Instructions to use XHToken/Spark-X2.5-4B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use XHToken/Spark-X2.5-4B with Transformers:

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

pipe = pipeline("text-generation", model="XHToken/Spark-X2.5-4B", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("XHToken/Spark-X2.5-4B", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use XHToken/Spark-X2.5-4B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "XHToken/Spark-X2.5-4B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "XHToken/Spark-X2.5-4B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/XHToken/Spark-X2.5-4B

SGLang

How to use XHToken/Spark-X2.5-4B 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 "XHToken/Spark-X2.5-4B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "XHToken/Spark-X2.5-4B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "XHToken/Spark-X2.5-4B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "XHToken/Spark-X2.5-4B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use XHToken/Spark-X2.5-4B with Docker Model Runner:
```
docker model run hf.co/XHToken/Spark-X2.5-4B
```

feat(attention): add KV-BSS (Key-Value Binding Softmax Sharpening) and attention haze suppression

#15

by F-Labs - opened 2 days ago

base: refs/heads/main

←

from: refs/pr/15

Discussion Files changed

+490

-483

F-Labs

2 days ago

This PR implements KV-BSS (Key-Value Binding Softmax Sharpening) in modeling_spark.py:

Backwards Compatible: Defaults to key_value_binding_sharpening=False and attention_sink_suppression=False, preserving 100% exact baseline behavior unless explicitly enabled in config.json.
Key-Value Binding Sharpening: When enabled (key_value_binding_sharpening=True), scales attention logits with kv_focus_factor (default 1.10) to steepen peak Softmax probabilities around exact antecedent bindings (e.g. structured JSON, dictionaries, AST variable mapping).
Attention Haze Truncation: Truncates parasitic background attention haze (< max_logits - 12.0) to -inf before Softmax, preventing probability mass diffusion over long sequence windows.

Refer to Community Discussion #14 for empirical derivations and benchmarks:
https://huggingface.co/XHToken/Spark-X2.5-4B/discussions/14
And proof-of-concept release:
https://huggingface.co/F-Labs/Spark-X2.5-4B-Hadamard-GSQ

feat(attention): add KV-BSS (Key-Value Binding Softmax Sharpening) and attention haze suppression943a7700

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

Tap or paste here to upload images

Ready to merge

This branch is ready to get merged automatically.

· Sign up or log in to comment