Outer Step 4. Inner Step 60. Batch Size 12

config.json

@@ -268,24 +268,23 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5149374,
-    5149378,
-    5149382,
-    5149386,
-    5149390,
-    5149394,
-    5149398,
-    5149402,
-    5149406,
-    5149409,
-    5149416
+    5149455,
+    5149459,
+    5149463,
+    5149467,
+    5149471,
+    5149475,
+    5149479,
+    5149482,
+    5149486,
+    5149490
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 50,
+  "inner_step": 60,
   "inner_steps": 0,
   "last_allreduce_block": 5148895,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a9a7ebde7b9dfaac1fcae302b13a98870da733cb8423f76b96e56a46a7385c37
+oid sha256:42ca73fb2c5abd4bb985f57e89ec6f916ec49ba03a090a2492510aabd4e29e32
 size 8081781770

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9674b237c26263704d7f13e4019e41d5d2555bd8e409f11b28ca3a7e16255c96
+oid sha256:753dbc4f1cc0b0c04feb282d6da400d64ee4bfac5df54433f0ef0c2c79166d88
 size 4040701744