update to new format

README.md

@@ -28,7 +28,7 @@ export HF_HUB_ENABLE_HF_TRANSFER=1
 huggingface-cli download --local-dir Llama-2-7b-chat-mlx mlx-llama/Llama-2-7b-chat-mlx
 
 # Run example
-python mlx-examples/llama/llama.py Llama-2-7b-chat-mlx/Llama-2-7b-chat.npz Llama-2-7b-chat-mlx/tokenizer.model "My name is "
+python mlx-examples/llama/llama.py Llama-2-7b-chat-mlx/ Llama-2-7b-chat-mlx/tokenizer.model "My name is "
 ```
 
 Please, refer to the [original model card](https://huggingface.co/meta-llama/Llama-2-7b-chat) for details on Llama 2.

params.json

@@ -0,0 +1 @@
+{"dim": 4096, "multiple_of": 256, "n_heads": 32, "n_layers": 32, "norm_eps": 1e-05, "vocab_size": -1}

Llama-2-7b-chat.npz → weights.npz

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bcc00a339f8578fde60b328824dc5fcfe0dbab098dc46ca4872cce83f9398d10
-size 13476918398
+oid sha256:68122068194be1051286a36314577f8b1910e4d36c8f280f3852bff9609d90de
+size 13476919308