Outer Step 8. Inner Step 145. Batch Size 380

config.json

@@ -268,23 +268,23 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5292227,
-    5292230,
-    5292234,
-    5292237,
-    5292241,
-    5292244,
-    5292248,
-    5292251,
-    5292255,
-    5292258
+    5292300,
+    5292304,
+    5292307,
+    5292311,
+    5292315,
+    5292318,
+    5292322,
+    5292325,
+    5292329,
+    5292332
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 134,
+  "inner_step": 145,
   "inner_steps": 0,
   "last_allreduce_block": 5283755,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:825cff515add06d1831397749b0b817ff8bfc6970bb9d0437fba7e974fbb7979
+oid sha256:faea15b8935051443d450a788f72d36ef90bed13c214ed0d7b52f76d2657517a
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:993384b219bcea3a9154eeb77344649b17a20140c4ffed7d7cf3ea88ca3eb13d
+oid sha256:a29e21fa94a6ddebf7a0d3ec18eac5b06a4190c941406671acab25d9f88bb6f7
 size 4040701744