Instructions to use sinnybb/exact_random_selfdistill_grpo with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use sinnybb/exact_random_selfdistill_grpo with Transformers:

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

pipe = pipeline("image-text-to-text", model="sinnybb/exact_random_selfdistill_grpo")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("sinnybb/exact_random_selfdistill_grpo")
model = AutoModelForMultimodalLM.from_pretrained("sinnybb/exact_random_selfdistill_grpo", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.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=256)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use sinnybb/exact_random_selfdistill_grpo with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "sinnybb/exact_random_selfdistill_grpo"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "sinnybb/exact_random_selfdistill_grpo",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/sinnybb/exact_random_selfdistill_grpo

SGLang

How to use sinnybb/exact_random_selfdistill_grpo 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 "sinnybb/exact_random_selfdistill_grpo" \
    --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": "sinnybb/exact_random_selfdistill_grpo",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "sinnybb/exact_random_selfdistill_grpo" \
        --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": "sinnybb/exact_random_selfdistill_grpo",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use sinnybb/exact_random_selfdistill_grpo with Docker Model Runner:
```
docker model run hf.co/sinnybb/exact_random_selfdistill_grpo
```

Teacher exact39K random curriculum Stage3 — GRPO

This repository contains the final, merged GRPO checkpoint of our_stage1_ep2_teacher_exact39k_random42_stage3_exact39k_ep1.

The starting checkpoint is the Teacher exact39K Stage3 model: Stage1 training for 2 epochs, a seed-42 random curriculum of three 13K subsets with 2 epochs per subset, followed by Stage3 training on teacher-correct examples for 1 epoch.

The published checkpoint is GRPO global step 12, the final saved step of the completed run. The four distributed actor shards were merged into four Hugging Face safetensors shards. Tokenizer and image processor files are included.

RL configuration

Setting	Value
Algorithm	GRPO
Training data	Thyme-RL, maximum 3,200 training examples
Epochs	1
GPUs	4
Latent size	10
Rollouts per prompt	8
Sampling temperature	0.5
Learning rate	1e-6
KL coefficient	0.01
Monet RL sigma	10.0
Final saved step	12

The architecture is Qwen2.5-VL. Monet latent reasoning requires the Monet inference/runtime implementation used by the training project.

Downloads last month: -

Safetensors

Model size

8B params

Tensor type

BF16