Can the Radeon RX 9070 XT run ZwZ2-4B?
ZwZ2-4B·4.5B·Radeon RX 9070 XT·16 GB VRAM
Yes. The Radeon RX 9070 XT runs ZwZ2-4B fully in VRAM at FP16 / BF16 (about 9.7 GB of 16 GB), at roughly 39 tokens/s.
- best quant
- FP16 / BF16
- needs
- 9.7 GB
- usable
- 16 GB
- est. speed
- ~39 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 Radeon RX 9070 XT
| Quant | Memory | Fit | Speed |
|---|---|---|---|
| Q2_K | 3.2 GB | fits | ~236 tok/s |
| Q3_K_M | 3.5 GB | fits | ~159 tok/s |
| Q4_K_M | 4.2 GB | fits | ~128 tok/s |
| Q5_K_M | 4.7 GB | fits | ~109 tok/s |
| Q6_K | 5.1 GB | fits | ~94 tok/s |
| Q8_0 | 5.9 GB | fits | ~73 tok/s |
| FP16 / BF16 pick | 9.7 GB | fits | ~39 tok/s |
Speed is a bandwidth-based estimate for generating text (640 GB/s); prompt processing and long contexts are slower.
Measured speed on the Radeon RX 9070 XT
No community reports for this pair yet.
Report your own speed →Run ZwZ2-4B on the Radeon RX 9070 XT
Using FP16 / BF16, the best fit for Radeon RX 9070 XT.
ollama run hf.co/inclusionAI/ZwZ2-4B-GGUF:F16docker run -d --name ollama --device /dev/kfd --device /dev/dri \
-v ollama:/root/.ollama -p 11434:11434 ollama/ollama:rocm
docker exec -it ollama ollama run hf.co/inclusionAI/ZwZ2-4B-GGUF:F16The image bundles the ROCm runtime; the host only needs the amdgpu kernel driver.
llama-server -hf inclusionAI/ZwZ2-4B-GGUF:F16Needs a HIP (ROCm) build of llama.cpp. Downloads the GGUF on first run.
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/ZwZ2-4B-GGUF:F16 --host 0.0.0.0 --port 8080 -ngl 99Server and web UI on port 8080.
lms get inclusionAI/ZwZ2-4B-GGUF@F16Then load it from the LM Studio app or with `lms load`.
vllm serve inclusionAI/ZwZ2-4BNeeds vLLM built for ROCm; it supports Instinct MI200/MI300 and Radeon RX 7900 / 9000 cards.
services:
vllm:
image: rocm/vllm:latest
command: vllm serve inclusionAI/ZwZ2-4B --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/huggingfaceAMD’s rocm/vllm image. API on port 8000.
sglang serve inclusionAI/ZwZ2-4B --host 0.0.0.0 --port 30000Needs SGLang’s ROCm build (Instinct MI300-class GPUs).
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 Radeon RX 9070 XT comfortably
All that fit →| Model | Params | Best quant | Memory | Speed |
|---|---|---|---|---|
| 4.7B | Q8_0 | ≈5.9 GBFits in memory | ~71 tok/s | |
| 4.3B | Q8_0 | ≈5.2 GBFits in memory | ~77 tok/s | |
| 4.4B | Q8_0 | ≈5.8 GBFits in memory | ~75 tok/s | |
| 5B | Q8_0 | ≈5.6 GBFits in memory | ~67 tok/s | |
| 4.3B | Q8_0 | ≈5.6 GBFits in memory | ~77 tok/s | |
| 3.8B | Q8_0 | ≈4.0 GBFits in memory | ~88 tok/s |
Other GPUs that run ZwZ2-4B
Frequently asked questions
- Can the Radeon RX 9070 XT run ZwZ2-4B?
- Yes. The Radeon RX 9070 XT runs ZwZ2-4B fully in VRAM at FP16 / BF16 (about 9.7 GB of 16 GB), at roughly 39 tokens/s.
- How fast does ZwZ2-4B run on the Radeon RX 9070 XT?
- An estimate from memory bandwidth gives about 39 tokens/s at FP16 / BF16.
- Which quantisation of ZwZ2-4B should I use on the Radeon RX 9070 XT?
- FP16 / BF16: 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.