Outer Step 4. Inner Step 62. Batch Size 296

config.json

@@ -268,23 +268,23 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5149348,
-    5149352,
-    5149356,
-    5149361,
-    5149365,
-    5149369,
-    5149373,
-    5149378,
-    5149382,
-    5149386
+    5149422,
+    5149426,
+    5149431,
+    5149435,
+    5149439,
+    5149443,
+    5149447,
+    5149452,
+    5149456,
+    5149461
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 52,
+  "inner_step": 62,
   "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 sha256:8cf5bfeb21ea74ff21977f3be4d51c323f9ec416d7b090256ac1af8154859260
+oid sha256:4cc8a934a6d5118aab5f0b9265fee9987df9c7029b80d1d174679ad6a36b2501
 size 8081781770

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7f51781bccebbfd47afb79cfdad2bcc4611541cd05fe0c5da6c60f60ca242da2
+oid sha256:d1f31fb0edb2a88020a0c1c3ebc3f8eda8bc70b052d074aa65819df9840549c4
 size 4040701744