Can I run Muse-Glimmer-30B on an NVIDIA H100?
Yes, just
Its published weights (59.6 GB) fit in the 72.0 GB an NVIDIA H100 can use, with room for ≈ 128K tokens of context.
Full analysis of Muse-Glimmer-30B
Everything that runs on NVIDIA H100
Get the weights on Hugging Face
hf download meta-models/Muse-Glimmer-30BFiles and context on NVIDIA H100
Each file of Muse-Glimmer-30B (and of GGUF versions the site has analyzed), and how much context fits next to it, with the context cache in 16-bit or 8-bit.
| File | Size | On NVIDIA H100 | Context (16-bit) | Context (8-bit) |
|---|---|---|---|---|
| As published | 59.6 GB | Tight | 128K | 128K |
| Q4_K_XL | 19.7 GB | Fits | 128K | 128K |
| Q4_K_M | 16.8 GB | Fits | 128K | 128K |
| BF16 | 55.7 GB | Fits | not known | not known |
| UD-Q8_K_XL | 32.3 GB | Fits | not known | not known |
| Q8_0 | 29.6 GB | Fits | not known | not known |
| UD-Q6_K_XL | 26.3 GB | Fits | not known | not known |
| UD-Q5_K_XL | 21.8 GB | Fits | not known | not known |
| UD-Q5_K_L | 19.8 GB | Fits | not known | not known |
| UD-Q5_K_M | 19.2 GB | Fits | not known | not known |
| UD-Q4_K_XL | 15.9 GB | Fits | not known | not known |
| UD-IQ3_M | 14.1 GB | Fits | not known | not known |
| UD-Q3_K_XL | 13.4 GB | Fits | not known | not known |
| UD-IQ3_XXS | 13.1 GB | Fits | not known | not known |
| UD-Q2_K_XL | 12.4 GB | Fits | not known | not known |
| UD-IQ2_M | 12.3 GB | Fits | not known | not known |
| UD-IQ2_XS | 11.5 GB | Fits | not known | not known |
| UD-IQ2_XXS | 10.7 GB | Fits | not known | not known |