Run 11. Outer Step 338. Inner Step 0.

config.json

@@ -264,8 +264,7 @@
   "attention_bias": false,
   "attention_dropout": 0.0,
   "block_list": [
-    6717350,
-    6717378
+    6717470
   ],
   "bos_token_id": 1,
   "dtype": "float32",
@@ -274,7 +273,7 @@
   "hidden_act": "silu",
   "hidden_size": 2048,
   "initializer_range": 0.02,
-  "inner_step": 36,
+  "inner_step": 0,
   "intermediate_size": 5632,
   "last_allreduce_block": 6717092,
   "max_position_embeddings": 2048,

gradients.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ae333b2807f837f77b3455323d4f112fedad29bc91113aa25b37cce0459ff73c
+oid sha256:ccaa8a461acad993713c02bcefbbe06ef68f0fe463b38d2160ac7dac14e77a51
 size 88389626

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:47304822314b181171d3caf6de78ce3af937fb1e8a988126b1d500ecfc54af51
+oid sha256:55755489c7763b079eaa363415d629b4f553cb46de41f558871f329a676fc011
 size 4924926906

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:32276f9f728cebc398c7bbe6c3d9ae7db011552fc32ee7dbe6d943c8222daa45
+oid sha256:ade3e4818681f0449df886babc377b9ff4c46f147dfd619b024567aeffacdd2a
 size 4400216536