download_falcon.md

Download Falcon weights

UAE's Technology Innovation Institute has open-sourced Falcon LLM. It is trained on RefinedWeb enhanced with curated corpora Weights are released under the Apache 2.0 license.

The first Falcon release includes a base model and an instruction tuned model of sizes 7B and 40B called falcon-7b-instruct and falcon-40b-instruct. Recently, checkpoints for 180B parameter models were added as well; the 180B instruction tuned model is called falcon-180B-chat and similar to the falcon-40b-instruct architecture except for its larger size.

To see all the available checkpoints for Falcon, run:

python scripts/download.py | grep falcon

which will print

tiiuae/falcon-7b
tiiuae/falcon-7b-instruct
tiiuae/falcon-40b
tiiuae/falcon-40b-instruct
tiiuae/falcon-180B
tiiuae/falcon-180B-chat

In order to use a specific Falcon checkpoint, for instance falcon-7b, download the weights and convert the checkpoint to the lit-gpt format:

pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'

python scripts/download.py --repo_id tiiuae/falcon-7b

python scripts/convert_hf_checkpoint.py --checkpoint_dir checkpoints/tiiuae/falcon-7b

By default, the convert_hf_checkpoint step will use the data type of the HF checkpoint's parameters. In cases where RAM or disk size is constrained, it might be useful to pass --dtype bfloat16 to convert all parameters into this smaller precision before continuing.

You're done! To execute the model just run:

pip install tokenizers

python generate/base.py --prompt "Hello, my name is" --checkpoint_dir checkpoints/tiiuae/falcon-7b

or