v2.0.0

What's Changed

• Fix loading tokenizers from s3 by @meilame-tayebjee in #71
• fix: use save_pretrained from PreTrainedTokenizerFast object by @meilame-tayebjee in #73
• Fix label att parametrization by @meilame-tayebjee in #74
• fix: adapt tests by @meilame-tayebjee in #75
• Enable raw inputs by @meilame-tayebjee in #83
• Refactor TextEmbedder by @meilame-tayebjee in #86
• doc: adapt read me by @meilame-tayebjee in #87
• chore(deps): update dependency uv_build to >=0.10.0,<0.11.0 by @renovate[bot] in #72
• chore(deps): update dependency astral-sh/uv to v0.11.7 by @renovate[bot] in #70
• chore(deps): update actions/upload-pages-artifact action to v5 by @renovate[bot] in #82
• chore(deps): update actions/deploy-pages action to v5 by @renovate[bot] in #80
• chore(deps): update dependency python to 3.14 by @renovate[bot] in #81
• chore(deps): update dependency uv_build to >=0.11.7,<0.12.0 by @renovate[bot] in #88
• Enable custom model construction by @meilame-tayebjee in #90
• fix: correct examples configuration and fix validation data never being checked by @julber95 in #91
• fix: correct LabelAttention mask by @julber95 in #92
• fix: align input tensors to model device in predict instead of cpu by @julber95 in #94
• fix: correctly mask padding tokens when computing label attention matrix by @julber95 in #96
• chore(deps): update dependency astral-sh/uv to v0.11.19 by @renovate[bot] in #89
• chore(deps): update astral-sh/setup-uv action to v8 by @renovate[bot] in #84
• Fix device mismatch in captum attribution target by @julber95 in #97
• feat: support per-sample loss weighting (sample_weights) by @meilame-tayebjee in #100

New Contributors

• @julber95 made their first contribution in #91

Full Changelog: v1.0.4...v2.0.0