[DRAFT] DO NOT MERGE Esm2 minifold

.devcontainer/devcontainer.json

@@ -12,7 +12,9 @@
         "source=${localEnv:HOME}/.cursor,target=/home/ubuntu/.cursor,type=bind,consistency=cached",
         "source=${localEnv:HOME}/.gnupg,target=/home/ubuntu/.gnupg,type=bind,consistency=cached",
         "source=${localEnv:HOME}/.netrc,target=/home/ubuntu/.netrc,type=bind,consistency=cached",
-        "source=${localEnv:HOME}/.ssh,target=/home/ubuntu/.ssh,readonly,type=bind,consistency=cached"
+        "source=${localEnv:HOME}/.ssh,target=/home/ubuntu/.ssh,readonly,type=bind,consistency=cached",
+        "source=${localEnv:HOME}/Projects/minifold,target=/workspaces/minifold,type=bind,consistency=cached"
+
     ],
     "postCreateCommand": ".devcontainer/postCreateCommand.sh",
     "initializeCommand": ".devcontainer/initializeCommand.sh",

.devcontainer/requirements.txt

@@ -4,6 +4,7 @@ deepspeed
 hydra-core
 lm-eval
 megatron-fsdp
+ml_collections
 peft
 pytest
 torch

bionemo-recipes/recipes/esm2_minifold_te/.dockerignore

@@ -0,0 +1,38 @@
+# Docker
+Dockerfile
+Dockerfile.*
+.dockerignore
+
+# Docs
+README.md
+
+# Python caches
+__pycache__
+.pytest_cache
+.ruff_cache
+.venv/
+
+# Linting
+.ruff.toml
+
+# Profiling & debugging artifacts
+memory_snapshots/
+nsight_profiling/
+*.nsys-rep
+*.sqlite
+logs/
+wandb/
+
+# Downloaded CIF files (reproduced by data/prepare_*.py scripts)
+data/cif_files/
+data/eval_cif_files/
+
+# Hydra / training outputs
+outputs/
+checkpoints/
+
+# Checkpoint export
+checkpoint_export/
+
+# Temp / scratch
+j/

bionemo-recipes/recipes/esm2_minifold_te/.ruff.toml

@@ -0,0 +1,26 @@
+extend = "../../../.ruff.toml"
+
+[lint]
+# Extend ignores for code ported from MiniFold
+ignore = [
+    "C901",    # Complexity
+    "D100",    # Missing module docstring
+    "D101",    # Missing class docstring
+    "D102",    # Missing method docstring
+    "D103",    # Missing function docstring
+    "D105",    # Missing magic method docstring
+    "D107",    # Missing __init__ docstring
+    "D205",    # 1 blank line required between summary and description
+    "D415",    # First line should end with period
+    "D417",    # Missing argument descriptions
+    "E501",    # Line too long
+    "E721",    # Type comparison
+    "E731",    # Lambda assignment
+    "E741",    # Ambiguous variable name
+    "F841",    # Unused variable
+    "N806",    # Uppercase variable in function
+    "N812",    # Lowercase imported as non-lowercase
+    "PLW2901", # Loop variable overwritten
+    "RUF005",  # Collection literal concatenation
+    "RUF010",  # Explicit f-string conversion
+]

bionemo-recipes/recipes/esm2_minifold_te/Dockerfile

@@ -0,0 +1,9 @@
+# syntax=docker/dockerfile:1.4
+FROM nvcr.io/nvidia/pytorch:26.02-py3
+
+RUN --mount=type=cache,target=/root/.cache/pip \
+    --mount=type=bind,source=requirements.txt,target=/requirements.txt \
+    PIP_CONSTRAINT= pip install -r /requirements.txt
+
+WORKDIR /workspace/bionemo
+COPY . .