Instructions to use multimolecule/deepsea with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MultiMolecule
How to use multimolecule/deepsea with MultiMolecule:
pip install multimolecule
from multimolecule import AutoModel, AutoTokenizer tokenizer = AutoTokenizer.from_pretrained("multimolecule/deepsea") model = AutoModel.from_pretrained("multimolecule/deepsea") inputs = tokenizer("ACTCCCCTGCCCTCAACAAGATGTTTTGCCAACTGGCCAAGACCTGCCCTGTGCAGCTGTGGGTTGATTCCACACCCCCGCCCGGCACCCGCGTCCGCGCCATGGCCATCTACAAGCAGTCACAGCACATGACGGAGGTTGTGAGGCGCTGCCCCCACCATGAGCGCTGCTCAGATAGCGATGG", return_tensors="pt") outputs = model(**inputs) embeddings = outputs.last_hidden_state - Notebooks
- Google Colab
- Kaggle
Download pytorch_model.bin from multimolecule/deepsea: direct link, hf CLI and curl.
- Browser
- Download file 211 MB
-
https://huggingface.co/multimolecule/deepsea/resolve/main/pytorch_model.bin
- Command line
-
hf download hf://multimolecule/deepsea/pytorch_model.bin
-
curl -L -o pytorch_model.bin https://huggingface.co/multimolecule/deepsea/resolve/main/pytorch_model.bin
211 MB
- Xet hash:
- 0e3c52a5183b275fabebef4b05163a258db6d0506bf7969c9983411ed07dc8d4
- Size of remote file:
- 211 MB
- SHA256:
- 7ccb5102d11e114425f678602b08d9d86973bc1bc9c1c62b7a54e3f4f4b4eb82
·
Xet efficiently stores Large Files inside Git, intelligently splitting files into unique chunks and accelerating uploads and downloads. More info.