Run 4. Outer Step 51. Inner Step 40.

config.json

@@ -268,20 +268,20 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5574633,
-    5574637,
-    5574641,
-    5574645,
-    5574649,
-    5574653,
-    5574657
+    5574662,
+    5574666,
+    5574670,
+    5574674,
+    5574677,
+    5574681,
+    5574685
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 39,
+  "inner_step": 40,
   "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 sha256:0ab9cbfd49a10625a8a4cf9b20eddc51ae02f6b6a85a242be059ec6c055f9bfa
+oid sha256:9a86f4dd417b9df704fcdc8f8daa74d53a148ae3a60904223137cc3f628a65b5
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:85b6bb11275ed9a24146ffb6805c8aa41536b3f896338925cb35996435f3848c
+oid sha256:8aba8b449482150bb2a4d8cf6f1e0053dd0a5c0f9eb142e73450d0d797271c8a
 size 4040701744