doc-to-lora/scripts/main_exp/eval/d2l.sh

14 lines
1.2 KiB
Bash
Raw Permalink Normal View History

2025-09-30 14:53:21 +00:00
# main results
# batched
2026-06-15 04:31:47 +00:00
WANDB_MODE=disabled uv run run_eval.py --checkpoint_path train_outputs/runs/$RUN_NAME/checkpoint-$step/pytorch_model.bin --datasets squad drop ropes longbench/qasper_e longbench/2wikimqa_e longbench/multifieldqa_en_e longbench/gov_report_e --split test --max_ctx_chunk_len 8192 --eval_batch_size_gen 1
2025-09-30 14:53:21 +00:00
# iterative
2026-06-15 04:31:47 +00:00
WANDB_MODE=disabled uv run run_eval.py --checkpoint_path train_outputs/runs/$RUN_NAME/checkpoint-$step/pytorch_model.bin --datasets squad drop ropes longbench/qasper_e longbench/2wikimqa_e longbench/multifieldqa_en_e longbench/gov_report_e --split test --max_ctx_chunk_len 8192 --eval_batch_size_gen 1 --use_iterative_mode
2025-09-30 14:53:21 +00:00
# query internalization
WANDB_MODE=disabled uv run run_eval.py --checkpoint_path train_outputs/runs/$RUN_NAME/checkpoint-$step/pytorch_model.bin --datasets squad --split test --eval_batch_size_gen=1 --flip_ctx_inp
2025-09-30 14:53:21 +00:00
# replaced squad context
WANDB_MODE=disabled uv run python run_eval.py --checkpoint_path train_outputs/runs/$RUN_NAME/checkpoint-$step/pytorch_model.bin --datasets squad_assistant_ctx_no_passage --split test
WANDB_MODE=disabled uv run python run_eval.py --checkpoint_path train_outputs/runs/$RUN_NAME/checkpoint-$step/pytorch_model.bin --datasets squad_negative_no_passage --split test