Run 4. Outer Step 50. Inner Step 9.

config.json

@@ -268,36 +268,30 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5571801,
-    5571803,
-    5571804,
-    5571806,
-    5571807,
-    5571808,
-    5571810,
-    5571811,
-    5571812,
-    5571813,
-    5571814,
-    5571816,
-    5571818,
-    5571820,
-    5571821,
-    5571823,
-    5571824,
-    5571824,
-    5571825,
-    5571827,
-    5571828,
-    5571829,
-    5571830
+    5571832,
+    5571834,
+    5571835,
+    5571836,
+    5571836,
+    5571838,
+    5571839,
+    5571840,
+    5571842,
+    5571843,
+    5571845,
+    5571846,
+    5571848,
+    5571850,
+    5571851,
+    5571853,
+    5571854
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 7,
+  "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 sha256:88c03c430c8c03d57ae1e2173aea9c670149f68fb9157bd74a80510057be40fd
+oid sha256:6af042dd22fdb5d636501c5541404ff58c6476ac83ac99c9bcc51be0a11bedc4
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:18de05fc3d1e3f03e1e5893c87ce32417ab8036e2da92411a919ed572a6825d3
+oid sha256:e0760b860d6d5ae102e2a6515d4a3f1e0203e471e39f81d51a92b12fe7c475c5
 size 4040701744