Outer Step 1. Inner Step 423. Batch Size 56

config.json

@@ -268,23 +268,23 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5146647,
-    5146651,
-    5146655,
-    5146659,
-    5146663,
-    5146667,
-    5146671,
-    5146675,
-    5146679,
-    5146683
+    5146715,
+    5146719,
+    5146723,
+    5146727,
+    5146731,
+    5146735,
+    5146739,
+    5146742,
+    5146746,
+    5146750
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 413,
+  "inner_step": 423,
   "inner_steps": 0,
   "last_allreduce_block": 5143202,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:59437e3d29ca2c9e101db4d4c62bc889892ca8626448a36df1c7dad4ed695df0
+oid sha256:3917755e3fe251dc65a3b1fcc57145ceaaf0524a1da94b05a9b3bc962442d287
 size 8081781770

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:76dcc4ca09accd0eb6cf66e740ab149988a20981fab95f5d157ab14c0a3f3ce5
+oid sha256:68744488b5af9d69e9f7343747b930eb7211ef6c33f413fa74966375476abedb
 size 4040701744