Update README.md

Correction - doesn't need dummy tokens since vocab was reverted back to 32000

Upload added_tokens.json

Create README.md

Upload llama 2 files

Upload tokenizer

Upload LlamaForCausalLM

initial commit