evaluating sewing instructions

• exploratoryanalysis.ipynb: overview of data
• model_input: pattern overview images, pattern pieces descriptions
• gptbaseline_promptingstrategies.ipynb: baseline prompts & improved prompting strategies
• model_output: model generated sewing instructions instructions
• reworkedsewinginstructions: manually crafted sewing instructions for evaluation purposes
• shufflesteps.ipynb: creating counterexamples
• traditional_eval.ipynb: BLEU, ROUGE-L, BERT-Score
• tree_eval.ipynb: tree evaluation
• tree_eval: manually annotated gold trees for each pattern, automatically extracted trees for each pattern for each set of generated instructions
• results.csv: all scores over all patterns plus manually annotated error count
• human_eval: human evaluation experiment data