Upload checkpoint 400
Browse files- README.md +4 -4
- adapter_config.json +1 -1
- loss.png +2 -2
- optimizer.pt +1 -1
- scheduler.pt +1 -1
- trainer_state.json +703 -3
README.md
CHANGED
|
@@ -2,7 +2,7 @@
|
|
| 2 |
base_model: Qwen/Qwen2.5-7B-Instruct
|
| 3 |
library_name: peft
|
| 4 |
---
|
| 5 |
-
# Gradience T1 7B (Step
|
| 6 |
|
| 7 |
> [!NOTE]
|
| 8 |
> Training in progress...
|
|
@@ -38,11 +38,11 @@ library_name: peft
|
|
| 38 |
</head>
|
| 39 |
<body>
|
| 40 |
<div style="width: 100%; background-color: #e0e0e0; border-radius: 25px; overflow: hidden; margin: 20px 0;">
|
| 41 |
-
<div style="height: 30px; width:
|
| 42 |
-
<!--
|
| 43 |
</div>
|
| 44 |
</div>
|
| 45 |
-
<p style="font-family: Arial, sans-serif; font-size: 16px;">Progress:
|
| 46 |
</body>
|
| 47 |
</html>
|
| 48 |
|
|
|
|
| 2 |
base_model: Qwen/Qwen2.5-7B-Instruct
|
| 3 |
library_name: peft
|
| 4 |
---
|
| 5 |
+
# Gradience T1 7B (Step 400 Checkpoint)
|
| 6 |
|
| 7 |
> [!NOTE]
|
| 8 |
> Training in progress...
|
|
|
|
| 38 |
</head>
|
| 39 |
<body>
|
| 40 |
<div style="width: 100%; background-color: #e0e0e0; border-radius: 25px; overflow: hidden; margin: 20px 0;">
|
| 41 |
+
<div style="height: 30px; width: 8.13%; background-color: #76c7c0; text-align: center; line-height: 30px; color: white; border-radius: 25px 0 0 25px;">
|
| 42 |
+
<!-- 8.13% -->
|
| 43 |
</div>
|
| 44 |
</div>
|
| 45 |
+
<p style="font-family: Arial, sans-serif; font-size: 16px;">Progress: 400 out of 4918 steps</p>
|
| 46 |
</body>
|
| 47 |
</html>
|
| 48 |
|
adapter_config.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
"alpha_pattern": {},
|
| 3 |
"auto_mapping": null,
|
| 4 |
-
"base_model_name_or_path": "
|
| 5 |
"bias": "none",
|
| 6 |
"eva_config": null,
|
| 7 |
"exclude_modules": null,
|
|
|
|
| 1 |
{
|
| 2 |
"alpha_pattern": {},
|
| 3 |
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "Qwen/Qwen2.5-7B-Instruct",
|
| 5 |
"bias": "none",
|
| 6 |
"eva_config": null,
|
| 7 |
"exclude_modules": null,
|
loss.png
CHANGED
|
Git LFS Details
|
|
Git LFS Details
|
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 82461044
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eefac9e8208f033b11163601df9a990821a006cdad2f5b671af5e786c9b60212
|
| 3 |
size 82461044
|
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 sha256:880e20dc665fe0d79037c2008c79e9fa46462cda4281c7356b7e603bf120e9f2
|
| 3 |
size 1064
|
trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -2108,6 +2108,706 @@
|
|
| 2108 |
"learning_rate": 0.00018803175249338492,
|
| 2109 |
"loss": 1.0508,
|
| 2110 |
"step": 300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2111 |
}
|
| 2112 |
],
|
| 2113 |
"logging_steps": 1,
|
|
@@ -2127,7 +2827,7 @@
|
|
| 2127 |
"attributes": {}
|
| 2128 |
}
|
| 2129 |
},
|
| 2130 |
-
"total_flos":
|
| 2131 |
"train_batch_size": 16,
|
| 2132 |
"trial_name": null,
|
| 2133 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.16263468184590363,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 400,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 2108 |
"learning_rate": 0.00018803175249338492,
|
| 2109 |
"loss": 1.0508,
|
| 2110 |
"step": 300
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 0.12238259808904249,
|
| 2114 |
+
"grad_norm": 0.09302002936601639,
|
| 2115 |
+
"learning_rate": 0.00018799104416853248,
|
| 2116 |
+
"loss": 1.0512,
|
| 2117 |
+
"step": 301
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 0.12278918479365725,
|
| 2121 |
+
"grad_norm": 0.09081271290779114,
|
| 2122 |
+
"learning_rate": 0.00018795033584368006,
|
| 2123 |
+
"loss": 0.9688,
|
| 2124 |
+
"step": 302
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 0.123195771498272,
|
| 2128 |
+
"grad_norm": 0.1059931218624115,
|
| 2129 |
+
"learning_rate": 0.0001879096275188276,
|
| 2130 |
+
"loss": 1.0483,
|
| 2131 |
+
"step": 303
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 0.12360235820288677,
|
| 2135 |
+
"grad_norm": 0.1018669605255127,
|
| 2136 |
+
"learning_rate": 0.0001878689191939752,
|
| 2137 |
+
"loss": 1.019,
|
| 2138 |
+
"step": 304
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 0.12400894490750153,
|
| 2142 |
+
"grad_norm": 0.1040007546544075,
|
| 2143 |
+
"learning_rate": 0.00018782821086912275,
|
| 2144 |
+
"loss": 1.037,
|
| 2145 |
+
"step": 305
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 0.12441553161211628,
|
| 2149 |
+
"grad_norm": 0.10204601287841797,
|
| 2150 |
+
"learning_rate": 0.0001877875025442703,
|
| 2151 |
+
"loss": 0.9816,
|
| 2152 |
+
"step": 306
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 0.12482211831673104,
|
| 2156 |
+
"grad_norm": 0.10591764748096466,
|
| 2157 |
+
"learning_rate": 0.00018774679421941788,
|
| 2158 |
+
"loss": 1.0939,
|
| 2159 |
+
"step": 307
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 0.1252287050213458,
|
| 2163 |
+
"grad_norm": 0.09306305646896362,
|
| 2164 |
+
"learning_rate": 0.00018770608589456544,
|
| 2165 |
+
"loss": 1.0476,
|
| 2166 |
+
"step": 308
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 0.12563529172596055,
|
| 2170 |
+
"grad_norm": 11.22681713104248,
|
| 2171 |
+
"learning_rate": 0.00018766537756971302,
|
| 2172 |
+
"loss": 1.0573,
|
| 2173 |
+
"step": 309
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 0.12604187843057532,
|
| 2177 |
+
"grad_norm": 0.09422402083873749,
|
| 2178 |
+
"learning_rate": 0.00018762466924486057,
|
| 2179 |
+
"loss": 0.9993,
|
| 2180 |
+
"step": 310
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 0.1264484651351901,
|
| 2184 |
+
"grad_norm": 0.0982229933142662,
|
| 2185 |
+
"learning_rate": 0.00018758396092000815,
|
| 2186 |
+
"loss": 0.9159,
|
| 2187 |
+
"step": 311
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 0.12685505183980483,
|
| 2191 |
+
"grad_norm": 0.12579265236854553,
|
| 2192 |
+
"learning_rate": 0.00018754325259515573,
|
| 2193 |
+
"loss": 1.0935,
|
| 2194 |
+
"step": 312
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 0.1272616385444196,
|
| 2198 |
+
"grad_norm": 0.10069390386343002,
|
| 2199 |
+
"learning_rate": 0.0001875025442703033,
|
| 2200 |
+
"loss": 1.0127,
|
| 2201 |
+
"step": 313
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 0.12766822524903434,
|
| 2205 |
+
"grad_norm": 0.10948827862739563,
|
| 2206 |
+
"learning_rate": 0.00018746183594545087,
|
| 2207 |
+
"loss": 1.0576,
|
| 2208 |
+
"step": 314
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 0.12807481195364911,
|
| 2212 |
+
"grad_norm": 0.09232445061206818,
|
| 2213 |
+
"learning_rate": 0.00018742112762059842,
|
| 2214 |
+
"loss": 0.9856,
|
| 2215 |
+
"step": 315
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 0.12848139865826388,
|
| 2219 |
+
"grad_norm": 0.08319563418626785,
|
| 2220 |
+
"learning_rate": 0.000187380419295746,
|
| 2221 |
+
"loss": 0.9172,
|
| 2222 |
+
"step": 316
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 0.12888798536287863,
|
| 2226 |
+
"grad_norm": 0.09697309136390686,
|
| 2227 |
+
"learning_rate": 0.00018733971097089356,
|
| 2228 |
+
"loss": 1.0567,
|
| 2229 |
+
"step": 317
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 0.1292945720674934,
|
| 2233 |
+
"grad_norm": 0.09254255145788193,
|
| 2234 |
+
"learning_rate": 0.0001872990026460411,
|
| 2235 |
+
"loss": 1.0177,
|
| 2236 |
+
"step": 318
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 0.12970115877210814,
|
| 2240 |
+
"grad_norm": 0.09254108369350433,
|
| 2241 |
+
"learning_rate": 0.0001872582943211887,
|
| 2242 |
+
"loss": 1.0079,
|
| 2243 |
+
"step": 319
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 0.1301077454767229,
|
| 2247 |
+
"grad_norm": 0.09095866233110428,
|
| 2248 |
+
"learning_rate": 0.00018721758599633625,
|
| 2249 |
+
"loss": 1.0633,
|
| 2250 |
+
"step": 320
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 0.13051433218133768,
|
| 2254 |
+
"grad_norm": 0.09073010087013245,
|
| 2255 |
+
"learning_rate": 0.00018717687767148383,
|
| 2256 |
+
"loss": 0.9059,
|
| 2257 |
+
"step": 321
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 0.13092091888595242,
|
| 2261 |
+
"grad_norm": 0.09842764586210251,
|
| 2262 |
+
"learning_rate": 0.00018713616934663138,
|
| 2263 |
+
"loss": 1.0766,
|
| 2264 |
+
"step": 322
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 0.1313275055905672,
|
| 2268 |
+
"grad_norm": 0.09325529634952545,
|
| 2269 |
+
"learning_rate": 0.00018709546102177896,
|
| 2270 |
+
"loss": 1.066,
|
| 2271 |
+
"step": 323
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 0.13173409229518196,
|
| 2275 |
+
"grad_norm": 0.09692969918251038,
|
| 2276 |
+
"learning_rate": 0.00018705475269692654,
|
| 2277 |
+
"loss": 0.9743,
|
| 2278 |
+
"step": 324
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 0.1321406789997967,
|
| 2282 |
+
"grad_norm": 0.09432708472013474,
|
| 2283 |
+
"learning_rate": 0.0001870140443720741,
|
| 2284 |
+
"loss": 1.0141,
|
| 2285 |
+
"step": 325
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 0.13254726570441147,
|
| 2289 |
+
"grad_norm": 0.09226994961500168,
|
| 2290 |
+
"learning_rate": 0.00018697333604722168,
|
| 2291 |
+
"loss": 0.9837,
|
| 2292 |
+
"step": 326
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 0.1329538524090262,
|
| 2296 |
+
"grad_norm": 0.10843974351882935,
|
| 2297 |
+
"learning_rate": 0.00018693262772236923,
|
| 2298 |
+
"loss": 1.0248,
|
| 2299 |
+
"step": 327
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 0.13336043911364098,
|
| 2303 |
+
"grad_norm": 0.09324774891138077,
|
| 2304 |
+
"learning_rate": 0.00018689191939751681,
|
| 2305 |
+
"loss": 1.0642,
|
| 2306 |
+
"step": 328
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 0.13376702581825575,
|
| 2310 |
+
"grad_norm": 0.08934729546308517,
|
| 2311 |
+
"learning_rate": 0.00018685121107266437,
|
| 2312 |
+
"loss": 0.9792,
|
| 2313 |
+
"step": 329
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 0.1341736125228705,
|
| 2317 |
+
"grad_norm": 0.09125274419784546,
|
| 2318 |
+
"learning_rate": 0.00018681050274781192,
|
| 2319 |
+
"loss": 1.0093,
|
| 2320 |
+
"step": 330
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 0.13458019922748526,
|
| 2324 |
+
"grad_norm": 0.09645108133554459,
|
| 2325 |
+
"learning_rate": 0.0001867697944229595,
|
| 2326 |
+
"loss": 0.9503,
|
| 2327 |
+
"step": 331
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 0.13498678593210003,
|
| 2331 |
+
"grad_norm": 0.09900861978530884,
|
| 2332 |
+
"learning_rate": 0.00018672908609810706,
|
| 2333 |
+
"loss": 0.9966,
|
| 2334 |
+
"step": 332
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 0.13539337263671478,
|
| 2338 |
+
"grad_norm": 0.09018311649560928,
|
| 2339 |
+
"learning_rate": 0.00018668837777325464,
|
| 2340 |
+
"loss": 0.965,
|
| 2341 |
+
"step": 333
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 0.13579995934132955,
|
| 2345 |
+
"grad_norm": 0.10296136885881424,
|
| 2346 |
+
"learning_rate": 0.00018664766944840222,
|
| 2347 |
+
"loss": 1.1011,
|
| 2348 |
+
"step": 334
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 0.1362065460459443,
|
| 2352 |
+
"grad_norm": 0.09104129672050476,
|
| 2353 |
+
"learning_rate": 0.00018660696112354977,
|
| 2354 |
+
"loss": 0.9814,
|
| 2355 |
+
"step": 335
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 0.13661313275055906,
|
| 2359 |
+
"grad_norm": 0.09881450235843658,
|
| 2360 |
+
"learning_rate": 0.00018656625279869736,
|
| 2361 |
+
"loss": 1.0989,
|
| 2362 |
+
"step": 336
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 0.13701971945517383,
|
| 2366 |
+
"grad_norm": 0.09691241383552551,
|
| 2367 |
+
"learning_rate": 0.0001865255444738449,
|
| 2368 |
+
"loss": 1.0967,
|
| 2369 |
+
"step": 337
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 0.13742630615978857,
|
| 2373 |
+
"grad_norm": 0.10152243077754974,
|
| 2374 |
+
"learning_rate": 0.0001864848361489925,
|
| 2375 |
+
"loss": 1.0951,
|
| 2376 |
+
"step": 338
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 0.13783289286440334,
|
| 2380 |
+
"grad_norm": 0.10802541673183441,
|
| 2381 |
+
"learning_rate": 0.00018644412782414005,
|
| 2382 |
+
"loss": 0.8742,
|
| 2383 |
+
"step": 339
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 0.13823947956901808,
|
| 2387 |
+
"grad_norm": 0.09942565858364105,
|
| 2388 |
+
"learning_rate": 0.0001864034194992876,
|
| 2389 |
+
"loss": 0.9961,
|
| 2390 |
+
"step": 340
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 0.13864606627363285,
|
| 2394 |
+
"grad_norm": 0.08618199825286865,
|
| 2395 |
+
"learning_rate": 0.00018636271117443518,
|
| 2396 |
+
"loss": 0.9645,
|
| 2397 |
+
"step": 341
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 0.13905265297824762,
|
| 2401 |
+
"grad_norm": 0.1056099608540535,
|
| 2402 |
+
"learning_rate": 0.00018632200284958273,
|
| 2403 |
+
"loss": 0.9885,
|
| 2404 |
+
"step": 342
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 0.13945923968286236,
|
| 2408 |
+
"grad_norm": 0.08862382173538208,
|
| 2409 |
+
"learning_rate": 0.00018628129452473032,
|
| 2410 |
+
"loss": 0.9316,
|
| 2411 |
+
"step": 343
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 0.13986582638747713,
|
| 2415 |
+
"grad_norm": 0.09923135489225388,
|
| 2416 |
+
"learning_rate": 0.00018624058619987787,
|
| 2417 |
+
"loss": 0.9959,
|
| 2418 |
+
"step": 344
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 0.1402724130920919,
|
| 2422 |
+
"grad_norm": 0.09120538830757141,
|
| 2423 |
+
"learning_rate": 0.00018619987787502545,
|
| 2424 |
+
"loss": 0.968,
|
| 2425 |
+
"step": 345
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 0.14067899979670664,
|
| 2429 |
+
"grad_norm": 0.09669141471385956,
|
| 2430 |
+
"learning_rate": 0.00018615916955017303,
|
| 2431 |
+
"loss": 1.085,
|
| 2432 |
+
"step": 346
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 0.1410855865013214,
|
| 2436 |
+
"grad_norm": 0.08598754554986954,
|
| 2437 |
+
"learning_rate": 0.00018611846122532059,
|
| 2438 |
+
"loss": 0.9504,
|
| 2439 |
+
"step": 347
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 0.14149217320593616,
|
| 2443 |
+
"grad_norm": 0.09238371253013611,
|
| 2444 |
+
"learning_rate": 0.00018607775290046817,
|
| 2445 |
+
"loss": 0.9742,
|
| 2446 |
+
"step": 348
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 0.14189875991055093,
|
| 2450 |
+
"grad_norm": 0.091258205473423,
|
| 2451 |
+
"learning_rate": 0.00018603704457561572,
|
| 2452 |
+
"loss": 0.9341,
|
| 2453 |
+
"step": 349
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 0.1423053466151657,
|
| 2457 |
+
"grad_norm": 0.10129548609256744,
|
| 2458 |
+
"learning_rate": 0.0001859963362507633,
|
| 2459 |
+
"loss": 1.0814,
|
| 2460 |
+
"step": 350
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 0.14271193331978044,
|
| 2464 |
+
"grad_norm": 0.09523019194602966,
|
| 2465 |
+
"learning_rate": 0.00018595562792591086,
|
| 2466 |
+
"loss": 0.9848,
|
| 2467 |
+
"step": 351
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 0.1431185200243952,
|
| 2471 |
+
"grad_norm": 0.09485248476266861,
|
| 2472 |
+
"learning_rate": 0.0001859149196010584,
|
| 2473 |
+
"loss": 0.9828,
|
| 2474 |
+
"step": 352
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 0.14352510672900995,
|
| 2478 |
+
"grad_norm": 0.09963666647672653,
|
| 2479 |
+
"learning_rate": 0.000185874211276206,
|
| 2480 |
+
"loss": 1.1075,
|
| 2481 |
+
"step": 353
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 0.14393169343362472,
|
| 2485 |
+
"grad_norm": 0.09067155420780182,
|
| 2486 |
+
"learning_rate": 0.00018583350295135355,
|
| 2487 |
+
"loss": 0.971,
|
| 2488 |
+
"step": 354
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"epoch": 0.1443382801382395,
|
| 2492 |
+
"grad_norm": 0.09153544157743454,
|
| 2493 |
+
"learning_rate": 0.00018579279462650113,
|
| 2494 |
+
"loss": 0.9405,
|
| 2495 |
+
"step": 355
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 0.14474486684285423,
|
| 2499 |
+
"grad_norm": 0.1024472787976265,
|
| 2500 |
+
"learning_rate": 0.00018575208630164868,
|
| 2501 |
+
"loss": 0.9967,
|
| 2502 |
+
"step": 356
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 0.145151453547469,
|
| 2506 |
+
"grad_norm": 0.09804495424032211,
|
| 2507 |
+
"learning_rate": 0.00018571137797679626,
|
| 2508 |
+
"loss": 0.9578,
|
| 2509 |
+
"step": 357
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 0.14555804025208377,
|
| 2513 |
+
"grad_norm": 0.099054716527462,
|
| 2514 |
+
"learning_rate": 0.00018567066965194384,
|
| 2515 |
+
"loss": 0.9999,
|
| 2516 |
+
"step": 358
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 0.1459646269566985,
|
| 2520 |
+
"grad_norm": 0.09781336784362793,
|
| 2521 |
+
"learning_rate": 0.0001856299613270914,
|
| 2522 |
+
"loss": 1.09,
|
| 2523 |
+
"step": 359
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 0.14637121366131328,
|
| 2527 |
+
"grad_norm": 0.08993211388587952,
|
| 2528 |
+
"learning_rate": 0.00018558925300223898,
|
| 2529 |
+
"loss": 1.0719,
|
| 2530 |
+
"step": 360
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"epoch": 0.14677780036592802,
|
| 2534 |
+
"grad_norm": 0.09146003425121307,
|
| 2535 |
+
"learning_rate": 0.00018554854467738653,
|
| 2536 |
+
"loss": 1.0008,
|
| 2537 |
+
"step": 361
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 0.1471843870705428,
|
| 2541 |
+
"grad_norm": 0.09643495827913284,
|
| 2542 |
+
"learning_rate": 0.00018550783635253411,
|
| 2543 |
+
"loss": 1.0791,
|
| 2544 |
+
"step": 362
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 0.14759097377515756,
|
| 2548 |
+
"grad_norm": 0.09078676998615265,
|
| 2549 |
+
"learning_rate": 0.00018546712802768167,
|
| 2550 |
+
"loss": 0.8641,
|
| 2551 |
+
"step": 363
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 0.1479975604797723,
|
| 2555 |
+
"grad_norm": 0.08719085901975632,
|
| 2556 |
+
"learning_rate": 0.00018542641970282922,
|
| 2557 |
+
"loss": 0.985,
|
| 2558 |
+
"step": 364
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 0.14840414718438708,
|
| 2562 |
+
"grad_norm": 0.09189736843109131,
|
| 2563 |
+
"learning_rate": 0.0001853857113779768,
|
| 2564 |
+
"loss": 0.9638,
|
| 2565 |
+
"step": 365
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 0.14881073388900182,
|
| 2569 |
+
"grad_norm": 0.09381456673145294,
|
| 2570 |
+
"learning_rate": 0.00018534500305312436,
|
| 2571 |
+
"loss": 1.0036,
|
| 2572 |
+
"step": 366
|
| 2573 |
+
},
|
| 2574 |
+
{
|
| 2575 |
+
"epoch": 0.1492173205936166,
|
| 2576 |
+
"grad_norm": 0.0922684445977211,
|
| 2577 |
+
"learning_rate": 0.00018530429472827194,
|
| 2578 |
+
"loss": 1.0391,
|
| 2579 |
+
"step": 367
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 0.14962390729823136,
|
| 2583 |
+
"grad_norm": 0.09465248882770538,
|
| 2584 |
+
"learning_rate": 0.0001852635864034195,
|
| 2585 |
+
"loss": 0.8874,
|
| 2586 |
+
"step": 368
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 0.1500304940028461,
|
| 2590 |
+
"grad_norm": 0.0938408225774765,
|
| 2591 |
+
"learning_rate": 0.00018522287807856707,
|
| 2592 |
+
"loss": 1.0269,
|
| 2593 |
+
"step": 369
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 0.15043708070746087,
|
| 2597 |
+
"grad_norm": 0.09377933293581009,
|
| 2598 |
+
"learning_rate": 0.00018518216975371466,
|
| 2599 |
+
"loss": 1.0142,
|
| 2600 |
+
"step": 370
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"epoch": 0.15084366741207564,
|
| 2604 |
+
"grad_norm": 0.1117277517914772,
|
| 2605 |
+
"learning_rate": 0.0001851414614288622,
|
| 2606 |
+
"loss": 1.0371,
|
| 2607 |
+
"step": 371
|
| 2608 |
+
},
|
| 2609 |
+
{
|
| 2610 |
+
"epoch": 0.15125025411669038,
|
| 2611 |
+
"grad_norm": 0.10293183475732803,
|
| 2612 |
+
"learning_rate": 0.0001851007531040098,
|
| 2613 |
+
"loss": 1.0,
|
| 2614 |
+
"step": 372
|
| 2615 |
+
},
|
| 2616 |
+
{
|
| 2617 |
+
"epoch": 0.15165684082130515,
|
| 2618 |
+
"grad_norm": 0.09216313809156418,
|
| 2619 |
+
"learning_rate": 0.00018506004477915734,
|
| 2620 |
+
"loss": 0.9703,
|
| 2621 |
+
"step": 373
|
| 2622 |
+
},
|
| 2623 |
+
{
|
| 2624 |
+
"epoch": 0.1520634275259199,
|
| 2625 |
+
"grad_norm": 0.09088669717311859,
|
| 2626 |
+
"learning_rate": 0.00018501933645430493,
|
| 2627 |
+
"loss": 0.8766,
|
| 2628 |
+
"step": 374
|
| 2629 |
+
},
|
| 2630 |
+
{
|
| 2631 |
+
"epoch": 0.15247001423053466,
|
| 2632 |
+
"grad_norm": 0.09916643798351288,
|
| 2633 |
+
"learning_rate": 0.00018497862812945248,
|
| 2634 |
+
"loss": 1.0958,
|
| 2635 |
+
"step": 375
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 0.15287660093514943,
|
| 2639 |
+
"grad_norm": 0.08404985070228577,
|
| 2640 |
+
"learning_rate": 0.00018493791980460003,
|
| 2641 |
+
"loss": 0.9602,
|
| 2642 |
+
"step": 376
|
| 2643 |
+
},
|
| 2644 |
+
{
|
| 2645 |
+
"epoch": 0.15328318763976417,
|
| 2646 |
+
"grad_norm": 0.10011377185583115,
|
| 2647 |
+
"learning_rate": 0.00018489721147974762,
|
| 2648 |
+
"loss": 1.0377,
|
| 2649 |
+
"step": 377
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 0.15368977434437894,
|
| 2653 |
+
"grad_norm": 0.09958089143037796,
|
| 2654 |
+
"learning_rate": 0.00018485650315489517,
|
| 2655 |
+
"loss": 1.0213,
|
| 2656 |
+
"step": 378
|
| 2657 |
+
},
|
| 2658 |
+
{
|
| 2659 |
+
"epoch": 0.15409636104899369,
|
| 2660 |
+
"grad_norm": 0.09488838911056519,
|
| 2661 |
+
"learning_rate": 0.00018481579483004275,
|
| 2662 |
+
"loss": 0.941,
|
| 2663 |
+
"step": 379
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 0.15450294775360846,
|
| 2667 |
+
"grad_norm": 0.09099314361810684,
|
| 2668 |
+
"learning_rate": 0.00018477508650519033,
|
| 2669 |
+
"loss": 0.8913,
|
| 2670 |
+
"step": 380
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 0.15490953445822322,
|
| 2674 |
+
"grad_norm": 0.0956854447722435,
|
| 2675 |
+
"learning_rate": 0.00018473437818033789,
|
| 2676 |
+
"loss": 1.1478,
|
| 2677 |
+
"step": 381
|
| 2678 |
+
},
|
| 2679 |
+
{
|
| 2680 |
+
"epoch": 0.15531612116283797,
|
| 2681 |
+
"grad_norm": 0.11225584149360657,
|
| 2682 |
+
"learning_rate": 0.00018469366985548547,
|
| 2683 |
+
"loss": 1.0795,
|
| 2684 |
+
"step": 382
|
| 2685 |
+
},
|
| 2686 |
+
{
|
| 2687 |
+
"epoch": 0.15572270786745274,
|
| 2688 |
+
"grad_norm": 0.11592987924814224,
|
| 2689 |
+
"learning_rate": 0.00018465296153063302,
|
| 2690 |
+
"loss": 1.0863,
|
| 2691 |
+
"step": 383
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 0.1561292945720675,
|
| 2695 |
+
"grad_norm": 0.09232570976018906,
|
| 2696 |
+
"learning_rate": 0.0001846122532057806,
|
| 2697 |
+
"loss": 0.9551,
|
| 2698 |
+
"step": 384
|
| 2699 |
+
},
|
| 2700 |
+
{
|
| 2701 |
+
"epoch": 0.15653588127668225,
|
| 2702 |
+
"grad_norm": 0.08860056847333908,
|
| 2703 |
+
"learning_rate": 0.00018457154488092816,
|
| 2704 |
+
"loss": 1.0206,
|
| 2705 |
+
"step": 385
|
| 2706 |
+
},
|
| 2707 |
+
{
|
| 2708 |
+
"epoch": 0.15694246798129702,
|
| 2709 |
+
"grad_norm": 0.10788331180810928,
|
| 2710 |
+
"learning_rate": 0.00018453083655607574,
|
| 2711 |
+
"loss": 0.9378,
|
| 2712 |
+
"step": 386
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 0.15734905468591176,
|
| 2716 |
+
"grad_norm": 0.10758615285158157,
|
| 2717 |
+
"learning_rate": 0.0001844901282312233,
|
| 2718 |
+
"loss": 1.1149,
|
| 2719 |
+
"step": 387
|
| 2720 |
+
},
|
| 2721 |
+
{
|
| 2722 |
+
"epoch": 0.15775564139052653,
|
| 2723 |
+
"grad_norm": 0.10551386326551437,
|
| 2724 |
+
"learning_rate": 0.00018444941990637085,
|
| 2725 |
+
"loss": 1.0729,
|
| 2726 |
+
"step": 388
|
| 2727 |
+
},
|
| 2728 |
+
{
|
| 2729 |
+
"epoch": 0.1581622280951413,
|
| 2730 |
+
"grad_norm": 0.08733198046684265,
|
| 2731 |
+
"learning_rate": 0.00018440871158151843,
|
| 2732 |
+
"loss": 1.0058,
|
| 2733 |
+
"step": 389
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 0.15856881479975604,
|
| 2737 |
+
"grad_norm": 0.1095399409532547,
|
| 2738 |
+
"learning_rate": 0.00018436800325666598,
|
| 2739 |
+
"loss": 1.0566,
|
| 2740 |
+
"step": 390
|
| 2741 |
+
},
|
| 2742 |
+
{
|
| 2743 |
+
"epoch": 0.1589754015043708,
|
| 2744 |
+
"grad_norm": 0.12356330454349518,
|
| 2745 |
+
"learning_rate": 0.00018432729493181356,
|
| 2746 |
+
"loss": 1.0173,
|
| 2747 |
+
"step": 391
|
| 2748 |
+
},
|
| 2749 |
+
{
|
| 2750 |
+
"epoch": 0.15938198820898555,
|
| 2751 |
+
"grad_norm": 0.09934639930725098,
|
| 2752 |
+
"learning_rate": 0.00018428658660696114,
|
| 2753 |
+
"loss": 1.1237,
|
| 2754 |
+
"step": 392
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 0.15978857491360032,
|
| 2758 |
+
"grad_norm": 0.09402013570070267,
|
| 2759 |
+
"learning_rate": 0.0001842458782821087,
|
| 2760 |
+
"loss": 1.0018,
|
| 2761 |
+
"step": 393
|
| 2762 |
+
},
|
| 2763 |
+
{
|
| 2764 |
+
"epoch": 0.1601951616182151,
|
| 2765 |
+
"grad_norm": 0.10511749237775803,
|
| 2766 |
+
"learning_rate": 0.00018420516995725628,
|
| 2767 |
+
"loss": 0.9844,
|
| 2768 |
+
"step": 394
|
| 2769 |
+
},
|
| 2770 |
+
{
|
| 2771 |
+
"epoch": 0.16060174832282983,
|
| 2772 |
+
"grad_norm": 0.11193688213825226,
|
| 2773 |
+
"learning_rate": 0.00018416446163240383,
|
| 2774 |
+
"loss": 0.9888,
|
| 2775 |
+
"step": 395
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 0.1610083350274446,
|
| 2779 |
+
"grad_norm": 0.09895443916320801,
|
| 2780 |
+
"learning_rate": 0.00018412375330755141,
|
| 2781 |
+
"loss": 1.1045,
|
| 2782 |
+
"step": 396
|
| 2783 |
+
},
|
| 2784 |
+
{
|
| 2785 |
+
"epoch": 0.16141492173205937,
|
| 2786 |
+
"grad_norm": 0.09660319238901138,
|
| 2787 |
+
"learning_rate": 0.00018408304498269897,
|
| 2788 |
+
"loss": 1.0457,
|
| 2789 |
+
"step": 397
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 0.16182150843667412,
|
| 2793 |
+
"grad_norm": 0.1339186728000641,
|
| 2794 |
+
"learning_rate": 0.00018404233665784655,
|
| 2795 |
+
"loss": 1.1266,
|
| 2796 |
+
"step": 398
|
| 2797 |
+
},
|
| 2798 |
+
{
|
| 2799 |
+
"epoch": 0.16222809514128889,
|
| 2800 |
+
"grad_norm": 0.1154564693570137,
|
| 2801 |
+
"learning_rate": 0.0001840016283329941,
|
| 2802 |
+
"loss": 1.0299,
|
| 2803 |
+
"step": 399
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 0.16263468184590363,
|
| 2807 |
+
"grad_norm": 0.09698904305696487,
|
| 2808 |
+
"learning_rate": 0.00018396092000814166,
|
| 2809 |
+
"loss": 1.1101,
|
| 2810 |
+
"step": 400
|
| 2811 |
}
|
| 2812 |
],
|
| 2813 |
"logging_steps": 1,
|
|
|
|
| 2827 |
"attributes": {}
|
| 2828 |
}
|
| 2829 |
},
|
| 2830 |
+
"total_flos": 1.2713704795934392e+18,
|
| 2831 |
"train_batch_size": 16,
|
| 2832 |
"trial_name": null,
|
| 2833 |
"trial_params": null
|