step 1200 | val_ppl=15.94

README.md

@@ -25,9 +25,9 @@ Small language model (9.9M parameters) trained from scratch.
 | Total parameters | 9.853M |
 
 ## Training
-- Tokens seen: 838,041,600
-- Val loss: 2.9375
-- Val PPL: 18.87
+- Tokens seen: 1,257,062,400
+- Val loss: 2.7687
+- Val PPL: 15.94
 
 ## Usage
 ```python

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0aa439042ac467021b8680b0cc1a440ca060e9d316df6edd36c34617b2c8aa38
+oid sha256:b6d23ec7d2a1a452230c62da1c076f1aba2a74d298b845cc1e44e11f8ea044e8
 size 39421616

resume/ckpt.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c8b846794e1ec343ec39fdbe203f94103c67d9a658c88c769590edaf29584d77
+oid sha256:abfe18901f1c58da3ef40548191c2d25d422dd60d7096ec4fcccfca2559b20bb
 size 119811943

resume/latest_step.txt

@@ -1 +1 @@
-800
+1200

training_meta.json

@@ -1,8 +1,8 @@
 {
-  "step": 800,
-  "val_loss": 2.9375060623168947,
-  "val_ppl": 18.868730147139708,
+  "step": 1200,
+  "val_loss": 2.7687056308746336,
+  "val_ppl": 15.937991009362028,
   "params_M": 9.853,
-  "pushed_at": "2026-06-17T11:23:42.505200",
-  "tokens_seen": 838041600
+  "pushed_at": "2026-06-17T13:08:06.578893",
+  "tokens_seen": 1257062400
 }