Request for the Dataset used in this project.

[night-fury-me]

Dear Sir,

From the readme section, I came to know that you are interested to share this dataset for research purpose. My group and I, are doing a project related to NLP, for which we are in need of a dataset of Bangla sentences. If you can share the dataset with us, it will be a great help.

Thanks in advance.

[firojalam]

Hi,
Thanks for your interest. Can you please write your details? Which institute are you from? We typically engage with the project that people are interested in. So please do let us know.

Firoj

[night-fury-me]

I am from North South University, Dept. of ECE. And It will be a course ( Neural Networks ) project. I have also emailed you regarding this issue.

[night-fury-me]

May I have your email? I would like to discuss about our research with you.

[firojalam]

Hi,
Sorry for the delayed reply. You can write us at [email protected] and [email protected].

Thank you

[Nesarul-Hoque]

How can I reduce the dimension from 300 to 100?

[firojalam]

There are several ways you can do that:

• Naive approach: just get the first 100,
• You can use PCA to reduce the dimension. Get top 100 principle dimensions,
• Train another model with 100 dimension

[Nesarul-Hoque]

Thank you for the quick response. I thought there might be a built-in function to reduce the dimensions like the fastText pre-trained model. Actually, I know about the three approaches to reducing the dimensions. However, the best way is to train the corpus again with 100 dimensions. So, can I get the link to the corpus which you used? [image: Mailtrack] <https://mailtrack.io?utm_source=gmail&utm_medium=signature&utm_campaign=signaturevirality11&> Sender notified by Mailtrack <https://mailtrack.io?utm_source=gmail&utm_medium=signature&utm_campaign=signaturevirality11&> 12/21/22, 11:14:08 AM

…

On Tue, Dec 20, 2022 at 6:44 PM Firoj Alam, Scientist, QCRI < ***@***.***> wrote: There are several ways you can do that: - Naive approach: just get the first 100, - You can use PCA to reduce the dimension. Get top 100 principle dimensions, - Train another model with 100 dimension — Reply to this email directly, view it on GitHub <#1 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A436L2JHVNLGVPBPTVFBDH3WOGS3FANCNFSM4FCWNEJA> . You are receiving this because you commented.Message ID: ***@***.*** com>

-- Md. Nesarul Hoque Assistant Professor Dept. of CSE Bangabandhu Sheikh Mujibur Rahman Science and Technology University Gopalganj 8100, Dhaka, Bangladesh Cell No: +8801712-820727

[firojalam]

Dear Nesarul-Hoque,
We do not have oroginal dataset now. And unfortunately, we could not share it before.

Firoj

[Nesarul-Hoque]

Thank you for your clarification. [image: Mailtrack] <https://mailtrack.io?utm_source=gmail&utm_medium=signature&utm_campaign=signaturevirality11&> Sender notified by Mailtrack <https://mailtrack.io?utm_source=gmail&utm_medium=signature&utm_campaign=signaturevirality11&> 12/27/22, 12:56:25 PM

…

On Mon, Dec 26, 2022 at 9:51 PM Firoj Alam, Scientist, QCRI < ***@***.***> wrote: Dear Nesarul-Hoque, We do not have oroginal dataset now. And unfortunately, we could not share it before. Firoj — Reply to this email directly, view it on GitHub <#1 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A436L2JXZIOCBDCZ4ZGKOBLWPG5JTANCNFSM4FCWNEJA> . You are receiving this because you commented.Message ID: ***@***.*** com>

-- Md. Nesarul Hoque Assistant Professor Dept. of CSE Bangabandhu Sheikh Mujibur Rahman Science and Technology University Gopalganj 8100, Dhaka, Bangladesh Cell No: +8801712-820727