Jun-jie-Huang / CoCLR

MIT License
43 stars 8 forks source link

The pre-trained model on the task of code search #1

Closed skye95git closed 3 years ago

skye95git commented 3 years ago

In the Model Checkpoint section, there is the link of the checkpoint with best code question answering results. Can you share your pre-trained model on the task of code search?

Jun-jie-Huang commented 3 years ago

Do you mean the checkpoint trained on CodeSearchNet or trained on CoSQA?

skye95git commented 3 years ago

Do you mean the checkpoint trained on CodeSearchNet or trained on CoSQA?

Yes, the checkpoint with best code search results.

Jun-jie-Huang commented 3 years ago

OK, I will upload the checkpoint asap.

skye95git commented 3 years ago

OK, I will upload the checkpoint asap.

Thanks. The work is cool.

Jun-jie-Huang commented 3 years ago

Hi, the checkpoint on the task of code search has been uploaded. Please refer to this.

skye95git commented 3 years ago

Hi, the checkpoint on the task of code search has been uploaded. Please refer to this.

Thanks! Was this checkpoint trained on CoSQA?

Jun-jie-Huang commented 3 years ago

Hi, the checkpoint on the task of code search has been uploaded. Please refer to this.

Thanks! Was this checkpoint trained on CoSQA?

Yes

skye95git commented 3 years ago

Hi, the checkpoint on the task of code search has been uploaded. Please refer to this.

Hi, I make dir ./model/codesearchnet-checkpoint and save the checkpoint trained on CodeSearchNet in it.

捕获

Then I make dir ./model/search_codebert_switch.

捕获

When I run the step3 in CoCLR on Code Search section, there is an error: image

I search the error on the internet:

捕获

But I don't konw which dimensionality should be changed. What should I do? Did I do something wrong?

I hava another question: I download the checkpoint with best code search results you provided and move then to ./model/:

捕获

Can this checkpoint be used directly for evaluation? Why is there an error when I use it?

捕获

Isn't the checkpoint stored in the ./model/?

Jun-jie-Huang commented 3 years ago

Hi, the checkpoint on the task of code search has been uploaded. Please refer to this.

Hi, I make dir ./model/codesearchnet-checkpoint and save the checkpoint trained on CodeSearchNet in it.

捕获

Then I make dir ./model/search_codebert_switch.

捕获

When I run the step3 in CoCLR on Code Search section, there is an error: image

I search the error on the internet:

捕获

But I don't konw which dimensionality should be changed. What should I do? Did I do something wrong?

I hava another question: I download the checkpoint with best code search results you provided and move then to ./model/:

捕获

Can this checkpoint be used directly for evaluation? Why is there an error when I use it?

捕获

Isn't the checkpoint stored in the ./model/?

Question 2: I think made a mistake that I uploaded a checkpoint for QA. I have updated the link now with the right code search model. Sorry about that..

skye95git commented 3 years ago

Hi, the checkpoint on the task of code search has been uploaded. Please refer to this.

Hi, I make dir ./model/codesearchnet-checkpoint and save the checkpoint trained on CodeSearchNet in it.

捕获

Then I make dir ./model/search_codebert_switch.

捕获

When I run the step3 in CoCLR on Code Search section, there is an error: image I search the error on the internet:

捕获

But I don't konw which dimensionality should be changed. What should I do? Did I do something wrong? I hava another question: I download the checkpoint with best code search results you provided and move then to ./model/:

捕获

Can this checkpoint be used directly for evaluation? Why is there an error when I use it?

捕获

Isn't the checkpoint stored in the ./model/?

Question 2: I think made a mistake that I uploaded a checkpoint for QA. I have updated the link now with the right code search model. Sorry about that..

Thanks for your reply! Are downloaded file checkpoint-test-mrr stored in the ./model folder or ./model/search_codebert_switch folder?

skye95git commented 3 years ago

Hi, the checkpoint on the task of code search has been uploaded. Please refer to this.

Hi, I make dir ./model/codesearchnet-checkpoint and save the checkpoint trained on CodeSearchNet in it.

捕获

Then I make dir ./model/search_codebert_switch.

捕获

When I run the step3 in CoCLR on Code Search section, there is an error: image I search the error on the internet:

捕获

But I don't konw which dimensionality should be changed. What should I do? Did I do something wrong? I hava another question: I download the checkpoint with best code search results you provided and move then to ./model/:

捕获

Can this checkpoint be used directly for evaluation? Why is there an error when I use it?

捕获

Isn't the checkpoint stored in the ./model/?

Question 2: I think made a mistake that I uploaded a checkpoint for QA. I have updated the link now with the right code search model. Sorry about that..

For the question1, I reloaded cuda and PyTorch and the model can now train properly.

skye95git commented 3 years ago

Hi, the checkpoint on the task of code search has been uploaded. Please refer to this.

Hi, I make dir ./model/codesearchnet-checkpoint and save the checkpoint trained on CodeSearchNet in it.

捕获

Then I make dir ./model/search_codebert_switch.

捕获

When I run the step3 in CoCLR on Code Search section, there is an error: image I search the error on the internet:

捕获

But I don't konw which dimensionality should be changed. What should I do? Did I do something wrong? I hava another question: I download the checkpoint with best code search results you provided and move then to ./model/:

捕获

Can this checkpoint be used directly for evaluation? Why is there an error when I use it?

捕获

Isn't the checkpoint stored in the ./model/?

Question 2: I think made a mistake that I uploaded a checkpoint for QA. I have updated the link now with the right code search model. Sorry about that..

Thanks for your reply! Are downloaded file checkpoint-test-mrr stored in the ./model folder or ./model/search_codebert_switch folder?

Thank you. I know the answer.