RoGemma-7b-Instruct needs about 8.4 GB of memory at IQ3_M (a 4.1 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 25 quants, from 2.2 GB (IQ1_S) to 17.1 GB.
RoGemma‑7b‑Instruct – Romanian‑focused 7B instruction model
RoGemma‑7b‑Instruct is a 7.54‑billion‑parameter generative model fine‑tuned for Romanian. It is built on the Gemma architecture and released under a cc‑by‑nc‑4.0 license. The model supports instruction‑driven dialogue and research tasks in Romanian.
Good at
- Instruction‑driven chat in Romanian
- Fine‑tuned Romanian NLP tasks
- Research‑grade Romanian language modeling
Limits
- No vision support
- License restricts commercial use
- Context limited to 8192 tokens
Best for
- Chat assistants for Romanian
- Research on Romanian language models
- Fine‑tuning Romanian NLP pipelines
Who it’s for. Researchers and developers working with Romanian language tasks who need an open‑source instruction model. The model is suitable for academic projects and non‑commercial applications.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ1_S | 2.2 GB | 6.4 GB | Fits | Get file → | |
| IQ1_M | 2.3 GB | 6.6 GB | Fits | Get file → | |
| IQ2_XXS | 2.6 GB | 6.8 GB | Fits | Get file → | |
| IQ2_XS | 2.8 GB | 7.1 GB | Fits | Get file → | |
| IQ2_S | 2.9 GB | 7.2 GB | Fits | Get file → | |
| IQ2_M | 3.1 GB | 7.4 GB | Tight | Get file → | |
| Q2_K_S | 3.2 GB | 7.5 GB | Tight | Get file → | |
| Q2_K | 3.5 GB | 7.7 GB | Tight | Get file → | |
| IQ3_XXS | 3.5 GB | 7.7 GB | Tight | Get file → | |
| IQ3_XS | 3.8 GB | 8.1 GB | Too big | Get file → | |
| IQ3_S | 4.0 GB | 8.2 GB | Too big | Get file → | |
| Q3_K_S | 4.0 GB | 8.2 GB | Too big | Get file → | |
| IQ3_M | 4.1 GB | 8.4 GB | Too big | Get file → | |
| Q3_K | 4.4 GB | 8.6 GB | Too big | Get file → | |
| Q3_K_L | 4.7 GB | 9.0 GB | Too big | Get file → | |
| IQ4_XS | 4.8 GB | 9.0 GB | Too big | Get file → | |
| IQ4_NL | 5.0 GB | 9.3 GB | Too big | Get file → | |
| Q4_K_S | 5.0 GB | 9.3 GB | Too big | Get file → | |
| Q4_K | 5.3 GB | 9.6 GB | Too big | Get file → | |
| Q5_K_S | 6.0 GB | 10.2 GB | Too big | Get file → | |
| Q5_K | 6.1 GB | 10.4 GB | Too big | Get file → | |
| Q6_K | 7.0 GB | 11.3 GB | Too big | Get file → | |
| Q8_0 | 9.1 GB | 13.3 GB | Too big | Get file → | |
| BF16 | 17.1 GB | 21.3 GB | Too big | Get file → | |
| — | 17.1 GB | 21.3 GB | Too big | Get file → |
For 8 GB VRAM, use IQ1_S (2.16 GB) or IQ1_M (2.32 GB). For 12 GB VRAM, IQ2_M (3.13 GB) or Q2_K (3.48 GB) are suitable. For 24 GB VRAM, Q5_K (6.14 GB) or Q6_K (7.01 GB) provide higher quality. Estimates are approximate.
What GPU do you need for RoGemma-7b-Instruct?
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_XXS (3.5 GB file, about 7.7 GB in use).
- 12 GB VRAM: the best quant that fits is Q6_K (7.0 GB file, about 11.3 GB in use).
- 16 GB VRAM: the best quant that fits is Q8_0 (9.1 GB file, about 13.3 GB in use).
- 24 GB VRAM: the best quant that fits is BF16 (17.1 GB file, about 21.3 GB in use).
Which RoGemma-7b-Instruct quant should you choose?
IQ3_M (4.1 GB) is the middle option in this repository. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
If memory is tight, the smallest file is IQ1_S at 2.2 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 RoGemma-7b-Instruct
How much VRAM does RoGemma-7b-Instruct need?
About 8.4 GB at IQ3_M with an 8K context window. That is the 4.1 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 RoGemma-7b-Instruct on an 8 GB GPU?
Yes, with IQ3_XXS: the 3.5 GB file needs about 7.7 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of RoGemma-7b-Instruct GGUF?
It depends on the quant: IQ1_S is 2.2 GB, IQ1_M is 2.3 GB, IQ2_XXS is 2.6 GB, IQ2_XS is 2.8 GB, IQ2_S is 2.9 GB, IQ2_M is 3.1 GB and 19 more.
What context length does RoGemma-7b-Instruct support?
The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 3.8 GB at 8K and 3.8 GB at 8K.
Where can I download RoGemma-7b-Instruct GGUF?
The files are on Hugging Face in the legraphista/RoGemma-7b-Instruct-IMat-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 RoGemma-7b-Instruct use?
Hugging Face lists the license as cc-by-nc-4.0. 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 RoGemma‑7b‑Instruct on 8 GB VRAM?
Yes, using IQ1_S or IQ1_M quantizations which require about 6.4–6.6 GB of VRAM at 8k context.
Is RoGemma‑7b‑Instruct suitable for commercial use?
No, the cc‑by‑nc‑4.0 license prohibits commercial use.
Does the model support vision tasks?
No, RoGemma‑7b‑Instruct is a text‑only model.