Outer Step 1. Inner Step 325. Batch Size 444

config.json

@@ -268,23 +268,23 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5146239,
-    5146243,
-    5146247,
-    5146252,
-    5146255,
-    5146260,
-    5146264,
-    5146268,
-    5146272,
-    5146277
+    5146315,
+    5146319,
+    5146324,
+    5146328,
+    5146332,
+    5146337,
+    5146341,
+    5146345,
+    5146350,
+    5146354
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 316,
+  "inner_step": 325,
   "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 sha256:3a2fccc46b9776fc4c9ccc2a5dc67bc81d59c758786e2f1770e516427d369c92
+oid sha256:546168e6acf1d23381b8d9676d4abf1eaba53893357fb6d2f9492791f92e11e6
 size 8081781770

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cbe2b49cdea6e70f28483aa35f91a4e270cb1952d4802f5e4236fd4e6641b8e7
+oid sha256:abc2ef11d12c353789f750ec955e9bc2c1fdac75d44ceb9634f0a5d930b59a9c
 size 4040701744