Run 3. Outer Step 9. Inner Step 975.

config.json

@@ -268,18 +268,18 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5420601,
-    5420605,
-    5420608,
-    5420611,
-    5420614
+    5420640,
+    5420643,
+    5420646,
+    5420649,
+    5420653
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 973,
+  "inner_step": 975,
   "inner_steps": 0,
   "last_allreduce_block": 5398844,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:decf2f1f08b5f02c1a6169806dbfef82043497e70abb1713433d3fbd8b6d2090
+oid sha256:8482d6d9326f5821c2bd44ac9c4c1d329a13670c33784892b9b5778ad0dd29be
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b05e2a1f4cd4df1fb8d1a644cb558ca0de37d56f34346f79d3063590ecf7ed46
+oid sha256:8764f1a900f7706ca4a7bc5cbae0c10500de62a9a8a9e12e2379301cb7140637
 size 4040701744