Apr 11, 2025 01:39 Upload data_augmentation_final.tsv Note: all of these are NEW queries. This does not include the queries we wrote earlier.
- took out the clippers
- took out nulls, 0 result queries that didn't make sense
- removed queries that mentioned championship season
overall got rid of ~200
will make another one just for the clippers due to the naming convention issue. 00:46 Fix bug on function evaluation Apr 10, 2025 23:07 Upload data_augmentation.tsv file (still need to go through queries and check if they are valid) 22:56 Modify path to relative path 20:50 Add relative path functionality for colab 20:37 colab support part two 20:28 Add support for colab 20:06 Add a source folder Apr 9, 2025 22:08 Move paht to a variable 19:59 Create demo_rag.py 19:49 Create rag_metadata.py 19:48 Create test_rag.py 19:13 Add column evaluator 05:20 added fine-tune testing on validation set with short queries only 04:09 Adding validation set testing of newly fine-tuned model Apr 8, 2025 21:12 switch fine-tuned evaluation script to use 8 rank model 20:58 trained 8 rank LoRA finetuned model, saving train and validation datasets to disk Apr 7, 2025 19:11 first cut at fine-tuning with masked training, working but very very slow for inference 02:36 fixed bug in test_finetuned notebook to read from proper directory 02:13 Storing 16 and 32 rank LoRA finetuned models 01:06 Increasing LoRA rank to 32 and updating loss function Apr 6, 2025 22:13 first cut at adding data augmentation python notebook Apr 4, 2025 03:57 added python notebook for testing finetuned model Apr 3, 2025 23:38 First decent attempt, seems to generate valid queries after pre-training. Needs evaluation still 19:26 First successful fine-tuning run, resulting model is no good but it does train successfully 19:07 Initial attempt at fine-tuning using LoRA with basic cross-entropy loss Apr 2, 2025 23:12 ran full testing with updated prompt, improves team and other_stats performance, slight drop on game performance 19:01 trying to prompt engineer fixes for other_stats table 04:36 Adding completed pre-training testing runs to python notebook 02:39 Created evaluation loop for running on full dataframes 02:15 Merge branch 'main' of https://huggingface.co/USC-Applied-NLP-Group/SQL-Generation into main 02:14 Updated evaluation function to give tolerance for slight floating point differences Apr 1, 2025 23:42 train data eda notebook 23:39 Classify training data 22:31 Test 22:14 fixed issue with inconsistent formatting in training data 21:51 updating prompt to include team name and abbreviations 21:47 Small tweaks to evaluation function 21:45 updating test pretrained script to evaluate model outputs 20:24 Renamed demo script and added initial pre-trained test python notebook 18:48 Adding SQL-NLP training data Mar 26, 2025 16:36 added better prompt engineering 05:19 tweaking prompt a little further so it only outputs SQL more consistently 05:00 Updated to use real nba database for test query Mar 25, 2025 23:40 added nba kaggle sqlite database and testing python notebook 03:06 Update README.md 03:04 Create README.md 03:03 Delete README.md 01:49 adding test code for deepseek coder model Mar 3, 2025 05:50 initial commit 
