step 3200 | val_ppl=12.02

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: 2.5094
-- Val PPL: 12.30
+- Tokens seen: 3,352,166,400
+- Val loss: 2.4868
+- Val PPL: 12.02
 
 ## Usage
 ```python

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6a6fbfcfd67990663b263b1e2770cb6843cd6d59679edb697794cda26b78ed95
+oid sha256:d29efa8f79fe0b356ca629f9e37e0cbd2d815b52f3f6c93503b1e0132100f376
 size 39421616

resume/ckpt.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:552f68536884f8f8a7e67e7a139ffbb96b419b5a790d1e0041db7bab37338cff
+oid sha256:acb7f948efa6048817aa1e549ef765990d65548a00b09c633ec65c4c12c6be8f
 size 119812327

resume/latest_step.txt

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

training_meta.json

@@ -1,8 +1,8 @@
 {
-  "step": 2800,
-  "val_loss": 2.5094214358520506,
-  "val_ppl": 12.297812928191703,
+  "step": 3200,
+  "val_loss": 2.4867598774719237,
+  "val_ppl": 12.022259351659718,
   "params_M": 9.853,
-  "pushed_at": "2026-06-20T20:08:39.474048",
-  "tokens_seen": 2933145600
+  "pushed_at": "2026-06-20T22:16:48.331353",
+  "tokens_seen": 3352166400
 }