Run 11. Outer Step 200. Inner Step 50.

config.json

@@ -264,11 +264,10 @@
   "attention_bias": false,
   "attention_dropout": 0.0,
   "block_list": [
-    6659167,
-    6659195,
-    6659214,
-    6659234,
-    6659255
+    6659276,
+    6659304,
+    6659325,
+    6659346
   ],
   "bos_token_id": 1,
   "dtype": "float32",
@@ -277,7 +276,7 @@
   "hidden_act": "silu",
   "hidden_size": 2048,
   "initializer_range": 0.02,
-  "inner_step": 31,
+  "inner_step": 50,
   "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:cc433d0ebe9776897654df7ac1ec7d0e67da88f2978d87cb4ef1f72f2898971b
+oid sha256:b5694ab5e36f1abe2ded91633fde1fb3803b9d9fb89ace1b6188209727028f37
 size 88390091

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3e55abf066ef40be3fb198a39e3f6bcdc74fb8554547137a8895fde890204e66
+oid sha256:e34182425cfd11dd666810e0e809b99144e0401892670e9118e9d00baf072b1e
 size 4924928111

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1b8d279bbb4ebd07cfff48e8360679680e8af7862fb0078cc59a3c6334801c7b
+oid sha256:84c7f718c9fed06097d57ea909974249946e5766a47d2aa05677b9d686d9362c
 size 4400216536