Update max_position_embeddings to 128k context size instead of 32k (#4)

Browse files

- Update max_position_embeddings to 128k context size instead of 32k(https://huggingface.co/mistralai/Mistral-Large-Instruct-2407/commit/5c9ce5b5f7a7ad62d03e8c66c719b66d586de26b) (aaf19738edbe4f83f277c63d00ba5cae17b16088)

Co-authored-by: Junjae Lee <qwopqwop@users.noreply.huggingface.co>

Files changed (1) hide show

config.json +1 -1

config.json CHANGED Viewed

@@ -11,7 +11,7 @@
   "hidden_size": 12288,
   "initializer_range": 0.02,
   "intermediate_size": 28672,
-  "max_position_embeddings": 32768,
   "model_type": "mistral",
   "num_attention_heads": 96,
   "num_hidden_layers": 88,

   "hidden_size": 12288,
   "initializer_range": 0.02,
   "intermediate_size": 28672,
+  "max_position_embeddings": 131072,
   "model_type": "mistral",
   "num_attention_heads": 96,
   "num_hidden_layers": 88,