Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch
Paper • 2311.03099 • Published • 37
How to use vishnukv/newmerge with Transformers:
# Use a pipeline as a high-level helper
from transformers import pipeline
pipe = pipeline("text-generation", model="vishnukv/newmerge") # Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("vishnukv/newmerge")
model = AutoModelForCausalLM.from_pretrained("vishnukv/newmerge", device_map="auto")How to use vishnukv/newmerge with vLLM:
# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "vishnukv/newmerge"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
-H "Content-Type: application/json" \
--data '{
"model": "vishnukv/newmerge",
"prompt": "Once upon a time,",
"max_tokens": 512,
"temperature": 0.5
}'docker model run hf.co/vishnukv/newmerge
How to use vishnukv/newmerge with SGLang:
# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
--model-path "vishnukv/newmerge" \
--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": "vishnukv/newmerge",
"prompt": "Once upon a time,",
"max_tokens": 512,
"temperature": 0.5
}'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 "vishnukv/newmerge" \
--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": "vishnukv/newmerge",
"prompt": "Once upon a time,",
"max_tokens": 512,
"temperature": 0.5
}'How to use vishnukv/newmerge with Docker Model Runner:
docker model run hf.co/vishnukv/newmerge
This is a merge of pre-trained language models created using mergekit.
This model was merged using the DARE TIES merge method using uukuguy/speechless-mistral-dolphin-orca-platypus-samantha-7b as a base.
The following models were included in the merge: