Run 4. Outer Step 38. Inner Step 3.

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": [
-    5552682,
-    5552686,
-    5552691,
-    5552695,
-    5552700
-  ],
+  "block_list": [],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 2,
+  "inner_step": 3,
   "inner_steps": 0,
   "last_allreduce_block": 5552497,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8e4e7516eab5ec83bedbc2a96b8b65318ccc33c000fb85574aa46826b3c5e0ee
+oid sha256:8dc1411befb8fd7bb9f200437b6ec85c0776fc6f18d6e0cf4b7d6672a694707e
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b4a015aceb47b9050791bda16077d213d06b7c207268cc36baf5556fffeaf074
+oid sha256:1e62fa68f710d07b27abee9facc10aa48796250f197ac082bf0e21a59edf8f67
 size 4040701744