We are unlocking the power of large language models. Our latest version of Llama is now accessible to individuals, creators, researchers, and businesses of all sizes so that they can experiment, innovate, and scale their ideas responsibly.
This release includes model weights and starting code for pre-trained and fine-tuned Llama language models — ranging from 7B to 70B parameters.
This repository is intended as a minimal example to load Llama 2 models and run inference. For more detailed examples leveraging Hugging Face, see llama-recipes.
See UPDATES.md. Also for a running list of frequently asked questions, see here.
In order to download the model weights and tokenizer, please visit the Meta website and accept our License.
Once your request is approved, you will receive a signed URL over email. Then run the download.sh script, passing the URL provided when prompted to start the download.
Pre-requisites: Make sure you have wget and md5sum installed. Then run the script: ./download.sh.
Keep in mind that the links expire after 24 hours and a certain amount of downloads. If you start seeing errors such as 403: Forbidden, you can always re-request a link.
We are also providing downloads on Hugging Face. You can request access to the models by acknowledging the license and filling the form in the model card of a repo. After doing so, you should get access to all the Llama models of a version (Code Llama, Llama 2, or Llama Guard) within 1 hour.
You can follow the steps below to quickly get up and running with Llama 2 models. These steps will let you run quick inference locally. For more examples, see the Llama 2 recipes repository.
-
In a conda env with PyTorch / CUDA available clone and download this repository.
-
In the top-level directory run:
pip install -e . -
Visit the Meta website and register to download the model/s.
-
Once registered, you will get an email with a URL to download the models. You will need this URL when you run the download.sh script.
-
Once you get the email, navigate to your downloaded llama repository and run the download.sh script.
- Make sure to grant execution permissions to the download.sh script
- During this process, you will be prompted to enter the URL from the email.
- Do not use the “Copy Link” option but rather make sure to manually copy the link from the email.
-
Once the model/s you want have been downloaded, you can run the model locally using the command below:
torchrun --nproc_per_node 1 example_chat_completion.py \
--ckpt_dir llama-2-7b-chat/ \
--tokenizer_path tokenizer.model \
--max_seq_len 512 --max_batch_size 6Note
- Replace
llama-2-7b-chat/with the path to your checkpoint directory andtokenizer.modelwith the path to your tokenizer model. - The
–nproc_per_nodeshould be set to the MP value for the model you are using. - Adjust the
max_seq_lenandmax_batch_sizeparameters as needed. - This example runs the example_chat_completion.py found in this repository but you can change that to a different .py file.
Different models require different model-parallel (MP) values:
| Model | MP |
|---|---|
| 7B | 1 |
| 13B | 2 |
| 70B | 8 |
All models support sequence length up to 4096 tokens, but we pre-allocate the cache according to max_seq_len and max_batch_size values. So set those according to your hardware.
These models are not finetuned for chat or Q&A. They should be prompted so that the expected answer is the natural continuation of the prompt.
See example_text_completion.py for some examples. To illustrate, see the command below to run it with the llama-2-7b model (nproc_per_node needs to be set to the MP value):
torchrun --nproc_per_node 1 example_text_completion.py \
--ckpt_dir llama-2-7b/ \
--tokenizer_path tokenizer.model \
--max_seq_len 128 --max_batch_size 4
The fine-tuned models were trained for dialogue applications. To get the expected features and performance for them, a specific formatting defined in chat_completion
needs to be followed, including the INST and <<SYS>> tags, BOS and EOS tokens, and the whitespaces and breaklines in between (we recommend calling strip() on inputs to avoid double-spaces).
You can also deploy additional classifiers for filtering out inputs and outputs that are deemed unsafe. See the llama-recipes repo for an example of how to add a safety checker to the inputs and outputs of your inference code.
Examples using llama-2-7b-chat:
torchrun --nproc_per_node 1 example_chat_completion.py \
--ckpt_dir llama-2-7b-chat/ \
--tokenizer_path tokenizer.model \
--max_seq_len 512 --max_batch_size 6
Llama 2 is a new technology that carries potential risks with use. Testing conducted to date has not — and could not — cover all scenarios. In order to help developers address these risks, we have created the Responsible Use Guide. More details can be found in our research paper as well.
Please report any software “bug”, or other problems with the models through one of the following means:
- Reporting issues with the model: github.com/facebookresearch/llama
- Reporting risky content generated by the model: developers.facebook.com/llama_output_feedback
- Reporting bugs and security concerns: facebook.com/whitehat/info
See MODEL_CARD.md.
Our model and weights are licensed for both researchers and commercial entities, upholding the principles of openness. Our mission is to empower individuals, and industry through this opportunity, while fostering an environment of discovery and ethical AI advancements.
See the LICENSE file, as well as our accompanying Acceptable Use Policy
For common questions, the FAQ can be found here which will be kept up to date over time as new questions arise.
The repo for the original llama release is in the llama_v1 branch.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent
