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

Libraries

How to use pianzhikuang/VisAlloc_Final with Transformers:

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

pipe = pipeline("image-text-to-text", model="pianzhikuang/VisAlloc_Final")
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("pianzhikuang/VisAlloc_Final")
model = AutoModelForMultimodalLM.from_pretrained("pianzhikuang/VisAlloc_Final", 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 pianzhikuang/VisAlloc_Final with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "pianzhikuang/VisAlloc_Final"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "pianzhikuang/VisAlloc_Final",
		"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/pianzhikuang/VisAlloc_Final

SGLang

How to use pianzhikuang/VisAlloc_Final 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 "pianzhikuang/VisAlloc_Final" \
    --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": "pianzhikuang/VisAlloc_Final",
		"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 "pianzhikuang/VisAlloc_Final" \
        --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": "pianzhikuang/VisAlloc_Final",
		"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 pianzhikuang/VisAlloc_Final with Docker Model Runner:
```
docker model run hf.co/pianzhikuang/VisAlloc_Final
```

VisAlloc_Final — Qwen3.5-9B

Full BF16 export of global step 320 from the single-node, 8-GPU visalloc-9b-arma-bundle-1x8-b16 training run. This was the latest fully saved checkpoint when this export was selected on October 13, 2026; training was still in progress at selection time.

The model contains 9,409,813,744 parameters. The eight FSDP actor shards were merged into a complete Hugging Face model with its tokenizer, image processor, generation configuration, and perception chat template.

Base model: Qwen/Qwen3.5-9B
Base revision: c202236235762e1c871ad0ccb60c8ee5ba337b9a
Training code: VisAlloc, revision d73ee6696494efdb18e6a62c849f0359394ab39e
Training data: 5,795 samples, using the repository's prepared ARMA prior bundle
Global prompt batch / PPO minibatch: 16 / 16
Checkpoint: global_step_320

Usage

The export was validated with Transformers 5.5.0 and PyTorch 2.10.0. Keep the supplied processor and perception chat template.

import torch
from PIL import Image
from transformers import AutoProcessor, Qwen3_5ForConditionalGeneration

model_id = "pianzhikuang/VisAlloc_Final"
processor = AutoProcessor.from_pretrained(model_id)
model = Qwen3_5ForConditionalGeneration.from_pretrained(
    model_id, dtype=torch.bfloat16, device_map="auto"
).eval()

messages = [{"role": "user", "content": [
    {"type": "image", "image": Image.open("image.jpg").convert("RGB")},
    {"type": "text", "text": "Describe this image."},
]}]
inputs = processor.apply_chat_template(
    messages, tokenize=True, add_generation_prompt=True,
    return_dict=True, return_tensors="pt",
).to(model.device)
with torch.inference_mode():
    output = model.generate(**inputs, max_new_tokens=256, do_sample=False)
print(processor.decode(
    output[0, inputs["input_ids"].shape[1]:], skip_special_tokens=True
))

Integrity

training_metadata.json records the checkpoint, base revision, parameter count, template hash, and export validation. The weight file's SHA-256 is recorded in SHA256SUMS. All 760 tensors were checked for expected names, shapes, BF16 dtype, and finite values. The processor and chat template were validated locally.

The base model's Apache-2.0 license is included in LICENSE.

Downloads last month: 29

Safetensors

Model size

9B params

Tensor type

BF16

Model tree for pianzhikuang/VisAlloc_Final

Base model

Qwen/Qwen3.5-9B-Base

Finetuned

Qwen/Qwen3.5-9B

Finetuned

(1148)

this model