search

HowieHwong / MetaTool

[ICLR 2024] MetaTool Benchmark for Large Language Models: Deciding Whether to Use Tools and Which to Use
MIT License
70 stars 8 forks source link

FileNotFoundError: [Errno 2] No such file or directory: 'tool/test_data/vicuna-7b' #6

Closed sian-chen closed 11 hours ago

sian-chen commented 10 months ago

Dear Authors,

In src/generation/run.py, there are many places requiring data from tool/test_data or tool/new_test_res. But these data are neither provided in the repository nor generated after prompt construction. Could you please confirm where can we access the data? Thanks.

HowieHwong commented 10 months ago

Dear Sian,

The file path can be defined by yourself. What you need to do is replace it with your file path of the dataset. I will update is in README.md ASAP.

sian-chen commented 10 months ago

Hi @HowieHwong,

Could you explain more about which file should we use? It is the file generated by prompt_construction.py? It would be very helpful if there are some scripts to reproduce the results in the paper. Thanks!

HowieHwong commented 10 months ago

Yes. That's it.

Hi @HowieHwong,

Could you explain more about which file should we use? It is the file generated by prompt_construction.py? It would be very helpful if there are some scripts to reproduce the results in the paper. Thanks!

HowieHwong commented 10 months ago

Moreover, I have uploaded the datasets in dataset/tmp_dataset. You can check about it.

Hi @HowieHwong,

Could you explain more about which file should we use? It is the file generated by prompt_construction.py? It would be very helpful if there are some scripts to reproduce the results in the paper. Thanks!

wenjunli-0 commented 4 months ago

Moreover, I have uploaded the datasets in dataset/tmp_dataset. You can check about it.

Hi @HowieHwong, Could you explain more about which file should we use? It is the file generated by prompt_construction.py? It would be very helpful if there are some scripts to reproduce the results in the paper. Thanks!

Hi @HowieHwong, thanks for the excellent work.

Can you give a specific setting for tool/test_data and tool/new_test_res/?

I saw your updated files in dataset/tmp_dataset/, but there are also .json files (under /prompt_data/) generated by prompt_construction.py. So I am a bit confused about how to map tool/test_data and tool/new_test_res/ to the new directories.

HowieHwong commented 4 months ago

Moreover, I have uploaded the datasets in dataset/tmp_dataset. You can check about it.

Hi @HowieHwong, Could you explain more about which file should we use? It is the file generated by prompt_construction.py? It would be very helpful if there are some scripts to reproduce the results in the paper. Thanks!

Hi @HowieHwong, thanks for the excellent work.

Can you give a specific setting for tool/test_data and tool/new_test_res/?

I saw your updated files in dataset/tmp_dataset/, but there are also .json files (under /prompt_data/) generated by prompt_construction.py. So I am a bit confused about how to map tool/test_data and tool/new_test_res/ to the new directories.

Hi Wenjun,

You can manually create the dictionary in your project. Since the time is a bit distant, one possible is that the two of them are the same dict. This may be due to my handwriting the wrong code.