Instructions to use SL-AI/GRaPE-Flash with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SL-AI/GRaPE-Flash with Transformers:

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

pipe = pipeline("text-generation", model="SL-AI/GRaPE-Flash")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("SL-AI/GRaPE-Flash")
model = AutoModelForCausalLM.from_pretrained("SL-AI/GRaPE-Flash")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use SL-AI/GRaPE-Flash with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/SL-AI/GRaPE-Flash

SGLang

How to use SL-AI/GRaPE-Flash 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 "SL-AI/GRaPE-Flash" \
    --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": "SL-AI/GRaPE-Flash",
		"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 "SL-AI/GRaPE-Flash" \
        --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": "SL-AI/GRaPE-Flash",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use SL-AI/GRaPE-Flash with Docker Model Runner:
```
docker model run hf.co/SL-AI/GRaPE-Flash
```

Sweaterdog commited on Feb 24

Commit

008b609

verified ·

1 Parent(s): 8115646

Update README.md

Browse files

Files changed (1) hide show

README.md +11 -5

README.md CHANGED Viewed

@@ -15,8 +15,8 @@ _The **G**eneral **R**easoning **A**gent (for) **P**roject **E**xploration_
 # The GRaPE Family
 | Attribute | Size | Modalities | Domain |
 | :--- | :--- | :--- | :--- |
-| **GRaPE Pro** | 16B A4B | Text + Image + Video in, Text out | Complex Reasoning Tasks |
-| **GRaPE Flash** (this model) | 7B A1B | Text in, Text out | High-Speed Applications |
 | **GRaPE Mini** | 3B | Text + Image + Video in, Text out | Edge Deployment |
 ***
@@ -29,7 +29,7 @@ The GRaPE Family was trained on about **14 billion** tokens of data after pre-tr
 # Architecture
-* GRaPE Pro: Built on the `Qwen3 VL MoE` Architecture, allowing for long-context understanding and reasoning over visual tasks, as well as any text-based task. Allowing for deep understanding.
 * GRaPE Flash: Built on the `OlMoE` Architecture, allowing for incredibly fast speeds where it matters. Allows for retaining factual information, but lacks in logical tasks.
@@ -38,7 +38,7 @@ The GRaPE Family was trained on about **14 billion** tokens of data after pre-tr
 ***
 # Demos
-> *The following demos were ran for the model in this repo, GRaPE Flash, which does not have any reasoning abilities*
 `Make me a landing page in HTML for a music creation software named "Aurora Beats". Output only the HTML file. Make it amazing, and extravagant`
 ![image](https://cdn-uploads.huggingface.co/production/uploads/66960602f0ffd8e3a381106a/QuKyKsI4XvFcn3_7wRe-1.png)
@@ -133,4 +133,10 @@ The F# Mixolydian Scale is often used in songs where the composer wants a sense
 contrast to other scales due to its unique character. Remember to always approach your music with emotion and soul, and let the notes
 flow from your heart!
 </details>
-GRaPE Flash got this question wrong. The correct answer is `F#, G#, A#, B, C#, D#, and E`. It had listed the notes in the F# Major Scale.

 # The GRaPE Family
 | Attribute | Size | Modalities | Domain |
 | :--- | :--- | :--- | :--- |
+| **GRaPE Pro** | 30B A3B | Text + Image + Video in, Text out | Complex Reasoning Tasks |
+| **GRaPE Flash** | 7B A1B | Text in, Text out | High-Speed Applications |
 | **GRaPE Mini** | 3B | Text + Image + Video in, Text out | Edge Deployment |
 ***
 # Architecture
+* GRaPE Pro: Built on the `Qwen3.5` Architecture, allowing for long-context understanding and reasoning over visual tasks, as well as any text-based task. Allowing for deep understanding.
 * GRaPE Flash: Built on the `OlMoE` Architecture, allowing for incredibly fast speeds where it matters. Allows for retaining factual information, but lacks in logical tasks.
 ***
 # Demos
+> *The following demos were ran for the model in this repo, GRaPE Flash Base, which does not have any reasoning abilities*
 `Make me a landing page in HTML for a music creation software named "Aurora Beats". Output only the HTML file. Make it amazing, and extravagant`
 ![image](https://cdn-uploads.huggingface.co/production/uploads/66960602f0ffd8e3a381106a/QuKyKsI4XvFcn3_7wRe-1.png)
 contrast to other scales due to its unique character. Remember to always approach your music with emotion and soul, and let the notes
 flow from your heart!
 </details>
+GRaPE Flash got this question wrong. The correct answer is `F#, G#, A#, B, C#, D#, and E`. It had listed the notes in the F# Major Scale.
+# Notes
+Unsloth was a huge part of making GRaPE possible. Massive shout-out to the work they do.
+![Sloth_with_grapes](https://cdn-uploads.huggingface.co/production/uploads/66960602f0ffd8e3a381106a/yG8uh8Hy0Q78uN2eDHf--.png)