-
I have trained a seq2seq NMT model (EN-DE) with 1M samples and saved the latest checkpoint. Now, I have some domain-specific data of 50K sentence pairs which has **not** been seen in previous training…
-
For eg given the sentence, "the acting was good but direction was poor", I want to infer the (aspect, aspects-sentiment) pairs such as (acting,good), (direction,poor). How can I do that?
Kind regar…
-
# Task Name
Heteronym differentiation in the English language
## Task Objective
A heteronym is a word that has different pronunciations and meanings but the same spelling. For example, “attri…
-
Dear maintainers,
Thank you for the great work on the Sentence Transformers library.
I am writing to ask for clarification regarding the MultipleNegativesRankingLoss class. In the class document…
-
### Deep Learning Simplified Repository (Proposing new issue)
:red_circle: **Project Title** : Autocorrect using DL model
:red_circle: **Aim** : To make a model that uses text as input and fixes any…
-
> WordLlama begins by extracting the token embedding codebook from a state-of-the-art LLM (e.g., LLama3 70B), and training a small context-less model in a general purpose embedding framework.
I rea…
-
Hi, I'm now trying to use training data in generated_data/data-clipped for fine-tuning. I'm now using Roberta, but I found that there are still many sentence pairs over the limit of 512. Now I process…
-
I want to use the Class "ParallelSentencesDataset" to load my very big parallel data to fine tune the pre-trained model "LaBSE". But when I used it , it seems that this Class "ParallelSentencesDatase…
-
# Prosody Naturalness
## Task Objective
Evaluate the prosodic understanding level of the models. The task is part of the metatask #140.
## Datasets and Evaluation Metric
### General idea
- De…
-
Hey!
I'm a bit confused about the relationship within the . Let me break it down: If I want to align sentences from 10 pairs of source and target, would it work like this? Under the "summaries" fol…