Run 4. Outer Step 54. Inner Step 95.

config.json

@@ -268,21 +268,18 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5580399,
-    5580402,
-    5580404,
-    5580406,
-    5580407,
-    5580408,
-    5580410,
-    5580413
+    5580416,
+    5580419,
+    5580421,
+    5580422,
+    5580425
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 94,
+  "inner_step": 95,
   "inner_steps": 0,
   "last_allreduce_block": 5577515,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7f0e7a8e93143189854656059c097d9c370aaa16555eb89d0f83e6c423d61fcc
+oid sha256:1461e715c2226a648c462ae2d8a6e48dbab0d5db2281f25a07c3430d00a14f76
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1fdc2e5254c96666260128e0d5aade11f0cd1e9a260db7e2ff39adf5a9766786
+oid sha256:3486f0e1d9359fe7b960bdfe48c2b07cbe83d4eb1fa01864914ebc9eca420c36
 size 4040701744