Run 8. Outer Step 75. Inner Step 121.

config.json

@@ -264,21 +264,23 @@
   "attention_bias": false,
   "attention_dropout": 0.0,
   "block_list": [
-    6396916,
-    6396924,
-    6396932,
-    6396941,
-    6396948,
-    6396957,
-    6396965,
-    6396973
+    6396998,
+    6397006,
+    6397013,
+    6397020,
+    6397027,
+    6397035,
+    6397043,
+    6397051,
+    6397060,
+    6397068
   ],
   "bos_token_id": 1,
   "eos_token_id": 2,
   "hidden_act": "silu",
   "hidden_size": 2048,
   "initializer_range": 0.02,
-  "inner_step": 95,
+  "inner_step": 121,
   "intermediate_size": 5632,
   "last_allreduce_block": 6396547,
   "max_position_embeddings": 2048,

gradients.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b069783f13b6baf942a6965531a7f7fea755262d3aa60817cd689dff0be36df3
+oid sha256:5d446fce34ebc2cc6a77b10d981e76a36b3756dd9828e98f85f8cc5051bcf5cd
 size 88389626

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:669b9be29a913dad75e66c5ec1898716231a6c5118b979fcbcbd28b497c74b55
+oid sha256:6283a2d03b387e951d48bd901f3f6fc89f408cf2b8c0b4b74ab98b479578ad65
 size 4924926906

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c824fa0e9b2cf9fe90f3cff59b526b2ef3e522080efc3e08c97f5ec7973d8604
+oid sha256:7f071a91bfd110c21e419a27ea54753f245e8a4390c80483330020dfe2d7104d
 size 4400216536