Run 4. Outer Step 50. Inner Step 3.

config.json

@@ -268,25 +268,47 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5571694,
-    5571695,
-    5571697,
-    5571698,
-    5571700,
-    5571702,
-    5571704,
-    5571705,
-    5571708,
-    5571710,
-    5571711,
-    5571713
+    5571715,
+    5571717,
+    5571718,
+    5571720,
+    5571721,
+    5571723,
+    5571724,
+    5571725,
+    5571727,
+    5571728,
+    5571729,
+    5571730,
+    5571731,
+    5571731,
+    5571733,
+    5571734,
+    5571736,
+    5571737,
+    5571739,
+    5571740,
+    5571742,
+    5571743,
+    5571744,
+    5571745,
+    5571745,
+    5571747,
+    5571748,
+    5571749,
+    5571749,
+    5571751,
+    5571752,
+    5571753,
+    5571754,
+    5571755
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 1,
+  "inner_step": 3,
   "inner_steps": 0,
   "last_allreduce_block": 5571611,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:596c1096fc49edc964513bf1a3d7435def59478639b32e4ccc996dcc870c232c
+oid sha256:afd30352a51857d9f166a8f94f781613d5a82f819491ba9f8dd9766848a5d9a5
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9594d83826dd7d866c138d748f93fda08e09d9a58e07adab8e2e38be72ee51cf
+oid sha256:e5130ab7de1d1487ba180c5fde83fc349237aa05c4b52372f524e0b77d16101b
 size 4040701744