Self_Explaining_Structures_Improve_NLP_Models
Self_Explaining_Structures_Improve_NLP_Models copied to clipboard
ShannonAI
I ran four times but the results were only around 57%. Could you tell me what hyperparameters are to reach 59.1
I have tried the training script in this repository to train a model for SNLI dataset, and found that when training process continued, the predicted result of all valid cases...
In your paper, you reported the following results on SST-5 for RoBERTa without Self-Explaining as a baseline: Model | Accuracy -------|-------- RoBERTa-base | 56.4% RoBERTa-large | 57.9% The original paper...
https://github.com/ShannonAI/Self_Explaining_Structures_Improve_NLP_Models/blob/d8b0511cd06964dbb23f1c4560f01996662b0b66/explain/trainer.py#L90 Hi, thanks for sharing your work. I am a little bit confused about how you calculate your final loss. In your paper, it says it should the sum but...
roberta-base currently has been updated since. Do I just add "num_labels": 5 into the json still? Also currently running into error involving unable to load weights from pytorch checkpoint when...
I am wondering what are the exact hyperparameters for the model to get the same results mentioned in the paper. Since the seed is set fixed, I wished running the...
Tried to allocate 20.00 MiB (GPU 0; 1.96 GiB total capacity; 1.24 GiB already allocated; 2.88 MiB free; 1.25 GiB reserved in total by PyTorch)
