Outer Step 0. Inner Step 51. Batch Size 272

config.json

@@ -268,23 +268,23 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5155164,
-    5155169,
-    5155173,
-    5155177,
-    5155182,
-    5155186,
-    5155190,
-    5155195,
-    5155200,
-    5155205
+    5155242,
+    5155247,
+    5155252,
+    5155257,
+    5155262,
+    5155267,
+    5155272,
+    5155276,
+    5155281,
+    5155285
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 42,
+  "inner_step": 51,
   "inner_steps": 0,
   "last_allreduce_block": 4016775,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bd625187949b982e8e9a0a341ee6413fb0b58c50dca976361d08604063e4f416
+oid sha256:bc11b825ea85525daaab978f837341377a886e0061e064182733bac007857185
 size 8081781770

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:48445af707c3282ed36027263ea3472f585cd1dd579acba94d336c9d314c33f3
+oid sha256:8f7bf5a8b9b4aa48987db2732fcbdf058588c43c27111358d8614f959b66c52b
 size 4040701744