File size: 814 Bytes
a46d242 9af596d a46d242 9af596d a46d242 |
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 |
https://github.com/albertan017/LLM4Decompile converted to gguf for use with https://github.com/ggerganov/llama.cpp
# Usage
```sh
./main --model ./models/llm4decompile-6.7b-uo-f16.gguf --threads 16 --color -c 2048 -n -1 --repeat-penalty 1.2 -ngl 33 --temp 0.7 -f prompts/llm4decompile.txt
```
`-ngl` and `--threads` values may be lowered to reduce gpu and cpu usage respectively
# Prompt Format
```
# This is the assembly code:
0000000000001139 <main>:
1139: push %rbp
113a: mov %rsp,%rbp
113d: sub $0x10,%rsp
1141: mov %edi,-0x4(%rbp)
1144: mov %rsi,-0x10(%rbp)
1148: lea 0xeb5(%rip),%rax
114f: mov %rax,%rdi
1152: call 1030 <puts@plt>
1157: mov $0x0,%eax
115c: leave
115d: ret
# What is the source code?
``` |