Instructions to use 0xSero/gemma-moe-reap with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use 0xSero/gemma-moe-reap with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="0xSero/gemma-moe-reap")# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("0xSero/gemma-moe-reap", dtype="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use 0xSero/gemma-moe-reap with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "0xSero/gemma-moe-reap" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "0xSero/gemma-moe-reap", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/0xSero/gemma-moe-reap
- SGLang
How to use 0xSero/gemma-moe-reap with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "0xSero/gemma-moe-reap" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "0xSero/gemma-moe-reap", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "0xSero/gemma-moe-reap" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "0xSero/gemma-moe-reap", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use 0xSero/gemma-moe-reap with Docker Model Runner:
docker model run hf.co/0xSero/gemma-moe-reap
A request if you are inclined
Hello helpful reaper fellow. I see you are doing the gemma moe. Fantastic! Thank you :)
Any chance you also do a 16-17b reap, like ~35%? Suspect this will fit on fully on 8gb cards like mine without too much additionally lobotomizing quantization.
If not, all good, and have a great day.
Hey! I will be trying a REAP model hopefully within the next 7 days just because of work.
Hey! I will be trying a REAP model hopefully within the next 7 days just because of work.
Sero has done a 20 and a 30. If you do happen to do a 35% reap, let us know! I feel like that's probs going to be optimal for my paltry 8gb. I got excited when I saw the Gemma 4 moe's because I though, just a little reap and that bad boy is going to fit in vram and run smooth.
(And probably be a stupid request given it's for work, but if you happen to heretic that thing afterward, even better, but do not expect anything at all).
Sigh, one day I'll be vram rich, lol. Thanks for chiming in :) Have a good one.