2023-07-18 19:29:56 +02:00
|
|
|
# LLaMA-v2
|
|
|
|
|
|
|
|
To convert LLaMA-v2 from the `.pth` format provided by Meta to transformers format, follow the steps below:
|
|
|
|
|
|
|
|
1) `cd` into your `llama` folder (the one containing `download.sh` and the models that you downloaded):
|
|
|
|
|
|
|
|
```
|
|
|
|
cd llama
|
|
|
|
```
|
|
|
|
|
2023-07-18 22:21:18 +02:00
|
|
|
2) Clone the transformers library:
|
2023-07-18 19:29:56 +02:00
|
|
|
|
|
|
|
```
|
2023-07-18 22:21:18 +02:00
|
|
|
git clone 'https://github.com/huggingface/transformers'
|
2023-07-18 19:29:56 +02:00
|
|
|
|
|
|
|
```
|
|
|
|
|
|
|
|
3) Create symbolic links from the downloaded folders to names that the conversion script can recognize:
|
|
|
|
|
|
|
|
```
|
|
|
|
ln -s llama-2-7b 7B
|
|
|
|
ln -s llama-2-13b 13B
|
|
|
|
```
|
|
|
|
|
|
|
|
4) Do the conversions:
|
|
|
|
|
|
|
|
```
|
|
|
|
mkdir llama-2-7b-hf llama-2-13b-hf
|
|
|
|
python ./transformers/src/transformers/models/llama/convert_llama_weights_to_hf.py --input_dir . --model_size 7B --output_dir llama-2-7b-hf --safe_serialization true
|
|
|
|
python ./transformers/src/transformers/models/llama/convert_llama_weights_to_hf.py --input_dir . --model_size 13B --output_dir llama-2-13b-hf --safe_serialization true
|
|
|
|
```
|
|
|
|
|
|
|
|
5) Move the output folders inside `text-generation-webui/models`
|
|
|
|
|
|
|
|
6) Have fun
|