Instructions to use RISys-Lab/RedSage-Qwen3-8B-Base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use RISys-Lab/RedSage-Qwen3-8B-Base with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="RISys-Lab/RedSage-Qwen3-8B-Base")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("RISys-Lab/RedSage-Qwen3-8B-Base")
model = AutoModelForCausalLM.from_pretrained("RISys-Lab/RedSage-Qwen3-8B-Base", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use RISys-Lab/RedSage-Qwen3-8B-Base with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "RISys-Lab/RedSage-Qwen3-8B-Base"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "RISys-Lab/RedSage-Qwen3-8B-Base",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/RISys-Lab/RedSage-Qwen3-8B-Base

SGLang

How to use RISys-Lab/RedSage-Qwen3-8B-Base 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 "RISys-Lab/RedSage-Qwen3-8B-Base" \
    --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": "RISys-Lab/RedSage-Qwen3-8B-Base",
		"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 "RISys-Lab/RedSage-Qwen3-8B-Base" \
        --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": "RISys-Lab/RedSage-Qwen3-8B-Base",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use RISys-Lab/RedSage-Qwen3-8B-Base with Docker Model Runner:
```
docker model run hf.co/RISys-Lab/RedSage-Qwen3-8B-Base
```

RedSage-Dump (~700M Tokens)，we could not locate a download link for RedSage-Dump

by quboshi - opened Sep 18

Discussion

quboshi

Sep 18

Thank you for sharing RedSage and releasing the models, datasets, and evaluation code.
We are preparing to reproduce the complete RedSage training pipeline and independently evaluate the resulting models. We have downloaded RedSage-CFW, RedSage-Seed, RedSage-Conv, and the released model checkpoints.

However, we could not locate a download link for RedSage-Dump, described in Appendix A.2 and Table 9 as approximately 459,473 documents and 700 million tokens. We understand that this corpus is used alongside RedSage-Seed in the continued pretraining stage that produces RedSage-Base.

Could you please share the dataset or advise us on how to request access? Ideally, we would like the exact processed snapshot used in the paper, together with any version information and preprocessing instructions needed for reproducibility.

If the complete corpus cannot be distributed, a source manifest, collection dates, and reconstruction scripts would also be helpful. We would additionally appreciate the training configuration for the Seed + Dump stage, if available.

Thank you for your time and support.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment