CodeLlama 13B MORepair needs about 15.1 GB of memory at Q4_K_M (a 7.9 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 25 quants, from 4.2 GB (IQ2_S) to 26.0 GB (F16).
CodeLlama 13B MORepair – a fine‑tuned program repair model
CodeLlama 13B MORepair is a 13‑billion‑parameter model fine‑tuned from CodeLlama‑13b‑instruct using the MOREPAIR framework. It focuses on automated program repair, primarily tested on C++ and Java. The model is released under the Llama 2 Community License.
Good at
- Automated program repair
- Code transformation with explanatory guidance
- Multi‑objective fine‑tuning
Limits
- No vision support
- Requires 8‑bit or higher quantization for efficient inference
- Large VRAM needed for higher‑quality quants
Best for
- Repairing buggy C++ code
- Fixing Java programs
- Generating repair explanations
Who it’s for. Developers and researchers needing a specialized tool for automated program repair. The model is suitable for those who can run llama.cpp or similar inference engines and have access to sufficient VRAM for the chosen quantization.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ2_S | 4.2 GB | 11.4 GB | Too big | Get file → | |
| IQ2_M | 4.5 GB | 11.7 GB | Too big | Get file → | |
| Q2_K | 4.9 GB | 12.1 GB | Too big | Get file → | |
| Q2_K_L | 5.0 GB | 12.2 GB | Too big | Get file → | |
| IQ3_XS | 5.4 GB | 12.6 GB | Too big | Get file → | |
| Q3_K_S | 5.7 GB | 12.9 GB | Too big | Get file → | |
| IQ3_M | 6.0 GB | 13.2 GB | Too big | Get file → | |
| Q3_K_M | 6.3 GB | 13.5 GB | Too big | Get file → | |
| Q3_K_L | 6.9 GB | 14.1 GB | Too big | Get file → | |
| IQ4_XS | 7.0 GB | 14.2 GB | Too big | Get file → | |
| Q3_K_XL | 7.1 GB | 14.3 GB | Too big | Get file → | |
| Q4_0 | 7.4 GB | 14.6 GB | Too big | Get file → | |
| Q4_0 | 7.4 GB | 14.6 GB | Too big | Get file → | |
| Q4_0 | 7.4 GB | 14.6 GB | Too big | Get file → | |
| Q4_0 | 7.4 GB | 14.6 GB | Too big | Get file → | |
| Q4_K_S | 7.4 GB | 14.6 GB | Too big | Get file → | |
| Q4_K_M | 7.9 GB | 15.1 GB | Too big | Get file → | |
| Q4_K_L | 8.0 GB | 15.2 GB | Too big | Get file → | |
| Q5_K_S | 9.0 GB | 16.2 GB | Too big | Get file → | |
| Q5_K_M | 9.2 GB | 16.4 GB | Too big | Get file → | |
| Q5_K_L | 9.3 GB | 16.5 GB | Too big | Get file → | |
| Q6_K | 10.7 GB | 17.9 GB | Too big | Get file → | |
| Q6_K_L | 10.8 GB | 18.0 GB | Too big | Get file → | |
| Q8_0 | 13.8 GB | 21.0 GB | Too big | Get file → | |
| F16 | 26.0 GB | 33.2 GB | Too big | Get file → |
For 8‑GB VRAM, use IQ2_S (4.2 GB file, 11.4 GB memory at 8k context) or IQ2_M (4.52 GB file, 11.7 GB memory). For 16‑GB VRAM, Q4_K_M (7.87 GB file, 15.1 GB memory) or Q5_K_S (8.97 GB file, 16.2 GB memory) provide a good balance of size and performance. Higher quants like Q6_K (10.68 GB file, 17.9 GB memory) require more VRAM.
What GPU do you need for CodeLlama 13B MORepair?
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_M (4.5 GB file, about 11.7 GB in use).
- 16 GB VRAM: the best quant that fits is Q4_K_L (8.0 GB file, about 15.2 GB in use).
- 24 GB VRAM: the best quant that fits is Q8_0 (13.8 GB file, about 21.0 GB in use).
- 32 GB VRAM: the best quant that fits is Q8_0 (13.8 GB file, about 21.0 GB in use).
- 48 GB VRAM: the best quant that fits is F16 (26.0 GB file, about 33.2 GB in use).
Which CodeLlama 13B MORepair quant should you choose?
Q4_K_M is the usual starting point for CodeLlama 13B MORepair: it balances file size and quality well. Here it is a 7.9 GB file.
The Q8_0 file is 13.8 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 IQ2_S at 4.2 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 CodeLlama 13B MORepair
How much VRAM does CodeLlama 13B MORepair need?
About 15.1 GB at Q4_K_M with an 8K context window. That is the 7.9 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 CodeLlama 13B MORepair on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 11.4 GB, so you would need a larger GPU or CPU offloading.
What is the file size of CodeLlama 13B MORepair GGUF?
It depends on the quant: IQ2_S is 4.2 GB, IQ2_M is 4.5 GB, Q2_K is 4.9 GB, Q2_K_L is 5.0 GB, IQ3_XS is 5.4 GB, Q3_K_S is 5.7 GB and 19 more.
What context length does CodeLlama 13B MORepair support?
The GGUF header lists a trained context length of 16,384 tokens (16K). The KV cache grows with the context you use: about 6.7 GB at 8K and 13.4 GB at 16K.
Where can I download CodeLlama 13B MORepair GGUF?
The files are on Hugging Face in the bartowski/CodeLlama-13B-MORepair-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 CodeLlama 13B MORepair use?
Hugging Face lists the license as llama2. 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 CodeLlama 13B MORepair on 8 GB VRAM?
Yes, by using a low‑bit quant such as IQ2_S or IQ2_M, which require about 11.4–11.7 GB memory at 8k context.
Does this model support vision tasks?
No, the model is designed for coding tasks only and has no vision capabilities.
What license applies to this model?
The model is released under the Llama 2 Community License.