Instructions to use recursionpharma/OpenPhenom with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use recursionpharma/OpenPhenom with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("feature-extraction", model="recursionpharma/OpenPhenom", trust_remote_code=True)# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("recursionpharma/OpenPhenom", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
File size: 460 Bytes
6ded986 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 | # © Recursion Pharmaceuticals 2024
loss:
_target_: torch.nn.MSELoss # combine with fourier loss weighted at 0.01 mixing factor for best results
reduction: none
optimizer:
_target_: timm.optim.lion.Lion
_partial_: true
lr: *lr 1e-4 # 1e-4 for <= ViT-B, and 3e-5 for ViT-L
weight_decay: 0.05
betas: [0.9, 0.95]
lr_scheduler:
_target_: torch.optim.lr_scheduler.OneCycleLR
_partial_: true
max_lr: @lr
pct_start: 0.1
anneal_strategy: cos |