Instructions to use cyankiwi/Qwen3.8-27B-AWQ-INT4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use cyankiwi/Qwen3.8-27B-AWQ-INT4 with Transformers:

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

pipe = pipeline("image-text-to-text", model="cyankiwi/Qwen3.8-27B-AWQ-INT4")
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("cyankiwi/Qwen3.8-27B-AWQ-INT4")
model = AutoModelForMultimodalLM.from_pretrained("cyankiwi/Qwen3.8-27B-AWQ-INT4", 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 cyankiwi/Qwen3.8-27B-AWQ-INT4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "cyankiwi/Qwen3.8-27B-AWQ-INT4"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "cyankiwi/Qwen3.8-27B-AWQ-INT4",
		"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/cyankiwi/Qwen3.8-27B-AWQ-INT4

SGLang

How to use cyankiwi/Qwen3.8-27B-AWQ-INT4 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 "cyankiwi/Qwen3.8-27B-AWQ-INT4" \
    --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": "cyankiwi/Qwen3.8-27B-AWQ-INT4",
		"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 "cyankiwi/Qwen3.8-27B-AWQ-INT4" \
        --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": "cyankiwi/Qwen3.8-27B-AWQ-INT4",
		"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 cyankiwi/Qwen3.8-27B-AWQ-INT4 with Docker Model Runner:
```
docker model run hf.co/cyankiwi/Qwen3.8-27B-AWQ-INT4
```

dived42 commited on 8 days ago

Commit

71ce7fd

verified ·

1 Parent(s): 63768c1

Fix MTP ignore names for SGLang fused Linear layers (`qkv_proj` / `gate_up_proj`)

Browse files

## Summary

Hi @cyankiwi, I noticed that the MTP head is incorrectly treated as quantized when this checkpoint is loaded by SGLang, which makes NEXTN speculative decoding nearly ineffective.

SGLang fuses QKV and gate/up at module construction (`self_attn.qkv_proj`, `mlp.gate_up_proj`). The ignore check uses these runtime names, not the names produced during `load_weights()`. The MTP entry has no `packed_modules_mapping`, so the fused names are not expanded back to the original shards. Both Linears then fall back to the default INT4 scheme while the checkpoint tensors are BF16.

## Fix

Add the fused runtime names to `ignore`:

```diff
"ignore": [
+ "mtp.layers.0.self_attn.qkv_proj",
+ "mtp.layers.0.mlp.gate_up_proj"
]

## Reproduction

Hardware: 4× RTX 4090, TP=2, PD disaggregation (1 prefill + 1 decode)

```bash
python -m sglang.launch_server \
--model-path <model> \
--tp-size 2 \
--speculative-algorithm NEXTN \
--speculative-num-steps 3 \
--speculative-eagle-topk 1 \
--speculative-num-draft-tokens 4 \
--speculative-attention-mode decode
```

Decode `accept len`: ~1.03 before the fix, ~3 after.

PD / HiCache / metrics / trace are orthogonal to the config issue and not required
to observe the difference.

Files changed (1) hide show

config.json +3 -1

config.json CHANGED Viewed

@@ -342,10 +342,12 @@
       "mtp.layers.0.mlp.down_proj",
       "mtp.layers.0.mlp.gate_proj",
       "mtp.layers.0.mlp.up_proj",
       "mtp.layers.0.self_attn.k_proj",
       "mtp.layers.0.self_attn.o_proj",
       "mtp.layers.0.self_attn.q_proj",
-      "mtp.layers.0.self_attn.v_proj"
     ],
     "kv_cache_scheme": null,
     "quant_method": "compressed-tensors",

       "mtp.layers.0.mlp.down_proj",
       "mtp.layers.0.mlp.gate_proj",
       "mtp.layers.0.mlp.up_proj",
+      "mtp.layers.0.mlp.gate_up_proj",
       "mtp.layers.0.self_attn.k_proj",
       "mtp.layers.0.self_attn.o_proj",
       "mtp.layers.0.self_attn.q_proj",
+      "mtp.layers.0.self_attn.v_proj",
+      "mtp.layers.0.self_attn.qkv_proj"
     ],
     "kv_cache_scheme": null,
     "quant_method": "compressed-tensors",