Run 11. Outer Step 232. Inner Step 0.

config.json

@@ -264,8 +264,7 @@
   "attention_bias": false,
   "attention_dropout": 0.0,
   "block_list": [
-    6671479,
-    6671508
+    6671540
   ],
   "bos_token_id": 1,
   "dtype": "float32",
@@ -274,7 +273,7 @@
   "hidden_act": "silu",
   "hidden_size": 2048,
   "initializer_range": 0.02,
-  "inner_step": 45,
+  "inner_step": 0,
   "intermediate_size": 5632,
   "last_allreduce_block": 6670848,
   "max_position_embeddings": 2048,

gradients.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:298ddb99a5866f0fb1c320295dda1e36d2c99f14837256a0fd9e3a787c3340fa
+oid sha256:9fc47a294a094da5259ef6a1c7a213121f9892eca22432bf5ce61783e906aa2e
 size 88389626

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e613455549e7997701240baa56ea918be979db29b4c734dd9867cd62c0e52827
+oid sha256:28264b601d397df5c39e084499342468a4a879b6f6c0e873f748d66f60b873ca
 size 4924926906

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b315d3b6592cf7d728260a48e104bbb849b3dd076320d8aff67a558545ba210c
+oid sha256:b1de035fd07577aeb7d1516567fd6fb081d82f4965db1ee314e5b7fbbfdde861
 size 4400216536