Run 11. Outer Step 203. Inner Step 0.

config.json

@@ -264,8 +264,8 @@
   "attention_bias": false,
   "attention_dropout": 0.0,
   "block_list": [
-    6660026,
-    6660058
+    6660105,
+    6660137
   ],
   "bos_token_id": 1,
   "dtype": "float32",
@@ -274,7 +274,7 @@
   "hidden_act": "silu",
   "hidden_size": 2048,
   "initializer_range": 0.02,
-  "inner_step": 45,
+  "inner_step": 0,
   "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 sha256:9550cfa3ac717f9ba6d9a262a1a71114e02e3a676810a2b7145855b10080039c
+oid sha256:f25b56802713a81e9d730e024f4ecaeff67783d739724049cef67f0e527bcee2
 size 88389626

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:eb4700cc6e3ae0d523d174f18f3efba32335ead2c3cd1a6738d7af9ffc15cbb2
+oid sha256:761b9f081d1d79bc72b8290592fcc506c6afbc0f1c0378df3bec1be4841db619
 size 4924926906

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b1fedd0709bb878c14bcfc5ddbd6ec3a0ccdde759ed2fd139b82ce8a6c93af86
+oid sha256:357116651638863537ce648c7ccf80966e0beaf4e0555d5397eeae81b2e4c678
 size 4400216536