Full training results & evaluation with BERTScore

Medium training run

Fixed compiling issue, added legnth penalty, and atttempting freezing encoder layers 0-5 to lower parameters and preserve T5's langauge understanding.

Improve emotion task architecture and fix visualizations

Clean up codebase and fix training bugs

Redesign Gradio demo with book/news browsing, update .gitignore

Update Gradio demo, inference factory, and evaluation results

Update LexiMind: improved training, model architecture, and evaluation

Improve summarization output quality

Fix Pylance type errors, add inductor compilation support

Full training run, code cleanup, mypy/ruff fixes

Training run: dev config 1 epoch results

feat: Add FLAN-T5 compatibility with relative position bias

Build: Pin ruff to v0.4.4 in pre-commit to match CI

Refactor: Fix B905 zip() strictness and resolve mypy error in evaluate.py

Fix: Resolve mypy type errors and configuration

Fix: Resolve mypy type errors and configuration

Chore: Update pre-commit hooks and fix formatting

Style: Fix formatting for models and tests

Style: Apply ruff formatting

Style: Fix linting errors and organize imports (ruff & mypy)

Refactor: Consolidate dependencies, improve testing, and add CI/CD

refactor: update training code, history and remove large checkpoint to fix storage limit

Added pretrained BART wights for the encoder/decoder while keeping custom architectiure and task heads

Summarization fix through pipeline

Summarization fix

Summarization fix

Summarization fix

Summarization fix

Refactor: Fix tokenization, pathing, and track visuals with LFS

Updated Greedy Decode w/ a minlen search

Updated Gradio Demo to include model visualizations

Fixing Model Pipeline Problem

chore: snapshot current refinements

Updated Summarizer, Preprocessor to run on my custom transformer and added basic streamlit frontend demo

Finsihed Decoder Implementation, as well as prediction heads and multitask

Implemented the following parts for the Transformer model: poisional_encoding, feedfoward, encoder, and skeleton of decoder. As well as Test cases for each class respectively and visualizations

Implemented ScaledDotProduct Attention and Multi-Head Attention

Reformatted Project Structure

Finished Preprocessor and Implemented Summarizer with PyTorch

Implemented Proprocessing for each dataset

Updating preprocessor.py and download_datasets.py to download and start processing data

LexiMind Initial Project Structure