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.

LiquidAI logo

Can the Arc B580 run LFM3-1.7B-Base?

LFM3-1.7B-Base·1.7B·Arc B580·12 GB VRAM

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

Yes. The Arc B580 runs LFM3-1.7B-Base fully in VRAM at FP16 / BF16 (about 4.3 GB of 12 GB), at roughly 73 tokens/s.

best quant
FP16 / BF16
needs
4.3 GB
usable
12 GB
est. speed
~73 tok/s

Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the LFM3-1.7B-Base page →

Every quantisation of LFM3-1.7B-Base on the Arc B580

QuantQualityMemoryFitSpeed
Q2_K Heavy loss1.6 GBfits~441 tok/s
Q3_K_M Noticeable loss1.8 GBfits~297 tok/s
Q4_K_M Small loss1.9 GBfits~239 tok/s
Q5_K_M Near-lossless2.1 GBfits~204 tok/s
Q6_K Near-lossless2.3 GBfits~176 tok/s
Q8_0 Lossless2.7 GBfits~137 tok/s
FP16 / BF16 pickLossless4.3 GBfits~73 tok/s

Speed is a bandwidth-based estimate for generating text (456 GB/s); prompt processing and long contexts are slower.

Measured speed on the Arc B580

No community reports for this pair yet.

Report your own speed →

Run LFM3-1.7B-Base on the Arc B580

Using FP16 / BF16, the best fit for Arc B580.

Ollama · From the Ollama librarydocs
ollama run lfm3:1.7b

The library build picks its own default quantisation. 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.

Ollama · F16 straight from Hugging Facedocs
ollama run hf.co/mradermacher/LFM3-1.7B-Base-GGUF:F16

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 mradermacher/LFM3-1.7B-Base-GGUF:F16

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 mradermacher/LFM3-1.7B-Base-GGUF:F16 --host 0.0.0.0 --port 8080 -ngl 99

Server and web UI on port 8080.

Transformers · Chat with the pipeline APIdocs
from transformers import pipeline

pipe = pipeline("text-generation", model="LiquidAI/LFM3-1.7B-Base", dtype="auto", device_map="auto")
messages = [{"role": "user", "content": "Explain self-hosting in one paragraph."}]
out = pipe(messages, max_new_tokens=512)
print(out[0]["generated_text"][-1]["content"])

Chat models that fit the Arc B580 comfortably

All that fit →
ModelParamsBest quantMemorySpeedLicence
LiquidAI logoLFM3-1.7BLiquidAI
1.7BQ8_0≈2.7 GBFits in memory~137 tok/sConditions
2BQ8_0≈3.2 GBFits in memory~116 tok/sCommercial OK
1.5BQ8_0≈2.6 GBFits in memory~153 tok/sCommercial OK
1.2BQ8_0≈2.0 GBFits in memory~191 tok/sConditions
1BQ8_0≈1.8 GBFits in memory~236 tok/sConditions
2.5BQ8_0≈3.5 GBFits in memory~94 tok/sCommercial OK

Other GPUs that run LFM3-1.7B-Base

Frequently asked questions

Can the Arc B580 run LFM3-1.7B-Base?
Yes. The Arc B580 runs LFM3-1.7B-Base fully in VRAM at FP16 / BF16 (about 4.3 GB of 12 GB), at roughly 73 tokens/s.
How fast does LFM3-1.7B-Base run on the Arc B580?
An estimate from memory bandwidth gives about 73 tokens/s at FP16 / BF16.
Which quantisation of LFM3-1.7B-Base should I use on the Arc B580?
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.