Sponsored by APIMart
Toolify Toolify logo
AI Models
Qwen3-VL-30B-A3B-Instruct.w4a16

EmbeddedLLM / Qwen3-VL-30B-A3B-Instruct.w4a16

Run on huggingface.co
huggingface.co
Total runs: 33
24-hour runs: 0
7-day runs: 5
30-day runs: 6
Model's Last Updated: January 13 2026
text-generation

Introduction of Qwen3-VL-30B-A3B-Instruct.w4a16

Model Details of Qwen3-VL-30B-A3B-Instruct.w4a16

Qwen3-VL-30B-A3B-Instruct.w4a16

Model Overview
  • Model Optimizations:
    • Weight quantization: INT4
Model Optimizations

This model was obtained by quantizing the weights of Qwen/Qwen3-VL-30B-A3B-Instruct to INT4 data type. This optimization reduces the number of bits per parameter from 16 to 4, reducing the disk size and GPU memory requirements by approximately 75%.

Only the weights of the linear operators within transformers blocks are quantized. Weights are quantized using a symmetric per-group scheme, with group size 128. The AutoRound algorithm is applied for quantization, as implemented as AutoRoundModifier in the llm-compressor library.

Deployment

This model can be deployed efficiently using the vLLM backend, as shown in the example below. 

vllm serve EmbeddedLLM/Qwen3-VL-30B-A3B-Instruct.w4a16
Creation
Creation details This model was created with [llm-compressor](https://github.com/vllm-project/llm-compressor) by running the code snippet below. 
from auto_round.calib_dataset import get_dataset
from transformers import AutoTokenizer, Qwen3VLMoeForConditionalGeneration, AutoProcessor

from llmcompressor import oneshot
from llmcompressor.modifiers.autoround import AutoRoundModifier
from llmcompressor.utils import dispatch_for_generation

# Select model and load it.
model_id = "Qwen/Qwen3-VL-30B-A3B-Instruct"
model = Qwen3VLMoeForConditionalGeneration.from_pretrained(model_id, dtype="auto")
tokenizer = AutoTokenizer.from_pretrained(model_id)
processor = AutoProcessor.from_pretrained(model_id)


# Select calibration dataset.
NUM_CALIBRATION_SAMPLES = 128
MAX_SEQUENCE_LENGTH = 2048
# Get aligned calibration dataset.
ds = get_dataset(
    tokenizer=tokenizer,
    seqlen=MAX_SEQUENCE_LENGTH,
    nsamples=NUM_CALIBRATION_SAMPLES,
)


# Configure the quantization algorithm to run.
#   * quantize the weights to 4 bit with AutoRound with a group size 128
recipe = AutoRoundModifier(
    targets="Linear",
    scheme="W4A16",
    ignore=[
        "re:.*lm_head",
        "re:visual.*",
        "re:model.visual.*",
        "re:.*mlp.gate$",
    ],
    iters=200
)

oneshot(
  model=model,
  dataset=ds,
  recipe=recipe,
  max_seq_length=MAX_SEQUENCE_LENGTH,
  num_calibration_samples=NUM_CALIBRATION_SAMPLES,
  # disable shuffling to get slightly better mmlu score
  shuffle_calibration_samples=False,
)

# Save to disk compressed.
SAVE_DIR = model_id.rstrip("/").split("/")[-1] + ".w4a16"
model.save_pretrained(SAVE_DIR, save_compressed=True)
tokenizer.save_pretrained(SAVE_DIR)
processor.save_pretrained(SAVE_DIR)
Evaluation

After started the vllm server as shown above, the model was evaluated on with the following commands.

Disclaimer : Results may differ from official benchmarks due to evaluation setup variations.

Evaluation details

lm-evaluation-harness 

lm_eval
  --model local-chat-completions \
  --model_args model="EmbeddedLLM/Qwen3-VL-30B-A3B-Instruct.w4a16",base_url=http://127.0.0.1:8000/v1/chat/completions,num_concurrent=64 \
  --tasks mmlu_pro \
  --apply_chat_template

lmms-eval 

lmms-eval \
  --model async_openai \
  --model_args model_version=EmbeddedLLM/Qwen3-VL-30B-A3B-Instruct.w4a16,base_url=http://127.0.0.1:8000/v1,is_qwen3_vl=True,api_key=DUMMY,num_cpus=8 \
  --tasks mmmu_val
Accuracy
Category Benchmark Qwen3-VL-30B-A3B-Instruct Qwen3-VL-30B-A3B-Instruct.w4a16
(this model) 		Recovery
Text MMLU_pro 71.28 69.55 97.6%
Vision MMMU_val 51.56 53.22 103.2%

Runs of EmbeddedLLM Qwen3-VL-30B-A3B-Instruct.w4a16 on huggingface.co

33
Total runs
0
24-hour runs
1
3-day runs
5
7-day runs
6
30-day runs

More Information About Qwen3-VL-30B-A3B-Instruct.w4a16 huggingface.co Model

More Qwen3-VL-30B-A3B-Instruct.w4a16 license Visit here:

https://choosealicense.com/licenses/apache-2.0

Qwen3-VL-30B-A3B-Instruct.w4a16 huggingface.co

Qwen3-VL-30B-A3B-Instruct.w4a16 huggingface.co is an AI model on huggingface.co that provides Qwen3-VL-30B-A3B-Instruct.w4a16's model effect (), which can be used instantly with this EmbeddedLLM Qwen3-VL-30B-A3B-Instruct.w4a16 model. huggingface.co supports a free trial of the Qwen3-VL-30B-A3B-Instruct.w4a16 model, and also provides paid use of the Qwen3-VL-30B-A3B-Instruct.w4a16. Support call Qwen3-VL-30B-A3B-Instruct.w4a16 model through api, including Node.js, Python, http.

Qwen3-VL-30B-A3B-Instruct.w4a16 huggingface.co Url

https://huggingface.co/EmbeddedLLM/Qwen3-VL-30B-A3B-Instruct.w4a16

EmbeddedLLM Qwen3-VL-30B-A3B-Instruct.w4a16 online free

Qwen3-VL-30B-A3B-Instruct.w4a16 huggingface.co is an online trial and call api platform, which integrates Qwen3-VL-30B-A3B-Instruct.w4a16's modeling effects, including api services, and provides a free online trial of Qwen3-VL-30B-A3B-Instruct.w4a16, you can try Qwen3-VL-30B-A3B-Instruct.w4a16 online for free by clicking the link below.

EmbeddedLLM Qwen3-VL-30B-A3B-Instruct.w4a16 online free url in huggingface.co:

https://huggingface.co/EmbeddedLLM/Qwen3-VL-30B-A3B-Instruct.w4a16

Qwen3-VL-30B-A3B-Instruct.w4a16 install

Qwen3-VL-30B-A3B-Instruct.w4a16 is an open source model from GitHub that offers a free installation service, and any user can find Qwen3-VL-30B-A3B-Instruct.w4a16 on GitHub to install. At the same time, huggingface.co provides the effect of Qwen3-VL-30B-A3B-Instruct.w4a16 install, users can directly use Qwen3-VL-30B-A3B-Instruct.w4a16 installed effect in huggingface.co for debugging and trial. It also supports api for free installation.

Qwen3-VL-30B-A3B-Instruct.w4a16 install url in huggingface.co:

https://huggingface.co/EmbeddedLLM/Qwen3-VL-30B-A3B-Instruct.w4a16

Url of Qwen3-VL-30B-A3B-Instruct.w4a16

Qwen3-VL-30B-A3B-Instruct.w4a16 huggingface.co Url

Provider of Qwen3-VL-30B-A3B-Instruct.w4a16 huggingface.co

EmbeddedLLM
ORGANIZATIONS
huggingface.co

Other API from EmbeddedLLM