Instructions to use jkim96/Qwen3.6-27B-DASHQ-INT4-g128 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use jkim96/Qwen3.6-27B-DASHQ-INT4-g128 with Transformers:

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

pipe = pipeline("image-text-to-text", model="jkim96/Qwen3.6-27B-DASHQ-INT4-g128", trust_remote_code=True)
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, AutoModelForCausalLM

processor = AutoProcessor.from_pretrained("jkim96/Qwen3.6-27B-DASHQ-INT4-g128", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("jkim96/Qwen3.6-27B-DASHQ-INT4-g128", trust_remote_code=True, 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 jkim96/Qwen3.6-27B-DASHQ-INT4-g128 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "jkim96/Qwen3.6-27B-DASHQ-INT4-g128"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jkim96/Qwen3.6-27B-DASHQ-INT4-g128",
		"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/jkim96/Qwen3.6-27B-DASHQ-INT4-g128

SGLang

How to use jkim96/Qwen3.6-27B-DASHQ-INT4-g128 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 "jkim96/Qwen3.6-27B-DASHQ-INT4-g128" \
    --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": "jkim96/Qwen3.6-27B-DASHQ-INT4-g128",
		"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 "jkim96/Qwen3.6-27B-DASHQ-INT4-g128" \
        --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": "jkim96/Qwen3.6-27B-DASHQ-INT4-g128",
		"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 jkim96/Qwen3.6-27B-DASHQ-INT4-g128 with Docker Model Runner:
```
docker model run hf.co/jkim96/Qwen3.6-27B-DASHQ-INT4-g128
```

jkim96 commited on May 9

Commit

98efcc6

verified ·

1 Parent(s): 7faf2fb

Upload dashq quantized checkpoint (INT4, g128, scale_zero_dtype=float16, zero-shot avg=71.2)

Browse files

Files changed (6) hide show

README.md +13 -10
dashq_config.json +15 -16
model-00002-of-00005.safetensors +1 -1
model-00003-of-00005.safetensors +1 -1
model-00004-of-00005.safetensors +1 -1
model-00005-of-00005.safetensors +1 -1

README.md CHANGED Viewed

@@ -38,6 +38,7 @@ model, tokenizer = load_quantized(
 | Base model | `Qwen/Qwen3.6-27B` |
 | Bits | `4` |
 | Group size | `128` |
 | Calibration dataset | `wikitext2` |
 | Calibration samples | `128` |
 | Sequence length | `2048` |
@@ -48,13 +49,15 @@ model, tokenizer = load_quantized(
 | Metric | Value |
 | --- | ---: |
-| `wikitext2_ppl` | 7.5355 |
-| `arc_challenge` | 59.9829 |
-| `arc_easy` | 76.3889 |
-| `commonsense_qa` | 85.9951 |
-| `hellaswag` | 83.3201 |
-| `lambada_openai` | 74.3450 |
-| `openbookqa` | 45.0000 |
-| `piqa` | 82.6986 |
-| `truthfulqa_mc2` | 56.6694 |
-| `winogrande` | 77.7427 |

 | Base model | `Qwen/Qwen3.6-27B` |
 | Bits | `4` |
 | Group size | `128` |
+| Scale/zero dtype | `float16` |
 | Calibration dataset | `wikitext2` |
 | Calibration samples | `128` |
 | Sequence length | `2048` |
 | Metric | Value |
 | --- | ---: |
+| `wikitext2_ppl` | 7.4794 |
+| `zero-shot accuracy avg` | 71.2425 |
+| `arc_challenge` | 59.5563 |
+| `arc_easy` | 76.0101 |
+| `commonsense_qa` | 85.8313 |
+| `hellaswag` | 83.4396 |
+| `lambada_openai` | 74.7720 |
+| `openbookqa` | 44.2000 |
+| `piqa` | 82.8618 |
+| `truthfulqa_mc2` | 56.6108 |
+| `winogrande` | 77.9006 |

dashq_config.json CHANGED Viewed

@@ -10,6 +10,7 @@
     "low_memory_optimization": false,
     "moe_hessian_scope": "shared",
     "n_samples": 128,
     "symmetric": false,
     "use_error_compensation": true,
     "use_optimal_shrinkage": true,
@@ -5974,20 +5975,18 @@
     "Model": "Qwen/Qwen3.6-27B",
     "ModelSizeGB": 18.948856152,
     "OriginalSizeGB": 55.5630064,
-    "PPL": 7.53549337387085,
-    "Params": "{'bits': 4, 'group_size': 128, 'n_samples': 128, 'moe_hessian_scope': 'shared', 'use_error_compensation': True, 'use_optimal_shrinkage': True, 'use_weighted_quantization': True, 'symmetric': False, 'low_memory_optimization': False}",
-    "QuantTime": 898.2053005695343,
-    "arc_challenge": 59.98293515358362,
-    "arc_easy": 76.38888888888889,
-    "boolq": 77.2782874617737,
-    "commonsense_qa": 85.995085995086,
-    "hellaswag": 83.3200557657837,
-    "lambada_openai": 74.345041723268,
-    "openbookqa": 45.0,
-    "piqa": 82.69858541893362,
-    "social_iqa": 65.83979328165374,
-    "truthfulqa_mc2": 56.669406789395374,
-    "winogrande": 77.7426992896606,
-    "zeroshot_avg": 71.38734361527521
   }
-}

     "low_memory_optimization": false,
     "moe_hessian_scope": "shared",
     "n_samples": 128,
+    "scale_zero_dtype": "float16",
     "symmetric": false,
     "use_error_compensation": true,
     "use_optimal_shrinkage": true,
     "Model": "Qwen/Qwen3.6-27B",
     "ModelSizeGB": 18.948856152,
     "OriginalSizeGB": 55.5630064,
+    "PPL": 7.479367256164551,
+    "Params": "{'bits': 4, 'group_size': 128, 'scale_zero_dtype': 'float16', 'n_samples': 128, 'moe_hessian_scope': 'shared', 'use_error_compensation': True, 'use_optimal_shrinkage': True, 'use_weighted_quantization': True, 'symmetric': False, 'low_memory_optimization': False}",
+    "QuantTime": 839.7647020816803,
+    "arc_challenge": 59.55631399317406,
+    "arc_easy": 76.01010101010101,
+    "commonsense_qa": 85.83128583128583,
+    "hellaswag": 83.43955387373033,
+    "lambada_openai": 74.77197748884146,
+    "openbookqa": 44.2,
+    "piqa": 82.86180631120783,
+    "truthfulqa_mc2": 56.610824269727146,
+    "winogrande": 77.90055248618785,
+    "zeroshot_avg": 71.24249058491728
   }
+}

model-00002-of-00005.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6bd1ef08a93ec894b46d7c7e3c84de93ee51627f49490319104054777dc23d02
 size 4997583600

 version https://git-lfs.github.com/spec/v1
+oid sha256:72cbb0bfce29be4bef20951afea7ffbbd29f9fd4e38b5c661c623a6db7c46a5a
 size 4997583600

model-00003-of-00005.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:385a8da079a4c98292d0527a1165d170f17f6fe3bf27153e930611e6e8f149ca
 size 4981282880

 version https://git-lfs.github.com/spec/v1
+oid sha256:9d0eb5a07bd954735323559955be5c33c1ea24a234491b8de7f1add388c7ad30
 size 4981282880

model-00004-of-00005.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:01e18b59d224c133b4bf441394ae629c7c72d62ad58ce4bc2837698742aa2179
 size 4962151088

 version https://git-lfs.github.com/spec/v1
+oid sha256:275636a91121fdbda2cd480cfb720066ac2b79a390a18a6ee69eaeea77f1b263
 size 4962151088

model-00005-of-00005.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f9042bbfb1dae6c639cc37e18bb1e70f99979379b0dde16b11507a7ade370512
 size 1465041656

 version https://git-lfs.github.com/spec/v1
+oid sha256:261d3088c8eb0fb6ed386b3db63447589a281f6b55cf5da72884b6ce45e68981
 size 1465041656