Run 4. Outer Step 20. Inner Step 9.

config.json

@@ -268,26 +268,27 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5525073,
-    5525074,
-    5525076,
-    5525078,
-    5525080,
-    5525082,
-    5525084,
-    5525086,
-    5525089,
-    5525091,
-    5525094,
-    5525096,
-    5525099
+    5525101,
+    5525102,
+    5525103,
+    5525104,
+    5525105,
+    5525107,
+    5525109,
+    5525111,
+    5525112,
+    5525114,
+    5525118,
+    5525120,
+    5525123,
+    5525126
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 8,
+  "inner_step": 9,
   "inner_steps": 0,
   "last_allreduce_block": 5523943,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bfaa0e9827f9bf86e19f2b264746ae81ce6c12da2173bb0329f97d8e82ebca67
+oid sha256:347af0386b4f0f2bcddbc5c454f65b44cb4befc77d86b3c2b37c1b2e65d52329
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:219210585206e773c671579e9ce8b04ca31b8c5222dcba6c868d82643d61bb08
+oid sha256:1aba6d151cfe813fdd47f106f21dab5f58a10ecb6c044b5d35146121332336b8
 size 4040701744