Run 4. Outer Step 47. Inner Step 66.

config.json

@@ -268,33 +268,38 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5568762,
-    5568764,
-    5568766,
-    5568767,
-    5568767,
-    5568769,
-    5568770,
-    5568772,
-    5568773,
-    5568773,
-    5568774,
-    5568776,
-    5568776,
-    5568779,
-    5568780,
-    5568781,
-    5568783,
-    5568784,
-    5568785,
-    5568787
+    5568789,
+    5568791,
+    5568792,
+    5568794,
+    5568795,
+    5568797,
+    5568799,
+    5568800,
+    5568801,
+    5568802,
+    5568803,
+    5568804,
+    5568805,
+    5568806,
+    5568807,
+    5568808,
+    5568809,
+    5568810,
+    5568811,
+    5568812,
+    5568813,
+    5568813,
+    5568816,
+    5568817,
+    5568818
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 64,
+  "inner_step": 66,
   "inner_steps": 0,
   "last_allreduce_block": 5566279,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e317a2963c576c618f84ab24a2adeff955bba91c272276fa59d25e21001b7b8c
+oid sha256:1ee5d30a9e15468178d0940ef2c285d37704b748968b009141402b1b36cfc463
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:20e5d8e993780e11e09772f77529486f4f8f1ef8f1f6c98b820a5e72ec209e43
+oid sha256:b50eaa8a6a19853e20a2709671e19558503692f88df5178f855cd4d9cd98646c
 size 4040701744