Phi 3 mini 4k instruct LLaMAfied 8xMoE raw needs about 16.6 GB of memory at Q4_K_M (a 12.9 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 11 quants, from 8.5 GB (IQ3_XXS) to 41.5 GB (F16).
Phi 3 mini 4k instruct LLaMAfied 8xMoE raw is a 20.7B-parameter Llama model in GGUF format, quantized by jamesburton. Its trained context length is 4K tokens. The repository has 11 quant files, from 8.5 GB (IQ3_XXS) to 41.5 GB (F16).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ3_XXS | 8.5 GB | 10.6 GB | Too big | Get file → | |
| IQ3_M | 9.8 GB | 11.9 GB | Too big | Get file → | |
| IQ4_XS | 11.4 GB | 13.5 GB | Too big | Get file → | |
| IQ4_NL | 12.0 GB | 14.1 GB | Too big | Get file → | |
| Q4_K_S | 12.1 GB | 14.2 GB | Too big | Get file → | |
| Q4_K_M | 12.9 GB | 15.0 GB | Too big | Get file → | |
| Q5_K_S | 14.5 GB | 16.6 GB | Too big | Get file → | |
| Q5_K_M | 14.9 GB | 17.0 GB | Too big | Get file → | |
| Q6_K | 17.2 GB | 19.3 GB | Too big | Get file → | |
| Q8_0 | 22.0 GB | 24.1 GB | Too big | Get file → | |
| F16 | 41.5 GB | 43.6 GB | Too big | Get file → |
What GPU do you need for Phi 3 mini 4k instruct LLaMAfied 8xMoE raw?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) 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_S (12.1 GB file, about 15.9 GB in use).
- 24 GB VRAM: the best quant that fits is Q6_K (17.2 GB file, about 20.9 GB in use).
- 32 GB VRAM: the best quant that fits is Q8_0 (22.0 GB file, about 25.8 GB in use).
- 48 GB VRAM: the best quant that fits is F16 (41.5 GB file, about 45.2 GB in use).
Which Phi 3 mini 4k instruct LLaMAfied 8xMoE raw quant should you choose?
Q4_K_M is the usual starting point for Phi 3 mini 4k instruct LLaMAfied 8xMoE raw: it balances file size and quality well. Here it is a 12.9 GB file.
The Q8_0 file is 22.0 GB, about 1.7 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is IQ3_XXS at 8.5 GB, with the quality trade-offs that come with fewer bits. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Frequently asked questions about Phi 3 mini 4k instruct LLaMAfied 8xMoE raw
How much VRAM does Phi 3 mini 4k instruct LLaMAfied 8xMoE raw need?
About 16.6 GB at Q4_K_M with an 8K context window. That is the 12.9 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 Phi 3 mini 4k instruct LLaMAfied 8xMoE raw on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 12.2 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Phi 3 mini 4k instruct LLaMAfied 8xMoE raw GGUF?
It depends on the quant: IQ3_XXS is 8.5 GB, IQ3_M is 9.8 GB, IQ4_XS is 11.4 GB, IQ4_NL is 12.0 GB, Q4_K_S is 12.1 GB, Q4_K_M is 12.9 GB and 5 more.
What context length does Phi 3 mini 4k instruct LLaMAfied 8xMoE raw support?
The GGUF header lists a trained context length of 4,096 tokens (4K). The KV cache grows with the context you use: about 3.2 GB at 8K and 1.6 GB at 4K.
Where can I download Phi 3 mini 4k instruct LLaMAfied 8xMoE raw GGUF?
The files are on Hugging Face in the jamesburton/Phi-3-mini-4k-instruct-LLaMAfied-8xMoE-raw-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does Phi 3 mini 4k instruct LLaMAfied 8xMoE raw use?
Hugging Face lists the license as mit. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.