Run 4. Outer Step 18. Inner Step 0.

config.json

@@ -268,19 +268,23 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5520384,
-    5520387,
-    5520391,
-    5520394,
-    5520398,
-    5520401
+    5520406,
+    5520409,
+    5520413,
+    5520417,
+    5520420,
+    5520424,
+    5520428,
+    5520431,
+    5520435,
+    5520438
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 53,
+  "inner_step": 0,
   "inner_steps": 0,
   "last_allreduce_block": 5509976,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2ae1fa4d8021ca268e5c60711e5d40afe6be8893638652235d9a4b30b4dc1484
+oid sha256:ce6f891d5da72887ae2b514c62798fde055a21684ffb0a512269135d4ad27a3a
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:940ef7464bf6943c9f5f1ec9648eea79b8b7624dacefa831d2943fe27f16712f
+oid sha256:b14c8889c612643b3f45148b7b56d55e0f0397141c6ee44cfd8679735c0a5d4c
 size 4040701744