selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

inclusionAI logo

Can the Arc A380 run ZwZ2-4B?

ZwZ2-4B·4.5B·Arc A380·6 GB VRAM

Saved in this browser, used on every model page.
Yes

Yes. The Arc A380 runs ZwZ2-4B fully in VRAM at Q6_K (about 5.1 GB of 6.0 GB), at roughly 27 tokens/s.

best quant
Q6_K
needs
5.1 GB
usable
6.0 GB
est. speed
~27 tok/s

Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the ZwZ2-4B page →

Every quantisation of ZwZ2-4B on the Arc A380

QuantQualityMemoryFitSpeed
Q2_K Heavy loss3.2 GBfits~69 tok/s
Q3_K_M Noticeable loss3.5 GBfits~46 tok/s
Q4_K_M Small loss4.2 GBfits~37 tok/s
Q5_K_M Near-lossless4.7 GBfits~32 tok/s
Q6_K pickNear-lossless5.1 GBfits~27 tok/s
Q8_0 Lossless5.9 GBtight~21 tok/s
FP16 / BF16 Lossless9.7 GBoffload~6.2 tok/s

Speed is a bandwidth-based estimate for generating text (186 GB/s); prompt processing and long contexts are slower.

Measured speed on the Arc A380

No community reports for this pair yet.

Report your own speed →

Run ZwZ2-4B on the Arc A380

Using Q6_K, the best fit for Arc A380.

Ollama · Q6_K straight from Hugging Facedocs
ollama run hf.co/bartowski/ZwZ2-4B-GGUF:Q6_K

Ollama’s ROCm/CUDA builds don’t cover this GPU. Recent Ollama releases use Vulkan when its Vulkan backend and driver are installed (older ones needed OLLAMA_VULKAN=1); llama.cpp or LM Studio with Vulkan are the safer bet.

llama.cpp · OpenAI-compatible server with a web UIdocs
llama-server -hf bartowski/ZwZ2-4B-GGUF:Q6_K

Needs a Vulkan build of llama.cpp. Downloads the GGUF on first run.

llama.cpp · Docker with Vulkandocs
docker run -d --name llama --device /dev/dri -p 8080:8080 \
  -v ~/.cache/llama.cpp:/root/.cache/llama.cpp \
  ghcr.io/ggml-org/llama.cpp:server-vulkan \
  -hf bartowski/ZwZ2-4B-GGUF:Q6_K --host 0.0.0.0 --port 8080 -ngl 99

Server and web UI on port 8080.

LM Studio · Download with the lms CLIdocs
lms get bartowski/ZwZ2-4B-GGUF@Q6_K

Then load it from the LM Studio app or with `lms load`.

Transformers · Ask about an imagedocs
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="inclusionAI/ZwZ2-4B", dtype="auto", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/bee.jpg"},
            {"type": "text", "text": "Describe this image."},
        ],
    }
]
out = pipe(text=messages, max_new_tokens=256)
print(out[0]["generated_text"][-1]["content"])

Vision models that fit the Arc A380 comfortably

All that fit →
ModelParamsBest quantMemorySpeedLicence
4.7BQ6_K≈4.9 GBFits in memory~27 tok/sCommercial OK
4.3BQ8_0≈5.2 GBFits in memory~22 tok/sConditions
4.4BQ6_K≈4.9 GBFits in memory~28 tok/sCommercial OK
4.3BQ8_0≈5.6 GBFits in memory~22 tok/sConditions
4.3BQ8_0≈5.6 GBFits in memory~22 tok/sConditions
5BQ8_0≈5.6 GBFits in memory~19 tok/sConditions

Other GPUs that run ZwZ2-4B

Frequently asked questions

Can the Arc A380 run ZwZ2-4B?
Yes. The Arc A380 runs ZwZ2-4B fully in VRAM at Q6_K (about 5.1 GB of 6.0 GB), at roughly 27 tokens/s.
How fast does ZwZ2-4B run on the Arc A380?
An estimate from memory bandwidth gives about 27 tokens/s at Q6_K.
Which quantisation of ZwZ2-4B should I use on the Arc A380?
Q6_K: the largest one that fits. Expected quality: near-lossless compared with the full-precision weights.

Memory and speed are estimates from the model's size and the card's memory bandwidth; real numbers depend on the runtime, drivers and context length. Community figures are medians of user reports.