v0.0.3
·
62 commits
to master
since this release
6856ad8
This commit was signed with the committer’s verified signature.
ZhiyuanChen
Zhiyuan Chen
Commits
- d99eeb5: run tests before publishing (Zhiyuan Chen)
- 89d52ae: fix rotary embedding in UtrLm and RnaFm (Zhiyuan Chen)
- 461ba05: fix attention mask of RnaMsm (Zhiyuan Chen)
- 3fcb921: fix head for utrlm and utrbert (Zhiyuan Chen)
- 2968292: unify classification head (Zhiyuan Chen)
- 9f8b2a7: rename Outputs of RnaMsm (Zhiyuan Chen)
- bde2098: add support of CrisprOffTarget tasks (Zhiyuan Chen)
- 83572c7: set default value of head in code (Zhiyuan Chen)
- 3747c05: improve token initialisation (Zhiyuan Chen)
- 9a41985: support delete_existing in convert_checkpoint (Zhiyuan Chen)
- 8011687: ensure pooler consistency (Zhiyuan Chen)
- 57039f7: rename tokenizers -> tokenisers to avoid conflicts (Zhiyuan Chen)
- e06fb22: relax the input_ids in ContactPredictionHead (Zhiyuan Chen)
- e2702d3: use strameline vocab by default if nmers > 1 (Zhiyuan Chen)
- 4b1cd97: add KMerTokenHead (Zhiyuan Chen)
- 3f9ec11: add head for nucleotide-level tasks (Zhiyuan Chen)
- 91b0201: add support of Nucleotide level tasks (Zhiyuan Chen)
- 449b56e: pytest now fails under 50 (Zhiyuan Chen)
- cb98098: set default transform_act to gelu (Zhiyuan Chen)
- 8e31573: fix type hints (Zhiyuan Chen)
- f74b91f: inspect if input_ids is NestedTensor (Zhiyuan Chen)
- 4ff51f9: ignore convert_checkpoint.py in testing (Zhiyuan Chen)
- c7cebc7: unify typehints (Zhiyuan Chen)
- 6c53582: unify criterion & build criterion into head (Zhiyuan Chen)
- f8c825b: make multimolecule.module for shared modulet diff develop (Zhiyuan Chen)
- 5dfc33c: fix token & nucleodite head (Zhiyuan Chen)
- b0c2b63: use mcrmse for multi regression tasks (Zhiyuan Chen)
- ce325f0: allow additional kwargs in Nucleotide-level tasks (Zhiyuan Chen)
- b61796b: add AutoModelForNucleotideClassification & AutoModelForCrisprOffTarget (Zhiyuan Chen)
- b981919: set model_max_length in covert_checkpoint (Zhiyuan Chen)
- 0c850cb: fix rnamsm output (Zhiyuan Chen)
- c229a7d: add copyright header (Zhiyuan Chen)
- c47578c: fix head config (Zhiyuan Chen)
- 6856ad8: fix model_max_length of RnaFm Tokenizer (Zhiyuan Chen)