README.md

---
license: cc-by-sa-4.0
---

## emozilla/landmark-llama-7b

This model is an out-of-the-box ready version of the LLaMA-7B variant of [Landmark Attention](https://arxiv.org/abs/2305.16300).
The original code is modified from the [Landmark GitHub](https://github.com/epfml/landmark-attention) and the weights from [here](https://huggingface.co/epfml/landmark-attention-llama7b-wdiff).

As a LLaMA variant, this model may be subject to the LLaMA license.

### To use

```python
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
import torch

tokenizer = AutoTokenizer.from_pretrained("emozilla/landmark-llama-7b", use_fast=False)
model = AutoModelForCausalLM.from_pretrained("emozilla/landmark-llama-7b", \
  torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto")
pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)

print(pipe("Somebody once told me the world is gonna roll me", \
           max_new_tokens=256, temperature=0.8, do_sample=True))
```

You can configure the Landmark parameters by editing `mem_freq`, `mem_top_k`, `mem_max_seq_len`, and `mem_max_cache_size`.

```python
config = AutoConfig.from_pretrained("emozilla/landmark-llama-7b", trust_remote_code=True)
config.mem_top_k = 6
model = AutoModelForCausalLM.from_pretrained("emozilla/landmark-llama-7b", \
  torch_dtype=torch.bfloat16, trust_remote_code=True, device_map="auto", config=config)
```