Run 4. Outer Step 50. Inner Step 9.

config.json

@@ -268,32 +268,31 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5571810,
-    5571812,
-    5571814,
-    5571815,
-    5571816,
-    5571818,
-    5571819,
-    5571821,
-    5571823,
-    5571825,
-    5571827,
-    5571829,
-    5571830,
-    5571831,
-    5571832,
-    5571833,
-    5571834,
-    5571835,
-    5571837
+    5571839,
+    5571841,
+    5571843,
+    5571845,
+    5571846,
+    5571847,
+    5571848,
+    5571849,
+    5571851,
+    5571852,
+    5571853,
+    5571855,
+    5571856,
+    5571858,
+    5571860,
+    5571862,
+    5571863,
+    5571865
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 8,
+  "inner_step": 9,
   "inner_steps": 0,
   "last_allreduce_block": 5571611,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:94937ec469a2553fb104a019c8edd4733e3847fd72cf81e0fe248bf0e8714af5
+oid sha256:fef197b3adba55e1f2227090cd620eed95e9cf2cdd26c6c667e9b2140dc2291a
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:af6df267b6480a12d5c699d1ecc6153a4f3b8122ca0d32ae7fdd8a35f8d89eec
+oid sha256:cc0b2622821773a0d2d3140772e952de6199fb1a86b983d8d7fe5bac18257f20
 size 4040701744