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.

nvidia logo

Can the Instinct MI50 32GB run NLD-Uni?

NLD-Uni·18.3B·Instinct MI50 32GB·32 GB VRAM

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

Yes. The Instinct MI50 32GB runs NLD-Uni fully in VRAM at Q8_0 (about 21 GB of 32 GB), at roughly 29 tokens/s.

best quant
Q8_0
needs
21 GB
usable
32 GB
est. speed
~29 tok/s

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

Every quantisation of NLD-Uni on the Instinct MI50 32GB

QuantQualityMemoryFitSpeed
Q2_K Heavy loss7.4 GBfits~94 tok/s
Q3_K_M Noticeable loss10 GBfits~63 tok/s
Q4_K_M Small loss12 GBfits~51 tok/s
Q5_K_M Near-lossless14 GBfits~43 tok/s
Q6_K Near-lossless16 GBfits~37 tok/s
Q8_0 pickLossless21 GBfits~29 tok/s
FP16 / BF16 Lossless37 GBoffload~4.7 tok/s

Speed is a bandwidth-based estimate for generating text (1024 GB/s); prompt processing and long contexts are slower.

Measured speed on the Instinct MI50 32GB

No community reports for this pair yet.

Report your own speed →

Omni models that fit the Instinct MI50 32GB comfortably

All that fit →
ModelParamsBest quantMemorySpeedLicence
17.5BQ8_0≈20 GBFits in memory~30 tok/sCommercial OK
12BQ8_0≈16 GBFits in memory~44 tok/sCommercial OK
24.3BQ8_0≈26 GBFits in memory~22 tok/sCommercial OK
ByteDance-Seed logoBAGEL-7B-MoTByteDance-Seed
14.7BQ8_0≈17 GBFits in memory~36 tok/sCommercial OK
10.7BQ8_0≈8.9 GBFits in memory~49 tok/sCommercial OK
35.3BA3BQ6_K≈30 GBFits in memory~228 tok/sCommercial OK

Other GPUs that run NLD-Uni

Frequently asked questions

Can the Instinct MI50 32GB run NLD-Uni?
Yes. The Instinct MI50 32GB runs NLD-Uni fully in VRAM at Q8_0 (about 21 GB of 32 GB), at roughly 29 tokens/s.
How fast does NLD-Uni run on the Instinct MI50 32GB?
An estimate from memory bandwidth gives about 29 tokens/s at Q8_0.
Which quantisation of NLD-Uni should I use on the Instinct MI50 32GB?
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.