Run 4. Outer Step 20. Inner Step 15.

config.json

@@ -268,26 +268,26 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5525130,
-    5525132,
-    5525134,
-    5525136,
-    5525138,
-    5525139,
-    5525141,
-    5525143,
-    5525145,
-    5525147,
-    5525148,
-    5525150,
-    5525152
+    5525155,
+    5525157,
+    5525159,
+    5525161,
+    5525162,
+    5525164,
+    5525166,
+    5525168,
+    5525170,
+    5525172,
+    5525173,
+    5525175,
+    5525177
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 14,
+  "inner_step": 15,
   "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 sha256:12cb61ecac14dab598511b0f5544e2034a29b336f32f6c432b8eb21dd6dd4091
+oid sha256:2a63054de502f2a4746ed1080c001df31630f1d9fe5fdd34a166cc5469b418ba
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e4b68c6380fabee4501e38a8d882e00f271d1da22492d095e75c3f426629ee83
+oid sha256:fe34c20e15d51b632443f22d1aa6aae66adab17937542eac0d431429cbdede3d
 size 4040701744