Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add the huggingface token parameter, and modify the file path in llama.cpp repo #761

Merged
merged 1 commit into from
Aug 19, 2024
Merged

Add the huggingface token parameter, and modify the file path in llama.cpp repo #761

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions convert_models/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,10 +19,10 @@ podman build -t converter .

## Quantize and Convert

You can run the conversion image directly with podman in the terminal. You just need to provide it with the huggingface model name you want to download, the quantization level you want to use and whether or not you want to keep the raw files after conversion.
You can run the conversion image directly with podman in the terminal. You just need to provide it with the huggingface model name you want to download, the quantization level you want to use and whether or not you want to keep the raw files after conversion. "HF_TOKEN" is optional, it is required for private models.

```bash
podman run -it --rm -v models:/converter/converted_models -e HF_MODEL_URL=<ORG/MODEL_NAME> -e QUANTIZATION=Q4_K_M -e KEEP_ORIGINAL_MODEL="False"
podman run -it --rm -v models:/converter/converted_models -e HF_MODEL_URL=<ORG/MODEL_NAME> -e HF_TOKEN=<YOUR_HF_TOKEN_ID> -e QUANTIZATION=Q4_K_M -e KEEP_ORIGINAL_MODEL="False" converter
```

You can also use the UI shown above to do the same.
Expand Down
2 changes: 2 additions & 0 deletions convert_models/download_huggingface.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,9 +3,11 @@

parser = argparse.ArgumentParser()
parser.add_argument("-m", "--model")
parser.add_argument("-t", "--token")
args = parser.parse_args()

snapshot_download(repo_id=args.model,
token=args.token,
local_dir=f"converted_models/{args.model}",
local_dir_use_symlinks=True,
cache_dir=f"converted_models/cache")
9 changes: 5 additions & 4 deletions convert_models/run.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,7 @@
#! /bin/bash

hf_model_url=${HF_MODEL_URL}
hf_token=${HF_TOKEN:="None"}
model_org=$(echo $hf_model_url | sed -n 's/\(.*\)\/\(.*\)/\1/p')
model_name=$(echo $hf_model_url | sed -n 's/\(.*\)\/\(.*\)/\2/p')
keep_orgi=${KEEP_ORIGINAL_MODEL}
Expand All @@ -15,11 +16,11 @@ if [ -e "/opt/app-root/src/converter/converted_models/cache/models--$model_org--
fi

echo "Downloading $hf_model_url"
python download_huggingface.py --model $hf_model_url
python llama.cpp/convert.py /opt/app-root/src/converter/converted_models/$hf_model_url
python llama.cpp/convert-hf-to-gguf.py /opt/app-root/src/converter/converted_models/$hf_model_url
python download_huggingface.py --model $hf_model_url --token $hf_token
python llama.cpp/examples/convert_legacy_llama.py /opt/app-root/src/converter/converted_models/$hf_model_url
python llama.cpp/convert_hf_to_gguf.py /opt/app-root/src/converter/converted_models/$hf_model_url
mkdir -p /opt/app-root/src/converter/converted_models/gguf/
llama.cpp/quantize /opt/app-root/src/converter/converted_models/$hf_model_url/ggml-model-f16.gguf /opt/app-root/src/converter/converted_models/gguf/$model_org-$model_name-${QUANTIZATION}.gguf ${QUANTIZATION}
llama.cpp/llama-quantize /opt/app-root/src/converter/converted_models/$hf_model_url/ggml-model-f16.gguf /opt/app-root/src/converter/converted_models/gguf/$model_org-$model_name-${QUANTIZATION}.gguf ${QUANTIZATION}
rm -rf /opt/app-root/src/converter/converted_models/$model_org

if [ $keep_orgi = "False" ]; then
Expand Down
8 changes: 6 additions & 2 deletions convert_models/ui.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,6 +22,9 @@

model_name = st.text_input(label="Enter a huggingface model url to convert",
placeholder="org/model_name")
token_id = st.text_input(label="Enter your huggingface token (optional)",
help="huggingface token is required for private model"
) or "None"
keep_files = st.checkbox("Keep huggingface model files after conversion?")
submit_button = st.button(label="submit")
if submit_button:
Expand All @@ -30,8 +33,9 @@
"run",
"-it",
"--rm",
"-v", f"{volume}:/opt/app-root/src/converter/converted_models",
"-e", f"HF_MODEL_URL={model_name}" ,
"-v", f"{volume}:/converter/converted_models",
"-e", f"HF_MODEL_URL={model_name}",
"-e", f"HF_TOKEN={token_id}",
"-e", f"QUANTIZATION={quantization}",
"-e", f"KEEP_ORIGINAL_MODEL={keep_files}",
"converter"],stdout=subprocess.PIPE)
Expand Down