Mol-1 mirror(国内加速镜像 / China-friendly mirror)

Unofficial byte-identical mirror of molence/Mol-1 at commit 7480cc41a7c64b5c517709184bdc81c08dbd5fdb (v1.0.0, Apache-2.0). Weights and reference data come from the official release bucket mol-1-release.oss-cn-shanghai.aliyuncs.com; see CHECKSUMS.sha256.

Layout

Path What it is
checkpoint/mol-1-base.pt base model checkpoint (2,393,615,173 bytes)
common/ CCD components + clustering/release-date reference data (output of scripts/database/download_mol_data.sh)
search_database/pdb_seqres_20261001.fasta template-search sequence database
source/mol-1-7480cc41a7c64b5c517709184bdc81c08dbd5fdb.tar.gz source archive of the pinned commit

Place checkpoint/ and common/ under your MOL_ROOT_DIR exactly as the upstream README describes, then run mol pred (or python -m mol.runner.inference).

Download in mainland China (hf-mirror.com)

pip install -U "huggingface_hub[cli]"
export HF_ENDPOINT=https://hf-mirror.com

MOL_ROOT_DIR=~/mol1-data   # any writable root
hf download RUIMA294264/Mol-1 \
  checkpoint/mol-1-base.pt \
  common/components.cif \
  common/components.cif.rdkit_mol.pkl \
  common/clusters-by-entity-40.txt \
  common/seq_to_pdb_index.json \
  common/release_date_cache.json \
  common/obsolete_to_successor.json \
  common/obsolete_release_date.csv \
  --repo-type model --local-dir "$MOL_ROOT_DIR"

huggingface.co works directly wherever it is reachable; hf-mirror.com serves the same LFS objects anonymously. Verify with sha256sum -c CHECKSUMS.sha256.

For the code, either use source/mol-1-*.tar.gz above or clone upstream GitHub.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support