Abhinav M Kulkarni

Relationship between Lexicon and LM vocab

4

Hi, Perhaps, this is a naive question, but do we need to retain non-Lexicon words in KenLM file? I haven't read the KenLM paper but do backoff and smoothing require...

Difference between sota/2019/am_tds_ctc and streaming_convnets/librispeech/am_500ms_future_context models?

4

### Question Hi, Other than the architecture what is the difference between [sota/2019/am_tds_ctc](https://github.com/facebookresearch/wav2letter/blob/master/recipes/sota/2019/librispeech/train_am_tds_ctc.cfg) and [streaming_convnets/librispeech/am_500ms_future_context](https://github.com/facebookresearch/wav2letter/blob/master/recipes/streaming_convnets/librispeech/train_am_500ms_future_context.cfg) models? I am able to convert the latter to FBGEMM streaming convnet using the conversion...

Streaming convnets with lexicon free decoder?

13

### Question Hi, Is there a way to run streaming convnets examples (Interactive/Streaming ASR examples) with a `LexiconFreeDecoder`? I do see a class for `LexiconFreeDecoder` [in the source code](https://github.com/facebookresearch/wav2letter/blob/v0.2/src/libraries/decoder/LexiconFreeDecoder.h). In...

### System Info ```$ uname -a Linux knockdhu 5.4.0-139-generic #156-Ubuntu SMP Fri Jan 20 17:27:18 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux ``` ### Who can help? @hwchase17 @agola11 @vowelparrot ###...

3X slow inference on GeForce RTX 3060 after 4bit-128g quantization

17

Hi, Thanks for the great work! You can find the code and the output in the attached Jupyter notebook: [quantize.ipynb](https://github.com/abhinavkulkarni/sharing-github/blob/main/quantize.ipynb) I am attaching a report of my experiment of quantizing...

Unable to copy/paste from Jypyter notebook cell to Continue text area

5

Hi, Lets say I have submitted a query to Continue and gotten a response. To further the conversation, I want to copy over some code (or error msg) from a...

Hi, I have the following in the list of context providers: ``` "contextProviders": [ { "name": "codebase", "params": { "nRetrieve": 25, "nFinal": 5, "useReranking": true } }, { "name": "search"...

Is it possible to add metadata filtering logic to ColBERTv2?

9

Hi, How difficult would it be to add metadata filtering logic to ColBERTv2? E.g., I would like to index books (along with metadata such as `publisher_name`, `publish_date`, etc.) and search...

A short tutorial on how to train a smaller biencoder model on custom dataset

1

Hi, I have seen a lot of comments asking how to create a custom dataset or how to use a smaller, different BERT base model for biencoder or how to...

How to generate embeddings for new candidates?

6

Hi, I am been going through the code, documentation and issues to figure out how to obtain embeddings for new candidates - however I have not been able to figure...