step 2000 | val_ppl=13.38

README.md

@@ -25,9 +25,9 @@ Small language model (9.9M parameters) trained from scratch.
 | Total parameters | 9.853M |
 
 ## Training
-- Tokens seen: 1,676,083,200
-- Val loss: 2.6657
-- Val PPL: 14.38
+- Tokens seen: 2,095,104,000
+- Val loss: 2.5937
+- Val PPL: 13.38
 
 ## Usage
 ```python

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0692f3ebb90934ebe0d7504caf6a2340535690e6e15117d093dc990727ef2f81
+oid sha256:ce31762a215b9af8d1be6e67b35ff25caaca6f4e7c4e88930d34527bbc38088b
 size 39421616

resume/ckpt.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f00fcbcd0047588c352e8025caceb5560dc778e003de9adc2c9f1f926741664
+oid sha256:d36c694f3dac51d63f3162c316b7666e7cca2b78a0b81ac3ebf7f526dd7a90e2
 size 119811943

resume/latest_step.txt

@@ -1 +1 @@
-1600
+2000

training_meta.json

@@ -1,8 +1,8 @@
 {
-  "step": 1600,
-  "val_loss": 2.6657352474212646,
-  "val_ppl": 14.378517428377364,
+  "step": 2000,
+  "val_loss": 2.593691735458374,
+  "val_ppl": 13.3790725406917,
   "params_M": 9.853,
-  "pushed_at": "2026-06-17T14:52:31.290553",
-  "tokens_seen": 1676083200
+  "pushed_at": "2026-06-17T16:36:36.179467",
+  "tokens_seen": 2095104000
 }