File size: 1,200 Bytes
4ff80da
 
 
 
 
463b175
 
 
4ff80da
 
463b175
4ff80da
463b175
 
 
4ff80da
463b175
 
 
 
4ff80da
463b175
4ff80da
463b175
4ff80da
463b175
4ff80da
463b175
4ff80da
463b175
4ff80da
463b175
4ff80da
463b175
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
---
license: apache-2.0
base_model: syntheticlab/fix-json
tags:
- llama-cpp
- gguf
- lora
- merged-model
---

# fix-json (LoRA Merged GGUF)

**Model creator:** [syntheticlab](https://huggingface.co/syntheticlab)  
**Base model required:** [syntheticlab/fix-json](https://huggingface.co/syntheticlab/fix-json) (Meta-Llama-3.1-8B-Instruct)  
**GGUF conversion & quantization:** by [CronoBJS](https://huggingface.co/CronoBJS) using [`llama.cpp`](https://github.com/ggerganov/llama.cpp)  

⚠️ **Important:**  
This file is a **merged LoRA adapter in GGUF format**. It is **not** a standalone model.  
You **must** also have the original base GGUF model (e.g., `Meta-Llama-3.1-8B-Instruct.Q4_K_M.gguf`) to run it.  
Use the `--lora` or `--lora-scaled` flags in `llama.cpp` to apply it at runtime.

---

## Special thanks

🙏 Thanks to [Georgi Gerganov](https://github.com/ggerganov) and the [llama.cpp team](https://github.com/ggerganov/llama.cpp) for making all of this possible.

---

## Running the Model

### 1️⃣ llama.cpp CLI
```bash
llama-cli -m "path/to/Meta-Llama-3.1-8B-Instruct.Q4_K_M.gguf" \
  --lora "path/to/syntheticlab-fix-json-lora.gguf" \
  -c 2048 -n 256 -p "Your prompt here"