agent=agent_filler train_key=e8f3355683e0a25f
Browse files
e8f3355683e0a25f/config.json
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"act_cache_key": "e4916bcae1881963",
|
| 3 |
+
"agent": "agent_filler",
|
| 4 |
+
"arch": "tsae_paper",
|
| 5 |
+
"arch_version": "1.0.0",
|
| 6 |
+
"datasource": "gemma_2_2b_it_l13_fineweb_24k128",
|
| 7 |
+
"saved_ts": "2026-05-05T15:17:06Z",
|
| 8 |
+
"seed": 1,
|
| 9 |
+
"train_key": "e8f3355683e0a25f",
|
| 10 |
+
"training_cfg": {
|
| 11 |
+
"batch_size": 1024,
|
| 12 |
+
"bricken_enabled": false,
|
| 13 |
+
"bricken_max_resample_fraction": 0.5,
|
| 14 |
+
"bricken_min_fires": 1,
|
| 15 |
+
"bricken_n_check": 2048,
|
| 16 |
+
"bricken_resample_every": 500,
|
| 17 |
+
"dead_threshold_tokens": 10000000,
|
| 18 |
+
"ema_auxk_alpha": 0.03125,
|
| 19 |
+
"learning_rate": 0.0003,
|
| 20 |
+
"n_steps": 20000,
|
| 21 |
+
"optimizer": "adam",
|
| 22 |
+
"plateau_early_stop": false,
|
| 23 |
+
"plateau_min_delta": 0.0001,
|
| 24 |
+
"plateau_patience": 5000,
|
| 25 |
+
"precision": "bf16",
|
| 26 |
+
"train_window_size": 2,
|
| 27 |
+
"warmup_steps": 1000
|
| 28 |
+
}
|
| 29 |
+
}
|
e8f3355683e0a25f/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9ae8aa67438cac25a1b8c13059f04a18cd94a4eca8999b10287e1b2119d343d0
|
| 3 |
+
size 302196344
|