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.

fastino logo

Can the Radeon RX 9070 XT run GLiDE-27B?

GLiDE-27B·26.1B·Radeon RX 9070 XT·16 GB VRAM

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

Yes, but only at Q3_K_M (about 15 GB of 16 GB), which costs noticeable quality, at roughly 28 tokens/s.

best quant
Q3_K_M
needs
15 GB
usable
16 GB
est. speed
~28 tok/s

Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the GLiDE-27B page →

Every quantisation of GLiDE-27B on the Radeon RX 9070 XT

QuantQualityMemoryFitSpeed
Q2_K Heavy loss11 GBfits~41 tok/s
Q3_K_M pickNoticeable loss15 GBfits~28 tok/s
Q4_K_M Small loss19 GBoffload~9.3 tok/s
Q5_K_M Near-lossless21 GBoffload~6 tok/s
Q6_K Near-lossless24 GBoffload~4 tok/s
Q8_0 Lossless30 GBoffload~2.3 tok/s
FP16 / BF16 Lossless54 GBtoo big—

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 GLiDE-27B on the Radeon RX 9070 XT

Using Q3_K_M, the best fit for Radeon RX 9070 XT.

llama.cpp · OpenAI-compatible server with a web UIdocs
llama-server -hf kingjones777/GLiDE-27B-ROCmFP4-STRIX-GGUF:Q3_K_M

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 kingjones777/GLiDE-27B-ROCmFP4-STRIX-GGUF:Q3_K_M --host 0.0.0.0 --port 8080 -ngl 99

Server and web UI on port 8080.

vLLM · Serve an OpenAI-compatible APIdocs
vllm serve fastino/GLiDE-27B

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 fastino/GLiDE-27B --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 fastino/GLiDE-27B --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="fastino/GLiDE-27B", 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 →
ModelParamsBest quantMemorySpeedLicence
12.2BQ6_K≈13 GBFits in memory~35 tok/sConditions
19.4BA3BQ4_K_M≈13 GBFits in memory~194 tok/sCommercial OK
9.7BQ8_0≈11 GBFits in memory~34 tok/sCommercial OK
13.2BQ6_K≈13 GBFits in memory~32 tok/sConditions
8.8BQ8_0≈10 GBFits in memory~38 tok/sCommercial OK
10.7BQ8_0≈12 GBFits in memory~31 tok/sConditions

Other GPUs that run GLiDE-27B

Frequently asked questions

Can the Radeon RX 9070 XT run GLiDE-27B?
Yes, but only at Q3_K_M (about 15 GB of 16 GB), which costs noticeable quality, at roughly 28 tokens/s.
How fast does GLiDE-27B run on the Radeon RX 9070 XT?
An estimate from memory bandwidth gives about 28 tokens/s at Q3_K_M.
Which quantisation of GLiDE-27B should I use on the Radeon RX 9070 XT?
Q3_K_M: the largest one that fits. Expected quality: noticeable loss 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.