step 8000 | val_ppl=9.88

README.md

@@ -25,9 +25,9 @@ Small language model (3.0M parameters) trained from scratch.
 | Total parameters | 3.000M |
 
 ## Training
-- Tokens seen: 7,961,395,200
-- Val loss: 2.3053
-- Val PPL: 10.03
+- Tokens seen: 8,380,416,000
+- Val loss: 2.2903
+- Val PPL: 9.88
 
 ## Usage
 ```python

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f01ae6ed9208eb764f5cacac4bda241de333f21ad0d9facb6d79f1dbf53aa0c0
+oid sha256:35e09a095d03f11e9d8ee0d63ab4dd4df7d9b22c3b5ba0b885a9d69c7e503581
 size 12006544

resume/ckpt.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ddc1ca06d675a1b7a4e5cb5a3d01c149acf2c891e32b2c90bd8d45831b681ae3
+oid sha256:216881e6ce1c92cc07a46976160be1b9ca24879407414cd1822a055788c3092e
 size 36728175

resume/latest_step.txt

@@ -1 +1 @@
-7600
+8000

training_meta.json

@@ -1,8 +1,8 @@
 {
-  "step": 7600,
-  "val_loss": 2.305310995101929,
-  "val_ppl": 10.027296207571576,
+  "step": 8000,
+  "val_loss": 2.290264854888916,
+  "val_ppl": 9.877553453080482,
   "params_M": 3.0,
-  "pushed_at": "2026-07-02T01:19:45.405244",
-  "tokens_seen": 7961395200
+  "pushed_at": "2026-07-02T02:05:39.788789",
+  "tokens_seen": 8380416000
 }