Run 4. Outer Step 20. Inner Step 14.

config.json

@@ -268,26 +268,30 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5525100,
-    5525102,
-    5525103,
-    5525105,
-    5525106,
-    5525108,
-    5525110,
-    5525112,
-    5525113,
-    5525116,
-    5525118,
-    5525120,
-    5525122
+    5525125,
+    5525127,
+    5525129,
+    5525130,
+    5525132,
+    5525133,
+    5525134,
+    5525136,
+    5525137,
+    5525138,
+    5525139,
+    5525141,
+    5525142,
+    5525143,
+    5525144,
+    5525146,
+    5525147
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 13,
+  "inner_step": 14,
   "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 sha256:3f751d3c80730db461b874dfe35cadc8a506e69d4e820077561f44794e5fa3d8
+oid sha256:d68c442fca57dbf2cdcebcdb5b000fe110ffe252f020fe3ba9be0289c7f6a4bd
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bfcdaff8ee2838153ba311bdaf24174cb77dc0e18444c584ca840cf007fee3e1
+oid sha256:40842c7187b2d1b310a7844853a8652452ff713afb59ecbe8ba1e5e2c4cf7d8e
 size 4040701744