h2m mhm 8x7B FrankenMoE v1.0 needs about 13.1 GB of memory at Q4_K_M (a 11.3 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 16 quants, from 11.3 GB (Q4_K_M) to 32.2 GB (Q5_0).
h2m mhm 8x7B FrankenMoE v1.0 is a 46.7B-parameter Mhm model in GGUF format, quantized by RichardErkhov. Its trained context length is 32K tokens. The repository has 16 quant files, from 11.3 GB (Q4_K_M) to 32.2 GB (Q5_0).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q4_K_M | 11.3 GB | 13.1 GB | Too big | Get file → | |
| Q2_K | 17.3 GB | 19.9 GB | Too big | Get file → | |
| IQ3_XS | 19.4 GB | 22.2 GB | Too big | Get file → | |
| IQ3_S | 20.4 GB | 23.4 GB | Too big | Get file → | |
| Q3_K_S | 20.4 GB | 23.4 GB | Too big | Get file → | |
| IQ3_M | 21.4 GB | 24.5 GB | Too big | Get file → | |
| Q3_K | 22.5 GB | 25.8 GB | Too big | Get file → | |
| Q3_K_M | 22.5 GB | 25.8 GB | Too big | Get file → | |
| Q3_K_L | 24.2 GB | 27.6 GB | Too big | Get file → | |
| IQ4_XS | 25.4 GB | 28.9 GB | Too big | Get file → | |
| Q4_K | 25.9 GB | 29.5 GB | Too big | Get file → | |
| Q4_0 | 26.4 GB | 30.1 GB | Too big | Get file → | |
| IQ4_NL | 26.7 GB | 30.5 GB | Too big | Get file → | |
| Q4_K_S | 26.7 GB | 30.5 GB | Too big | Get file → | |
| Q4_1 | 29.3 GB | 33.4 GB | Too big | Get file → | |
| Q5_0 | 32.2 GB | 36.6 GB | Too big | Get file → |
What GPU do you need for h2m mhm 8x7B FrankenMoE v1.0?
These figures add the file size, an fp16 KV cache at an 8K context window and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: no quant fits at an 8K context.
- 6 GB VRAM: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: no quant fits at an 8K context.
- 16 GB VRAM: the best quant that fits is Q4_K_M (11.3 GB file, about 13.1 GB in use).
- 24 GB VRAM: the best quant that fits is IQ3_S (20.4 GB file, about 23.4 GB in use).
- 32 GB VRAM: the best quant that fits is IQ4_NL (26.7 GB file, about 30.5 GB in use).
- 48 GB VRAM: the best quant that fits is Q5_0 (32.2 GB file, about 36.6 GB in use).
Which h2m mhm 8x7B FrankenMoE v1.0 quant should you choose?
Q4_K_M is the usual starting point for h2m mhm 8x7B FrankenMoE v1.0: it balances file size and quality well. Here it is a 11.3 GB file.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Frequently asked questions about h2m mhm 8x7B FrankenMoE v1.0
How much VRAM does h2m mhm 8x7B FrankenMoE v1.0 need?
About 13.1 GB at Q4_K_M with an 8K context window. That is the 11.3 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run h2m mhm 8x7B FrankenMoE v1.0 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 13.1 GB, so you would need a larger GPU or CPU offloading.
What is the file size of h2m mhm 8x7B FrankenMoE v1.0 GGUF?
It depends on the quant: Q4_K_M is 11.3 GB, Q2_K is 17.3 GB, IQ3_XS is 19.4 GB, IQ3_S is 20.4 GB, Q3_K_S is 20.4 GB, IQ3_M is 21.4 GB and 10 more.
What context length does h2m mhm 8x7B FrankenMoE v1.0 support?
The GGUF header lists a trained context length of 32,768 tokens (32K).
Where can I download h2m mhm 8x7B FrankenMoE v1.0 GGUF?
The files are on Hugging Face in the RichardErkhov/h2m_-_mhm-8x7B-FrankenMoE-v1.0-gguf repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.