step 7600 | val_ppl=10.03

README.md

@@ -25,9 +25,9 @@ Small language model (3.0M parameters) trained from scratch.
 | Total parameters | 3.000M |
 
 ## Training
-- Tokens seen: 7,542,374,400
-- Val loss: 2.3293
-- Val PPL: 10.27
+- Tokens seen: 7,961,395,200
+- Val loss: 2.3053
+- Val PPL: 10.03
 
 ## Usage
 ```python

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7d183cded60bf7dd3ae0910bb83c421b49c95e3be77b444f3e6bd28322f3fc8f
+oid sha256:f01ae6ed9208eb764f5cacac4bda241de333f21ad0d9facb6d79f1dbf53aa0c0
 size 12006544

resume/ckpt.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0f40aef193b715f4af8e44ced8dd5d9e870ce4ad35773f446788c538fbdb08be
+oid sha256:ddc1ca06d675a1b7a4e5cb5a3d01c149acf2c891e32b2c90bd8d45831b681ae3
 size 36728175

resume/latest_step.txt

@@ -1 +1 @@
-7200
+7600

training_meta.json

@@ -1,8 +1,8 @@
 {
-  "step": 7200,
-  "val_loss": 2.329314754486084,
-  "val_ppl": 10.270901032226934,
+  "step": 7600,
+  "val_loss": 2.305310995101929,
+  "val_ppl": 10.027296207571576,
   "params_M": 3.0,
-  "pushed_at": "2026-07-02T00:33:50.567532",
-  "tokens_seen": 7542374400
+  "pushed_at": "2026-07-02T01:19:45.405244",
+  "tokens_seen": 7961395200
 }