Run 4. Outer Step 21. Inner Step 28.

config.json

@@ -268,25 +268,26 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5526013,
-    5526014,
-    5526017,
-    5526017,
-    5526019,
-    5526021,
-    5526023,
-    5526026,
-    5526028,
-    5526029,
-    5526030,
-    5526031
+    5526034,
+    5526036,
+    5526037,
+    5526037,
+    5526038,
+    5526040,
+    5526041,
+    5526044,
+    5526045,
+    5526047,
+    5526049,
+    5526051,
+    5526053
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 27,
+  "inner_step": 28,
   "inner_steps": 0,
   "last_allreduce_block": 5525369,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e22fd975fbbfcb0f50e2f6b46751711c2d710c5cbbee3ee1a7352c37e2f05361
+oid sha256:3084ae7f166422414449f84b90d3d3bc0b3767d12b84429c22b0ddf0477ba37e
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:199534b71f15bf3a9a09952713586c44a7653055c57b933554343cc323b51cbe
+oid sha256:36f977f671a918de0bca1fde66d8144ff11c4e6da4d480a16eece6fd938eaf0a
 size 4040701744