File size: 3,524 Bytes
373e2bb | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 | # Docker Installation
Use Docker for GPU inference on a Linux host with an NVIDIA GPU. All examples
below are one-shot `docker run` commands executed from the host. For non-Docker
installation, see [inference_instructions.md](./inference_instructions.md).
## 1. Verify Docker GPU support
Install Docker and the
[NVIDIA Container Toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html),
then verify that containers can see the GPU:
```bash
docker run --rm --gpus all nvidia/cuda:12.6.3-base-ubuntu24.04 nvidia-smi
```
## 2. Get the image
Pull the prebuilt image:
```bash
docker pull aurekaresearch/opendde:v1
```
Or build from the repository root:
```bash
docker build -t aurekaresearch/opendde:v1 .
```
## 3. Prepare runtime data
OpenDDE reads checkpoints and runtime data from `OPENDDE_ROOT_DIR`. If you have
local checkpoints, place `opendde.pt` and/or `opendde_abag.pt` under
`checkpoint/` in that directory:
```bash
export OPENDDE_ROOT_DIR="$PWD/opendde_data"
mkdir -p "$OPENDDE_ROOT_DIR/checkpoint"
cp /absolute/path/to/opendde.pt "$OPENDDE_ROOT_DIR/checkpoint/opendde.pt"
```
Released checkpoints:
| Checkpoint | Use case | Download |
| --- | --- | --- |
| `opendde.pt` | General-purpose OpenDDE checkpoint. | [opendde.pt](https://huggingface.co/aurekaresearch/OpenDDE/resolve/main/opendde.pt) |
| `opendde_abag.pt` | ABAG-optimized checkpoint for antibody-antigen complexes. | [opendde_abag.pt](https://huggingface.co/aurekaresearch/OpenDDE/resolve/main/opendde_abag.pt) |
For the default Docker command below, place the general-purpose checkpoint at
`$OPENDDE_ROOT_DIR/checkpoint/opendde.pt`. When using the ABAG-optimized
checkpoint, add this to the `opendde pred` command:
```bash
--load_checkpoint_path /opendde_data/checkpoint/opendde_abag.pt
```
Download one checkpoint directly into the default host path:
```bash
# General-purpose checkpoint:
curl -L \
-o "$OPENDDE_ROOT_DIR/checkpoint/opendde.pt" \
https://huggingface.co/aurekaresearch/OpenDDE/resolve/main/opendde.pt
# ABAG-optimized checkpoint:
curl -L \
-o "$OPENDDE_ROOT_DIR/checkpoint/opendde_abag.pt" \
https://huggingface.co/aurekaresearch/OpenDDE/resolve/main/opendde_abag.pt
```
Download or verify the remaining runtime files with Docker:
```bash
docker run --rm \
-v "$OPENDDE_ROOT_DIR":/opendde_data \
aurekaresearch/opendde:v1 \
bash scripts/download_opendde_data.sh \
--root /opendde_data
```
For protein-only smoke tests that disable MSA/template/RNA-MSA preprocessing, you
can skip search databases:
```bash
docker run --rm \
-v "$OPENDDE_ROOT_DIR":/opendde_data \
aurekaresearch/opendde:v1 \
bash scripts/download_opendde_data.sh \
--root /opendde_data \
--skip-search-database
```
## 4. Run inference
The command below assumes `tiny.json` exists in the current host directory. See
[../README.md](../README.md) for the minimal input example.
```bash
mkdir -p output
docker run --rm --gpus all --shm-size=4g \
-e OPENDDE_ROOT_DIR=/opendde_data \
-v "$OPENDDE_ROOT_DIR":/opendde_data:ro \
-v "$PWD":/workspace \
-v "$PWD/output":/output \
aurekaresearch/opendde:v1 \
opendde pred \
-i /workspace/tiny.json \
-o /output \
-n opendde_v1 \
--use_msa false \
--use_template false \
--use_rna_msa false \
--sample 1 \
--step 200 \
--cycle 10
```
For production inference options, MSA/template preprocessing, and checkpoint
configuration, see [inference_instructions.md](./inference_instructions.md).
|