Wei Zheng comments

Results 4 comments of


                                            Wei Zheng

statisticsgen visualization does not work in jupyter under container

some information: I ran it on google colab and things work fine. I ran it locally (no docker), visualization tool is blank. I rank it on different browsers, same problem....

The loss curve exhibits a stair-step pattern of descent.

Hi, Does anyone notice that the eval loss diverge? I had many runs and most of them diverges. In some cases, the overfitted checkpoint produces better response (i.e. dulcet-shape-11 below,...

Poor results when fine-tuning with `alpaca_data.json` and suggested settings.

> > > The reason why it generated "### instruction" is because your fine-tuning is inefficient. In this case, we put a eos_token_id=2 into the tensor for each instance before...

Evaluation on super-NI (cannot reproduce official model's performance published on hf)

Hi @oleksost, I also saw the commit history on hf for the published model. The latest commit note was “actually masked loss”. That leads me to believe that the published...