Sidharth Pulipaka

psidharth567

·

AI & ML interests

None yet

Organizations

New activity in ai4bharat/IndicBERT-v3-270M 2 months ago

attention_mask not honoured: real-token hidden states drift ~10x RMS with batch padding

#1 opened 2 months ago by

New activity in ai4bharat/Cadence 9 months ago

Update README.md

#18 opened 9 months ago by

New activity in ai4bharat/Cadence 10 months ago

Introducing Cadence-Fast

#16 opened 10 months ago by

Fixed annoying bug that caused it to not work right-of-the-bat

#15 opened 10 months ago by

Prevent Cadence from defaulting to Hindi punctuation for Marathi transcripts

#12 opened about 1 year ago by

New activity in ai4bharat/Cadence about 1 year ago

Updated config. Removed "use_cache": true and also "cache_implementation". They were interfering in some versions of transformers.

#13 opened about 1 year ago by

Updated the docs of the python package as we now officially support passing attn_implementation too.

#11 opened about 1 year ago by

Completely overhauled the attention implementation. Using the existing Gemma-3 attention implementation rather than custom monkey-patched implementation.

#10 opened about 1 year ago by

Changed file path "arxiv_logo.svg" to "Cadence/arxiv_logo.svg". One last try.

#9 opened about 1 year ago by

Upload arxiv_logo.svg

#8 opened about 1 year ago by

Update README.md

#7 opened about 1 year ago by

That previous logo from the link did not render correctly

#6 opened about 1 year ago by

Added arXiv logo beside the paper name.

#5 opened about 1 year ago by

Update README.md

#4 opened about 1 year ago by

Added the paper link

#3 opened about 1 year ago by

Update README.md

#2 opened about 1 year ago by

Update README.md

#1 opened about 1 year ago by