Outer Step 8. Inner Step 222. Batch Size 352

config.json

@@ -268,23 +268,23 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5292839,
-    5292839,
-    5292844,
-    5292850,
-    5292851,
-    5292856,
-    5292861,
-    5292862,
-    5292868,
-    5292868
+    5292910,
+    5292916,
+    5292921,
+    5292922,
+    5292927,
+    5292932,
+    5292933,
+    5292939,
+    5292940,
+    5292945
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 211,
+  "inner_step": 222,
   "inner_steps": 0,
   "last_allreduce_block": 5283755,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ad0e5e128a9fae18c31aaaf977e9cc513a5bd124dc9c6b068c876a5029884aee
+oid sha256:5bdf8bcac2c68c16724a2e9fafe4f0d42a61892598d341a3a2e4b2b92b502858
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d8bb27c84dce99dc5558dca17fcbaf6334c74f175c01b6fdb2d50f71aece4341
+oid sha256:d3f4fb52a7c2727dd18befe968712ff267b8a24ead4f5eafb2a2819a3d2477b2
 size 4040701744