⚠ INFORMATION
This page was translated by an AI (LLM) with a cursory human check and is awaiting full review.

AlphaFold 3 on Jean Zay

Licence Restriction

DeepMind has open-sourced the AlphaFold 3 code and made it available on GitHub. However, the model weights are subject to a licence that prevents them from being made available to all IDRIS users.

https://github.com/google-deepmind/alphafold3/blob/main/WEIGHTS_TERMS_OF_USE.md

Indeed, a number of people using IDRIS resources are not part of the academic world.

The procedure for using AlphaFold 3 is:

• IDRIS installs the code so that it is accessible to all users.
• Each person wishing to use AlphaFold 3 must obtain authorisation from DeepMind via the following form: https://forms.gle/svvpY4u2jsHEwWYS6. Please note that the response can take up to 2 weeks.
• The weights must be installed locally on the account of each person wishing to use AlphaFold 3.

Sequence Alignments

The sequence alignment tools are not ported to GPU. As this phase can be quite long, it is recommended not to use a GPU partition to avoid wasting resources. However, you can request an allocation on the CPU partition or use the prepost partition.

Useful Links

• 🌐 Official website: https://www.deepmind.com
• 📘 Documentation: https://github.com/google-deepmind/alphafold3
• 🧬 Git repository: https://github.com/google-deepmind/alphafold3

Submission Script Examples

V100 Partition

#!/usr/bin/env bash
#SBATCH --account=<project_id>@v100

#SBATCH --job-name=AlphaFold3
#SBATCH --output=%x.%j
#SBATCH --error=%x.%j

## For inference
#SBATCH --nodes=1 # Number of nodes
#SBATCH --ntasks-per-node=1 # Number of tasks per node 
#SBATCH --cpus-per-task=10 # Number of cores per Task
#SBATCH --gpus-per-node=1 # Number of GPUs per node
#SBATCH --constraint=v100-32g # We recommend using the 32GB V100 GPUs

# Load the modules
module purge
module load alphafold/3.0.1

# Define the path needed to run
## where are stored the weights of AlphaFold3 (directory containing af3.bin)
weight_dir=${WORK}/AlphaFold3/weights

python $(which run_alphafold.py) \
    --json_path=msa_data.json \
    --model_dir=${weight_dir} \
    --output_dir=<output_dir> \
    --db_dir=${DB_DIR} \ # DBDIR is defined when the module is loaded
    --flash_attention_implementation=xla \ # Needed for V100 GPUs
    --run_data_pipeline=false \
    --run_inference=true

A100 Partition

#!/usr/bin/env bash
#SBATCH --account=<project_id>@a100

#SBATCH --job-name=AlphaFold3
#SBATCH --output=%x.%j
#SBATCH --error=%x.%j

## For inference
#SBATCH --nodes=1 # Number of nodes
#SBATCH --ntasks-per-node=1 # Number of tasks per node 
#SBATCH --cpus-per-task=8 # Number of cores per Task
#SBATCH --gpus-per-node=1 # Number of GPUs per node
#SBATCH --constraint=a100 # To access the A100 partition

# Load the modules
module purge
module load arch/a100
module load alphafold/3.0.1

# Define the path needed to run
## where are stored the weights of AlphaFold3 (directory containing af3.bin)
weight_dir=${WORK}/AlphaFold3/weights

python $(which run_alphafold.py) \
    --json_path=msa_data.json \
    --model_dir=${weight_dir} \
    --output_dir=<output_dir> \
    --db_dir=${DB_DIR} \ # DBDIR is defined when the module is loaded
    --run_data_pipeline=false \
    --run_inference=true

H100 Partition

#!/usr/bin/env bash
#SBATCH --account=<project_id>@h100

#SBATCH --job-name=AlphaFold3
#SBATCH --output=%x.%j
#SBATCH --error=%x.%j

## For inference
#SBATCH --nodes=1 # Number of nodes
#SBATCH --ntasks-per-node=1 # Number of tasks per node 
#SBATCH --cpus-per-task=24 # Number of cores per Task
#SBATCH --gpus-per-node=1 # Number of GPUs per node
#SBATCH --constraint=h100 # To access the H100 partition

# Load the modules
module purge
module load arch/h100
module load alphafold/3.0.1

# Define the path needed to run
## where are stored the weights of AlphaFold3 (directory containing af3.bin)
weight_dir=${WORK}/AlphaFold3/weights

python $(which run_alphafold.py) \
    --json_path=msa_data.json \
    --model_dir=${weight_dir} \
    --output_dir=<output_dir> \
    --db_dir=${DB_DIR} \ # DBDIR is defined when the module is loaded
    --run_data_pipeline=false \
    --run_inference=true