Run 3. Outer Step 6. Inner Step 0.

config.json

@@ -267,19 +267,13 @@
     "AutoConfig": "distributed/optimized-gpt2-500m--configuration_gpt_optimized.GPTOptimConfig",
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
-  "block_list": [
-    5384665,
-    5384669,
-    5384672,
-    5384676,
-    5384680
-  ],
+  "block_list": [],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 273,
+  "inner_step": 0,
   "inner_steps": 0,
   "last_allreduce_block": 5378990,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:81e9816fb385e77beacb1a9d7c8e2780e545a7069e425b7786165f85554732b5
+oid sha256:91074fcebac8cb9e44af7f0c432ab1d75be2b0c6f2edde6db4db7e726759416f
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3a5fd29f1fa6da1ef309f7ed0654e4db67db755e60b296619b381585aca20027
+oid sha256:ca0236b46615f797360b088a3bcfe5dd20a85f99747917208b8843811f8bd5b4
 size 4040701744