bash scripts/train_tokenizer.sh python500k 400000 /opt/tiger/byte-lingua/superbpe/subsample_python.jsonl 1
bash scripts/train_tokenizer.sh python500k 400000 /opt/tiger/byte-lingua/superbpe/subsample_python.jsonl 2 tokenizer_json/python500k_vocab400K_stage1 250000


bash scripts/train_tokenizer.sh python500k 800000 /opt/tiger/byte-lingua/superbpe/subsample_python.jsonl 1
bash scripts/train_tokenizer.sh python500k 800000 /opt/tiger/byte-lingua/superbpe/subsample_python.jsonl 2 tokenizer_json/python500k_vocab800K_stage1 160000


bash scripts/train_tokenizer.sh opencoder300k 400000 /opt/tiger/byte-lingua/superbpe/subsample_opencoder.jsonl 1
bash scripts/train_tokenizer.sh opencoder300k 400000 /opt/tiger/byte-lingua/superbpe/subsample_opencoder.jsonl 2 tokenizer_json/opencoder300k_vocab400K_stage1 250000


bash scripts/train_tokenizer.sh opencoder300k 800000 /opt/tiger/byte-lingua/superbpe/subsample_opencoder.jsonl 1
bash scripts/train_tokenizer.sh opencoder300k 800000 /opt/tiger/byte-lingua/superbpe/subsample_opencoder.jsonl 2 tokenizer_json/opencoder300k_vocab800K_stage1 320000


python scripts/evaluate_compression_rate.py --eval_data_path eval_python.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/python500k_vocab400K_stage1/tokenizer.json
python scripts/evaluate_compression_rate.py --eval_data_path eval_python.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/python500k_vocab400K_from250K_stage2/tokenizer.json

python scripts/evaluate_compression_rate.py --eval_data_path eval_python.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/python500k_vocab800K_stage1/tokenizer.json
python scripts/evaluate_compression_rate.py --eval_data_path eval_python.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/python500k_vocab800K_from80K_stage2/tokenizer.json
python scripts/evaluate_compression_rate.py --eval_data_path eval_python.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/python500k_vocab800K_from160K_stage2/tokenizer.json
python scripts/evaluate_compression_rate.py --eval_data_path eval_python.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/python500k_vocab800K_from320K_stage2/tokenizer.json
python scripts/evaluate_compression_rate.py --eval_data_path eval_python.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/python500k_vocab800K_from480K_stage2/tokenizer.json
python scripts/evaluate_compression_rate.py --eval_data_path eval_python.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/python500k_vocab800K_from640K_stage2/tokenizer.json



# python scripts/evaluate_compression_rate.py --eval_data_path eval_opencoder.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/opencoder300k_vocab400K_stage1/tokenizer.json
# python scripts/evaluate_compression_rate.py --eval_data_path eval_opencoder.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/opencoder300k_vocab400K_from250K_stage2/tokenizer.json
# python scripts/evaluate_compression_rate.py --eval_data_path eval_opencoder.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/opencoder300k_vocab800K_stage1/tokenizer.json
# python scripts/evaluate_compression_rate.py --eval_data_path eval_opencoder.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/opencoder300k_vocab800K_from80K_stage2/tokenizer.json
# python scripts/evaluate_compression_rate.py --eval_data_path eval_opencoder.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/opencoder300k_vocab800K_from160K_stage2/tokenizer.json
# python scripts/evaluate_compression_rate.py --eval_data_path eval_opencoder.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/opencoder300k_vocab800K_from320K_stage2/tokenizer.json
# python scripts/evaluate_compression_rate.py --eval_data_path eval_opencoder.jsonl --tokenizer_path /opt/tiger/byte-lingua/superbpe/tokenizer_json/opencoder300k_vocab800K_from480K_stage2/tokenizer.json