Run 11. Outer Step 410. Inner Step 9.

config.json

@@ -264,8 +264,8 @@
   "attention_bias": false,
   "attention_dropout": 0.0,
   "block_list": [
-    6753566,
-    6769512
+    6769719,
+    6769747
   ],
   "bos_token_id": 1,
   "dtype": "float32",
@@ -274,7 +274,7 @@
   "hidden_act": "silu",
   "hidden_size": 2048,
   "initializer_range": 0.02,
-  "inner_step": 654,
+  "inner_step": 9,
   "intermediate_size": 5632,
   "last_allreduce_block": 6750947,
   "max_position_embeddings": 2048,
@@ -292,4 +292,4 @@
   "transformers_version": "4.39.3",
   "use_cache": false,
   "vocab_size": 32000
-}
+}

gradients.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bfd3b39477dbdfe5e7a5f9e9686723a0396d966060a85ec48b0b7be174e44d76
+oid sha256:3232e750eb467880fc063712244136c31c20455109e8e097ee0a2c2f4c023c05
 size 88389626

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2e8d6dcd6c27780a97ec218cdcb3fb03178ed59ca8c2e4fa65d83fba54a66b24
+oid sha256:d868652da4c3268b036216f2d6950cef53634bdfbd9f8c6af56c516ea82e05a9
 size 4924926906

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cb296835b9f82b8454862090b378c3b0538404b5b77dba74a4c33bbaa31f17c5
+oid sha256:2cb04097cae139c5b5fd4ada27c51493bc9fa9f2c46bc7ebdea0227b2466c25f
 size 4400216536