model annealed to lr 0

config.json

@@ -14,7 +14,7 @@
   "d_model": 4096,
   "embedding_dropout": 0.0,
   "embedding_size": 50304,
-  "eos_token_id": 0,
+  "eos_token_id": 50279,
   "flash_attention": true,
   "include_bias": false,
   "init_cutoff_factor": null,

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dfce4e7f9c66302437c7596f0f1ccffc92feea5764f37876079a7289afe705a5
+oid sha256:c691ecdf9ec32368e950af49421a699e321b6efad9d21387011c0ba9e985706f
 size 27552427238

special_tokens_map.json

@@ -1,4 +1,4 @@
 {
-  "eos_token": "|||IP_ADDRESS|||",
+  "eos_token": "<|endoftext|>",
   "pad_token": "<|padding|>"
 }

tokenizer_config.json

@@ -226,7 +226,7 @@
     }
   },
   "clean_up_tokenization_spaces": true,
-  "eos_token": "|||IP_ADDRESS|||",
+  "eos_token": "<|endoftext|>",
   "max_length": null,
   "model_max_length": 1000000000000000019884624838656,
   "pad_token": "<|padding|>",