Run 4. Outer Step 20. Inner Step 19.

config.json

@@ -268,30 +268,28 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5525202,
-    5525202,
-    5525204,
-    5525205,
-    5525206,
-    5525207,
-    5525208,
-    5525210,
-    5525211,
-    5525212,
-    5525213,
-    5525214,
-    5525215,
-    5525216,
-    5525218,
-    5525220,
-    5525222
+    5525225,
+    5525227,
+    5525228,
+    5525230,
+    5525231,
+    5525232,
+    5525233,
+    5525233,
+    5525235,
+    5525236,
+    5525237,
+    5525239,
+    5525241,
+    5525243,
+    5525245
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 18,
+  "inner_step": 19,
   "inner_steps": 0,
   "last_allreduce_block": 5523943,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:04d17d5620181d48753998a00253cbd62bbd5ad1f25cafd036ab2d03bb273d58
+oid sha256:2af6f97b396dda87209e407dc0f71e49cd1732c86288981ae40512b6764da5ba
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d374bd79a383532460827fff4e97b7afd64ec19e93671e7b358f1b5f72b1601c
+oid sha256:8a27a0a07a271c36d247fe15aaedf40207e196a04850d3fb6c0f0de3a11a3c9e
 size 4040701744