Run 4. Outer Step 60. Inner Step 51.

config.json

@@ -268,24 +268,20 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5608634,
-    5608641,
-    5608645,
-    5608649,
-    5608653,
-    5608657,
-    5608661,
-    5608665,
-    5608668,
-    5608672,
-    5608676
+    5608680,
+    5608684,
+    5608688,
+    5608696,
+    5608700,
+    5608703,
+    5608707
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 50,
+  "inner_step": 51,
   "inner_steps": 0,
   "last_allreduce_block": 5607244,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3bac865e4bcebc01766fa0c9bc768f9732af7b2b9daf5060db789b1c84d9b323
+oid sha256:0f7a781b45289c146a956fecf5790add5550d09c32852df90e74278821f78714
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:52b10dafe59340222235e01cc3d1ee84d812d584e01117b84658cfb5361c8bc0
+oid sha256:5cabb337a625c5756f84ded92fc4fa149ad8c211d5f7147328e337d1def64de1
 size 4040701744