Ahma 7B needs about 9.1 GB of memory at Q4_K_M (a 4.3 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 15 quants, from 2.7 GB (Q2_K) to 14.0 GB (F16).
Ahma 7B – a 7B parameter Finnish language model
Ahma 7B is a 7B parameter decoder‑only transformer based on Meta Llama v1. It was pretrained from scratch on Finnish text and includes a second stage with single‑turn instruction examples. The model can generate text and be fine‑tuned for downstream tasks.
Good at
- Text generation in Finnish
- Instruction following without fine‑tuning
- Low‑resource language support
Limits
- No vision support
- Limited to 2048 context tokens
- Apache‑2.0 license
Best for
- Generating Finnish stories
- Fine‑tuning for Finnish chatbots
- Research on data‑constrained models
Who it’s for. Researchers and developers working with Finnish language models who need a decoder‑only architecture. The model is suitable for those who want to experiment with instruction following or fine‑tune for specific Finnish applications.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 2.7 GB | 4.3 GB | Fits | Get file → | |
| IQ3_XS | 3.0 GB | 4.5 GB | Fits | Get file → | |
| IQ3_S | 3.1 GB | 4.7 GB | Fits | Get file → | |
| Q3_K_S | 3.1 GB | 4.7 GB | Fits | Get file → | |
| IQ3_M | 3.3 GB | 4.9 GB | Fits | Get file → | |
| Q3_K_M | 3.5 GB | 5.0 GB | Fits | Get file → | |
| Q3_K_L | 3.8 GB | 5.3 GB | Fits | Get file → | |
| IQ4_XS | 3.8 GB | 5.4 GB | Fits | Get file → | |
| Q4_K_S | 4.0 GB | 5.6 GB | Fits | Get file → | |
| Q4_K_M | 4.3 GB | 5.8 GB | Fits | Get file → | |
| Q5_K_S | 4.9 GB | 6.4 GB | Fits | Get file → | |
| Q5_K_M | 5.0 GB | 6.6 GB | Fits | Get file → | |
| Q6_K | 5.7 GB | 7.3 GB | Tight | Get file → | |
| Q8_0 | 7.4 GB | 9.0 GB | Too big | Get file → | |
| F16 | 14.0 GB | 15.6 GB | Too big | Get file → |
For 8 GB VRAM, use Q2_K (2.69 GB, 7.5 GB memory) or IQ3_XS (2.96 GB, 7.8 GB memory). With 12 GB VRAM, Q4_K_M (4.26 GB, 9.1 GB memory) or Q5_K_S (4.85 GB, 9.6 GB memory) are good choices. Higher precision quants like F16 (14.01 GB, 18.8 GB memory) require more than 18 GB VRAM.
What GPU do you need for Ahma 7B?
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: the best quant that fits is IQ3_S (3.1 GB file, about 7.9 GB in use).
- 12 GB VRAM: the best quant that fits is Q6_K (5.7 GB file, about 10.5 GB in use).
- 16 GB VRAM: the best quant that fits is Q8_0 (7.4 GB file, about 12.2 GB in use).
- 24 GB VRAM: the best quant that fits is F16 (14.0 GB file, about 18.8 GB in use).
Which Ahma 7B quant should you choose?
Q4_K_M is the usual starting point for Ahma 7B: it balances file size and quality well. Here it is a 4.3 GB file.
The Q8_0 file is 7.4 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 Q2_K at 2.7 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 Ahma 7B
How much VRAM does Ahma 7B need?
About 9.1 GB at Q4_K_M with an 8K context window. That is the 4.3 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 Ahma 7B on an 8 GB GPU?
Yes, with IQ3_S: the 3.1 GB file needs about 7.9 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of Ahma 7B GGUF?
It depends on the quant: Q2_K is 2.7 GB, IQ3_XS is 3.0 GB, IQ3_S is 3.1 GB, Q3_K_S is 3.1 GB, IQ3_M is 3.3 GB, Q3_K_M is 3.5 GB and 9 more.
What context length does Ahma 7B support?
The GGUF header lists a trained context length of 2,048 tokens (2K). The KV cache grows with the context you use: about 4.3 GB at 8K and 1.1 GB at 2K.
Where can I download Ahma 7B GGUF?
The files are on Hugging Face in the mradermacher/Ahma-7B-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 Ahma 7B 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.
Can I run Ahma 7B on 8 GB VRAM?
Yes, using Q2_K or IQ3_XS quants will fit within 8 GB VRAM with approximate memory usage of 7.5–7.8 GB.
Does Ahma 7B support image inputs?
No, the model is a text‑only decoder and does not support vision inputs.
What license applies to Ahma 7B?
The model is released under the Apache‑2.0 license.