Run 4. Outer Step 47. Inner Step 41.

config.json

@@ -268,34 +268,36 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5568285,
-    5568287,
-    5568288,
-    5568289,
-    5568291,
-    5568293,
-    5568294,
-    5568295,
-    5568296,
-    5568297,
-    5568299,
-    5568300,
-    5568302,
-    5568303,
-    5568305,
-    5568306,
-    5568306,
-    5568309,
-    5568310,
-    5568311,
-    5568313
+    5568315,
+    5568318,
+    5568319,
+    5568321,
+    5568323,
+    5568324,
+    5568325,
+    5568327,
+    5568328,
+    5568330,
+    5568331,
+    5568333,
+    5568334,
+    5568334,
+    5568335,
+    5568336,
+    5568338,
+    5568339,
+    5568340,
+    5568341,
+    5568342,
+    5568343,
+    5568343
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 39,
+  "inner_step": 41,
   "inner_steps": 0,
   "last_allreduce_block": 5566279,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ff8ff9ca0c19cac34746ecbbb26fb82fae87da33f0d3e1913ec9a70728654b27
+oid sha256:e1f8e55114a53566a75080f0d65603e46bb06670e71f011420882b54a766c0ad
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:38c5f6e16ac7b6297a1101fad8910422aad59d6963948b004816e01095a44172
+oid sha256:98f2a64f486c75fa085cc28437a2685b1a271d5dc36500c3b843f6cc292d88e9
 size 4040701744