Training in progress, step 432, 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 671149168
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0b3baf7f7a16380fb028f475ca35fb4a1485b5449798007811d88895bd00a770
|
| 3 |
size 671149168
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 341314644
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5e22ca7071dd076252b43b473fa1704a21a4785031a90f0b9028e8c83e8ba89a
|
| 3 |
size 341314644
|
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 sha256:6d78c488bb947774f6f4ca54910a40ea5bafc41a037da391b9eb2e717aae29e0
|
| 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 sha256:25fb31c3275f1b8913e16332b6b15556ec3b58d9bf65928efef64fedc142b95b
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 0.3306281268596649,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-300",
|
| 4 |
-
"epoch":
|
| 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 +2131,930 @@
|
|
| 2131 |
"eval_samples_per_second": 23.374,
|
| 2132 |
"eval_steps_per_second": 5.851,
|
| 2133 |
"step": 300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2134 |
}
|
| 2135 |
],
|
| 2136 |
"logging_steps": 1,
|
|
@@ -2154,12 +3078,12 @@
|
|
| 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": 0.3306281268596649,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-300",
|
| 4 |
+
"epoch": 1.0011580775911986,
|
| 5 |
"eval_steps": 150,
|
| 6 |
+
"global_step": 432,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2131 |
"eval_samples_per_second": 23.374,
|
| 2132 |
"eval_steps_per_second": 5.851,
|
| 2133 |
"step": 300
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 0.6971627099015634,
|
| 2137 |
+
"grad_norm": 2.477572441101074,
|
| 2138 |
+
"learning_rate": 2.293880768178576e-05,
|
| 2139 |
+
"loss": 1.0048,
|
| 2140 |
+
"step": 301
|
| 2141 |
+
},
|
| 2142 |
+
{
|
| 2143 |
+
"epoch": 0.6994788650839606,
|
| 2144 |
+
"grad_norm": 2.3567538261413574,
|
| 2145 |
+
"learning_rate": 2.2619003103344606e-05,
|
| 2146 |
+
"loss": 1.1464,
|
| 2147 |
+
"step": 302
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 0.7017950202663579,
|
| 2151 |
+
"grad_norm": 2.297689199447632,
|
| 2152 |
+
"learning_rate": 2.2300790559367552e-05,
|
| 2153 |
+
"loss": 0.9767,
|
| 2154 |
+
"step": 303
|
| 2155 |
+
},
|
| 2156 |
+
{
|
| 2157 |
+
"epoch": 0.704111175448755,
|
| 2158 |
+
"grad_norm": 2.396148204803467,
|
| 2159 |
+
"learning_rate": 2.1984188551932512e-05,
|
| 2160 |
+
"loss": 0.9108,
|
| 2161 |
+
"step": 304
|
| 2162 |
+
},
|
| 2163 |
+
{
|
| 2164 |
+
"epoch": 0.7064273306311523,
|
| 2165 |
+
"grad_norm": 2.255086660385132,
|
| 2166 |
+
"learning_rate": 2.166921548947466e-05,
|
| 2167 |
+
"loss": 0.6876,
|
| 2168 |
+
"step": 305
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 0.7087434858135495,
|
| 2172 |
+
"grad_norm": 2.383619546890259,
|
| 2173 |
+
"learning_rate": 2.1355889685716224e-05,
|
| 2174 |
+
"loss": 0.7666,
|
| 2175 |
+
"step": 306
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 0.7110596409959468,
|
| 2179 |
+
"grad_norm": 2.329786777496338,
|
| 2180 |
+
"learning_rate": 2.1044229358601542e-05,
|
| 2181 |
+
"loss": 0.813,
|
| 2182 |
+
"step": 307
|
| 2183 |
+
},
|
| 2184 |
+
{
|
| 2185 |
+
"epoch": 0.7133757961783439,
|
| 2186 |
+
"grad_norm": 2.474313497543335,
|
| 2187 |
+
"learning_rate": 2.0734252629237894e-05,
|
| 2188 |
+
"loss": 0.6183,
|
| 2189 |
+
"step": 308
|
| 2190 |
+
},
|
| 2191 |
+
{
|
| 2192 |
+
"epoch": 0.7156919513607412,
|
| 2193 |
+
"grad_norm": 2.5291049480438232,
|
| 2194 |
+
"learning_rate": 2.0425977520841838e-05,
|
| 2195 |
+
"loss": 0.5384,
|
| 2196 |
+
"step": 309
|
| 2197 |
+
},
|
| 2198 |
+
{
|
| 2199 |
+
"epoch": 0.7180081065431384,
|
| 2200 |
+
"grad_norm": 1.997194766998291,
|
| 2201 |
+
"learning_rate": 2.011942195769122e-05,
|
| 2202 |
+
"loss": 0.5236,
|
| 2203 |
+
"step": 310
|
| 2204 |
+
},
|
| 2205 |
+
{
|
| 2206 |
+
"epoch": 0.7203242617255357,
|
| 2207 |
+
"grad_norm": 1.2435768842697144,
|
| 2208 |
+
"learning_rate": 1.9814603764083112e-05,
|
| 2209 |
+
"loss": 0.2937,
|
| 2210 |
+
"step": 311
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 0.7226404169079328,
|
| 2214 |
+
"grad_norm": 1.364814281463623,
|
| 2215 |
+
"learning_rate": 1.9511540663297285e-05,
|
| 2216 |
+
"loss": 0.3747,
|
| 2217 |
+
"step": 312
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 0.72495657209033,
|
| 2221 |
+
"grad_norm": 1.3511894941329956,
|
| 2222 |
+
"learning_rate": 1.921025027656587e-05,
|
| 2223 |
+
"loss": 0.4068,
|
| 2224 |
+
"step": 313
|
| 2225 |
+
},
|
| 2226 |
+
{
|
| 2227 |
+
"epoch": 0.7272727272727273,
|
| 2228 |
+
"grad_norm": 1.3470934629440308,
|
| 2229 |
+
"learning_rate": 1.8910750122048637e-05,
|
| 2230 |
+
"loss": 0.3442,
|
| 2231 |
+
"step": 314
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 0.7295888824551245,
|
| 2235 |
+
"grad_norm": 1.4333728551864624,
|
| 2236 |
+
"learning_rate": 1.8613057613814584e-05,
|
| 2237 |
+
"loss": 0.3411,
|
| 2238 |
+
"step": 315
|
| 2239 |
+
},
|
| 2240 |
+
{
|
| 2241 |
+
"epoch": 0.7319050376375217,
|
| 2242 |
+
"grad_norm": 1.348252296447754,
|
| 2243 |
+
"learning_rate": 1.831719006082924e-05,
|
| 2244 |
+
"loss": 0.3321,
|
| 2245 |
+
"step": 316
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"epoch": 0.7342211928199189,
|
| 2249 |
+
"grad_norm": 0.9950866103172302,
|
| 2250 |
+
"learning_rate": 1.8023164665948456e-05,
|
| 2251 |
+
"loss": 0.1945,
|
| 2252 |
+
"step": 317
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"epoch": 0.7365373480023162,
|
| 2256 |
+
"grad_norm": 1.1160790920257568,
|
| 2257 |
+
"learning_rate": 1.7730998524917957e-05,
|
| 2258 |
+
"loss": 0.263,
|
| 2259 |
+
"step": 318
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 0.7388535031847133,
|
| 2263 |
+
"grad_norm": 1.3591563701629639,
|
| 2264 |
+
"learning_rate": 1.7440708625379505e-05,
|
| 2265 |
+
"loss": 0.3155,
|
| 2266 |
+
"step": 319
|
| 2267 |
+
},
|
| 2268 |
+
{
|
| 2269 |
+
"epoch": 0.7411696583671106,
|
| 2270 |
+
"grad_norm": 1.3873182535171509,
|
| 2271 |
+
"learning_rate": 1.7152311845883095e-05,
|
| 2272 |
+
"loss": 0.2736,
|
| 2273 |
+
"step": 320
|
| 2274 |
+
},
|
| 2275 |
+
{
|
| 2276 |
+
"epoch": 0.7434858135495078,
|
| 2277 |
+
"grad_norm": 1.4288737773895264,
|
| 2278 |
+
"learning_rate": 1.686582495490554e-05,
|
| 2279 |
+
"loss": 0.3022,
|
| 2280 |
+
"step": 321
|
| 2281 |
+
},
|
| 2282 |
+
{
|
| 2283 |
+
"epoch": 0.7458019687319051,
|
| 2284 |
+
"grad_norm": 1.7382503747940063,
|
| 2285 |
+
"learning_rate": 1.658126460987558e-05,
|
| 2286 |
+
"loss": 0.2989,
|
| 2287 |
+
"step": 322
|
| 2288 |
+
},
|
| 2289 |
+
{
|
| 2290 |
+
"epoch": 0.7481181239143022,
|
| 2291 |
+
"grad_norm": 1.7087494134902954,
|
| 2292 |
+
"learning_rate": 1.6298647356205254e-05,
|
| 2293 |
+
"loss": 0.3102,
|
| 2294 |
+
"step": 323
|
| 2295 |
+
},
|
| 2296 |
+
{
|
| 2297 |
+
"epoch": 0.7504342790966995,
|
| 2298 |
+
"grad_norm": 1.5120712518692017,
|
| 2299 |
+
"learning_rate": 1.601798962632799e-05,
|
| 2300 |
+
"loss": 0.2155,
|
| 2301 |
+
"step": 324
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 0.7527504342790967,
|
| 2305 |
+
"grad_norm": 1.5661240816116333,
|
| 2306 |
+
"learning_rate": 1.5739307738743057e-05,
|
| 2307 |
+
"loss": 0.1865,
|
| 2308 |
+
"step": 325
|
| 2309 |
+
},
|
| 2310 |
+
{
|
| 2311 |
+
"epoch": 0.755066589461494,
|
| 2312 |
+
"grad_norm": 1.2175641059875488,
|
| 2313 |
+
"learning_rate": 1.546261789706686e-05,
|
| 2314 |
+
"loss": 0.1658,
|
| 2315 |
+
"step": 326
|
| 2316 |
+
},
|
| 2317 |
+
{
|
| 2318 |
+
"epoch": 0.7573827446438911,
|
| 2319 |
+
"grad_norm": 1.4670747518539429,
|
| 2320 |
+
"learning_rate": 1.5187936189090669e-05,
|
| 2321 |
+
"loss": 0.1604,
|
| 2322 |
+
"step": 327
|
| 2323 |
+
},
|
| 2324 |
+
{
|
| 2325 |
+
"epoch": 0.7596988998262884,
|
| 2326 |
+
"grad_norm": 1.8161323070526123,
|
| 2327 |
+
"learning_rate": 1.491527858584535e-05,
|
| 2328 |
+
"loss": 0.2243,
|
| 2329 |
+
"step": 328
|
| 2330 |
+
},
|
| 2331 |
+
{
|
| 2332 |
+
"epoch": 0.7620150550086856,
|
| 2333 |
+
"grad_norm": 1.3583159446716309,
|
| 2334 |
+
"learning_rate": 1.4644660940672627e-05,
|
| 2335 |
+
"loss": 0.1281,
|
| 2336 |
+
"step": 329
|
| 2337 |
+
},
|
| 2338 |
+
{
|
| 2339 |
+
"epoch": 0.7643312101910829,
|
| 2340 |
+
"grad_norm": 1.2247012853622437,
|
| 2341 |
+
"learning_rate": 1.4376098988303405e-05,
|
| 2342 |
+
"loss": 0.1084,
|
| 2343 |
+
"step": 330
|
| 2344 |
+
},
|
| 2345 |
+
{
|
| 2346 |
+
"epoch": 0.76664736537348,
|
| 2347 |
+
"grad_norm": 1.2522218227386475,
|
| 2348 |
+
"learning_rate": 1.4109608343942854e-05,
|
| 2349 |
+
"loss": 0.1137,
|
| 2350 |
+
"step": 331
|
| 2351 |
+
},
|
| 2352 |
+
{
|
| 2353 |
+
"epoch": 0.7689635205558772,
|
| 2354 |
+
"grad_norm": 1.0144122838974,
|
| 2355 |
+
"learning_rate": 1.384520450236244e-05,
|
| 2356 |
+
"loss": 0.0753,
|
| 2357 |
+
"step": 332
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 0.7712796757382745,
|
| 2361 |
+
"grad_norm": 1.0482442378997803,
|
| 2362 |
+
"learning_rate": 1.35829028369991e-05,
|
| 2363 |
+
"loss": 0.0872,
|
| 2364 |
+
"step": 333
|
| 2365 |
+
},
|
| 2366 |
+
{
|
| 2367 |
+
"epoch": 0.7735958309206716,
|
| 2368 |
+
"grad_norm": 1.6394468545913696,
|
| 2369 |
+
"learning_rate": 1.3322718599061251e-05,
|
| 2370 |
+
"loss": 0.1127,
|
| 2371 |
+
"step": 334
|
| 2372 |
+
},
|
| 2373 |
+
{
|
| 2374 |
+
"epoch": 0.7759119861030689,
|
| 2375 |
+
"grad_norm": 1.1458487510681152,
|
| 2376 |
+
"learning_rate": 1.306466691664216e-05,
|
| 2377 |
+
"loss": 0.1022,
|
| 2378 |
+
"step": 335
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 0.7782281412854661,
|
| 2382 |
+
"grad_norm": 1.6174955368041992,
|
| 2383 |
+
"learning_rate": 1.2808762793840201e-05,
|
| 2384 |
+
"loss": 0.1394,
|
| 2385 |
+
"step": 336
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 0.7805442964678634,
|
| 2389 |
+
"grad_norm": 1.2268593311309814,
|
| 2390 |
+
"learning_rate": 1.2555021109886589e-05,
|
| 2391 |
+
"loss": 0.1056,
|
| 2392 |
+
"step": 337
|
| 2393 |
+
},
|
| 2394 |
+
{
|
| 2395 |
+
"epoch": 0.7828604516502605,
|
| 2396 |
+
"grad_norm": 1.2561674118041992,
|
| 2397 |
+
"learning_rate": 1.2303456618280141e-05,
|
| 2398 |
+
"loss": 0.0783,
|
| 2399 |
+
"step": 338
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 0.7851766068326578,
|
| 2403 |
+
"grad_norm": 1.1064704656600952,
|
| 2404 |
+
"learning_rate": 1.2054083945929535e-05,
|
| 2405 |
+
"loss": 0.0762,
|
| 2406 |
+
"step": 339
|
| 2407 |
+
},
|
| 2408 |
+
{
|
| 2409 |
+
"epoch": 0.787492762015055,
|
| 2410 |
+
"grad_norm": 1.174625277519226,
|
| 2411 |
+
"learning_rate": 1.1806917592302762e-05,
|
| 2412 |
+
"loss": 0.0862,
|
| 2413 |
+
"step": 340
|
| 2414 |
+
},
|
| 2415 |
+
{
|
| 2416 |
+
"epoch": 0.7898089171974523,
|
| 2417 |
+
"grad_norm": 1.6088016033172607,
|
| 2418 |
+
"learning_rate": 1.1561971928584159e-05,
|
| 2419 |
+
"loss": 0.0962,
|
| 2420 |
+
"step": 341
|
| 2421 |
+
},
|
| 2422 |
+
{
|
| 2423 |
+
"epoch": 0.7921250723798494,
|
| 2424 |
+
"grad_norm": 1.1425096988677979,
|
| 2425 |
+
"learning_rate": 1.1319261196838782e-05,
|
| 2426 |
+
"loss": 0.0699,
|
| 2427 |
+
"step": 342
|
| 2428 |
+
},
|
| 2429 |
+
{
|
| 2430 |
+
"epoch": 0.7944412275622467,
|
| 2431 |
+
"grad_norm": 1.6147550344467163,
|
| 2432 |
+
"learning_rate": 1.1078799509184246e-05,
|
| 2433 |
+
"loss": 0.0919,
|
| 2434 |
+
"step": 343
|
| 2435 |
+
},
|
| 2436 |
+
{
|
| 2437 |
+
"epoch": 0.7967573827446439,
|
| 2438 |
+
"grad_norm": 1.4701851606369019,
|
| 2439 |
+
"learning_rate": 1.0840600846970334e-05,
|
| 2440 |
+
"loss": 0.0891,
|
| 2441 |
+
"step": 344
|
| 2442 |
+
},
|
| 2443 |
+
{
|
| 2444 |
+
"epoch": 0.7990735379270412,
|
| 2445 |
+
"grad_norm": 1.4786357879638672,
|
| 2446 |
+
"learning_rate": 1.0604679059965922e-05,
|
| 2447 |
+
"loss": 0.0638,
|
| 2448 |
+
"step": 345
|
| 2449 |
+
},
|
| 2450 |
+
{
|
| 2451 |
+
"epoch": 0.8013896931094383,
|
| 2452 |
+
"grad_norm": 2.1240947246551514,
|
| 2453 |
+
"learning_rate": 1.0371047865553846e-05,
|
| 2454 |
+
"loss": 0.1207,
|
| 2455 |
+
"step": 346
|
| 2456 |
+
},
|
| 2457 |
+
{
|
| 2458 |
+
"epoch": 0.8037058482918356,
|
| 2459 |
+
"grad_norm": 1.4813666343688965,
|
| 2460 |
+
"learning_rate": 1.0139720847933166e-05,
|
| 2461 |
+
"loss": 0.0871,
|
| 2462 |
+
"step": 347
|
| 2463 |
+
},
|
| 2464 |
+
{
|
| 2465 |
+
"epoch": 0.8060220034742328,
|
| 2466 |
+
"grad_norm": 1.3989876508712769,
|
| 2467 |
+
"learning_rate": 9.91071145732948e-06,
|
| 2468 |
+
"loss": 0.0964,
|
| 2469 |
+
"step": 348
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"epoch": 0.80833815865663,
|
| 2473 |
+
"grad_norm": 1.3372840881347656,
|
| 2474 |
+
"learning_rate": 9.684033009212752e-06,
|
| 2475 |
+
"loss": 0.0714,
|
| 2476 |
+
"step": 349
|
| 2477 |
+
},
|
| 2478 |
+
{
|
| 2479 |
+
"epoch": 0.8106543138390272,
|
| 2480 |
+
"grad_norm": 2.03450870513916,
|
| 2481 |
+
"learning_rate": 9.459698683523204e-06,
|
| 2482 |
+
"loss": 0.1074,
|
| 2483 |
+
"step": 350
|
| 2484 |
+
},
|
| 2485 |
+
{
|
| 2486 |
+
"epoch": 0.8129704690214244,
|
| 2487 |
+
"grad_norm": 1.326187252998352,
|
| 2488 |
+
"learning_rate": 9.237721523904891e-06,
|
| 2489 |
+
"loss": 1.0425,
|
| 2490 |
+
"step": 351
|
| 2491 |
+
},
|
| 2492 |
+
{
|
| 2493 |
+
"epoch": 0.8152866242038217,
|
| 2494 |
+
"grad_norm": 1.3223427534103394,
|
| 2495 |
+
"learning_rate": 9.018114436947373e-06,
|
| 2496 |
+
"loss": 0.8088,
|
| 2497 |
+
"step": 352
|
| 2498 |
+
},
|
| 2499 |
+
{
|
| 2500 |
+
"epoch": 0.8176027793862188,
|
| 2501 |
+
"grad_norm": 1.4911216497421265,
|
| 2502 |
+
"learning_rate": 8.80089019143524e-06,
|
| 2503 |
+
"loss": 0.784,
|
| 2504 |
+
"step": 353
|
| 2505 |
+
},
|
| 2506 |
+
{
|
| 2507 |
+
"epoch": 0.8199189345686161,
|
| 2508 |
+
"grad_norm": 1.5768686532974243,
|
| 2509 |
+
"learning_rate": 8.586061417605668e-06,
|
| 2510 |
+
"loss": 0.7212,
|
| 2511 |
+
"step": 354
|
| 2512 |
+
},
|
| 2513 |
+
{
|
| 2514 |
+
"epoch": 0.8222350897510133,
|
| 2515 |
+
"grad_norm": 1.8095353841781616,
|
| 2516 |
+
"learning_rate": 8.373640606414096e-06,
|
| 2517 |
+
"loss": 0.5331,
|
| 2518 |
+
"step": 355
|
| 2519 |
+
},
|
| 2520 |
+
{
|
| 2521 |
+
"epoch": 0.8245512449334106,
|
| 2522 |
+
"grad_norm": 1.7539920806884766,
|
| 2523 |
+
"learning_rate": 8.163640108807896e-06,
|
| 2524 |
+
"loss": 0.5814,
|
| 2525 |
+
"step": 356
|
| 2526 |
+
},
|
| 2527 |
+
{
|
| 2528 |
+
"epoch": 0.8268674001158077,
|
| 2529 |
+
"grad_norm": 1.8924403190612793,
|
| 2530 |
+
"learning_rate": 7.956072135008336e-06,
|
| 2531 |
+
"loss": 0.6049,
|
| 2532 |
+
"step": 357
|
| 2533 |
+
},
|
| 2534 |
+
{
|
| 2535 |
+
"epoch": 0.829183555298205,
|
| 2536 |
+
"grad_norm": 1.558434009552002,
|
| 2537 |
+
"learning_rate": 7.750948753800507e-06,
|
| 2538 |
+
"loss": 0.3521,
|
| 2539 |
+
"step": 358
|
| 2540 |
+
},
|
| 2541 |
+
{
|
| 2542 |
+
"epoch": 0.8314997104806022,
|
| 2543 |
+
"grad_norm": 1.900602102279663,
|
| 2544 |
+
"learning_rate": 7.548281891831716e-06,
|
| 2545 |
+
"loss": 0.3716,
|
| 2546 |
+
"step": 359
|
| 2547 |
+
},
|
| 2548 |
+
{
|
| 2549 |
+
"epoch": 0.8338158656629994,
|
| 2550 |
+
"grad_norm": 2.0248496532440186,
|
| 2551 |
+
"learning_rate": 7.348083332917926e-06,
|
| 2552 |
+
"loss": 0.3998,
|
| 2553 |
+
"step": 360
|
| 2554 |
+
},
|
| 2555 |
+
{
|
| 2556 |
+
"epoch": 0.8361320208453966,
|
| 2557 |
+
"grad_norm": 1.3187469244003296,
|
| 2558 |
+
"learning_rate": 7.150364717358698e-06,
|
| 2559 |
+
"loss": 0.2192,
|
| 2560 |
+
"step": 361
|
| 2561 |
+
},
|
| 2562 |
+
{
|
| 2563 |
+
"epoch": 0.8384481760277939,
|
| 2564 |
+
"grad_norm": 1.5400038957595825,
|
| 2565 |
+
"learning_rate": 6.955137541260287e-06,
|
| 2566 |
+
"loss": 0.2471,
|
| 2567 |
+
"step": 362
|
| 2568 |
+
},
|
| 2569 |
+
{
|
| 2570 |
+
"epoch": 0.8407643312101911,
|
| 2571 |
+
"grad_norm": 1.604353666305542,
|
| 2572 |
+
"learning_rate": 6.7624131558672756e-06,
|
| 2573 |
+
"loss": 0.2206,
|
| 2574 |
+
"step": 363
|
| 2575 |
+
},
|
| 2576 |
+
{
|
| 2577 |
+
"epoch": 0.8430804863925883,
|
| 2578 |
+
"grad_norm": 1.945562481880188,
|
| 2579 |
+
"learning_rate": 6.572202766902569e-06,
|
| 2580 |
+
"loss": 0.2775,
|
| 2581 |
+
"step": 364
|
| 2582 |
+
},
|
| 2583 |
+
{
|
| 2584 |
+
"epoch": 0.8453966415749855,
|
| 2585 |
+
"grad_norm": 2.460178852081299,
|
| 2586 |
+
"learning_rate": 6.384517433915793e-06,
|
| 2587 |
+
"loss": 0.4032,
|
| 2588 |
+
"step": 365
|
| 2589 |
+
},
|
| 2590 |
+
{
|
| 2591 |
+
"epoch": 0.8477127967573828,
|
| 2592 |
+
"grad_norm": 1.6772760152816772,
|
| 2593 |
+
"learning_rate": 6.199368069640343e-06,
|
| 2594 |
+
"loss": 0.253,
|
| 2595 |
+
"step": 366
|
| 2596 |
+
},
|
| 2597 |
+
{
|
| 2598 |
+
"epoch": 0.85002895193978,
|
| 2599 |
+
"grad_norm": 1.7364951372146606,
|
| 2600 |
+
"learning_rate": 6.016765439358774e-06,
|
| 2601 |
+
"loss": 0.2342,
|
| 2602 |
+
"step": 367
|
| 2603 |
+
},
|
| 2604 |
+
{
|
| 2605 |
+
"epoch": 0.8523451071221771,
|
| 2606 |
+
"grad_norm": 1.791764259338379,
|
| 2607 |
+
"learning_rate": 5.83672016027697e-06,
|
| 2608 |
+
"loss": 0.1775,
|
| 2609 |
+
"step": 368
|
| 2610 |
+
},
|
| 2611 |
+
{
|
| 2612 |
+
"epoch": 0.8546612623045744,
|
| 2613 |
+
"grad_norm": 1.4430023431777954,
|
| 2614 |
+
"learning_rate": 5.659242700906719e-06,
|
| 2615 |
+
"loss": 0.1663,
|
| 2616 |
+
"step": 369
|
| 2617 |
+
},
|
| 2618 |
+
{
|
| 2619 |
+
"epoch": 0.8569774174869716,
|
| 2620 |
+
"grad_norm": 1.4145193099975586,
|
| 2621 |
+
"learning_rate": 5.484343380457125e-06,
|
| 2622 |
+
"loss": 0.1669,
|
| 2623 |
+
"step": 370
|
| 2624 |
+
},
|
| 2625 |
+
{
|
| 2626 |
+
"epoch": 0.8592935726693689,
|
| 2627 |
+
"grad_norm": 1.4928686618804932,
|
| 2628 |
+
"learning_rate": 5.312032368234526e-06,
|
| 2629 |
+
"loss": 0.19,
|
| 2630 |
+
"step": 371
|
| 2631 |
+
},
|
| 2632 |
+
{
|
| 2633 |
+
"epoch": 0.861609727851766,
|
| 2634 |
+
"grad_norm": 1.48283851146698,
|
| 2635 |
+
"learning_rate": 5.1423196830513e-06,
|
| 2636 |
+
"loss": 0.1979,
|
| 2637 |
+
"step": 372
|
| 2638 |
+
},
|
| 2639 |
+
{
|
| 2640 |
+
"epoch": 0.8639258830341633,
|
| 2641 |
+
"grad_norm": 1.4714758396148682,
|
| 2642 |
+
"learning_rate": 4.975215192643246e-06,
|
| 2643 |
+
"loss": 0.1778,
|
| 2644 |
+
"step": 373
|
| 2645 |
+
},
|
| 2646 |
+
{
|
| 2647 |
+
"epoch": 0.8662420382165605,
|
| 2648 |
+
"grad_norm": 1.187078595161438,
|
| 2649 |
+
"learning_rate": 4.81072861309591e-06,
|
| 2650 |
+
"loss": 0.1187,
|
| 2651 |
+
"step": 374
|
| 2652 |
+
},
|
| 2653 |
+
{
|
| 2654 |
+
"epoch": 0.8685581933989577,
|
| 2655 |
+
"grad_norm": 1.1508448123931885,
|
| 2656 |
+
"learning_rate": 4.648869508279613e-06,
|
| 2657 |
+
"loss": 0.1269,
|
| 2658 |
+
"step": 375
|
| 2659 |
+
},
|
| 2660 |
+
{
|
| 2661 |
+
"epoch": 0.8708743485813549,
|
| 2662 |
+
"grad_norm": 0.9376459717750549,
|
| 2663 |
+
"learning_rate": 4.489647289293369e-06,
|
| 2664 |
+
"loss": 0.1059,
|
| 2665 |
+
"step": 376
|
| 2666 |
+
},
|
| 2667 |
+
{
|
| 2668 |
+
"epoch": 0.8731905037637522,
|
| 2669 |
+
"grad_norm": 1.0897644758224487,
|
| 2670 |
+
"learning_rate": 4.333071213917722e-06,
|
| 2671 |
+
"loss": 0.0853,
|
| 2672 |
+
"step": 377
|
| 2673 |
+
},
|
| 2674 |
+
{
|
| 2675 |
+
"epoch": 0.8755066589461494,
|
| 2676 |
+
"grad_norm": 0.9136444926261902,
|
| 2677 |
+
"learning_rate": 4.179150386076424e-06,
|
| 2678 |
+
"loss": 0.0743,
|
| 2679 |
+
"step": 378
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 0.8778228141285466,
|
| 2683 |
+
"grad_norm": 0.8426498174667358,
|
| 2684 |
+
"learning_rate": 4.027893755307144e-06,
|
| 2685 |
+
"loss": 0.0606,
|
| 2686 |
+
"step": 379
|
| 2687 |
+
},
|
| 2688 |
+
{
|
| 2689 |
+
"epoch": 0.8801389693109438,
|
| 2690 |
+
"grad_norm": 0.7755717635154724,
|
| 2691 |
+
"learning_rate": 3.879310116241042e-06,
|
| 2692 |
+
"loss": 0.0617,
|
| 2693 |
+
"step": 380
|
| 2694 |
+
},
|
| 2695 |
+
{
|
| 2696 |
+
"epoch": 0.8824551244933411,
|
| 2697 |
+
"grad_norm": 1.0966241359710693,
|
| 2698 |
+
"learning_rate": 3.733408108091485e-06,
|
| 2699 |
+
"loss": 0.0724,
|
| 2700 |
+
"step": 381
|
| 2701 |
+
},
|
| 2702 |
+
{
|
| 2703 |
+
"epoch": 0.8847712796757383,
|
| 2704 |
+
"grad_norm": 0.8539220094680786,
|
| 2705 |
+
"learning_rate": 3.5901962141516977e-06,
|
| 2706 |
+
"loss": 0.0714,
|
| 2707 |
+
"step": 382
|
| 2708 |
+
},
|
| 2709 |
+
{
|
| 2710 |
+
"epoch": 0.8870874348581355,
|
| 2711 |
+
"grad_norm": 1.0539127588272095,
|
| 2712 |
+
"learning_rate": 3.4496827613015202e-06,
|
| 2713 |
+
"loss": 0.0585,
|
| 2714 |
+
"step": 383
|
| 2715 |
+
},
|
| 2716 |
+
{
|
| 2717 |
+
"epoch": 0.8894035900405327,
|
| 2718 |
+
"grad_norm": 0.7426398396492004,
|
| 2719 |
+
"learning_rate": 3.3118759195232275e-06,
|
| 2720 |
+
"loss": 0.0527,
|
| 2721 |
+
"step": 384
|
| 2722 |
+
},
|
| 2723 |
+
{
|
| 2724 |
+
"epoch": 0.89171974522293,
|
| 2725 |
+
"grad_norm": 0.9449732899665833,
|
| 2726 |
+
"learning_rate": 3.176783701426528e-06,
|
| 2727 |
+
"loss": 0.0689,
|
| 2728 |
+
"step": 385
|
| 2729 |
+
},
|
| 2730 |
+
{
|
| 2731 |
+
"epoch": 0.8940359004053272,
|
| 2732 |
+
"grad_norm": 0.9856773614883423,
|
| 2733 |
+
"learning_rate": 3.0444139617826607e-06,
|
| 2734 |
+
"loss": 0.0671,
|
| 2735 |
+
"step": 386
|
| 2736 |
+
},
|
| 2737 |
+
{
|
| 2738 |
+
"epoch": 0.8963520555877244,
|
| 2739 |
+
"grad_norm": 0.826977014541626,
|
| 2740 |
+
"learning_rate": 2.91477439706771e-06,
|
| 2741 |
+
"loss": 0.0554,
|
| 2742 |
+
"step": 387
|
| 2743 |
+
},
|
| 2744 |
+
{
|
| 2745 |
+
"epoch": 0.8986682107701216,
|
| 2746 |
+
"grad_norm": 0.7119414210319519,
|
| 2747 |
+
"learning_rate": 2.787872545015069e-06,
|
| 2748 |
+
"loss": 0.0525,
|
| 2749 |
+
"step": 388
|
| 2750 |
+
},
|
| 2751 |
+
{
|
| 2752 |
+
"epoch": 0.9009843659525189,
|
| 2753 |
+
"grad_norm": 1.1008604764938354,
|
| 2754 |
+
"learning_rate": 2.663715784177201e-06,
|
| 2755 |
+
"loss": 0.0819,
|
| 2756 |
+
"step": 389
|
| 2757 |
+
},
|
| 2758 |
+
{
|
| 2759 |
+
"epoch": 0.903300521134916,
|
| 2760 |
+
"grad_norm": 1.2398359775543213,
|
| 2761 |
+
"learning_rate": 2.542311333496622e-06,
|
| 2762 |
+
"loss": 0.0695,
|
| 2763 |
+
"step": 390
|
| 2764 |
+
},
|
| 2765 |
+
{
|
| 2766 |
+
"epoch": 0.9056166763173132,
|
| 2767 |
+
"grad_norm": 0.6894932985305786,
|
| 2768 |
+
"learning_rate": 2.423666251886114e-06,
|
| 2769 |
+
"loss": 0.0404,
|
| 2770 |
+
"step": 391
|
| 2771 |
+
},
|
| 2772 |
+
{
|
| 2773 |
+
"epoch": 0.9079328314997105,
|
| 2774 |
+
"grad_norm": 1.0042632818222046,
|
| 2775 |
+
"learning_rate": 2.307787437818365e-06,
|
| 2776 |
+
"loss": 0.063,
|
| 2777 |
+
"step": 392
|
| 2778 |
+
},
|
| 2779 |
+
{
|
| 2780 |
+
"epoch": 0.9102489866821077,
|
| 2781 |
+
"grad_norm": 1.2008248567581177,
|
| 2782 |
+
"learning_rate": 2.194681628924816e-06,
|
| 2783 |
+
"loss": 0.0826,
|
| 2784 |
+
"step": 393
|
| 2785 |
+
},
|
| 2786 |
+
{
|
| 2787 |
+
"epoch": 0.9125651418645049,
|
| 2788 |
+
"grad_norm": 1.0816729068756104,
|
| 2789 |
+
"learning_rate": 2.0843554016039326e-06,
|
| 2790 |
+
"loss": 0.0553,
|
| 2791 |
+
"step": 394
|
| 2792 |
+
},
|
| 2793 |
+
{
|
| 2794 |
+
"epoch": 0.9148812970469021,
|
| 2795 |
+
"grad_norm": 1.0502961874008179,
|
| 2796 |
+
"learning_rate": 1.976815170638802e-06,
|
| 2797 |
+
"loss": 0.0598,
|
| 2798 |
+
"step": 395
|
| 2799 |
+
},
|
| 2800 |
+
{
|
| 2801 |
+
"epoch": 0.9171974522292994,
|
| 2802 |
+
"grad_norm": 1.179869532585144,
|
| 2803 |
+
"learning_rate": 1.8720671888242059e-06,
|
| 2804 |
+
"loss": 0.1096,
|
| 2805 |
+
"step": 396
|
| 2806 |
+
},
|
| 2807 |
+
{
|
| 2808 |
+
"epoch": 0.9195136074116966,
|
| 2809 |
+
"grad_norm": 1.1720819473266602,
|
| 2810 |
+
"learning_rate": 1.7701175466029895e-06,
|
| 2811 |
+
"loss": 0.0732,
|
| 2812 |
+
"step": 397
|
| 2813 |
+
},
|
| 2814 |
+
{
|
| 2815 |
+
"epoch": 0.9218297625940938,
|
| 2816 |
+
"grad_norm": 1.741204857826233,
|
| 2817 |
+
"learning_rate": 1.6709721717120042e-06,
|
| 2818 |
+
"loss": 0.0805,
|
| 2819 |
+
"step": 398
|
| 2820 |
+
},
|
| 2821 |
+
{
|
| 2822 |
+
"epoch": 0.924145917776491,
|
| 2823 |
+
"grad_norm": 1.379157304763794,
|
| 2824 |
+
"learning_rate": 1.5746368288373947e-06,
|
| 2825 |
+
"loss": 0.0735,
|
| 2826 |
+
"step": 399
|
| 2827 |
+
},
|
| 2828 |
+
{
|
| 2829 |
+
"epoch": 0.9264620729588883,
|
| 2830 |
+
"grad_norm": 1.8489305973052979,
|
| 2831 |
+
"learning_rate": 1.4811171192794627e-06,
|
| 2832 |
+
"loss": 0.0918,
|
| 2833 |
+
"step": 400
|
| 2834 |
+
},
|
| 2835 |
+
{
|
| 2836 |
+
"epoch": 0.9287782281412854,
|
| 2837 |
+
"grad_norm": 1.2802451848983765,
|
| 2838 |
+
"learning_rate": 1.3904184806269704e-06,
|
| 2839 |
+
"loss": 0.9387,
|
| 2840 |
+
"step": 401
|
| 2841 |
+
},
|
| 2842 |
+
{
|
| 2843 |
+
"epoch": 0.9310943833236827,
|
| 2844 |
+
"grad_norm": 1.3554513454437256,
|
| 2845 |
+
"learning_rate": 1.3025461864409394e-06,
|
| 2846 |
+
"loss": 0.9213,
|
| 2847 |
+
"step": 402
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 0.9334105385060799,
|
| 2851 |
+
"grad_norm": 1.2956500053405762,
|
| 2852 |
+
"learning_rate": 1.2175053459481e-06,
|
| 2853 |
+
"loss": 0.6561,
|
| 2854 |
+
"step": 403
|
| 2855 |
+
},
|
| 2856 |
+
{
|
| 2857 |
+
"epoch": 0.9357266936884772,
|
| 2858 |
+
"grad_norm": 1.4841207265853882,
|
| 2859 |
+
"learning_rate": 1.1353009037437523e-06,
|
| 2860 |
+
"loss": 0.4452,
|
| 2861 |
+
"step": 404
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 0.9380428488708743,
|
| 2865 |
+
"grad_norm": 1.3164417743682861,
|
| 2866 |
+
"learning_rate": 1.0559376395043285e-06,
|
| 2867 |
+
"loss": 0.4071,
|
| 2868 |
+
"step": 405
|
| 2869 |
+
},
|
| 2870 |
+
{
|
| 2871 |
+
"epoch": 0.9403590040532716,
|
| 2872 |
+
"grad_norm": 1.8103063106536865,
|
| 2873 |
+
"learning_rate": 9.794201677094162e-07,
|
| 2874 |
+
"loss": 0.3799,
|
| 2875 |
+
"step": 406
|
| 2876 |
+
},
|
| 2877 |
+
{
|
| 2878 |
+
"epoch": 0.9426751592356688,
|
| 2879 |
+
"grad_norm": 1.640889286994934,
|
| 2880 |
+
"learning_rate": 9.05752937373533e-07,
|
| 2881 |
+
"loss": 0.3303,
|
| 2882 |
+
"step": 407
|
| 2883 |
+
},
|
| 2884 |
+
{
|
| 2885 |
+
"epoch": 0.944991314418066,
|
| 2886 |
+
"grad_norm": 1.410726547241211,
|
| 2887 |
+
"learning_rate": 8.349402317873789e-07,
|
| 2888 |
+
"loss": 0.2026,
|
| 2889 |
+
"step": 408
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 0.9473074696004632,
|
| 2893 |
+
"grad_norm": 1.7915242910385132,
|
| 2894 |
+
"learning_rate": 7.669861682688239e-07,
|
| 2895 |
+
"loss": 0.2115,
|
| 2896 |
+
"step": 409
|
| 2897 |
+
},
|
| 2898 |
+
{
|
| 2899 |
+
"epoch": 0.9496236247828604,
|
| 2900 |
+
"grad_norm": 1.855295181274414,
|
| 2901 |
+
"learning_rate": 7.018946979234997e-07,
|
| 2902 |
+
"loss": 0.2876,
|
| 2903 |
+
"step": 410
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 0.9519397799652577,
|
| 2907 |
+
"grad_norm": 1.1143068075180054,
|
| 2908 |
+
"learning_rate": 6.396696054150719e-07,
|
| 2909 |
+
"loss": 0.1409,
|
| 2910 |
+
"step": 411
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 0.9542559351476549,
|
| 2914 |
+
"grad_norm": 1.3461558818817139,
|
| 2915 |
+
"learning_rate": 5.803145087451945e-07,
|
| 2916 |
+
"loss": 0.1312,
|
| 2917 |
+
"step": 412
|
| 2918 |
+
},
|
| 2919 |
+
{
|
| 2920 |
+
"epoch": 0.9565720903300521,
|
| 2921 |
+
"grad_norm": 1.343173623085022,
|
| 2922 |
+
"learning_rate": 5.238328590431162e-07,
|
| 2923 |
+
"loss": 0.1521,
|
| 2924 |
+
"step": 413
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 0.9588882455124493,
|
| 2928 |
+
"grad_norm": 1.174078106880188,
|
| 2929 |
+
"learning_rate": 4.7022794036505335e-07,
|
| 2930 |
+
"loss": 0.1421,
|
| 2931 |
+
"step": 414
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 0.9612044006948466,
|
| 2935 |
+
"grad_norm": 1.0609129667282104,
|
| 2936 |
+
"learning_rate": 4.1950286950321327e-07,
|
| 2937 |
+
"loss": 0.0897,
|
| 2938 |
+
"step": 415
|
| 2939 |
+
},
|
| 2940 |
+
{
|
| 2941 |
+
"epoch": 0.9635205558772437,
|
| 2942 |
+
"grad_norm": 1.1571158170700073,
|
| 2943 |
+
"learning_rate": 3.716605958046071e-07,
|
| 2944 |
+
"loss": 0.0685,
|
| 2945 |
+
"step": 416
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 0.965836711059641,
|
| 2949 |
+
"grad_norm": 0.8536246418952942,
|
| 2950 |
+
"learning_rate": 3.267039009995199e-07,
|
| 2951 |
+
"loss": 0.0694,
|
| 2952 |
+
"step": 417
|
| 2953 |
+
},
|
| 2954 |
+
{
|
| 2955 |
+
"epoch": 0.9681528662420382,
|
| 2956 |
+
"grad_norm": 0.7544981837272644,
|
| 2957 |
+
"learning_rate": 2.846353990398065e-07,
|
| 2958 |
+
"loss": 0.0442,
|
| 2959 |
+
"step": 418
|
| 2960 |
+
},
|
| 2961 |
+
{
|
| 2962 |
+
"epoch": 0.9704690214244355,
|
| 2963 |
+
"grad_norm": 0.7445719242095947,
|
| 2964 |
+
"learning_rate": 2.4545753594688583e-07,
|
| 2965 |
+
"loss": 0.0545,
|
| 2966 |
+
"step": 419
|
| 2967 |
+
},
|
| 2968 |
+
{
|
| 2969 |
+
"epoch": 0.9727851766068326,
|
| 2970 |
+
"grad_norm": 0.631585955619812,
|
| 2971 |
+
"learning_rate": 2.0917258966953733e-07,
|
| 2972 |
+
"loss": 0.0462,
|
| 2973 |
+
"step": 420
|
| 2974 |
+
},
|
| 2975 |
+
{
|
| 2976 |
+
"epoch": 0.9751013317892299,
|
| 2977 |
+
"grad_norm": 0.6365635991096497,
|
| 2978 |
+
"learning_rate": 1.7578266995142978e-07,
|
| 2979 |
+
"loss": 0.0554,
|
| 2980 |
+
"step": 421
|
| 2981 |
+
},
|
| 2982 |
+
{
|
| 2983 |
+
"epoch": 0.9774174869716271,
|
| 2984 |
+
"grad_norm": 0.7627168297767639,
|
| 2985 |
+
"learning_rate": 1.4528971820846893e-07,
|
| 2986 |
+
"loss": 0.0552,
|
| 2987 |
+
"step": 422
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 0.9797336421540244,
|
| 2991 |
+
"grad_norm": 0.8680822849273682,
|
| 2992 |
+
"learning_rate": 1.1769550741592139e-07,
|
| 2993 |
+
"loss": 0.0643,
|
| 2994 |
+
"step": 423
|
| 2995 |
+
},
|
| 2996 |
+
{
|
| 2997 |
+
"epoch": 0.9820497973364215,
|
| 2998 |
+
"grad_norm": 1.076411485671997,
|
| 2999 |
+
"learning_rate": 9.300164200530814e-08,
|
| 3000 |
+
"loss": 0.0644,
|
| 3001 |
+
"step": 424
|
| 3002 |
+
},
|
| 3003 |
+
{
|
| 3004 |
+
"epoch": 0.9843659525188188,
|
| 3005 |
+
"grad_norm": 1.0168262720108032,
|
| 3006 |
+
"learning_rate": 7.120955777112914e-08,
|
| 3007 |
+
"loss": 0.0862,
|
| 3008 |
+
"step": 425
|
| 3009 |
+
},
|
| 3010 |
+
{
|
| 3011 |
+
"epoch": 0.986682107701216,
|
| 3012 |
+
"grad_norm": 1.153483271598816,
|
| 3013 |
+
"learning_rate": 5.2320521787385667e-08,
|
| 3014 |
+
"loss": 0.0703,
|
| 3015 |
+
"step": 426
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 0.9889982628836133,
|
| 3019 |
+
"grad_norm": 0.9161254167556763,
|
| 3020 |
+
"learning_rate": 3.633563233388926e-08,
|
| 3021 |
+
"loss": 0.0481,
|
| 3022 |
+
"step": 427
|
| 3023 |
+
},
|
| 3024 |
+
{
|
| 3025 |
+
"epoch": 0.9913144180660104,
|
| 3026 |
+
"grad_norm": 0.8356172442436218,
|
| 3027 |
+
"learning_rate": 2.3255818832423894e-08,
|
| 3028 |
+
"loss": 0.0574,
|
| 3029 |
+
"step": 428
|
| 3030 |
+
},
|
| 3031 |
+
{
|
| 3032 |
+
"epoch": 0.9936305732484076,
|
| 3033 |
+
"grad_norm": 1.117587924003601,
|
| 3034 |
+
"learning_rate": 1.3081841792694783e-08,
|
| 3035 |
+
"loss": 0.0647,
|
| 3036 |
+
"step": 429
|
| 3037 |
+
},
|
| 3038 |
+
{
|
| 3039 |
+
"epoch": 0.9959467284308049,
|
| 3040 |
+
"grad_norm": 0.9720324873924255,
|
| 3041 |
+
"learning_rate": 5.814292768108187e-09,
|
| 3042 |
+
"loss": 0.0588,
|
| 3043 |
+
"step": 430
|
| 3044 |
+
},
|
| 3045 |
+
{
|
| 3046 |
+
"epoch": 0.998262883613202,
|
| 3047 |
+
"grad_norm": 1.3996336460113525,
|
| 3048 |
+
"learning_rate": 1.453594321393359e-09,
|
| 3049 |
+
"loss": 0.0588,
|
| 3050 |
+
"step": 431
|
| 3051 |
+
},
|
| 3052 |
+
{
|
| 3053 |
+
"epoch": 1.0011580775911986,
|
| 3054 |
+
"grad_norm": 2.588228702545166,
|
| 3055 |
+
"learning_rate": 0.0,
|
| 3056 |
+
"loss": 0.4998,
|
| 3057 |
+
"step": 432
|
| 3058 |
}
|
| 3059 |
],
|
| 3060 |
"logging_steps": 1,
|
|
|
|
| 3078 |
"should_evaluate": false,
|
| 3079 |
"should_log": false,
|
| 3080 |
"should_save": true,
|
| 3081 |
+
"should_training_stop": true
|
| 3082 |
},
|
| 3083 |
"attributes": {}
|
| 3084 |
}
|
| 3085 |
},
|
| 3086 |
+
"total_flos": 3.292330885612831e+17,
|
| 3087 |
"train_batch_size": 8,
|
| 3088 |
"trial_name": null,
|
| 3089 |
"trial_params": null
|