Run 8. Outer Step 73. Inner Step 16.

config.json

@@ -264,15 +264,17 @@
   "attention_bias": false,
   "attention_dropout": 0.0,
   "block_list": [
-    6395209,
-    6395228
+    6395266,
+    6395284,
+    6395302,
+    6395319
   ],
   "bos_token_id": 1,
   "eos_token_id": 2,
   "hidden_act": "silu",
   "hidden_size": 2048,
   "initializer_range": 0.02,
-  "inner_step": 4,
+  "inner_step": 16,
   "intermediate_size": 5632,
   "last_allreduce_block": 6395177,
   "max_position_embeddings": 2048,

gradients.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:daa855b681c3ff8bae560400361edcc42e7ada138362d4662002e947763a10ac
+oid sha256:417afbb2ff8fe73fdf39226a35ac78aebef79401d9d5b65f985ccb1e365ff7a5
 size 88389626

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3c91a7daebec057e008dd31bcc538184c190628acbac4b11eee6b3078bdc30bb
+oid sha256:8e5e28fb72ac482a9f3b68a6283ae94e5a6a910133b71192d08a0132b15c8ab3
 size 4924926906

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2847bc5fb38f7726f681c3c4e65b26c6f076f6b6bfb5166910bf8f5d5e59d2a5
+oid sha256:4aeca1c3dca63af775a5aa3d069132e46fc639b42d32fdd759decb717a03a946
 size 4400216536