llama3 42b v0 imat needs about 28.2 GB of memory at Q4_K_M (a 26.1 GB file) with an 8K context window, so it fits on a 32 GB GPU. This repository has 17 quants, from 9.6 GB (IQ1_S) to 45.9 GB (Q8_0).
llama3 42b v0 imat is a 43.2B-parameter Llama model in GGUF format, quantized by NyxKrage. Its trained context length is 8K tokens. The repository has 17 quant files, from 9.6 GB (IQ1_S) to 45.9 GB (Q8_0).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ1_S | 9.6 GB | 11.7 GB | Too big | Get file → | |
| IQ1_M | 10.5 GB | 12.6 GB | Too big | Get file → | |
| IQ2_XXS | 11.9 GB | 14.0 GB | Too big | Get file → | |
| IQ2_XS | 13.1 GB | 15.2 GB | Too big | Get file → | |
| IQ2_S | 13.8 GB | 15.9 GB | Too big | Get file → | |
| IQ2_M | 14.9 GB | 17.1 GB | Too big | Get file → | |
| Q2_K_S | 15.1 GB | 17.2 GB | Too big | Get file → | |
| Q3_K_S | 19.0 GB | 21.1 GB | Too big | Get file → | |
| Q3_K_M | 21.0 GB | 23.2 GB | Too big | Get file → | |
| Q3_K_L | 22.8 GB | 24.9 GB | Too big | Get file → | |
| IQ4_XS | 23.3 GB | 25.4 GB | Too big | Get file → | |
| Q4_K_S | 24.7 GB | 26.9 GB | Too big | Get file → | |
| Q4_K_M | 26.1 GB | 28.2 GB | Too big | Get file → | |
| Q5_K_S | 29.8 GB | 31.9 GB | Too big | Get file → | |
| Q5_K_M | 30.6 GB | 32.7 GB | Too big | Get file → | |
| Q6_K | 35.4 GB | 37.5 GB | Too big | Get file → | |
| Q8_0 | 45.9 GB | 48.0 GB | Too big | Get file → |
What GPU do you need for llama3 42b v0 imat?
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: the best quant that fits is IQ1_S (9.6 GB file, about 11.7 GB in use).
- 16 GB VRAM: the best quant that fits is IQ2_S (13.8 GB file, about 15.9 GB in use).
- 24 GB VRAM: the best quant that fits is Q3_K_M (21.0 GB file, about 23.2 GB in use).
- 32 GB VRAM: the best quant that fits is Q5_K_S (29.8 GB file, about 31.9 GB in use).
- 48 GB VRAM: the best quant that fits is Q8_0 (45.9 GB file, about 48.0 GB in use).
Which llama3 42b v0 imat quant should you choose?
Q4_K_M is the usual starting point for llama3 42b v0 imat: it balances file size and quality well. Here it is a 26.1 GB file.
The Q8_0 file is 45.9 GB, about 1.8 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 IQ1_S at 9.6 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about llama3 42b v0 imat
How much VRAM does llama3 42b v0 imat need?
About 28.2 GB at Q4_K_M with an 8K context window. That is the 26.1 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 llama3 42b v0 imat on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 11.7 GB, so you would need a larger GPU or CPU offloading.
What is the file size of llama3 42b v0 imat GGUF?
It depends on the quant: IQ1_S is 9.6 GB, IQ1_M is 10.5 GB, IQ2_XXS is 11.9 GB, IQ2_XS is 13.1 GB, IQ2_S is 13.8 GB, IQ2_M is 14.9 GB and 11 more.
What context length does llama3 42b v0 imat support?
The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 1.6 GB at 8K and 1.6 GB at 8K.
Where can I download llama3 42b v0 imat GGUF?
The files are on Hugging Face in the NyxKrage/llama3-42b-v0-imat-gguf repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.