PMA 1.0 needs about 0.6 GB of memory at Q4_K_M (a 0.0 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 12 quants, from 0.0 GB (Q1_0) to 0.1 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 0.4 GB.
PMA 1.0 is a 64M-parameter Pma model in GGUF format, quantized by patriotmemory-ai. Its trained context length is 32K tokens. The repository has 12 quant files, from 0.0 GB (Q1_0) to 0.1 GB (F16).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q1_0 | 0.0 GB | 0.6 GB | Fits | Get file → | |
| Q2_K | 0.0 GB | 0.6 GB | Fits | Get file → | |
| Q3_K_S | 0.0 GB | 0.6 GB | Fits | Get file → | |
| Q3_K_M | 0.0 GB | 0.6 GB | Fits | Get file → | |
| Q3_K_L | 0.0 GB | 0.6 GB | Fits | Get file → | |
| IQ4_XS | 0.0 GB | 0.6 GB | Fits | Get file → | |
| IQ4_NL | 0.0 GB | 0.6 GB | Fits | Get file → | |
| Q4_K_M | 0.0 GB | 0.6 GB | Fits | Get file → | |
| Q5_K_M | 0.0 GB | 0.6 GB | Fits | Get file → | |
| Q6_K | 0.1 GB | 0.7 GB | Fits | Get file → | |
| Q8_0 | 0.1 GB | 0.7 GB | Fits | Get file → | |
| F16 | 0.1 GB | 0.7 GB | Fits | Get file → |
What GPU do you need for PMA 1.0?
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: the best quant that fits is F16 (0.1 GB file, about 0.7 GB in use).
Which PMA 1.0 quant should you choose?
Q4_K_M is the usual starting point for PMA 1.0: it balances file size and quality well. Here it is a 0.0 GB file.
The Q8_0 file is 0.1 GB, about 1.6 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 Q1_0 at 0.0 GB, with the quality trade-offs that come with fewer bits. Lower bit counts mean smaller files and more quality loss.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about PMA 1.0
How much VRAM does PMA 1.0 need?
About 0.6 GB at Q4_K_M with an 8K context window. That is the 0.0 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 PMA 1.0 on an 8 GB GPU?
Yes, with F16: the 0.1 GB file needs about 0.7 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of PMA 1.0 GGUF?
It depends on the quant: Q1_0 is 0.0 GB, Q2_K is 0.0 GB, Q3_K_S is 0.0 GB, Q3_K_M is 0.0 GB, Q3_K_L is 0.0 GB, IQ4_XS is 0.0 GB and 6 more.
What context length does PMA 1.0 support?
The GGUF header lists a trained context length of 32,768 tokens (32K). The KV cache grows with the context you use: about 0.1 GB at 8K and 0.4 GB at 32K.
Where can I download PMA 1.0 GGUF?
The files are on Hugging Face in the patriotmemory-ai/PMA-1.0-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 PMA 1.0 use?
Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.