Instructions to use Eigeen/Xwin-LM-13B-V0.2-exl2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Eigeen/Xwin-LM-13B-V0.2-exl2 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="Eigeen/Xwin-LM-13B-V0.2-exl2")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("Eigeen/Xwin-LM-13B-V0.2-exl2") model = AutoModelForCausalLM.from_pretrained("Eigeen/Xwin-LM-13B-V0.2-exl2", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use Eigeen/Xwin-LM-13B-V0.2-exl2 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Eigeen/Xwin-LM-13B-V0.2-exl2" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Eigeen/Xwin-LM-13B-V0.2-exl2", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/Eigeen/Xwin-LM-13B-V0.2-exl2
- SGLang
How to use Eigeen/Xwin-LM-13B-V0.2-exl2 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 "Eigeen/Xwin-LM-13B-V0.2-exl2" \ --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": "Eigeen/Xwin-LM-13B-V0.2-exl2", "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 "Eigeen/Xwin-LM-13B-V0.2-exl2" \ --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": "Eigeen/Xwin-LM-13B-V0.2-exl2", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use Eigeen/Xwin-LM-13B-V0.2-exl2 with Docker Model Runner:
docker model run hf.co/Eigeen/Xwin-LM-13B-V0.2-exl2
Can you make a 5.5 bpw?
#2
by BigStationW - opened
With flash attention 2, it's now possible to go bigger than 5bpw on 12gb vram cards but 6bpw is too big, the 5.5 bpw would be the sweet spot!
Good idea, I'll do that.
Updated 5.5bpw.
Thanks dude I really appreciate!
Updated 5.5bpw.
It'd be nice to list it in the model card too as it's hard to notice.
And maybe to point out that 3bpw is the main branch to ensure people don't accidentally download it thinking that default was 4bpw.