Outer Step 9. Inner Step 309. Batch Size 60

config.json

@@ -268,23 +268,23 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5318898,
-    5318900,
-    5318903,
-    5318905,
-    5318906,
-    5318909,
-    5318912,
-    5318915,
-    5318918,
-    5318921
+    5318948,
+    5318951,
+    5318953,
+    5318954,
+    5318957,
+    5318958,
+    5318960,
+    5318960,
+    5318964,
+    5318966
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 302,
+  "inner_step": 309,
   "inner_steps": 0,
   "last_allreduce_block": 5297756,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2f8bfddb1fdeb5aec156bbf15318620ca9447d9f24d3f347d13df99ee46ea591
+oid sha256:e4ad2a01d7a7480ef1ad23f35175507a6523d614d34747041c6c62220aec2f24
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3a5e6f8a49b5775b4e8231a71904d8b6842a90fda88c40a22f124b38b797a226
+oid sha256:3360c3e6d065290c7d028764c62ddd21d6ca581661cf53115eb9607e5743c13d
 size 4040701744