Download scripts/01_train_tokenizer.sh from lollix91/boccaccioAI: direct link, hf CLI and curl.
- Browser
- Download file 353 Bytes
-
https://huggingface.co/lollix91/boccaccioAI/resolve/main/scripts/01_train_tokenizer.sh
- Command line
-
hf download hf://lollix91/boccaccioAI/scripts/01_train_tokenizer.sh
-
curl -L -o 01_train_tokenizer.sh https://huggingface.co/lollix91/boccaccioAI/resolve/main/scripts/01_train_tokenizer.sh
353 Bytes
| # BoccaccioAI - Fase 1: Train BPE Tokenizer | |
| # De Lauretis Tech | |
| set -euo pipefail | |
| echo "=== BoccaccioAI - Training Tokenizer BPE 32K ===" | |
| python -m src.tokenizer.train_tokenizer \ | |
| --config configs/tokenizer.yaml \ | |
| --output-dir tokenizer/ \ | |
| --vocab-size 32000 \ | |
| --corpus-size-gb 5 | |
| echo "=== Tokenizer training complete ===" | |