Run 11. Outer Step 202. Inner Step 9.

config.json

@@ -264,9 +264,8 @@
   "attention_bias": false,
   "attention_dropout": 0.0,
   "block_list": [
-    6659667,
-    6659702,
-    6659732
+    6659770,
+    6659803
   ],
   "bos_token_id": 1,
   "dtype": "float32",
@@ -275,7 +274,7 @@
   "hidden_act": "silu",
   "hidden_size": 2048,
   "initializer_range": 0.02,
-  "inner_step": 0,
+  "inner_step": 8,
   "intermediate_size": 5632,
   "last_allreduce_block": 6659044,
   "max_position_embeddings": 2048,

gradients.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:06b945b73bcbcc4ac11c4563f398e1879ca97b8d199dbe52a7086e4c10f1a9a9
+oid sha256:7f0bee3d0294330abfb1bd28f7b27290e8c5f90b926ad9c8c33f4f1382cc40b7
 size 88389626

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:72b91011e037b46fcd99062edbb0cc1e78cb5702f5270fcbc10549bd3bdf68f3
+oid sha256:cd20aecb1cfce5617469bb4264db675bedf734c9e54efca37dd15fc687b53f64
 size 4924926906

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5e92c87aa97880e363e5afbf75ed777463ef7aa24be3fd79833b5065f173576e
+oid sha256:ed464682be36e41f3db5672fb7692a310b63c0b7d7a6ff54a9fe7e0ab37a563d
 size 4400216536