Run 4. Outer Step 50. Inner Step 3.

config.json

@@ -268,27 +268,33 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5571713,
-    5571715,
-    5571717,
-    5571718,
-    5571719,
-    5571720,
-    5571721,
-    5571723,
-    5571723,
-    5571725,
-    5571727,
-    5571728,
-    5571730,
-    5571731
+    5571733,
+    5571735,
+    5571737,
+    5571738,
+    5571739,
+    5571740,
+    5571742,
+    5571744,
+    5571745,
+    5571746,
+    5571748,
+    5571749,
+    5571751,
+    5571752,
+    5571753,
+    5571754,
+    5571755,
+    5571756,
+    5571757,
+    5571758
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 2,
+  "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 sha256:988e16549447b81817e6fab7844a11a7727780dfbe9be3e467642cdbc9f3ef1d
+oid sha256:51d8b668a7316570d35335f26cb6b133a5a5e83467d551fb38a389c25dd6cf64
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e44b4a758ec4db046cec4380872f3b6e7abe5a077c9dd2c1b7ec88cf5451949c
+oid sha256:e851124c033dda1be3ab6d2eb7977966e8a1dd071c668caee4162011c993cc85
 size 4040701744