Sarashina 13B Finetuned v2 needs about 15.6 GB of memory at Q4_K_M (a 8.4 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 24 quants, from 3.3 GB (IQ1_S) to 11.3 GB (Q6_K).
Sarashina 13B Finetuned v2 – bilingual Japanese/English math and technical Q&A
Sarashina 13B Finetuned v2 is a 13.7B parameter causal LM trained on Japanese and English. It is fine‑tuned on math datasets and instruction data. The model is released under the MIT license.
Good at
- Math problem solving
- Technical question answering
- Bilingual Japanese/English responses
Limits
- No vision support
- Potential hallucinations
- Requires verification of critical outputs
Best for
- Educational tutoring
- Technical documentation assistance
- Code explanation
Who it’s for. Developers and educators needing a bilingual model for math and technical queries. The model is suitable for research and production use where MIT licensing is acceptable.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ1_S | 3.3 GB | 7.1 GB | Fits | Get file → | |
| IQ1_M | 3.5 GB | 7.4 GB | Tight | Get file → | |
| IQ2_XXS | 3.9 GB | 7.8 GB | Tight | Get file → | |
| IQ2_XS | 4.3 GB | 8.1 GB | Too big | Get file → | |
| IQ2_S | 4.6 GB | 8.5 GB | Too big | Get file → | |
| Q2_K_S | 4.9 GB | 8.7 GB | Too big | Get file → | |
| IQ2_M | 4.9 GB | 8.8 GB | Too big | Get file → | |
| Q2_K | 5.3 GB | 9.1 GB | Too big | Get file → | |
| IQ3_XXS | 5.4 GB | 9.2 GB | Too big | Get file → | |
| IQ3_XS | 5.8 GB | 9.7 GB | Too big | Get file → | |
| IQ3_S | 6.1 GB | 10.0 GB | Too big | Get file → | |
| Q3_K_S | 6.1 GB | 10.0 GB | Too big | Get file → | |
| IQ3_M | 6.4 GB | 10.3 GB | Too big | Get file → | |
| Q3_K_M | 6.8 GB | 10.6 GB | Too big | Get file → | |
| Q3_K_L | 7.4 GB | 11.2 GB | Too big | Get file → | |
| IQ4_XS | 7.5 GB | 11.3 GB | Too big | Get file → | |
| IQ4_NL | 7.9 GB | 11.7 GB | Too big | Get file → | |
| Q4_0 | 7.9 GB | 11.7 GB | Too big | Get file → | |
| Q4_K_S | 7.9 GB | 11.8 GB | Too big | Get file → | |
| Q4_K_M | 8.4 GB | 12.2 GB | Too big | Get file → | |
| Q4_1 | 8.7 GB | 12.5 GB | Too big | Get file → | |
| Q5_K_S | 9.5 GB | 13.4 GB | Too big | Get file → | |
| Q5_K_M | 9.8 GB | 13.6 GB | Too big | Get file → | |
| Q6_K | 11.3 GB | 15.1 GB | Too big | Get file → |
For 8 GB VRAM, use IQ1_S (3.27 GB, 10.5 GB memory at 8k context) or IQ1_M (3.51 GB, 10.7 GB). For 12 GB VRAM, Q4_K_M (8.37 GB, 15.6 GB) or Q5_K_M (9.78 GB, 17 GB) provide a good balance of size and memory usage. Memory estimates are approximate.
What GPU do you need for Sarashina 13B Finetuned v2?
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 IQ2_S (4.6 GB file, about 11.8 GB in use).
- 16 GB VRAM: the best quant that fits is Q4_1 (8.7 GB file, about 15.9 GB in use).
- 24 GB VRAM: the best quant that fits is Q6_K (11.3 GB file, about 18.5 GB in use).
Which Sarashina 13B Finetuned v2 quant should you choose?
Q4_K_M is the usual starting point for Sarashina 13B Finetuned v2: it balances file size and quality well. Here it is a 8.4 GB file.
If memory is tight, the smallest file is IQ1_S at 3.3 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 Sarashina 13B Finetuned v2
How much VRAM does Sarashina 13B Finetuned v2 need?
About 15.6 GB at Q4_K_M with an 8K context window. That is the 8.4 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 Sarashina 13B Finetuned v2 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 10.5 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Sarashina 13B Finetuned v2 GGUF?
It depends on the quant: IQ1_S is 3.3 GB, IQ1_M is 3.5 GB, IQ2_XXS is 3.9 GB, IQ2_XS is 4.3 GB, IQ2_S is 4.6 GB, Q2_K_S is 4.9 GB and 18 more.
What context length does Sarashina 13B Finetuned v2 support?
The GGUF header lists a trained context length of 4,096 tokens (4K). The KV cache grows with the context you use: about 6.7 GB at 8K and 3.4 GB at 4K.
Where can I download Sarashina 13B Finetuned v2 GGUF?
The files are on Hugging Face in the mradermacher/sarashina-13B-finetuned-v2-i1-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 Sarashina 13B Finetuned v2 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.
Can I run this model on 8 GB VRAM?
Yes, use IQ1_S or IQ1_M quantizations, which require about 10.5–10.7 GB memory at 8k context.
Does the model support image inputs?
No, the model does not have vision capabilities.
What license does the model use?
The model is released under the MIT license.