Instructions to use cyankiwi/Qwen3.5-35B-A3B-AWQ-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use cyankiwi/Qwen3.5-35B-A3B-AWQ-4bit with Transformers:

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

pipe = pipeline("image-text-to-text", model="cyankiwi/Qwen3.5-35B-A3B-AWQ-4bit")
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/Qwen3.5-35B-A3B-AWQ-4bit")
model = AutoModelForMultimodalLM.from_pretrained("cyankiwi/Qwen3.5-35B-A3B-AWQ-4bit")
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/Qwen3.5-35B-A3B-AWQ-4bit with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "cyankiwi/Qwen3.5-35B-A3B-AWQ-4bit"
# 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/Qwen3.5-35B-A3B-AWQ-4bit",
		"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/Qwen3.5-35B-A3B-AWQ-4bit

SGLang

How to use cyankiwi/Qwen3.5-35B-A3B-AWQ-4bit 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/Qwen3.5-35B-A3B-AWQ-4bit" \
    --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/Qwen3.5-35B-A3B-AWQ-4bit",
		"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/Qwen3.5-35B-A3B-AWQ-4bit" \
        --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/Qwen3.5-35B-A3B-AWQ-4bit",
		"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/Qwen3.5-35B-A3B-AWQ-4bit with Docker Model Runner:
```
docker model run hf.co/cyankiwi/Qwen3.5-35B-A3B-AWQ-4bit
```

cpatonn commited on Mar 26

Commit

affb52d

verified ·

1 Parent(s): 858208a

Upload folder using huggingface_hub

Browse files

Files changed (8) hide show

config.json +44 -40
model-00001-of-00005.safetensors +2 -2
model-00002-of-00005.safetensors +2 -2
model-00003-of-00005.safetensors +2 -2
model-00004-of-00005.safetensors +2 -2
model-00005-of-00005.safetensors +2 -2
model.safetensors.index.json +2 -2
recipe.yaml +40 -0

config.json CHANGED Viewed

@@ -172,6 +172,10 @@
         "model.language_model.layers.2.mlp.shared_expert.up_proj",
         "model.language_model.layers.2.mlp.shared_expert.down_proj",
         "model.language_model.layers.2.mlp.shared_expert_gate",
         "model.language_model.layers.3.mlp.shared_expert.gate_proj",
         "model.language_model.layers.3.mlp.shared_expert.up_proj",
         "model.language_model.layers.3.mlp.shared_expert.down_proj",
@@ -206,6 +210,10 @@
         "model.language_model.layers.6.mlp.shared_expert.up_proj",
         "model.language_model.layers.6.mlp.shared_expert.down_proj",
         "model.language_model.layers.6.mlp.shared_expert_gate",
         "model.language_model.layers.7.mlp.shared_expert.gate_proj",
         "model.language_model.layers.7.mlp.shared_expert.up_proj",
         "model.language_model.layers.7.mlp.shared_expert.down_proj",
@@ -240,6 +248,10 @@
         "model.language_model.layers.10.mlp.shared_expert.up_proj",
         "model.language_model.layers.10.mlp.shared_expert.down_proj",
         "model.language_model.layers.10.mlp.shared_expert_gate",
         "model.language_model.layers.11.mlp.shared_expert.gate_proj",
         "model.language_model.layers.11.mlp.shared_expert.up_proj",
         "model.language_model.layers.11.mlp.shared_expert.down_proj",
@@ -274,6 +286,10 @@
         "model.language_model.layers.14.mlp.shared_expert.up_proj",
         "model.language_model.layers.14.mlp.shared_expert.down_proj",
         "model.language_model.layers.14.mlp.shared_expert_gate",
         "model.language_model.layers.15.mlp.shared_expert.gate_proj",
         "model.language_model.layers.15.mlp.shared_expert.up_proj",
         "model.language_model.layers.15.mlp.shared_expert.down_proj",
@@ -308,6 +324,10 @@
         "model.language_model.layers.18.mlp.shared_expert.up_proj",
         "model.language_model.layers.18.mlp.shared_expert.down_proj",
         "model.language_model.layers.18.mlp.shared_expert_gate",
         "model.language_model.layers.19.mlp.shared_expert.gate_proj",
         "model.language_model.layers.19.mlp.shared_expert.up_proj",
         "model.language_model.layers.19.mlp.shared_expert.down_proj",
@@ -342,6 +362,10 @@
         "model.language_model.layers.22.mlp.shared_expert.up_proj",
         "model.language_model.layers.22.mlp.shared_expert.down_proj",
         "model.language_model.layers.22.mlp.shared_expert_gate",
         "model.language_model.layers.23.mlp.shared_expert.gate_proj",
         "model.language_model.layers.23.mlp.shared_expert.up_proj",
         "model.language_model.layers.23.mlp.shared_expert.down_proj",
@@ -376,6 +400,10 @@
         "model.language_model.layers.26.mlp.shared_expert.up_proj",
         "model.language_model.layers.26.mlp.shared_expert.down_proj",
         "model.language_model.layers.26.mlp.shared_expert_gate",
         "model.language_model.layers.27.mlp.shared_expert.gate_proj",
         "model.language_model.layers.27.mlp.shared_expert.up_proj",
         "model.language_model.layers.27.mlp.shared_expert.down_proj",
@@ -410,6 +438,10 @@
         "model.language_model.layers.30.mlp.shared_expert.up_proj",
         "model.language_model.layers.30.mlp.shared_expert.down_proj",
         "model.language_model.layers.30.mlp.shared_expert_gate",
         "model.language_model.layers.31.mlp.shared_expert.gate_proj",
         "model.language_model.layers.31.mlp.shared_expert.up_proj",
         "model.language_model.layers.31.mlp.shared_expert.down_proj",
@@ -444,6 +476,10 @@
         "model.language_model.layers.34.mlp.shared_expert.up_proj",
         "model.language_model.layers.34.mlp.shared_expert.down_proj",
         "model.language_model.layers.34.mlp.shared_expert_gate",
         "model.language_model.layers.35.mlp.shared_expert.gate_proj",
         "model.language_model.layers.35.mlp.shared_expert.up_proj",
         "model.language_model.layers.35.mlp.shared_expert.down_proj",
@@ -478,56 +514,24 @@
         "model.language_model.layers.38.mlp.shared_expert.up_proj",
         "model.language_model.layers.38.mlp.shared_expert.down_proj",
         "model.language_model.layers.38.mlp.shared_expert_gate",
         "model.language_model.layers.39.mlp.shared_expert.gate_proj",
         "model.language_model.layers.39.mlp.shared_expert.up_proj",
         "model.language_model.layers.39.mlp.shared_expert.down_proj",
         "model.language_model.layers.39.mlp.shared_expert_gate",
-        "model.language_model.layers.0.mlp.gate",
-        "model.language_model.layers.1.mlp.gate",
-        "model.language_model.layers.2.mlp.gate",
-        "model.language_model.layers.3.mlp.gate",
-        "model.language_model.layers.4.mlp.gate",
-        "model.language_model.layers.5.mlp.gate",
-        "model.language_model.layers.6.mlp.gate",
-        "model.language_model.layers.7.mlp.gate",
-        "model.language_model.layers.8.mlp.gate",
-        "model.language_model.layers.9.mlp.gate",
-        "model.language_model.layers.10.mlp.gate",
-        "model.language_model.layers.11.mlp.gate",
-        "model.language_model.layers.12.mlp.gate",
-        "model.language_model.layers.13.mlp.gate",
-        "model.language_model.layers.14.mlp.gate",
-        "model.language_model.layers.15.mlp.gate",
-        "model.language_model.layers.16.mlp.gate",
-        "model.language_model.layers.17.mlp.gate",
-        "model.language_model.layers.18.mlp.gate",
-        "model.language_model.layers.19.mlp.gate",
-        "model.language_model.layers.20.mlp.gate",
-        "model.language_model.layers.21.mlp.gate",
-        "model.language_model.layers.22.mlp.gate",
-        "model.language_model.layers.23.mlp.gate",
-        "model.language_model.layers.24.mlp.gate",
-        "model.language_model.layers.25.mlp.gate",
-        "model.language_model.layers.26.mlp.gate",
-        "model.language_model.layers.27.mlp.gate",
-        "model.language_model.layers.28.mlp.gate",
-        "model.language_model.layers.29.mlp.gate",
-        "model.language_model.layers.30.mlp.gate",
-        "model.language_model.layers.31.mlp.gate",
-        "model.language_model.layers.32.mlp.gate",
-        "model.language_model.layers.33.mlp.gate",
-        "model.language_model.layers.34.mlp.gate",
-        "model.language_model.layers.35.mlp.gate",
-        "model.language_model.layers.36.mlp.gate",
-        "model.language_model.layers.37.mlp.gate",
-        "model.language_model.layers.38.mlp.gate",
-        "model.language_model.layers.39.mlp.gate",
         "mtp.fc",
         "mtp.layers.0.mlp.gate",
         "mtp.layers.0.mlp.shared_expert.gate_up_proj",
         "mtp.layers.0.mlp.shared_expert.gate_proj",
         "mtp.layers.0.mlp.shared_expert.up_proj",
         "mtp.layers.0.mlp.shared_expert.down_proj",
         "lm_head"
       ],
       "kv_cache_scheme": null,

         "model.language_model.layers.2.mlp.shared_expert.up_proj",
         "model.language_model.layers.2.mlp.shared_expert.down_proj",
         "model.language_model.layers.2.mlp.shared_expert_gate",
+        "model.language_model.layers.3.self_attn.q_proj",
+        "model.language_model.layers.3.self_attn.k_proj",
+        "model.language_model.layers.3.self_attn.v_proj",
+        "model.language_model.layers.3.self_attn.o_proj",
         "model.language_model.layers.3.mlp.shared_expert.gate_proj",
         "model.language_model.layers.3.mlp.shared_expert.up_proj",
         "model.language_model.layers.3.mlp.shared_expert.down_proj",
         "model.language_model.layers.6.mlp.shared_expert.up_proj",
         "model.language_model.layers.6.mlp.shared_expert.down_proj",
         "model.language_model.layers.6.mlp.shared_expert_gate",
+        "model.language_model.layers.7.self_attn.q_proj",
+        "model.language_model.layers.7.self_attn.k_proj",
+        "model.language_model.layers.7.self_attn.v_proj",
+        "model.language_model.layers.7.self_attn.o_proj",
         "model.language_model.layers.7.mlp.shared_expert.gate_proj",
         "model.language_model.layers.7.mlp.shared_expert.up_proj",
         "model.language_model.layers.7.mlp.shared_expert.down_proj",
         "model.language_model.layers.10.mlp.shared_expert.up_proj",
         "model.language_model.layers.10.mlp.shared_expert.down_proj",
         "model.language_model.layers.10.mlp.shared_expert_gate",
+        "model.language_model.layers.11.self_attn.q_proj",
+        "model.language_model.layers.11.self_attn.k_proj",
+        "model.language_model.layers.11.self_attn.v_proj",
+        "model.language_model.layers.11.self_attn.o_proj",
         "model.language_model.layers.11.mlp.shared_expert.gate_proj",
         "model.language_model.layers.11.mlp.shared_expert.up_proj",
         "model.language_model.layers.11.mlp.shared_expert.down_proj",
         "model.language_model.layers.14.mlp.shared_expert.up_proj",
         "model.language_model.layers.14.mlp.shared_expert.down_proj",
         "model.language_model.layers.14.mlp.shared_expert_gate",
+        "model.language_model.layers.15.self_attn.q_proj",
+        "model.language_model.layers.15.self_attn.k_proj",
+        "model.language_model.layers.15.self_attn.v_proj",
+        "model.language_model.layers.15.self_attn.o_proj",
         "model.language_model.layers.15.mlp.shared_expert.gate_proj",
         "model.language_model.layers.15.mlp.shared_expert.up_proj",
         "model.language_model.layers.15.mlp.shared_expert.down_proj",
         "model.language_model.layers.18.mlp.shared_expert.up_proj",
         "model.language_model.layers.18.mlp.shared_expert.down_proj",
         "model.language_model.layers.18.mlp.shared_expert_gate",
+        "model.language_model.layers.19.self_attn.q_proj",
+        "model.language_model.layers.19.self_attn.k_proj",
+        "model.language_model.layers.19.self_attn.v_proj",
+        "model.language_model.layers.19.self_attn.o_proj",
         "model.language_model.layers.19.mlp.shared_expert.gate_proj",
         "model.language_model.layers.19.mlp.shared_expert.up_proj",
         "model.language_model.layers.19.mlp.shared_expert.down_proj",
         "model.language_model.layers.22.mlp.shared_expert.up_proj",
         "model.language_model.layers.22.mlp.shared_expert.down_proj",
         "model.language_model.layers.22.mlp.shared_expert_gate",
+        "model.language_model.layers.23.self_attn.q_proj",
+        "model.language_model.layers.23.self_attn.k_proj",
+        "model.language_model.layers.23.self_attn.v_proj",
+        "model.language_model.layers.23.self_attn.o_proj",
         "model.language_model.layers.23.mlp.shared_expert.gate_proj",
         "model.language_model.layers.23.mlp.shared_expert.up_proj",
         "model.language_model.layers.23.mlp.shared_expert.down_proj",
         "model.language_model.layers.26.mlp.shared_expert.up_proj",
         "model.language_model.layers.26.mlp.shared_expert.down_proj",
         "model.language_model.layers.26.mlp.shared_expert_gate",
+        "model.language_model.layers.27.self_attn.q_proj",
+        "model.language_model.layers.27.self_attn.k_proj",
+        "model.language_model.layers.27.self_attn.v_proj",
+        "model.language_model.layers.27.self_attn.o_proj",
         "model.language_model.layers.27.mlp.shared_expert.gate_proj",
         "model.language_model.layers.27.mlp.shared_expert.up_proj",
         "model.language_model.layers.27.mlp.shared_expert.down_proj",
         "model.language_model.layers.30.mlp.shared_expert.up_proj",
         "model.language_model.layers.30.mlp.shared_expert.down_proj",
         "model.language_model.layers.30.mlp.shared_expert_gate",
+        "model.language_model.layers.31.self_attn.q_proj",
+        "model.language_model.layers.31.self_attn.k_proj",
+        "model.language_model.layers.31.self_attn.v_proj",
+        "model.language_model.layers.31.self_attn.o_proj",
         "model.language_model.layers.31.mlp.shared_expert.gate_proj",
         "model.language_model.layers.31.mlp.shared_expert.up_proj",
         "model.language_model.layers.31.mlp.shared_expert.down_proj",
         "model.language_model.layers.34.mlp.shared_expert.up_proj",
         "model.language_model.layers.34.mlp.shared_expert.down_proj",
         "model.language_model.layers.34.mlp.shared_expert_gate",
+        "model.language_model.layers.35.self_attn.q_proj",
+        "model.language_model.layers.35.self_attn.k_proj",
+        "model.language_model.layers.35.self_attn.v_proj",
+        "model.language_model.layers.35.self_attn.o_proj",
         "model.language_model.layers.35.mlp.shared_expert.gate_proj",
         "model.language_model.layers.35.mlp.shared_expert.up_proj",
         "model.language_model.layers.35.mlp.shared_expert.down_proj",
         "model.language_model.layers.38.mlp.shared_expert.up_proj",
         "model.language_model.layers.38.mlp.shared_expert.down_proj",
         "model.language_model.layers.38.mlp.shared_expert_gate",
+        "model.language_model.layers.39.self_attn.q_proj",
+        "model.language_model.layers.39.self_attn.k_proj",
+        "model.language_model.layers.39.self_attn.v_proj",
+        "model.language_model.layers.39.self_attn.o_proj",
         "model.language_model.layers.39.mlp.shared_expert.gate_proj",
         "model.language_model.layers.39.mlp.shared_expert.up_proj",
         "model.language_model.layers.39.mlp.shared_expert.down_proj",
         "model.language_model.layers.39.mlp.shared_expert_gate",
         "mtp.fc",
         "mtp.layers.0.mlp.gate",
         "mtp.layers.0.mlp.shared_expert.gate_up_proj",
         "mtp.layers.0.mlp.shared_expert.gate_proj",
         "mtp.layers.0.mlp.shared_expert.up_proj",
         "mtp.layers.0.mlp.shared_expert.down_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",
         "lm_head"
       ],
       "kv_cache_scheme": null,

model-00001-of-00005.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:511407ab1269c0f4a3e391e1984b1ef1cd7a101579771dba5b4cf937cf56c8dc
-size 5370509296

 version https://git-lfs.github.com/spec/v1
+oid sha256:2070b3db53220b1051d7382f95b66c138a32bf2324abf39054aa950ef458f91d
+size 5374122648

model-00002-of-00005.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:20a165dfe961808f24b2b7dbfb4c22f373a6f9b0c464df229df5c244ac56a23d
-size 5372110480

 version https://git-lfs.github.com/spec/v1
+oid sha256:ba41d54eb8a2d49531f27c5a493838ffd413c2575df9cf6afedff0db791e9fd2
+size 5370188016

model-00003-of-00005.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8f43aa8efa84f00e13232e0bdd7b97318ba90db2466b8db9a74ba12d59e650b5
-size 5371922376

 version https://git-lfs.github.com/spec/v1
+oid sha256:6421d2cdab03682bc5da07c70db47729125812f769b2f3f61ec0a6b7b2bdde2b
+size 5370183944

model-00004-of-00005.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:54fe09d56dfbbeedc1622f6c3a7a54f220c473f232d4418886dba256c7862cc7
-size 5372104024

 version https://git-lfs.github.com/spec/v1
+oid sha256:14454ee08bf3344919cecf4106eef287cdcc90d42a350f80b6d89815e562e590
+size 5359586560

model-00005-of-00005.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2aee2973c5a296ba8a64a7255bf3e01777ee28299d1e2352b1fb3e7b3bd1b915
-size 2536847080

 version https://git-lfs.github.com/spec/v1
+oid sha256:3885571ec31627a8f0b0bc67243eab0132350de24e343f508cb28502e2b29b92
+size 2980536416

model.safetensors.index.json CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:43a769307382aa933d9e2cda6b1879ec965a518b0d40b65ea29cb99401dfcf88
-size 10637583

 version https://git-lfs.github.com/spec/v1
+oid sha256:8274494060f6e950c1a3cea599c32334584a38a46ac84dfcf087901aa0b86720
+size 10628250

recipe.yaml ADDED Viewed

	@@ -0,0 +1,40 @@

+default_stage:
+  default_modifiers:
+    AWQModifier:
+      config_groups:
+        group_0:
+          targets: [Linear]
+          weights:
+            num_bits: 4
+            type: int
+            symmetric: true
+            group_size: 32
+            strategy: group
+            block_structure: null
+            dynamic: false
+            actorder: null
+            scale_dtype: null
+            zp_dtype: null
+            observer: mse
+            observer_kwargs: {}
+          input_activations: null
+          output_activations: null
+          format: null
+      targets: [Linear]
+      ignore: ['re:.*embed_tokens', 're:.*linear_attn.*', 're:.*shared_expert.*', 're:.*shared_expert_gate$',
+        're:.*mlp[.]gate$', 're:.*self_attn.*', 're:model[.]visual.*', 're:mtp.*', lm_head]
+      bypass_divisibility_checks: false
+      mappings:
+      - smooth_layer: re:model.*post_attention_layernorm$
+        balance_layers: ['re:model.*mlp[.]experts.*gate_proj$', 're:model.*mlp[.]experts.*up_proj$',
+          're:model.*mlp[.]shared_expert[.]gate_proj$', 're:model.*mlp[.]shared_expert[.]up_proj$',
+          're:model.*mlp[.]gate$', 're:model.*mlp[.]shared_expert_gate$']
+        activation_hook_target: null
+        balance_exponent: 1
+      - smooth_layer: re:model.*mlp[.]experts.*up_proj$
+        balance_layers: ['re:model.*mlp[.]experts.*down_proj$']
+        activation_hook_target: null
+        balance_exponent: 1
+      offload_device: !!python/object/apply:torch.device [cpu]
+      duo_scaling: true
+      n_grid: 20