Run 9. Outer Step 11. Inner Step 11.

config.json

@@ -20,16 +20,16 @@
   "attention_bias": false,
   "attention_dropout": 0.0,
   "block_list": [
-    6454764,
-    6454771,
-    6454786
+    6454804,
+    6454826,
+    6454844
   ],
   "bos_token_id": 1,
   "eos_token_id": 2,
   "hidden_act": "silu",
   "hidden_size": 2048,
   "initializer_range": 0.02,
-  "inner_step": 3,
+  "inner_step": 11,
   "intermediate_size": 5632,
   "last_allreduce_block": 5428935,
   "max_position_embeddings": 2048,

gradients.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:afefec6e73aa6a80bb4da4c56a786f6fd751dfc871addf6df96fe0d00010c210
+oid sha256:f19f11b6b481e09236810e57bf09b23edc0eba6ad6c34d630b2527e82a32325f
 size 88389626

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d7d2bf8aec216ec6dcde6c1927ce5e47e8965b7fd2f8d4def9d9fbe28b8f560b
+oid sha256:49d7ee149941f8754f550fd87279f5d40cdcda78707af88b39838ae288984632
 size 4924926906

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d047803526c79718b81ef6c8f0e7c6b17fbff2338728fa0657ad3333d20c74b4
+oid sha256:fc7de7bd71353816576dc8b98e616b468a4ccad4649debe1a1c4082877a40882
 size 4400216536