Run 10. Outer Step 12. Inner Step 128.

config.json

@@ -264,16 +264,16 @@
   "attention_bias": false,
   "attention_dropout": 0.0,
   "block_list": [
-    6474345,
-    6474359,
-    6474371
+    6474396,
+    6474408,
+    6474421
   ],
   "bos_token_id": 1,
   "eos_token_id": 2,
   "hidden_act": "silu",
   "hidden_size": 2048,
   "initializer_range": 0.02,
-  "inner_step": 120,
+  "inner_step": 128,
   "intermediate_size": 5632,
   "last_allreduce_block": 6473589,
   "max_position_embeddings": 2048,

gradients.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:13ad39d73a86fa9c96db18822ef520c04c306a62c699a6ad4f1e1fecff871757
+oid sha256:df0061775343bd50a7c6633fd99e1359a58d38ceaa09cb50c11992eda65c212b
 size 88390091

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b0e831fe3839b5dec5836f239b3fccdd75257a6797b23f3e27785b176aee979b
+oid sha256:0203918dd5ce6513be520bbff688c53996c0dcdf4fd677ca13fa35335ca3b6e7
 size 4924928111

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:998c9a89c6aa0fcf7fa0f5faeb977072a55200c8f8876edacd122d888e83d3df
+oid sha256:b3b5ff5e2d49a90803604db70974903fc029fcb341130adafdcdaa04e2be4628
 size 4400216536