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 Mar 18

Commit

52f6c82

verified ·

1 Parent(s): d922bf1

Update README.md

Browse files

Files changed (1) hide show

README.md +11 -3

README.md CHANGED Viewed

@@ -24,14 +24,20 @@ _The **G**eneral **R**easoning **A**gent (for) **P**roject **E**xploration_
 # Capabilities
 The GRaPE Family was trained on about **14 billion** tokens of data after pre-training. About half was code related tasks, with the rest being heavy on STEAM. Ensuring the model has a sound logical basis.
-> Disclaimer: *GRaPE Flash does not have thinking capabilities, primarily in favor of instant responses.*
 ***
 # Architecture
 * 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.
-* GRaPE Mini: Built on the `Qwen3 VL` Architecture, allowing for edge case deployments, where logic nor speed can be sacrificed.
 * GRaPE Nano: Built on the `LFM 2` Architecture, allowing for the fastest speed, and the most knowledge in the tiniest package.
@@ -43,4 +49,6 @@ The GRaPE Family started all the way back in August of 2025, meaning these model
 GRaPE 2 will come sooner than the GRaPE 1 family had, and will show multiple improvements.
-There are no benchmarks for GRaPE 1 Models due to the costly nature of running them, as well as prioritization of newer models.

 # Capabilities
 The GRaPE Family was trained on about **14 billion** tokens of data after pre-training. About half was code related tasks, with the rest being heavy on STEAM. Ensuring the model has a sound logical basis.
+> *GRaPE Flash does not have thinking capabilities, primarily in favor of instant responses.*
 ***
+GRaPE Flash and Nano are monomodal models, only accepting text. GRaPE Mini being trained most recently supports image and video inputs.
+# GRaPE Flash as a Model
+GRaPE Flash was designed for one thing: Speed. If you need a model that can quickly fill in tons of JSON data, this is your model. GRaPE Flash was chosen to **not** recieve thinking training as the model architecture would not benefit from it.
 # Architecture
 * 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.
+* GRaPE Mini: Built on the `Qwen3 VL` Architecture, allowing for edge case deployments, where logic cannot be sacrificed.
 * GRaPE Nano: Built on the `LFM 2` Architecture, allowing for the fastest speed, and the most knowledge in the tiniest package.
 GRaPE 2 will come sooner than the GRaPE 1 family had, and will show multiple improvements.
+There are no benchmarks for GRaPE 1 Models due to the costly nature of running them, as well as prioritization of newer models.
+Updates for GRaPE 2 models will be posted here on Huggingface, as well as [Skinnertopia](https://www.skinnertopia.com/)