Run 4. Outer Step 20. Inner Step 13.

config.json

@@ -268,29 +268,26 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5525076,
-    5525077,
-    5525080,
-    5525082,
-    5525083,
-    5525085,
-    5525086,
-    5525088,
-    5525090,
-    5525091,
-    5525092,
-    5525093,
-    5525094,
-    5525095,
-    5525096,
-    5525098
+    5525100,
+    5525102,
+    5525103,
+    5525105,
+    5525106,
+    5525108,
+    5525110,
+    5525112,
+    5525113,
+    5525116,
+    5525118,
+    5525120,
+    5525122
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 11,
+  "inner_step": 13,
   "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 sha256:29d205cf43af6e67b1f809c2d5a64dfc1c2ff30330c09ec2f8b69b786bc3a00a
+oid sha256:3f751d3c80730db461b874dfe35cadc8a506e69d4e820077561f44794e5fa3d8
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a994fa317803e43263b9a3cf7aa831a93248e483c55bd14f6e5362a1319c3737
+oid sha256:bfcdaff8ee2838153ba311bdaf24174cb77dc0e18444c584ca840cf007fee3e1
 size 4040701744