Run 11. Outer Step 213. Inner Step 18.

config.json

@@ -264,8 +264,8 @@
   "attention_bias": false,
   "attention_dropout": 0.0,
   "block_list": [
-    6664219,
-    6664243
+    6664264,
+    6664292
   ],
   "bos_token_id": 1,
   "dtype": "float32",
@@ -274,7 +274,7 @@
   "hidden_act": "silu",
   "hidden_size": 2048,
   "initializer_range": 0.02,
-  "inner_step": 8,
+  "inner_step": 17,
   "intermediate_size": 5632,
   "last_allreduce_block": 6664151,
   "max_position_embeddings": 2048,

gradients.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b0e57b8fd61e1665b0c2a4812e4fe77603e59c02b79294c4c35d26cee21c9d9f
+oid sha256:f33de7dddfcde711b0c1df1e159f1d526d3542e81453a0378d4d7df530319c41
 size 88390091

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b0cc9b5f27af39e78bad053d647db3b057bc1d79fd1e2f47b33b477d8254c156
+oid sha256:6d55741573fdb797f0ac320d506a67350ef524f886ee6a8db2f16f12a1856e9d
 size 4924928111

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1266a52b2ec7f7e65cd6d67de02647ba6a675dba73b04a95db7467d83961ae17
+oid sha256:790fdac4c7776fb199c225c88056dea1da2122ee1860ae16f86a12ef37d9da9b
 size 4400216536