Outer Step 10. Inner Step 47. Batch Size 276

config.json

@@ -268,23 +268,23 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5319645,
-    5319647,
-    5319650,
-    5319652,
-    5319654,
-    5319656,
-    5319658,
-    5319661,
-    5319664,
-    5319666
+    5319691,
+    5319694,
+    5319695,
+    5319699,
+    5319701,
+    5319702,
+    5319705,
+    5319706,
+    5319708,
+    5319711
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 41,
+  "inner_step": 47,
   "inner_steps": 0,
   "last_allreduce_block": 5297756,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:884d14a05c95e8edc83b25612aae1bd30d4388fcb1c5a28bdaf4e3ed685595c5
+oid sha256:517bb49a72f9ed1a7b4ec5f20b890a92e45a809717d6785a553814887e25d66a
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:62c8de06fcb72351e9670f7c93a09b7e8240b3f8c5bf18ac4c3b74e930fa1f1d
+oid sha256:055c0195f28a4c1f052d067275b528785df8e0b03dd0f3d96a2b69a9c5b4477e
 size 4040701744