Outer Step 0. Inner Step 39. Batch Size 112

config.json

@@ -268,23 +268,23 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5139723,
-    5139731,
-    5139737,
-    5139752,
-    5139762,
-    5139768,
-    5139773,
-    5139779,
-    5139794,
-    5139804
+    5139829,
+    5139844,
+    5139857,
+    5139862,
+    5139869,
+    5139874,
+    5139881,
+    5139890,
+    5139901,
+    5139918
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 32,
+  "inner_step": 39,
   "inner_steps": 0,
   "last_allreduce_block": 5072405,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:71365e0c53f1b62c9efbeb9dfbeaabf0c6145595c028843633448e4001780c86
+oid sha256:0e759c414b75ea3ee97a5c9682e0f81480da88a18c6cd1ee2d44b98461050481
 size 8081781770

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ca9cdf28e4c35d7c4531cedc122a3dcd820fa1cda098ec2eec433d7437ae5628
+oid sha256:a769ad1f3c9941016cd47ad7c7a47cafe8caa6f0381ca4548c5f941884968df3
 size 4040701744