LLM4Decompile icon indicating copy to clipboard operation
LLM4Decompile copied to clipboard

Update and rename run_evaluation_llm4decompile_singleGPU.py to update…

Open thisisvk45 opened this issue 1 year ago • 1 comments

…d_run_evaluation_llm4decompile_singleGPU.py

The optimized code improves efficiency by minimizing file operations and subprocess calls. It consolidates file writes and uses temporary files more effectively, reducing redundant operations. By combining compile and run commands and handling exceptions concisely, the code avoids unnecessary subprocess calls.

String operations are streamlined using str.splitlines() and str.strip() to handle and process strings more efficiently. Additionally, the prompt preparation and decoding steps are simplified to speed up data processing. These changes collectively enhance the time and space efficiency of the code.

thisisvk45 avatar Aug 16 '24 20:08 thisisvk45

Thank you for your interest. However, we haven't observed any speed improvements with the script. It would be beneficial if you could use tempfile to handle the files, as implemented in run_evaluation_llm4decompile_vllm.py.

albertan017 avatar Sep 21 '24 10:09 albertan017