|
| 2773 |
+
{
|
| 2774 |
+
"epoch": 0.07790008467400508,
|
| 2775 |
+
"grad_norm": 0.26658689975738525,
|
| 2776 |
+
"learning_rate": 1.2193045910983863e-05,
|
| 2777 |
+
"loss": 0.6945,
|
| 2778 |
+
"step": 391
|
| 2779 |
+
},
|
| 2780 |
+
{
|
| 2781 |
+
"epoch": 0.07809931762713553,
|
| 2782 |
+
"grad_norm": 0.27355584502220154,
|
| 2783 |
+
"learning_rate": 1.1979701719998453e-05,
|
| 2784 |
+
"loss": 0.7023,
|
| 2785 |
+
"step": 392
|
| 2786 |
+
},
|
| 2787 |
+
{
|
| 2788 |
+
"epoch": 0.07829855058026598,
|
| 2789 |
+
"grad_norm": 0.24372565746307373,
|
| 2790 |
+
"learning_rate": 1.1767986192049984e-05,
|
| 2791 |
+
"loss": 0.6686,
|
| 2792 |
+
"step": 393
|
| 2793 |
+
},
|
| 2794 |
+
{
|
| 2795 |
+
"epoch": 0.07849778353339643,
|
| 2796 |
+
"grad_norm": 0.28914278745651245,
|
| 2797 |
+
"learning_rate": 1.1557908396327028e-05,
|
| 2798 |
+
"loss": 0.7444,
|
| 2799 |
+
"step": 394
|
| 2800 |
+
},
|
| 2801 |
+
{
|
| 2802 |
+
"epoch": 0.07869701648652687,
|
| 2803 |
+
"grad_norm": 0.26138660311698914,
|
| 2804 |
+
"learning_rate": 1.134947733186315e-05,
|
| 2805 |
+
"loss": 0.6624,
|
| 2806 |
+
"step": 395
|
| 2807 |
+
},
|
| 2808 |
+
{
|
| 2809 |
+
"epoch": 0.07889624943965733,
|
| 2810 |
+
"grad_norm": 0.25570860505104065,
|
| 2811 |
+
"learning_rate": 1.1142701927151456e-05,
|
| 2812 |
+
"loss": 0.6746,
|
| 2813 |
+
"step": 396
|
| 2814 |
+
},
|
| 2815 |
+
{
|
| 2816 |
+
"epoch": 0.07909548239278777,
|
| 2817 |
+
"grad_norm": 0.2897097170352936,
|
| 2818 |
+
"learning_rate": 1.0937591039762085e-05,
|
| 2819 |
+
"loss": 0.7515,
|
| 2820 |
+
"step": 397
|
| 2821 |
+
},
|
| 2822 |
+
{
|
| 2823 |
+
"epoch": 0.07929471534591821,
|
| 2824 |
+
"grad_norm": 0.2693999111652374,
|
| 2825 |
+
"learning_rate": 1.0734153455962765e-05,
|
| 2826 |
+
"loss": 0.7385,
|
| 2827 |
+
"step": 398
|
| 2828 |
+
},
|
| 2829 |
+
{
|
| 2830 |
+
"epoch": 0.07949394829904867,
|
| 2831 |
+
"grad_norm": 0.2895209491252899,
|
| 2832 |
+
"learning_rate": 1.0532397890342505e-05,
|
| 2833 |
+
"loss": 0.7491,
|
| 2834 |
+
"step": 399
|
| 2835 |
+
},
|
| 2836 |
+
{
|
| 2837 |
+
"epoch": 0.07969318125217911,
|
| 2838 |
+
"grad_norm": 0.2964557707309723,
|
| 2839 |
+
"learning_rate": 1.0332332985438248e-05,
|
| 2840 |
+
"loss": 0.7342,
|
| 2841 |
+
"step": 400
|
| 2842 |
+
},
|
| 2843 |
+
{
|
| 2844 |
+
"epoch": 0.07969318125217911,
|
| 2845 |
+
"eval_loss": 0.6768385171890259,
|
| 2846 |
+
"eval_runtime": 265.4226,
|
| 2847 |
+
"eval_samples_per_second": 31.851,
|
| 2848 |
+
"eval_steps_per_second": 7.965,
|
| 2849 |
+
"step": 400
|
| 2850 |
}
|
| 2851 |
],
|
| 2852 |
"logging_steps": 1,
|
|
|
|
| 2875 |
"attributes": {}
|
| 2876 |
}
|
| 2877 |
},
|
| 2878 |
+
"total_flos": 1.3302341221692211e+17,
|
| 2879 |
"train_batch_size": 8,
|
| 2880 |
"trial_name": null,
|
| 2881 |
"trial_params": null
|