-
Hello,
(Cross posting this between [SetFit](https://github.com/huggingface/setfit/issues/500) and sentence-transformers)
We're investigating the possibility to use SetFit for customer service me…
-
Hi, i have a fine tuned Cross Encoder model and i want to use the model for inference on sentence pairs, so reading the Docs, i found that there is a parameter called device when initializing the cro…
-
Hi,
Thanks for this wonderful repository. I am interested by semantic similarity between french sentences only for the moment and found the "stsb-xlm-r-multilingual" model to work pretty well.
I…
-
# Benchmarking embedding alignment with GPTs 🤖 and CLIPs 📎
Contacts: Mike (S), Marc
Participants:
## Goals and deliverable
1. Developing [AstroPT](https://github.com/Smith42/astroPT) to the p…
-
We might be able to find open source toxicity lists.
-
Hi everyone,
This is a small question related to how models are fine-tuned during the first step of training. I see that the default loss function is `losses.CosineSimilarityLoss`. But when generat…
-
Hello and thanks for the repo,
I think I have an issue with text validation. I have created a folia Doc with sentence elements and their text. I am machine translating it and then reading the machi…
-
Some language pairs are oddly missing from ...WikiMatrix/list_of_bitexts.txt, against my intuitions on which ones would have more data and thus more matching sentences.
For example, Armenian (`hy`)…
-
The task is to build a sequence-to-sequence (Seq2Seq) model with attention mechanism for language translation. The model will take a sentence in one language (e.g., English) as input and generate a tr…
-
> Remove sentences with source/target word ratio > 9
> Remove sentences with source/target word ratio < 0.1111
I am not sure what this means. I am looking at your repo for ideas on a Japanese to …