shamanez/bert-like-is-all-you-need issues and pull requests

#25 - Incorrect metrics calculation for mosi/mosei

Issue - State: closed - Opened by cyZhu98 about 3 years ago - 2 comments

#24 - achived low accuracy on MOSI dataset

Issue - State: closed - Opened by cyZhu98 about 3 years ago - 4 comments

#23 - Question about requirements

Issue - State: closed - Opened by jiminbot20 about 3 years ago - 1 comment

#22 - about the pre-trained model fine tuning

Issue - State: closed - Opened by 245789064 over 3 years ago

#21 - Use of Pre-trained Speech-BERT

Issue - State: closed - Opened by mingzi151 over 3 years ago - 6 comments

#20 - If you use the following command, it seems a binary accuracy command.

Issue - State: closed - Opened by 245789064 over 3 years ago - 2 comments

#19 - ValueError: tokens exceeds maximum length: 550 > 512

Issue - State: closed - Opened by Bharath1991999 over 3 years ago - 6 comments

#18 - Does this repository not contain IEMOCAP with 4-class?

Issue - State: closed - Opened by Chenpuh over 3 years ago - 10 comments

#17 - When I did the validation, preds was [1, 0, 0, 1].

Issue - State: closed - Opened by FriedaSmith over 3 years ago - 2 comments

#16 - how to fine-tune the model in your paper

Issue - State: closed - Opened by 245789064 over 3 years ago - 1 comment

#15 - The training accuracy is more than 100.

Issue - State: closed - Opened by FriedaSmith over 3 years ago - 6 comments

#14 - CUDA out of memory when trying to train on MOSI dataset

Issue - State: closed - Opened by shuangwei123 over 3 years ago - 3 comments

#13 - Fusion

Issue - State: closed - Opened by mahsaShafaei over 3 years ago - 1 comment

#12 - Why only use the data of dimension 0?

Issue - State: closed - Opened by FriedaSmith over 3 years ago - 2 comments

#11 - query about speech feature

Issue - State: closed - Opened by jun0wanan over 3 years ago

#10 - envrionment information and training commands of other datasets

Issue - State: closed - Opened by melodymelodymelody over 3 years ago - 5 comments

#9 - How to process the raw audio in IEMOCAP?

Issue - State: closed - Opened by FriedaSmith almost 4 years ago - 2 comments

#8 - You use GPT-2 for raw text preprocessing in your paper .I can not find it in your script. Can you tell me which script is for it? Thank you!

Issue - State: closed - Opened by 245789064 almost 4 years ago - 3 comments

#7 - RuntimeError: DataLoader worker (pid 3019) is killed by signal: Killed.

Issue - State: closed - Opened by 245789064 almost 4 years ago - 2 comments

#6 - about input for convert_aud_to_token.py

Issue - State: closed - Opened by Nie-Yingying almost 4 years ago - 11 comments

#5 - data

Issue - State: closed - Opened by cn-boop almost 4 years ago - 1 comment

#4 - pretrained roberta model and dataset size mismatch

Issue - State: closed - Opened by sniperwrb about 4 years ago - 2 comments

#3 - AssertionError: "misalignment of data"

Issue - State: closed - Opened by cuczy about 4 years ago - 5 comments

#2 - error: argument --arch/-a: invalid choice: 'robertaEMO'

Issue - State: closed - Opened by cuczy about 4 years ago - 16 comments

#1 - The preprocessed datasets and their labels can not be accessed.

Issue - State: closed - Opened by mazicwong about 4 years ago - 6 comments

Ecosyste.ms: Issues

GitHub / shamanez/bert-like-is-all-you-need issues and pull requests