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 Radeon RX 7800 XT run ZwZ2-9B?

ZwZ2-9B·9.4B·Radeon RX 7800 XT·16 GB VRAM

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

Yes. The Radeon RX 7800 XT runs ZwZ2-9B fully in VRAM at Q8_0 (about 11 GB of 16 GB), at roughly 34 tokens/s.

best quant
Q8_0
needs
11 GB
usable
16 GB
est. speed
~34 tok/s

Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the ZwZ2-9B page →

Every quantisation of ZwZ2-9B on the Radeon RX 7800 XT

QuantQualityMemoryFitSpeed
Q2_K Heavy loss4.5 GBfits~111 tok/s
Q3_K_M Noticeable loss6.0 GBfits~75 tok/s
Q4_K_M Small loss7.0 GBfits~60 tok/s
Q5_K_M Near-lossless7.8 GBfits~51 tok/s
Q6_K Near-lossless8.7 GBfits~44 tok/s
Q8_0 pickLossless11 GBfits~34 tok/s
FP16 / BF16 Lossless19 GBoffload~7.3 tok/s

Speed is a bandwidth-based estimate for generating text (624 GB/s); prompt processing and long contexts are slower.

Measured speed on the Radeon RX 7800 XT

No community reports for this pair yet.

Report your own speed →

Run ZwZ2-9B on the Radeon RX 7800 XT

Using Q8_0, the best fit for Radeon RX 7800 XT.

llama.cpp · OpenAI-compatible server with a web UIdocs
llama-server -hf inclusionAI/ZwZ2-9B-GGUF:Q8_0

Needs a HIP (ROCm) build of llama.cpp. Downloads the GGUF on first run.

llama.cpp · Docker with ROCm (AMD)docs
docker run -d --name llama --device /dev/kfd --device /dev/dri -p 8080:8080 \
  -v ~/.cache/llama.cpp:/root/.cache/llama.cpp \
  ghcr.io/ggml-org/llama.cpp:server-rocm \
  -hf inclusionAI/ZwZ2-9B-GGUF:Q8_0 --host 0.0.0.0 --port 8080 -ngl 99

Server and web UI on port 8080.

vLLM · Serve an OpenAI-compatible APIdocs
vllm serve inclusionAI/ZwZ2-9B

Needs vLLM built for ROCm; it supports Instinct MI200/MI300 and Radeon RX 7900 / 9000 cards.

vLLM · Docker Compose (AMD ROCm)docs
services:
  vllm:
    image: rocm/vllm:latest
    command: vllm serve inclusionAI/ZwZ2-9B --host 0.0.0.0 --port 8000
    ports:
      - "8000:8000"
    ipc: host
    devices:
      - /dev/kfd
      - /dev/dri
    group_add:
      - video
    security_opt:
      - seccomp=unconfined
    volumes:
      - ~/.cache/huggingface:/root/.cache/huggingface

AMD’s rocm/vllm image. API on port 8000.

SGLang · Serve an OpenAI-compatible APIdocs
sglang serve inclusionAI/ZwZ2-9B --host 0.0.0.0 --port 30000

Needs SGLang’s ROCm build (Instinct MI300-class GPUs).

Transformers · Ask about an imagedocs
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="inclusionAI/ZwZ2-9B", 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 Radeon RX 7800 XT comfortably

All that fit →
ModelParamsBest quantMemorySpeedLicence
9.7BQ8_0≈11 GBFits in memory~33 tok/sCommercial OK
8.8BQ8_0≈10 GBFits in memory~37 tok/sCommercial OK
9.4BQ8_0≈11 GBFits in memory~34 tok/sCommercial OK
8.3BQ8_0≈8.9 GBFits in memory~39 tok/sCommercial OK
Cloudflare logoclef-flashCloudflare
9.4BQ8_0≈11 GBFits in memory~34 tok/sCommercial OK
12.2BQ6_K≈13 GBFits in memory~34 tok/sConditions

Other GPUs that run ZwZ2-9B

Frequently asked questions

Can the Radeon RX 7800 XT run ZwZ2-9B?
Yes. The Radeon RX 7800 XT runs ZwZ2-9B fully in VRAM at Q8_0 (about 11 GB of 16 GB), at roughly 34 tokens/s.
How fast does ZwZ2-9B run on the Radeon RX 7800 XT?
An estimate from memory bandwidth gives about 34 tokens/s at Q8_0.
Which quantisation of ZwZ2-9B should I use on the Radeon RX 7800 XT?
Q8_0: the largest one that fits. Expected quality: 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.