Outer Step 2. Inner Step 90. Batch Size 52

config.json

@@ -268,23 +268,23 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5147483,
-    5147486,
-    5147492,
-    5147496,
-    5147500,
-    5147504,
-    5147508,
-    5147512,
-    5147516,
-    5147520
+    5147551,
+    5147555,
+    5147559,
+    5147564,
+    5147569,
+    5147573,
+    5147577,
+    5147581,
+    5147586,
+    5147590
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 80,
+  "inner_step": 90,
   "inner_steps": 0,
   "last_allreduce_block": 5146835,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6b51b13a298b5bb9b5fe33f1a6613bec103308d60e651fae0b091ef17ab129c8
+oid sha256:c30a4852477cdfe305a80f086a07db2630c7f5f83a15bd207b7b9568270f7222
 size 8081781770

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:78c98f66859dba294a9019542d1c70797fb968a5848478692da8e408d182c854
+oid sha256:3eee91d86437e019f72eba3e1705d8b5649d79bea9d5754279db7e48dcd2a03c
 size 4040701744