Instructions to use unsloth/gemma-3-27b-it-unsloth-bnb-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use unsloth/gemma-3-27b-it-unsloth-bnb-4bit with Transformers:

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

pipe = pipeline("image-text-to-text", model="unsloth/gemma-3-27b-it-unsloth-bnb-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("unsloth/gemma-3-27b-it-unsloth-bnb-4bit")
model = AutoModelForMultimodalLM.from_pretrained("unsloth/gemma-3-27b-it-unsloth-bnb-4bit", 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 unsloth/gemma-3-27b-it-unsloth-bnb-4bit with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "unsloth/gemma-3-27b-it-unsloth-bnb-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": "unsloth/gemma-3-27b-it-unsloth-bnb-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/unsloth/gemma-3-27b-it-unsloth-bnb-4bit

SGLang

How to use unsloth/gemma-3-27b-it-unsloth-bnb-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 "unsloth/gemma-3-27b-it-unsloth-bnb-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": "unsloth/gemma-3-27b-it-unsloth-bnb-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 "unsloth/gemma-3-27b-it-unsloth-bnb-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": "unsloth/gemma-3-27b-it-unsloth-bnb-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"
						}
					}
				]
			}
		]
	}'

Unsloth Desktop
Docker Model Runner
How to use unsloth/gemma-3-27b-it-unsloth-bnb-4bit with Docker Model Runner:
```
docker model run hf.co/unsloth/gemma-3-27b-it-unsloth-bnb-4bit
```

danielhanchen commited on Apr 10, 2025

Commit

afb9f44

verified ·

1 Parent(s): 990be35

Add files using upload-large-folder tool

Browse files

Files changed (7) hide show

config.json +42 -41
generation_config.json +4 -2
model-00001-of-00004.safetensors +2 -2
model-00002-of-00004.safetensors +2 -2
model-00003-of-00004.safetensors +2 -2
model-00004-of-00004.safetensors +2 -2
model.safetensors.index.json +0 -0

config.json CHANGED Viewed

@@ -25,64 +25,65 @@
       "multi_modal_projector",
       "merger",
       "modality_projection",
-      "vision_tower.vision_model.encoder.layers.24.self_attn",
-      "vision_tower.vision_model.encoder.layers.26.self_attn",
       "vision_tower.vision_model.encoder.layers.22.self_attn",
-      "language_model.model.layers.27.mlp",
       "vision_tower.vision_model.encoder.layers.23.self_attn",
       "vision_tower.vision_model.encoder.layers.25.self_attn",
-      "vision_tower.vision_model.encoder.layers.23.mlp",
       "vision_tower.vision_model.encoder.layers.25.mlp",
-      "vision_tower.vision_model.encoder.layers.18.self_attn",
       "vision_tower.vision_model.encoder.layers.17.self_attn",
-      "vision_tower.vision_model.encoder.layers.15.mlp",
-      "vision_tower.vision_model.encoder.layers.20.self_attn",
       "vision_tower.vision_model.encoder.layers.21.self_attn",
-      "vision_tower.vision_model.encoder.layers.19.mlp",
       "vision_tower.vision_model.encoder.layers.16.mlp",
-      "vision_tower.vision_model.encoder.layers.24.mlp",
       "vision_tower.vision_model.encoder.layers.13.mlp",
       "vision_tower.vision_model.encoder.layers.18.mlp",
-      "vision_tower.vision_model.encoder.layers.8.mlp",
-      "vision_tower.vision_model.encoder.layers.22.mlp",
-      "vision_tower.vision_model.encoder.layers.4.mlp",
-      "vision_tower.vision_model.encoder.layers.16.self_attn",
-      "vision_tower.vision_model.encoder.layers.21.mlp",
-      "vision_tower.vision_model.encoder.layers.9.mlp",
-      "vision_tower.vision_model.encoder.layers.13.self_attn",
       "vision_tower.vision_model.encoder.layers.10.mlp",
-      "vision_tower.vision_model.encoder.layers.19.self_attn",
       "vision_tower.vision_model.encoder.layers.11.self_attn",
-      "vision_tower.vision_model.encoder.layers.1.mlp",
-      "vision_tower.vision_model.encoder.layers.17.mlp",
-      "vision_tower.vision_model.encoder.layers.15.self_attn",
       "vision_tower.vision_model.encoder.layers.12.self_attn",
-      "vision_tower.vision_model.encoder.layers.20.mlp",
       "vision_tower.vision_model.encoder.layers.10.self_attn",
-      "vision_tower.vision_model.encoder.layers.0.self_attn",
-      "vision_tower.vision_model.encoder.layers.3.mlp",
-      "vision_tower.vision_model.encoder.layers.9.self_attn",
-      "vision_tower.vision_model.encoder.layers.1.self_attn",
-      "vision_tower.vision_model.encoder.layers.14.mlp",
-      "vision_tower.vision_model.encoder.layers.0.mlp",
-      "language_model.model.layers.27.self_attn",
-      "vision_tower.vision_model.encoder.layers.5.mlp",
-      "language_model.model.layers.5.mlp",
       "vision_tower.vision_model.encoder.layers.11.mlp",
-      "vision_tower.vision_model.encoder.layers.7.mlp",
-      "vision_tower.vision_model.encoder.layers.7.self_attn",
-      "vision_tower.vision_model.encoder.layers.8.self_attn",
-      "language_model.model.layers.23.self_attn",
-      "vision_tower.vision_model.encoder.layers.26.mlp",
-      "vision_tower.vision_model.encoder.layers.2.mlp",
       "vision_tower.vision_model.encoder.layers.6.self_attn",
-      "vision_tower.vision_model.encoder.layers.14.self_attn",
-      "vision_tower.vision_model.encoder.layers.12.mlp",
-      "vision_tower.vision_model.encoder.layers.3.self_attn",
       "vision_tower.vision_model.encoder.layers.4.self_attn",
       "vision_tower.vision_model.encoder.layers.6.mlp",
       "vision_tower.vision_model.encoder.layers.5.self_attn",
-      "vision_tower.vision_model.encoder.layers.2.self_attn"
     ],
     "llm_int8_threshold": 6.0,
     "load_in_4bit": true,
@@ -120,7 +121,7 @@
     "vocab_size": 262208
   },
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.50.0.dev0",
   "unsloth_fixed": true,
   "vision_config": {
     "attention_dropout": 0.0,

       "multi_modal_projector",
       "merger",
       "modality_projection",
+      "language_model.model.layers.10.mlp",
+      "language_model.model.layers.22.self_attn",
+      "language_model.model.layers.11.mlp",
+      "language_model.model.layers.18.mlp",
+      "language_model.model.layers.17.mlp",
       "vision_tower.vision_model.encoder.layers.22.self_attn",
+      "vision_tower.vision_model.encoder.layers.26.self_attn",
       "vision_tower.vision_model.encoder.layers.23.self_attn",
       "vision_tower.vision_model.encoder.layers.25.self_attn",
+      "vision_tower.vision_model.encoder.layers.24.self_attn",
+      "vision_tower.vision_model.encoder.layers.19.self_attn",
+      "vision_tower.vision_model.encoder.layers.16.self_attn",
+      "vision_tower.vision_model.encoder.layers.19.mlp",
       "vision_tower.vision_model.encoder.layers.25.mlp",
       "vision_tower.vision_model.encoder.layers.17.self_attn",
+      "vision_tower.vision_model.encoder.layers.24.mlp",
       "vision_tower.vision_model.encoder.layers.21.self_attn",
+      "vision_tower.vision_model.encoder.layers.17.mlp",
+      "vision_tower.vision_model.encoder.layers.20.self_attn",
+      "vision_tower.vision_model.encoder.layers.15.mlp",
+      "vision_tower.vision_model.encoder.layers.15.self_attn",
       "vision_tower.vision_model.encoder.layers.16.mlp",
+      "vision_tower.vision_model.encoder.layers.20.mlp",
+      "vision_tower.vision_model.encoder.layers.23.mlp",
+      "vision_tower.vision_model.encoder.layers.14.self_attn",
+      "vision_tower.vision_model.encoder.layers.18.self_attn",
+      "vision_tower.vision_model.encoder.layers.21.mlp",
+      "vision_tower.vision_model.encoder.layers.22.mlp",
       "vision_tower.vision_model.encoder.layers.13.mlp",
       "vision_tower.vision_model.encoder.layers.18.mlp",
+      "vision_tower.vision_model.encoder.layers.14.mlp",
       "vision_tower.vision_model.encoder.layers.10.mlp",
+      "vision_tower.vision_model.encoder.layers.12.mlp",
       "vision_tower.vision_model.encoder.layers.11.self_attn",
+      "vision_tower.vision_model.encoder.layers.8.mlp",
+      "vision_tower.vision_model.encoder.layers.13.self_attn",
+      "vision_tower.vision_model.encoder.layers.7.mlp",
       "vision_tower.vision_model.encoder.layers.12.self_attn",
+      "vision_tower.vision_model.encoder.layers.9.mlp",
       "vision_tower.vision_model.encoder.layers.10.self_attn",
       "vision_tower.vision_model.encoder.layers.11.mlp",
+      "vision_tower.vision_model.encoder.layers.4.mlp",
+      "vision_tower.vision_model.encoder.layers.5.mlp",
+      "vision_tower.vision_model.encoder.layers.9.self_attn",
       "vision_tower.vision_model.encoder.layers.6.self_attn",
+      "vision_tower.vision_model.encoder.layers.8.self_attn",
+      "vision_tower.vision_model.encoder.layers.3.mlp",
+      "vision_tower.vision_model.encoder.layers.7.self_attn",
       "vision_tower.vision_model.encoder.layers.4.self_attn",
       "vision_tower.vision_model.encoder.layers.6.mlp",
       "vision_tower.vision_model.encoder.layers.5.self_attn",
+      "vision_tower.vision_model.encoder.layers.3.self_attn",
+      "vision_tower.vision_model.encoder.layers.1.mlp",
+      "vision_tower.vision_model.encoder.layers.2.mlp",
+      "vision_tower.vision_model.encoder.layers.1.self_attn",
+      "vision_tower.vision_model.encoder.layers.26.mlp",
+      "vision_tower.vision_model.encoder.layers.2.self_attn",
+      "vision_tower.vision_model.encoder.layers.0.self_attn",
+      "vision_tower.vision_model.encoder.layers.0.mlp"
     ],
     "llm_int8_threshold": 6.0,
     "load_in_4bit": true,
     "vocab_size": 262208
   },
   "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.0",
   "unsloth_fixed": true,
   "vision_config": {
     "attention_dropout": 0.0,

generation_config.json CHANGED Viewed

@@ -1,11 +1,13 @@
 {
-  "_from_model_config": true,
   "bos_token_id": 2,
   "cache_implementation": "hybrid",
   "eos_token_id": [
     1,
     106
   ],
   "pad_token_id": 0,
-  "transformers_version": "4.50.0.dev0"
 }

 {
   "bos_token_id": 2,
   "cache_implementation": "hybrid",
+  "do_sample": true,
   "eos_token_id": [
     1,
     106
   ],
   "pad_token_id": 0,
+  "top_k": 64,
+  "top_p": 0.95,
+  "transformers_version": "4.51.0"
 }

model-00001-of-00004.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f7b5321b65e6e421cd20d02040856faa0ae716464064a478a7cbbafcc38decb3
-size 4995986199

 version https://git-lfs.github.com/spec/v1
+oid sha256:cbaa109f7ddc56f24322f9b241ce14563d158601166b5fef0c2ec1c26f7de8f2
+size 4977824200

model-00002-of-00004.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f13777589a2a8485ba9c2e236e048e4f0db79e967543d60230c084dd4dc4d1a1
-size 4974849837

 version https://git-lfs.github.com/spec/v1
+oid sha256:072aa0221017ee40d58b253fceb5c5a6b1ed2942bc70dd80dcf4806b64726660
+size 4947779253

model-00003-of-00004.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4659054666352a019f4e661054e92dfd306a3404590cee60fe1572fbd8b25a0b
-size 4967504061

 version https://git-lfs.github.com/spec/v1
+oid sha256:8813874764f0ac81b260955d2e61b0daca6ccdae1bb43b1cbd7ab6826304b719
+size 4998165790

model-00004-of-00004.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d05890dad6b5ddba02f3565a6f7c1689c21f27d2a45ac03d89c66b7c535185c9
-size 3161660722

 version https://git-lfs.github.com/spec/v1
+oid sha256:c5d696938bede1ecf49eb7cd4a6233bf7bcc4930ef4b8ed9a98745fda56b2536
+size 4107627920

model.safetensors.index.json CHANGED Viewed

The diff for this file is too large to render. See raw diff