Run 4. Outer Step 15. Inner Step 72.

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "distributed/optimized-gpt2-1b",
+  "_name_or_path": "thieu86/SN3812",
   "activation_function": "gelu_new",
   "all_reduce_scores": {
     "0": "NON_PARTICIPATING",
@@ -268,19 +268,18 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5515236,
-    5515239,
-    5515243,
-    5515247,
-    5515250,
-    5515254
+    5515284,
+    5515314,
+    5515318,
+    5515321,
+    5515325
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 0,
+  "inner_step": 72,
   "inner_steps": 0,
   "last_allreduce_block": 5513324,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d4e34473898e08e1885f44f15e1e842cceba2db4b7020e993ebcf42722ee2e51
+oid sha256:aea427d6c263209a47573a332dee9836fd9dc8f3e559f9cc3089ce9a3251a521
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0ec0e9250a2a6d8a75fdcb947e49656cdf726fbb28721e58eaa2bf73f730d1b0
+oid sha256:bbef75260785502960ec235987b66c3eb208402b972b74402f121320679d451f
 size 4040701744