step 3200 | val_ppl=11.88

README.md

@@ -25,9 +25,9 @@ Small language model (9.9M parameters) trained from scratch.
 | Total parameters | 9.853M |
 
 ## Training
-- Tokens seen: 2,933,145,600
-- Val loss: 2.5037
-- Val PPL: 12.23
+- Tokens seen: 3,352,166,400
+- Val loss: 2.4750
+- Val PPL: 11.88
 
 ## Usage
 ```python

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:70f36a0d1c5fa8de1eae21bcfee204dd2780cadca653a4548205bfd2352c306a
+oid sha256:5c2ecf240f59ef5bb582242b08f1eebf095d8c945a7d01f4ac6535ec190ad990
 size 39421616

resume/ckpt.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2676b6dbebe68dbdceeda5fc55b938075b759b4c1fd90df1d98a5f3ba168107f
+oid sha256:2dce5b25f1680e6920a5e01a353e7c3f60aa9a3d1f50a681f21aaf0971f9d494
 size 119812327

resume/latest_step.txt

@@ -1 +1 @@
-2800
+3200

training_meta.json

@@ -1,8 +1,8 @@
 {
-  "step": 2800,
-  "val_loss": 2.503683595275879,
-  "val_ppl": 12.227452090847526,
+  "step": 3200,
+  "val_loss": 2.4750425384521484,
+  "val_ppl": 11.882212553279116,
   "params_M": 9.853,
-  "pushed_at": "2026-06-17T20:37:43.963248",
-  "tokens_seen": 2933145600
+  "pushed_at": "2026-06-17T22:32:27.223238",
+  "tokens_seen": 3352166400
 }