Instructions to use jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound with Transformers:

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

pipe = pipeline("text-generation", model="jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound")
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)

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound")
model = AutoModelForMultimodalLM.from_pretrained("jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound", 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=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound

SGLang

How to use jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound 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 "jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound" \
    --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": "jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound",
		"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 "jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound" \
        --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": "jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound with Docker Model Runner:
```
docker model run hf.co/jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound
```

KAT-Coder-V2.5-Dev W8A128 with AutoRound int8

Weight-only 8-bit quant of KAT-Coder-V2.5-Dev using AutoRound v0.15.0, quantized at int8 against a python focused sampleset.

Under the hood

256-expert MoE on Qwen3.5. 40 layers, 30 use linear attention and 10 use full attention (every 4th). Shared expert gates kept at FP16.

Quantization is symmetric INT8, group size 128. Calibrated on 384 samples over 400 iterations with sequence length 4096 instead of the default 2048.

KL	Top-1 match	Top-1 in ref top-5	Tokens
0.0047	96.84%	99.97%	16,336

(Measured against the original BF16 checkpoint on Wikitext-103 validation)

Hardware

Fits on two 3090s with headroom. This quant was created largely to fit this into two 24gb cards while maintaining speed. On my machine, this retains enough space for 3 max-num-seqs at full context. Tinker as you see fit to get the number of parallel slots you'd like to serve.

Inference

vLLM V1 engine:

PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True \
vllm serve \
  ./KAT-Coder-V2.5-Dev-W8A128 \
  --port 5001 \
  --tensor-parallel-size 2 \
  --gpu-memory-utilization 0.975 \
  --max-num-seqs 3 \
  --enable-chunked-prefill \
  --enable-prefix-caching

Notes

I had to alter the auto-round library hella in order to get this to work end to end, but in the end, vanilla VLLM serves this just fine :)

Upstream PRs to auto-round to come, it seems to do a lot of double work and underutilizes gpu capability when quantizing some models.

License

Apache 2.0

Downloads last month: -

Safetensors

Model size

10B params

Tensor type

I32

BF16

F16

Model tree for jpbwin/KAT-Coder-V2.5-Dev-int8-AutoRound

Base model

Kwaipilot/KAT-Coder-V2.5-Dev

Quantized

(30)

this model