Run 4. Outer Step 53. Inner Step 5.

config.json

@@ -267,13 +267,33 @@
     "AutoConfig": "distributed/optimized-gpt2-500m--configuration_gpt_optimized.GPTOptimConfig",
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
-  "block_list": [],
+  "block_list": [
+    5587831,
+    5587835,
+    5587839,
+    5587843,
+    5587848,
+    5587852,
+    5587856,
+    5587861,
+    5587864,
+    5587869,
+    5587874,
+    5587878,
+    5587882,
+    5587887,
+    5587891,
+    5587895,
+    5587900,
+    5587905,
+    5587910
+  ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 2,
+  "inner_step": 5,
   "inner_steps": 0,
   "last_allreduce_block": 5576204,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:109edbcea12cad753c2456cacf5c48c6ed8cc4c096e3c92869624eaf34d814f8
+oid sha256:53f2475eda455593b642c06628f0291d85327de59d0ac9eebf38491235fad2a9
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d01f1e5f73292a02bda479878337ad5b8f1adc40b41fbe06f7eee7bcb6a079bf
+oid sha256:ef55032dfe71f3daf4c65fe4fcd7a5c3cf43a95e9092e737865f37142e607f25
 size 4040701744