Run 4. Outer Step 47. Inner Step 63.

config.json

@@ -268,33 +268,31 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5568708,
-    5568710,
-    5568711,
-    5568713,
-    5568714,
-    5568716,
-    5568717,
-    5568718,
-    5568719,
-    5568719,
-    5568719,
-    5568722,
-    5568723,
-    5568724,
-    5568725,
-    5568727,
-    5568728,
-    5568729,
-    5568730,
-    5568733
+    5568735,
+    5568737,
+    5568738,
+    5568739,
+    5568740,
+    5568741,
+    5568742,
+    5568744,
+    5568746,
+    5568747,
+    5568748,
+    5568750,
+    5568751,
+    5568753,
+    5568755,
+    5568757,
+    5568759,
+    5568760
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 61,
+  "inner_step": 63,
   "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 sha256:19a9dfa0b45a59b50c755fdfd2819e3a07a0a75782241e71773871b97be4002a
+oid sha256:6af579fb6d3b0a80e73e76b2df445d36cbe5ec0b19ea6076d69f27fb71904d33
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:24ccaaf56d2876653e4ff685adc6b068733af4dc849690dc0dad44058d0ec477
+oid sha256:5e9a41830162a2d391f508dc35ac0b53934f33ba10dcb9590c60bcefbd2b5806
 size 4040701744