Outer Step 2. Inner Step 136. Batch Size 380

config.json

@@ -268,23 +268,23 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5148044,
-    5148048,
-    5148053,
-    5148058,
-    5148063,
-    5148067,
-    5148072,
-    5148076,
-    5148081,
-    5148085
+    5148118,
+    5148122,
+    5148126,
+    5148131,
+    5148135,
+    5148139,
+    5148145,
+    5148149,
+    5148154,
+    5148158
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 127,
+  "inner_step": 136,
   "inner_steps": 0,
   "last_allreduce_block": 5143202,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1c1e8a50ac7f2e7c570425a8dedb8e3eff8348170534b0b8c15b73d4f16884e8
+oid sha256:e3143ed433c3ded4058ac8a644e697243937fc3cfa156858a00f8017b905ee59
 size 8081781770

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c83e036ab75974b6b235c617fa43447891ddee63f675a58dfd2a8933e9ab646e
+oid sha256:65ce7bbccde7339db741261c8130759054c16b88f60d59f70deeeaa80786f896
 size 4040701744