fix: paths and ipd readme (#656)

fix: update paths; add IPD readme

README.md

@@ -56,6 +56,9 @@ rf3 fold models/rf3/tests/data/5vht_from_json.json
 
 Details on the exact formatting of the json files are available [here](models/rf3/README.md).
 
+> [!NOTE]
+> **IPD Users**: If you're running within the IPD environment, please see [IPD_USAGE.md](docs/IPD_USAGE.md) for instructions on using the shebang system, apptainer containers, and SLURM job submission for training, inference, and validation.
+
 ## Development
 
 ### Package Structure

docs/IPD_USAGE.md

@@ -0,0 +1,224 @@
+# Running ModelForge at IPD
+
+This guide covers how to run training, inference, and validation within the IPD environment using our shebang and apptainer infrastructure.
+
+## Table of Contents
+
+- [Overview](#overview)
+- [The Shebang System](#the-shebang-system)
+- [Running Inference](#running-inference)
+- [Running Training](#running-training)
+- [Running Validation](#running-validation)
+- [Container Information](#container-information)
+- [Debugging](#debugging)
+
+## Overview
+
+The IPD environment provides two complementary approaches for running models:
+
+1. **Shebang System (Development)**: Executes Python scripts inside containers while allowing you to edit code on the host. Changes to your code are immediately reflected without rebuilding containers.
+
+2. **Pre-built Containers (Production)**: Fully containerized execution using `apptainer exec` with code baked into the container.
+
+For active development, the shebang approach is recommended as it provides the fastest iteration cycle.
+
+## The Shebang System
+
+### How It Works
+
+The shebang system allows you to execute Python scripts directly (e.g., `./models/rf3/src/rf3/train.py`) without explicitly calling `apptainer exec`. Here's what happens behind the scenes:
+
+1. **Entry Point Scripts** (train.py, validate.py, inference.py) include a special shebang line:
+   ```bash
+   #!/usr/bin/env -S /bin/sh -c '"$(dirname "$0")/../../../../.ipd/shebang/rf3_exec.sh" "$0" "$@"'
+   ```
+
+2. **The rf3_exec.sh Script** (`.ipd/shebang/rf3_exec.sh`) then:
+   - Locates the repository root by searching for `.project-root`
+   - Sets up PYTHONPATH to include modelhub, rf3, and atomworks
+   - Finds the development container at `.ipd/apptainer/rf3-dev.sif`
+   - Detects GPU support (uses `--nvccli` if available, falls back to `--nv`, or runs without GPU)
+   - Executes your script inside the container with the repository bind-mounted
+
+3. **Your Code Runs** inside the container but reads from the host filesystem, so any edits you make are immediately active.
+
+### Usage
+
+Simply make the script executable and run it directly:
+
+```bash
+# Run inference
+./models/rf3/src/rf3/inference.py inputs=example.json
+
+# Run training
+./models/rf3/src/rf3/train.py experiment=my_experiment
+
+# Run validation
+./models/rf3/src/rf3/validate.py experiment=my_experiment
+```
+
+You can also use the `rf3` CLI command if the package is installed in your environment, but the shebang approach ensures you're always using the container environment.
+
+## Running Inference
+
+### Method 1: Development (Shebang - Recommended)
+
+Execute the inference script directly:
+
+```bash
+./models/rf3/src/rf3/inference.py \
+  inputs=./path/to/input.json \
+  out_dir=./output \
+  ckpt_path=/net/software/containers/versions/modelhub_inference/ckpts/rf3-w-conf-run10-ep922-remapped.ckpt \
+  diffusion_batch_size=5 \
+  n_recycles=10 \
+  num_steps=200
+```
+
+> Inference arguments for RF3 are provided [here](models/rf3/README.md).
+
+### Method 2: Production (Pre-built Container)
+
+For production inference or when you don't need to modify code:
+
+```bash
+apptainer exec --nvccli \
+  /net/software/containers/versions/modelhub_inference/rf3.sif \
+  rf3 fold \
+  inputs=./path/to/input.json \
+  out_dir=./output \
+  ckpt_path=/net/software/containers/versions/modelhub_inference/ckpts/rf3-w-conf-run10-ep922-remapped.ckpt \
+  diffusion_batch_size=1 \
+  one_model_per_file=True \
+  annotate_b_factor_with_plddt=True
+```
+
+This approach is ideal for:
+- Large-scale inference jobs (e.g., distillation, design campaigns)
+- When you want to ensure exact reproducibility
+
+## Running Training
+
+### Method 1: Direct Execution (Development)
+
+For quick training runs or debugging:
+
+```bash
+./models/rf3/src/rf3/train.py \
+  experiment=quick-rf3 \
+  debug=default
+```
+
+The experiment name determines which config is loaded from `models/rf3/configs/experiment/`.
+
+### Method 2: SLURM Job Submission (Recommended for Production Training)
+
+For multi-GPU or multi-node training, use SLURM:
+
+```bash
+# Submit a training job
+sbatch .ipd/slurm/launch_rf3.sh
+```
+
+#### Customizing SLURM Jobs
+
+Edit `.ipd/slurm/launch_rf3.sh` or create a custom launch script:
+
+```bash
+#!/bin/bash
+#SBATCH -p gpu-train                    # Partition
+#SBATCH --nodes 1                       # Number of nodes
+#SBATCH --gres=gpu:l40:8               # GPUs per node
+#SBATCH --ntasks-per-node 8            # Tasks per node (usually = GPUs)
+#SBATCH -c 4                           # CPUs per task
+#SBATCH --mem=512g                     # Memory per node
+#SBATCH -t 1-00:00:00                  # Time limit (1 day)
+#SBATCH -J my_experiment               # Job name
+
+# Environment setup
+export OMP_NUM_THREADS=$SLURM_CPUS_PER_TASK
+export NCCL_DEBUG=INFO
+export NCCL_P2P_DISABLE=1              # For L40 GPUs (no NVLink)
+export PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True
+
+# Calculate gradient accumulation
+EFFECTIVE_BATCH_SIZE=32                 # Total effective batch size
+BATCH_SIZE_PER_GPU=1                    # Batch size per GPU
+DEVICES_PER_NODE=$SLURM_GPUS_ON_NODE
+TOTAL_GPUS=$((SLURM_NNODES * DEVICES_PER_NODE))
+GRAD_ACCUM_STEPS=$((EFFECTIVE_BATCH_SIZE / (BATCH_SIZE_PER_GPU * TOTAL_GPUS)))
+
+# Launch training
+srun --kill-on-bad-exit \
+  ../../models/rf3/src/rf3/train.py \
+  experiment=$SLURM_JOB_NAME \
+  ++trainer.devices_per_node=$DEVICES_PER_NODE \
+  ++trainer.num_nodes=$SLURM_NNODES \
+  ++trainer.grad_accum_steps=$GRAD_ACCUM_STEPS
+```
+
+#### Key SLURM Configuration Options
+
+**Resource Allocation:**
+- `-p gpu-train`: GPU training partition
+- `--nodes`: Number of nodes (1 for single-node, >1 for distributed)
+- `--gres=gpu:l40:8`: GPU type and count 
+- `--ntasks-per-node`: Number of processes per node (typically equals GPU count)
+- `-c`: CPU cores per task
+- `--mem`: Memory per node
+
+**Time and Naming:**
+- `-t`: Time limit (format: days-hours:minutes:seconds)
+- `-J`: Job name (used as experiment name if you use `$SLURM_JOB_NAME`)
+
+**Environment Variables:**
+- `OMP_NUM_THREADS`: OpenMP threads (usually set to CPUs per task)
+- `NCCL_P2P_DISABLE=1`: Disable peer-to-peer for GPUs without NVLink (which we don't have at the IPD)
+- `PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True`: Enable expandable CUDA memory
+
+
+## Running Validation
+
+Validate a trained model or checkpoint:
+
+```bash
+# Using shebang (development)
+./models/rf3/src/rf3/validate.py \
+  experiment=my_experiment \
+  ckpt_path=/path/to/checkpoint.ckpt
+```
+
+Validation datasets are configured in `models/rf3/configs/datasets/val/`.
+
+## Debugging
+
+### Remote Debugging with debugpy
+
+The shebang system supports remote debugging:
+
+```bash
+# Set DEBUG_PORT environment variable
+DEBUG_PORT=5678 ./models/rf3/src/rf3/train.py experiment=quick-rf3
+```
+
+This will:
+1. Launch debugpy server on the specified port
+2. Wait for a debugger to attach
+3. Print connection instructions
+
+Then attach your IDE debugger to `localhost:5678`.
+
+### Viewing Logs
+
+Training logs are saved to the experiment directory. For SLURM jobs:
+
+```bash
+# View real-time logs
+tail -f slurm-<job_id>.out
+
+# Check job status
+squeue -u $USER
+
+# View completed job info
+sacct -j <job_id> --format=JobID,JobName,Partition,State,ExitCode,Elapsed
+```

models/rf3/configs/datasets/train/disorder_distillation.yaml

@@ -2,7 +2,7 @@
 
 disorder_distillation:
   dataset:
-    _target_: atomworks.ml.datasets.datasets.StructuralDatasetWrapper
+    _target_: atomworks.ml.datasets.StructuralDatasetWrapper
     save_failed_examples_to_dir: null
 
     # cif parser arguments
@@ -18,7 +18,7 @@ disorder_distillation:
 
     # metadata dataset
     dataset:
-      _target_: atomworks.ml.datasets.datasets.PandasDataset
+      _target_: atomworks.ml.datasets.PandasDataset
       name: pdb_diso_distillation
       id_column: example_id
       data: ${paths.data.disorder_distill_parquet_dir}/disorderDistillation.csv

models/rf3/configs/datasets/train/domain_distillation.yaml

@@ -18,7 +18,7 @@ multidomain_distillation:
 
     # metadata dataset
     dataset:
-      _target_: atomworks.ml.datasets.datasets.PandasDataset
+      _target_: atomworks.ml.datasets.PandasDataset
       name: multidomain_distillation
       id_column: example_id
       data: /projects/ml/datahub/dfs/domain_domain/domain_domain_dataset.DIGS.parquet

models/rf3/configs/datasets/train/monomer_distillation.yaml

@@ -2,7 +2,7 @@
 
 monomer_distillation:
   dataset:
-    _target_: atomworks.ml.datasets.datasets.StructuralDatasetWrapper
+    _target_: atomworks.ml.datasets.StructuralDatasetWrapper
     save_failed_examples_to_dir: ${paths.data.failed_examples_dir}
 
     # cif parser arguments
@@ -18,7 +18,7 @@ monomer_distillation:
 
     # metadata dataset
     dataset:
-      _target_: atomworks.ml.datasets.datasets.PandasDataset
+      _target_: atomworks.ml.datasets.PandasDataset
       name: af2fb_distillation
       id_column: example_id
       data: ${paths.data.monomer_distillation_parquet_dir}/af2_distillation_facebook.parquet

models/rf3/configs/datasets/train/na_complex_distillation.yaml

@@ -2,7 +2,7 @@
 
 na_complex_distillation:
   dataset:
-    _target_: atomworks.ml.datasets.datasets.StructuralDatasetWrapper
+    _target_: atomworks.ml.datasets.StructuralDatasetWrapper
     save_failed_examples_to_dir: null
 
     # cif parser
@@ -19,7 +19,7 @@ na_complex_distillation:
 
     # metadata dataset
     dataset:
-      _target_: atomworks.ml.datasets.datasets.PandasDataset
+      _target_: atomworks.ml.datasets.PandasDataset
       name: tf_distillation
       id_column: example_id
       data: ${paths.data.na_complex_distillation_parquet_dir}/transcriptionFactor_distillation_rf3.newDL.csv

models/rf3/configs/datasets/train/pdb/base.yaml

@@ -1,12 +1,12 @@
 dataset:
-  _target_: atomworks.ml.datasets.datasets.StructuralDatasetWrapper
+  _target_: atomworks.ml.datasets.StructuralDatasetWrapper
   save_failed_examples_to_dir: ${paths.data.failed_examples_dir}
   cif_parser_args:
     cache_dir: null
     load_from_cache: false
     save_to_cache: false
   dataset:
-    _target_: atomworks.ml.datasets.datasets.PandasDataset
+    _target_: atomworks.ml.datasets.PandasDataset
     # we will use the example_id as the unique column
     id_column: example_id
   transform:

models/rf3/configs/datasets/train/rna_monomer_distillation.yaml

@@ -2,7 +2,7 @@
 
 rna_monomer_distillation:
   dataset:
-    _target_: atomworks.ml.datasets.datasets.StructuralDatasetWrapper
+    _target_: atomworks.ml.datasets.StructuralDatasetWrapper
     save_failed_examples_to_dir: ${paths.data.failed_examples_dir}
 
     # cif parser arguments
@@ -18,7 +18,7 @@ rna_monomer_distillation:
 
     # metadata dataset
     dataset:
-      _target_: atomworks.ml.datasets.datasets.PandasDataset
+      _target_: atomworks.ml.datasets.PandasDataset
       name: rna_monomer_distillation
       id_column: example_id
       data: /projects/ml/afavor/rna_distillation/rna_distillation_filtered_df.parquet

models/rf3/configs/datasets/val/af3_ab_set.yaml

@@ -6,6 +6,6 @@ dataset:
     _target_: atomworks.ml.datasets.parsers.ValidationDFParserLikeAF3
     base_dir: /projects/ml/frozen_pdb_copies/2025_07_13_pdb
   dataset:
-    _target_: atomworks.ml.datasets.datasets.PandasDataset
+    _target_: atomworks.ml.datasets.PandasDataset
     name: af3_validation
     data: /net/scratch/rib7/rf3_ab_splits/entry_level_val_df.parquet

models/rf3/configs/datasets/val/af3_validation.yaml

@@ -6,6 +6,6 @@ dataset:
     _target_: atomworks.ml.datasets.parsers.ValidationDFParserLikeAF3
     base_dir: /projects/ml/frozen_pdb_copies/2025_07_13_pdb
   dataset:
-    _target_: atomworks.ml.datasets.datasets.PandasDataset
+    _target_: atomworks.ml.datasets.PandasDataset
     name: af3_validation
     data: ${paths.data.pdb_data_dir}/entry_level_val_df.parquet

models/rf3/configs/datasets/val/base.yaml

@@ -1,12 +1,12 @@
 dataset:
-  _target_: atomworks.ml.datasets.datasets.StructuralDatasetWrapper
+  _target_: atomworks.ml.datasets.StructuralDatasetWrapper
   save_failed_examples_to_dir: ${paths.data.failed_examples_dir}
   cif_parser_args:
     cache_dir: null
     load_from_cache: False
     save_to_cache: False
   dataset:
-    _target_: atomworks.ml.datasets.datasets.PandasDataset
+    _target_: atomworks.ml.datasets.PandasDataset
     # we will use the example_id as the unique column
     id_column: example_id
   # return all keys (do not subset)

models/rf3/configs/datasets/val/runs_and_poses.yaml

@@ -5,7 +5,7 @@ dataset:
   dataset_parser:
     _target_: atomworks.ml.datasets.parsers.ValidationDFParserLikeAF3
   dataset:
-    _target_: atomworks.ml.datasets.datasets.PandasDataset
+    _target_: atomworks.ml.datasets.PandasDataset
     name: af3_validation
     data: /projects/ml/datahub/dfs/af3_splits/2024_12_16/runs_n_poses_entry_level_df.parquet
     filters:

models/rf3/src/rf3/data/paired_msa.py

@@ -7,8 +7,7 @@ from typing import Any
 import numpy as np
 from atomworks.common import exists
 from atomworks.enums import ChainType
-from atomworks.ml.datasets import logger
-from atomworks.ml.datasets.datasets import StructuralDatasetWrapper
+from atomworks.ml.datasets import logger, StructuralDatasetWrapper
 from atomworks.ml.datasets.parsers import (
     MetadataRowParser,
     load_example_from_metadata_row,

models/rf3/tests/data/inference_regression_tests/5vht_from_file/5vht_from_file_metrics.csv

@@ -1,2 +1,2 @@
 example_id,ptm.ptm_0,ptm.ptm_1,ptm.ptm_2,ptm.ptm_3,ptm.ptm_4,iptm.iptm_0,iptm.iptm_1,iptm.iptm_2,iptm.iptm_3,iptm.iptm_4,iptm.iptm_protein_protein_0,iptm.iptm_protein_protein_1,iptm.iptm_protein_protein_2,iptm.iptm_protein_protein_3,iptm.iptm_protein_protein_4,iptm.iptm_protein_ligand_0,iptm.iptm_protein_ligand_1,iptm.iptm_protein_ligand_2,iptm.iptm_protein_ligand_3,iptm.iptm_protein_ligand_4,iptm.iptm_ligand_ligand_0,iptm.iptm_ligand_ligand_1,iptm.iptm_ligand_ligand_2,iptm.iptm_ligand_ligand_3,iptm.iptm_ligand_ligand_4,count_clashing_chains.has_clash_0,count_clashing_chains.has_clash_1,count_clashing_chains.has_clash_2,count_clashing_chains.has_clash_3,count_clashing_chains.has_clash_4
-5vht_from_file,0.8497557,0.8497257,0.8498095,0.8496232,0.84974766,0.85205853,0.85195255,0.85201293,0.8519251,0.8520231,0.85205853,0.85195255,0.85201293,0.8519251,0.8520231,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0,0,0,0,0
+5vht_from_file,0.9095256,0.90964013,0.90937024,0.9094581,0.9096535,0.909933,0.9100767,0.9099717,0.91014886,0.91009957,0.909933,0.9100767,0.9099717,0.91014886,0.91009957,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0,0,0,0,0

models/rf3/tests/data/inference_regression_tests/5vht_from_file/5vht_from_file_score.csv

@@ -1,16 +1,16 @@
 example_id,chain_chainwise,chainwise_plddt,chainwise_pde,chainwise_pae,overall_plddt,overall_pde,overall_pae,batch_idx,chain_i_interface,chain_j_interface,pae_interface,pde_interface,min_pae_interface,min_pde_interface
-5vht_from_file,A_1,0.8646008968353271,0.9278936982154846,4.9273810386657715,0.8644909858703613,0.9399122595787048,5.076437950134277,0,,,,,,
-5vht_from_file,B_1,0.8643810749053955,0.9313923716545105,4.932040691375732,0.8644909858703613,0.9399122595787048,5.076437950134277,0,,,,,,
-5vht_from_file,A_1,0.864599883556366,0.9278801679611206,4.928858757019043,0.864488422870636,0.9398555159568787,5.077895641326904,1,,,,,,
-5vht_from_file,B_1,0.8643767833709717,0.9312825202941895,4.932775020599365,0.864488422870636,0.9398555159568787,5.077895641326904,1,,,,,,
-5vht_from_file,A_1,0.8645994663238525,0.92825847864151,4.927615165710449,0.8644875884056091,0.9400468468666077,5.076131820678711,2,,,,,,
-5vht_from_file,B_1,0.8643757104873657,0.9314727187156677,4.931051731109619,0.8644875884056091,0.9400468468666077,5.076131820678711,2,,,,,,
-5vht_from_file,A_1,0.8645936250686646,0.9283772110939026,4.932724952697754,0.8644911646842957,0.9403422474861145,5.080663681030273,3,,,,,,
-5vht_from_file,B_1,0.864388644695282,0.9315438866615295,4.935116291046143,0.8644911646842957,0.9403422474861145,5.080663681030273,3,,,,,,
-5vht_from_file,A_1,0.8645904064178467,0.9278889894485474,4.9294209480285645,0.864478349685669,0.9397901296615601,5.077117919921875,4,,,,,,
-5vht_from_file,B_1,0.864366352558136,0.9307928085327148,4.931059837341309,0.864478349685669,0.9397901296615601,5.077117919921875,4,,,,,,
-5vht_from_file,,,,,0.8644909858703613,0.9399122595787048,5.076437950134277,0,A_1,B_1,5.223164081573486,0.9501814842224121,1.6886605024337769,0.3923574686050415
-5vht_from_file,,,,,0.864488422870636,0.9398555159568787,5.077895641326904,1,A_1,B_1,5.224973201751709,0.9501296281814575,1.6891288757324219,0.39237290620803833
-5vht_from_file,,,,,0.8644875884056091,0.9400468468666077,5.076131820678711,2,A_1,B_1,5.22292947769165,0.9502280354499817,1.6894645690917969,0.3924662470817566
-5vht_from_file,,,,,0.8644911646842957,0.9403422474861145,5.080663681030273,3,A_1,B_1,5.227405548095703,0.9507239460945129,1.688165545463562,0.3923729658126831
-5vht_from_file,,,,,0.864478349685669,0.9397901296615601,5.077117919921875,4,A_1,B_1,5.223995208740234,0.9502394795417786,1.6886507272720337,0.39236772060394287
+5vht_from_file,A_1,0.8418933153152466,0.7738104462623596,4.3772172927856445,0.8415207266807556,0.8318539261817932,4.677595138549805,0,,,,,,
+5vht_from_file,B_1,0.8411479592323303,0.7790139317512512,4.396655082702637,0.8415207266807556,0.8318539261817932,4.677595138549805,0,,,,,,
+5vht_from_file,A_1,0.8418263792991638,0.7724140882492065,4.372034072875977,0.8414488434791565,0.8303730487823486,4.672811985015869,1,,,,,,
+5vht_from_file,B_1,0.8410713076591492,0.7779868841171265,4.392614364624023,0.8414488434791565,0.8303730487823486,4.672811985015869,1,,,,,,
+5vht_from_file,A_1,0.8417466878890991,0.7801038026809692,4.385378837585449,0.8413788676261902,0.8371773958206177,4.68752908706665,2,,,,,,
+5vht_from_file,B_1,0.8410109877586365,0.7858399748802185,4.405368804931641,0.8413788676261902,0.8371773958206177,4.68752908706665,2,,,,,,
+5vht_from_file,A_1,0.8417938947677612,0.77611243724823,4.379835605621338,0.8414207100868225,0.8332493901252747,4.680868625640869,3,,,,,,
+5vht_from_file,B_1,0.841047465801239,0.781791627407074,4.400188446044922,0.8414207100868225,0.8332493901252747,4.680868625640869,3,,,,,,
+5vht_from_file,A_1,0.841849148273468,0.7727293372154236,4.370999813079834,0.8414701819419861,0.8308433294296265,4.672425746917725,4,,,,,,
+5vht_from_file,B_1,0.8410910964012146,0.7783981561660767,4.391624450683594,0.8414701819419861,0.8308433294296265,4.672425746917725,4,,,,,,
+5vht_from_file,,,,,0.8415207266807556,0.8318539261817932,4.677595138549805,0,A_1,B_1,4.968254089355469,0.8872956037521362,0.7079706192016602,0.28655317425727844
+5vht_from_file,,,,,0.8414488434791565,0.8303730487823486,4.672811985015869,1,A_1,B_1,4.963299751281738,0.8855456709861755,0.6997964382171631,0.2865622043609619
+5vht_from_file,,,,,0.8413788676261902,0.8371773958206177,4.68752908706665,2,A_1,B_1,4.979684352874756,0.8913829326629639,0.7079532742500305,0.2865457534790039
+5vht_from_file,,,,,0.8414207100868225,0.8332493901252747,4.680868625640869,3,A_1,B_1,4.971724987030029,0.8875466585159302,0.6996998190879822,0.2865447700023651
+5vht_from_file,,,,,0.8414701819419861,0.8308433294296265,4.672425746917725,4,A_1,B_1,4.9635396003723145,0.8861227631568909,0.7079214453697205,0.286527156829834

models/rf3/tests/data/inference_regression_tests/8vkf_from_file/8vkf_from_file.score

@@ -1,76 +0,0 @@
-example_id,chain_chainwise,chainwise_plddt,chainwise_pde,chainwise_pae,overall_plddt,overall_pde,overall_pae,batch_idx,chain_i_interface,chain_j_interface,pae_interface,pde_interface,min_pae_interface,min_pde_interface
-8vkf_from_file,A_1,0.903170108795166,0.7459312081336975,4.248835563659668,0.9036211371421814,0.796832799911499,4.3507561683654785,0,,,,,,
-8vkf_from_file,B_1,0.9326015710830688,1.3645237684249878,5.259715557098389,0.9036211371421814,0.796832799911499,4.3507561683654785,0,,,,,,
-8vkf_from_file,C_1,0.9183257818222046,1.8940019607543945,6.616949081420898,0.9036211371421814,0.796832799911499,4.3507561683654785,0,,,,,,
-8vkf_from_file,D_1,0.9201919436454773,15.452116966247559,13.454815864562988,0.9036211371421814,0.796832799911499,4.3507561683654785,0,,,,,,
-8vkf_from_file,E_1,0.8646432757377625,17.121997833251953,13.2661714553833,0.9036211371421814,0.796832799911499,4.3507561683654785,0,,,,,,
-8vkf_from_file,A_1,0.9031737446784973,0.7429665327072144,4.2377729415893555,0.9036248922348022,0.793869137763977,4.341085910797119,1,,,,,,
-8vkf_from_file,B_1,0.9326193332672119,1.3646689653396606,5.2598371505737305,0.9036248922348022,0.793869137763977,4.341085910797119,1,,,,,,
-8vkf_from_file,C_1,0.9183292984962463,1.8951787948608398,6.617890357971191,0.9036248922348022,0.793869137763977,4.341085910797119,1,,,,,,
-8vkf_from_file,D_1,0.9202033877372742,15.478745460510254,13.45727825164795,0.9036248922348022,0.793869137763977,4.341085910797119,1,,,,,,
-8vkf_from_file,E_1,0.8645620346069336,17.12875747680664,13.272069931030273,0.9036248922348022,0.793869137763977,4.341085910797119,1,,,,,,
-8vkf_from_file,A_1,0.9031793475151062,0.7455524206161499,4.247676849365234,0.9036300778388977,0.7963874936103821,4.349856376647949,2,,,,,,
-8vkf_from_file,B_1,0.9325892329216003,1.3646607398986816,5.259612560272217,0.9036300778388977,0.7963874936103821,4.349856376647949,2,,,,,,
-8vkf_from_file,C_1,0.918330192565918,1.8964067697525024,6.61790132522583,0.9036300778388977,0.7963874936103821,4.349856376647949,2,,,,,,
-8vkf_from_file,D_1,0.9201703667640686,15.475332260131836,13.452682495117188,0.9036300778388977,0.7963874936103821,4.349856376647949,2,,,,,,
-8vkf_from_file,E_1,0.8645628690719604,17.12711524963379,13.264180183410645,0.9036300778388977,0.7963874936103821,4.349856376647949,2,,,,,,
-8vkf_from_file,A_1,0.903174638748169,0.7434669733047485,4.240329742431641,0.9036257863044739,0.7943676114082336,4.343367099761963,3,,,,,,
-8vkf_from_file,B_1,0.9326135516166687,1.3644018173217773,5.259875297546387,0.9036257863044739,0.7943676114082336,4.343367099761963,3,,,,,,
-8vkf_from_file,C_1,0.9183347225189209,1.8946807384490967,6.61772346496582,0.9036257863044739,0.7943676114082336,4.343367099761963,3,,,,,,
-8vkf_from_file,D_1,0.9202176332473755,15.453478813171387,13.452422142028809,0.9036257863044739,0.7943676114082336,4.343367099761963,3,,,,,,
-8vkf_from_file,E_1,0.8648159503936768,17.128904342651367,13.272899627685547,0.9036257863044739,0.7943676114082336,4.343367099761963,3,,,,,,
-8vkf_from_file,A_1,0.903176486492157,0.7440198659896851,4.24072265625,0.903627336025238,0.7949872612953186,4.343382358551025,4,,,,,,
-8vkf_from_file,B_1,0.9325862526893616,1.3641865253448486,5.259721279144287,0.903627336025238,0.7949872612953186,4.343382358551025,4,,,,,,
-8vkf_from_file,C_1,0.9183405637741089,1.893341302871704,6.615909099578857,0.903627336025238,0.7949872612953186,4.343382358551025,4,,,,,,
-8vkf_from_file,D_1,0.9201603531837463,15.452801704406738,13.456074714660645,0.903627336025238,0.7949872612953186,4.343382358551025,4,,,,,,
-8vkf_from_file,E_1,0.8648535013198853,17.13027000427246,13.264931678771973,0.903627336025238,0.7949872612953186,4.343382358551025,4,,,,,,
-8vkf_from_file,,,,,0.9036211371421814,0.796832799911499,4.3507561683654785,0,A_1,B_1,4.297342300415039,0.7376539707183838,1.812349796295166,0.30747416615486145
-8vkf_from_file,,,,,0.9036211371421814,0.796832799911499,4.3507561683654785,0,A_1,C_1,4.811293125152588,0.5839160084724426,1.8787881135940552,0.3352779746055603
-8vkf_from_file,,,,,0.9036211371421814,0.796832799911499,4.3507561683654785,0,A_1,D_1,10.173449516296387,2.466259241104126,7.62661600112915,1.7930353879928589
-8vkf_from_file,,,,,0.9036211371421814,0.796832799911499,4.3507561683654785,0,A_1,E_1,11.331687927246094,11.929487228393555,9.426876068115234,10.728137016296387
-8vkf_from_file,,,,,0.9036211371421814,0.796832799911499,4.3507561683654785,0,B_1,C_1,4.411619186401367,0.6202328205108643,2.9504830837249756,0.32363060116767883
-8vkf_from_file,,,,,0.9036211371421814,0.796832799911499,4.3507561683654785,0,B_1,D_1,10.569169044494629,2.8293044567108154,8.373913764953613,1.6293067932128906
-8vkf_from_file,,,,,0.9036211371421814,0.796832799911499,4.3507561683654785,0,B_1,E_1,10.919745445251465,10.450433731079102,9.151620864868164,9.581561088562012
-8vkf_from_file,,,,,0.9036211371421814,0.796832799911499,4.3507561683654785,0,C_1,D_1,10.661324501037598,3.1657612323760986,8.570053100585938,2.204336166381836
-8vkf_from_file,,,,,0.9036211371421814,0.796832799911499,4.3507561683654785,0,C_1,E_1,11.19679069519043,10.691766738891602,9.548834800720215,9.894548416137695
-8vkf_from_file,,,,,0.9036211371421814,0.796832799911499,4.3507561683654785,0,D_1,E_1,12.625629425048828,12.142257690429688,12.593172073364258,12.142263412475586
-8vkf_from_file,,,,,0.9036248922348022,0.793869137763977,4.341085910797119,1,A_1,B_1,4.291936874389648,0.7353562116622925,1.8156046867370605,0.3068144917488098
-8vkf_from_file,,,,,0.9036248922348022,0.793869137763977,4.341085910797119,1,A_1,C_1,4.804727554321289,0.581565797328949,1.8812065124511719,0.33534058928489685
-8vkf_from_file,,,,,0.9036248922348022,0.793869137763977,4.341085910797119,1,A_1,D_1,10.1722993850708,2.465243339538574,7.652758598327637,1.793786883354187
-8vkf_from_file,,,,,0.9036248922348022,0.793869137763977,4.341085910797119,1,A_1,E_1,11.337690353393555,11.883279800415039,9.425646781921387,10.76948356628418
-8vkf_from_file,,,,,0.9036248922348022,0.793869137763977,4.341085910797119,1,B_1,C_1,4.411491870880127,0.6201822757720947,2.95219087600708,0.32362139225006104
-8vkf_from_file,,,,,0.9036248922348022,0.793869137763977,4.341085910797119,1,B_1,D_1,10.568760871887207,2.826165199279785,8.375665664672852,1.628724217414856
-8vkf_from_file,,,,,0.9036248922348022,0.793869137763977,4.341085910797119,1,B_1,E_1,10.899205207824707,10.385791778564453,9.06670093536377,9.45239543914795
-8vkf_from_file,,,,,0.9036248922348022,0.793869137763977,4.341085910797119,1,C_1,D_1,10.661279678344727,3.166980266571045,8.5703125,2.2031681537628174
-8vkf_from_file,,,,,0.9036248922348022,0.793869137763977,4.341085910797119,1,C_1,E_1,11.18617057800293,10.668017387390137,9.607718467712402,10.071332931518555
-8vkf_from_file,,,,,0.9036248922348022,0.793869137763977,4.341085910797119,1,D_1,E_1,12.579813003540039,12.155464172363281,12.550373077392578,12.15546989440918
-8vkf_from_file,,,,,0.9036300778388977,0.7963874936103821,4.349856376647949,2,A_1,B_1,4.296799659729004,0.7375437021255493,1.8264133930206299,0.3068671226501465
-8vkf_from_file,,,,,0.9036300778388977,0.7963874936103821,4.349856376647949,2,A_1,C_1,4.811502456665039,0.5840663909912109,1.8806697130203247,0.3353065848350525
-8vkf_from_file,,,,,0.9036300778388977,0.7963874936103821,4.349856376647949,2,A_1,D_1,10.173911094665527,2.467724561691284,7.651202201843262,1.7933560609817505
-8vkf_from_file,,,,,0.9036300778388977,0.7963874936103821,4.349856376647949,2,A_1,E_1,11.350998878479004,11.898008346557617,9.472217559814453,10.83096694946289
-8vkf_from_file,,,,,0.9036300778388977,0.7963874936103821,4.349856376647949,2,B_1,C_1,4.411766529083252,0.62007737159729,2.9522931575775146,0.3235628306865692
-8vkf_from_file,,,,,0.9036300778388977,0.7963874936103821,4.349856376647949,2,B_1,D_1,10.568456649780273,2.8267123699188232,8.381253242492676,1.6290395259857178
-8vkf_from_file,,,,,0.9036300778388977,0.7963874936103821,4.349856376647949,2,B_1,E_1,10.902305603027344,10.377313613891602,9.157639503479004,9.32485580444336
-8vkf_from_file,,,,,0.9036300778388977,0.7963874936103821,4.349856376647949,2,C_1,D_1,10.662564277648926,3.1713716983795166,8.566250801086426,2.2024450302124023
-8vkf_from_file,,,,,0.9036300778388977,0.7963874936103821,4.349856376647949,2,C_1,E_1,11.18061637878418,10.643763542175293,9.567172050476074,10.075090408325195
-8vkf_from_file,,,,,0.9036300778388977,0.7963874936103821,4.349856376647949,2,D_1,E_1,12.614690780639648,12.20622444152832,12.574191093444824,12.206230163574219
-8vkf_from_file,,,,,0.9036257863044739,0.7943676114082336,4.343367099761963,3,A_1,B_1,4.293015480041504,0.7358640432357788,1.8161183595657349,0.3068287968635559
-8vkf_from_file,,,,,0.9036257863044739,0.7943676114082336,4.343367099761963,3,A_1,C_1,4.807739734649658,0.5823869705200195,1.880342960357666,0.33527007699012756
-8vkf_from_file,,,,,0.9036257863044739,0.7943676114082336,4.343367099761963,3,A_1,D_1,10.171187400817871,2.4647483825683594,7.647757053375244,1.7936530113220215
-8vkf_from_file,,,,,0.9036257863044739,0.7943676114082336,4.343367099761963,3,A_1,E_1,11.339066505432129,11.882888793945312,9.489239692687988,10.621736526489258
-8vkf_from_file,,,,,0.9036257863044739,0.7943676114082336,4.343367099761963,3,B_1,C_1,4.411233901977539,0.619956374168396,2.951002597808838,0.3235902786254883
-8vkf_from_file,,,,,0.9036257863044739,0.7943676114082336,4.343367099761963,3,B_1,D_1,10.567977905273438,2.8275861740112305,8.379467964172363,1.6287676095962524
-8vkf_from_file,,,,,0.9036257863044739,0.7943676114082336,4.343367099761963,3,B_1,E_1,10.905926704406738,10.406522750854492,9.048287391662598,9.378450393676758
-8vkf_from_file,,,,,0.9036257863044739,0.7943676114082336,4.343367099761963,3,C_1,D_1,10.660465240478516,3.170450210571289,8.571593284606934,2.205432653427124
-8vkf_from_file,,,,,0.9036257863044739,0.7943676114082336,4.343367099761963,3,C_1,E_1,11.15646743774414,10.600381851196289,9.579748153686523,10.07177448272705
-8vkf_from_file,,,,,0.9036257863044739,0.7943676114082336,4.343367099761963,3,D_1,E_1,12.597926139831543,12.024985313415527,12.551345825195312,12.024991035461426
-8vkf_from_file,,,,,0.903627336025238,0.7949872612953186,4.343382358551025,4,A_1,B_1,4.292372703552246,0.7358219027519226,1.8162869215011597,0.3068479299545288
-8vkf_from_file,,,,,0.903627336025238,0.7949872612953186,4.343382358551025,4,A_1,C_1,4.804841995239258,0.5825182199478149,1.8806352615356445,0.3352975845336914
-8vkf_from_file,,,,,0.903627336025238,0.7949872612953186,4.343382358551025,4,A_1,D_1,10.172637939453125,2.4656307697296143,7.650763511657715,1.7933474779129028
-8vkf_from_file,,,,,0.903627336025238,0.7949872612953186,4.343382358551025,4,A_1,E_1,11.329042434692383,11.928260803222656,9.406120300292969,10.859692573547363
-8vkf_from_file,,,,,0.903627336025238,0.7949872612953186,4.343382358551025,4,B_1,C_1,4.409976482391357,0.6200796961784363,2.9528026580810547,0.3236019015312195
-8vkf_from_file,,,,,0.903627336025238,0.7949872612953186,4.343382358551025,4,B_1,D_1,10.567740440368652,2.826847791671753,8.38717269897461,1.6291062831878662
-8vkf_from_file,,,,,0.903627336025238,0.7949872612953186,4.343382358551025,4,B_1,E_1,10.893219947814941,10.442112922668457,9.190901756286621,9.42338752746582
-8vkf_from_file,,,,,0.903627336025238,0.7949872612953186,4.343382358551025,4,C_1,D_1,10.659784317016602,3.155189275741577,8.569608688354492,2.2047157287597656
-8vkf_from_file,,,,,0.903627336025238,0.7949872612953186,4.343382358551025,4,C_1,E_1,11.20004940032959,10.708091735839844,9.58100700378418,10.21774673461914
-8vkf_from_file,,,,,0.903627336025238,0.7949872612953186,4.343382358551025,4,D_1,E_1,12.568294525146484,12.032037734985352,12.539318084716797,12.03204345703125

models/rf3/tests/data/inference_regression_tests/8vkf_from_file/8vkf_from_file_metrics.csv

@@ -1,2 +1,2 @@
 example_id,ptm.ptm_0,ptm.ptm_1,ptm.ptm_2,ptm.ptm_3,ptm.ptm_4,iptm.iptm_0,iptm.iptm_1,iptm.iptm_2,iptm.iptm_3,iptm.iptm_4,iptm.iptm_protein_protein_0,iptm.iptm_protein_protein_1,iptm.iptm_protein_protein_2,iptm.iptm_protein_protein_3,iptm.iptm_protein_protein_4,iptm.iptm_protein_ligand_0,iptm.iptm_protein_ligand_1,iptm.iptm_protein_ligand_2,iptm.iptm_protein_ligand_3,iptm.iptm_protein_ligand_4,iptm.iptm_ligand_ligand_0,iptm.iptm_ligand_ligand_1,iptm.iptm_ligand_ligand_2,iptm.iptm_ligand_ligand_3,iptm.iptm_ligand_ligand_4,count_clashing_chains.has_clash_0,count_clashing_chains.has_clash_1,count_clashing_chains.has_clash_2,count_clashing_chains.has_clash_3,count_clashing_chains.has_clash_4
-8vkf_from_file,0.90172315,0.9019547,0.90182585,0.9019781,0.9019156,0.89104724,0.89113414,0.8910541,0.8912014,0.8911338,0.0,0.0,0.0,0.0,0.0,0.89104724,0.89113414,0.8910541,0.8912014,0.8911338,0.82330704,0.82325506,0.823536,0.8234215,0.8234704,0,0,0,0,0
+8vkf_from_file,0.9397696,0.94022816,0.93984985,0.9406191,0.94042176,0.9700439,0.96997976,0.96992123,0.9700142,0.9700208,0.0,0.0,0.0,0.0,0.0,0.9700439,0.96997976,0.96992123,0.9700142,0.9700208,0.93240464,0.9323982,0.93233144,0.93221885,0.93262637,0,0,0,0,0

models/rf3/tests/data/inference_regression_tests/8vkf_from_file/8vkf_from_file_score.csv

@@ -0,0 +1,76 @@
+example_id,chain_chainwise,chainwise_plddt,chainwise_pde,chainwise_pae,overall_plddt,overall_pde,overall_pae,batch_idx,chain_i_interface,chain_j_interface,pae_interface,pde_interface,min_pae_interface,min_pde_interface
+8vkf_from_file,A_1,0.9320401549339294,0.9453674554824829,3.9356305599212646,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,,,,,,
+8vkf_from_file,B_1,0.9495342969894409,0.3442486524581909,1.6652324199676514,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,,,,,,
+8vkf_from_file,C_1,0.9272092580795288,0.3786144256591797,1.5746772289276123,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,,,,,,
+8vkf_from_file,D_1,0.9011284112930298,4.034590721130371,7.017383098602295,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,,,,,,
+8vkf_from_file,E_1,0.8282343745231628,13.62786865234375,8.54509449005127,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,,,,,,
+8vkf_from_file,A_1,0.932106077671051,0.93052077293396,3.9099152088165283,0.9322792887687683,0.8681702613830566,3.882534980773926,1,,,,,,
+8vkf_from_file,B_1,0.9495639204978943,0.34412628412246704,1.6654282808303833,0.9322792887687683,0.8681702613830566,3.882534980773926,1,,,,,,
+8vkf_from_file,C_1,0.9272972345352173,0.3788919746875763,1.574660062789917,0.9322792887687683,0.8681702613830566,3.882534980773926,1,,,,,,
+8vkf_from_file,D_1,0.9011081457138062,4.034159183502197,7.01707124710083,0.9322792887687683,0.8681702613830566,3.882534980773926,1,,,,,,
+8vkf_from_file,E_1,0.8281877636909485,13.573341369628906,8.574715614318848,0.9322792887687683,0.8681702613830566,3.882534980773926,1,,,,,,
+8vkf_from_file,A_1,0.932047963142395,0.9424294233322144,3.931513547897339,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,,,,,,
+8vkf_from_file,B_1,0.94952392578125,0.3441573977470398,1.6650241613388062,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,,,,,,
+8vkf_from_file,C_1,0.9272950291633606,0.37870317697525024,1.5746617317199707,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,,,,,,
+8vkf_from_file,D_1,0.9011602401733398,4.035135746002197,7.019920349121094,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,,,,,,
+8vkf_from_file,E_1,0.8280327320098877,13.63625717163086,8.625662803649902,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,,,,,,
+8vkf_from_file,A_1,0.9320767521858215,0.9201368093490601,3.8943328857421875,0.9322513937950134,0.859204888343811,3.8692407608032227,3,,,,,,
+8vkf_from_file,B_1,0.9496347904205322,0.344124972820282,1.6652103662490845,0.9322513937950134,0.859204888343811,3.8692407608032227,3,,,,,,
+8vkf_from_file,C_1,0.92729252576828,0.37906670570373535,1.5739272832870483,0.9322513937950134,0.859204888343811,3.8692407608032227,3,,,,,,
+8vkf_from_file,D_1,0.9012048244476318,4.029945373535156,7.012320518493652,0.9322513937950134,0.859204888343811,3.8692407608032227,3,,,,,,
+8vkf_from_file,E_1,0.8280571103096008,13.597679138183594,8.608673095703125,0.9322513937950134,0.859204888343811,3.8692407608032227,3,,,,,,
+8vkf_from_file,A_1,0.9320934414863586,0.9275834560394287,3.903271198272705,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,,,,,,
+8vkf_from_file,B_1,0.9496190547943115,0.344104528427124,1.6647707223892212,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,,,,,,
+8vkf_from_file,C_1,0.9273494482040405,0.37881240248680115,1.5747339725494385,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,,,,,,
+8vkf_from_file,D_1,0.9010905623435974,4.033513069152832,7.0145745277404785,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,,,,,,
+8vkf_from_file,E_1,0.8283198475837708,13.572347640991211,8.580305099487305,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,,,,,,
+8vkf_from_file,,,,,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,A_1,B_1,3.4717438220977783,0.605563223361969,0.6303903460502625,0.2651037275791168
+8vkf_from_file,,,,,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,A_1,C_1,4.8907575607299805,0.6995029449462891,0.7152637839317322,0.28492048382759094
+8vkf_from_file,,,,,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,A_1,D_1,4.7494916915893555,0.8386649489402771,2.0623931884765625,0.46002161502838135
+8vkf_from_file,,,,,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,A_1,E_1,9.049422264099121,3.8051230907440186,5.0449137687683105,2.8304648399353027
+8vkf_from_file,,,,,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,B_1,C_1,2.736272096633911,0.36354562640190125,1.1683330535888672,0.2732650637626648
+8vkf_from_file,,,,,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,B_1,D_1,3.264293909072876,0.48301342129707336,2.083875894546509,0.36788517236709595
+8vkf_from_file,,,,,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,B_1,E_1,7.824339389801025,2.5479378700256348,5.2585883140563965,1.8988345861434937
+8vkf_from_file,,,,,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,C_1,D_1,4.22511625289917,0.5288299918174744,2.2046451568603516,0.4303090572357178
+8vkf_from_file,,,,,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,C_1,E_1,9.384553909301758,2.672896146774292,7.508654594421387,2.3068904876708984
+8vkf_from_file,,,,,0.9322137236595154,0.8812259435653687,3.9045498371124268,0,D_1,E_1,9.492976188659668,4.036022186279297,8.316246032714844,4.03602409362793
+8vkf_from_file,,,,,0.9322792887687683,0.8681702613830566,3.882534980773926,1,A_1,B_1,3.459951162338257,0.5984316468238831,0.630462110042572,0.26509952545166016
+8vkf_from_file,,,,,0.9322792887687683,0.8681702613830566,3.882534980773926,1,A_1,C_1,4.877954006195068,0.6918987035751343,0.7154307961463928,0.2848895192146301
+8vkf_from_file,,,,,0.9322792887687683,0.8681702613830566,3.882534980773926,1,A_1,D_1,4.736978530883789,0.8306165933609009,2.0626161098480225,0.4596916437149048
+8vkf_from_file,,,,,0.9322792887687683,0.8681702613830566,3.882534980773926,1,A_1,E_1,9.124225616455078,3.7755820751190186,5.10176944732666,2.8448750972747803
+8vkf_from_file,,,,,0.9322792887687683,0.8681702613830566,3.882534980773926,1,B_1,C_1,2.73687744140625,0.3636227250099182,1.169067144393921,0.27333831787109375
+8vkf_from_file,,,,,0.9322792887687683,0.8681702613830566,3.882534980773926,1,B_1,D_1,3.262706756591797,0.4828460216522217,2.0824265480041504,0.36787867546081543
+8vkf_from_file,,,,,0.9322792887687683,0.8681702613830566,3.882534980773926,1,B_1,E_1,7.864833354949951,2.5528337955474854,5.208006858825684,1.8787106275558472
+8vkf_from_file,,,,,0.9322792887687683,0.8681702613830566,3.882534980773926,1,C_1,D_1,4.22160530090332,0.5286768078804016,2.1941330432891846,0.4300752282142639
+8vkf_from_file,,,,,0.9322792887687683,0.8681702613830566,3.882534980773926,1,C_1,E_1,9.301826477050781,2.655578374862671,7.413575172424316,2.298459053039551
+8vkf_from_file,,,,,0.9322792887687683,0.8681702613830566,3.882534980773926,1,D_1,E_1,9.516148567199707,3.908454418182373,8.317085266113281,3.908456325531006
+8vkf_from_file,,,,,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,A_1,B_1,3.4694740772247314,0.6041339039802551,0.6303707361221313,0.2650924623012543
+8vkf_from_file,,,,,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,A_1,C_1,4.888820171356201,0.698183000087738,0.7153699994087219,0.28490278124809265
+8vkf_from_file,,,,,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,A_1,D_1,4.746122360229492,0.8371659517288208,2.0624828338623047,0.4598386585712433
+8vkf_from_file,,,,,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,A_1,E_1,9.238248825073242,3.801729202270508,5.277400970458984,2.8974578380584717
+8vkf_from_file,,,,,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,B_1,C_1,2.7369372844696045,0.36360710859298706,1.1681721210479736,0.2720290422439575
+8vkf_from_file,,,,,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,B_1,D_1,3.2642736434936523,0.48311784863471985,2.0822458267211914,0.36787092685699463
+8vkf_from_file,,,,,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,B_1,E_1,7.957021236419678,2.5393381118774414,5.358416557312012,1.888516902923584
+8vkf_from_file,,,,,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,C_1,D_1,4.223407745361328,0.5279184579849243,2.2048254013061523,0.4302324652671814
+8vkf_from_file,,,,,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,C_1,E_1,9.441669464111328,2.6686744689941406,7.508003234863281,2.2565596103668213
+8vkf_from_file,,,,,0.9322216510772705,0.8786524534225464,3.9016854763031006,2,D_1,E_1,9.5327730178833,3.8588662147521973,8.329977989196777,3.85886812210083
+8vkf_from_file,,,,,0.9322513937950134,0.859204888343811,3.8692407608032227,3,A_1,B_1,3.4515013694763184,0.5935373902320862,0.6304340958595276,0.2650916874408722
+8vkf_from_file,,,,,0.9322513937950134,0.859204888343811,3.8692407608032227,3,A_1,C_1,4.87033224105835,0.6863856911659241,0.715640127658844,0.2848808169364929
+8vkf_from_file,,,,,0.9322513937950134,0.859204888343811,3.8692407608032227,3,A_1,D_1,4.7307820320129395,0.8252673745155334,2.062479019165039,0.45959168672561646
+8vkf_from_file,,,,,0.9322513937950134,0.859204888343811,3.8692407608032227,3,A_1,E_1,9.2240629196167,3.7983005046844482,5.149802207946777,2.77072811126709
+8vkf_from_file,,,,,0.9322513937950134,0.859204888343811,3.8692407608032227,3,B_1,C_1,2.734774589538574,0.36362209916114807,1.167860746383667,0.27200546860694885
+8vkf_from_file,,,,,0.9322513937950134,0.859204888343811,3.8692407608032227,3,B_1,D_1,3.2621781826019287,0.4826110005378723,2.07240629196167,0.3676457107067108
+8vkf_from_file,,,,,0.9322513937950134,0.859204888343811,3.8692407608032227,3,B_1,E_1,7.96647834777832,2.5434417724609375,5.3074846267700195,1.9209169149398804
+8vkf_from_file,,,,,0.9322513937950134,0.859204888343811,3.8692407608032227,3,C_1,D_1,4.221768379211426,0.5291066765785217,2.2043018341064453,0.4300277829170227
+8vkf_from_file,,,,,0.9322513937950134,0.859204888343811,3.8692407608032227,3,C_1,E_1,9.462154388427734,2.6579153537750244,7.632106304168701,2.2554261684417725
+8vkf_from_file,,,,,0.9322513937950134,0.859204888343811,3.8692407608032227,3,D_1,E_1,9.495095252990723,3.771697521209717,8.34188461303711,3.7716994285583496
+8vkf_from_file,,,,,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,A_1,B_1,3.453770637512207,0.5964496731758118,0.6305050253868103,0.26513877511024475
+8vkf_from_file,,,,,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,A_1,C_1,4.868809700012207,0.6897773146629333,0.7071991562843323,0.2849225699901581
+8vkf_from_file,,,,,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,A_1,D_1,4.734051704406738,0.828239917755127,2.0547072887420654,0.45981019735336304
+8vkf_from_file,,,,,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,A_1,E_1,9.135178565979004,3.786433219909668,5.102996826171875,2.852632761001587
+8vkf_from_file,,,,,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,B_1,C_1,2.734165668487549,0.36359599232673645,1.1680809259414673,0.273250013589859
+8vkf_from_file,,,,,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,B_1,D_1,3.2627716064453125,0.48312124609947205,2.0826122760772705,0.36937710642814636
+8vkf_from_file,,,,,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,B_1,E_1,7.866828441619873,2.5545129776000977,5.217214584350586,1.8813700675964355
+8vkf_from_file,,,,,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,C_1,D_1,4.222236156463623,0.528620183467865,2.2064924240112305,0.43011271953582764
+8vkf_from_file,,,,,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,C_1,E_1,9.318958282470703,2.6771984100341797,7.419789791107178,2.3011441230773926
+8vkf_from_file,,,,,0.9322678446769714,0.8655266165733337,3.8760082721710205,4,D_1,E_1,9.404740333557129,3.7400143146514893,8.203256607055664,3.740015983581543

models/rf3/tests/test_inference_regression.py

@@ -31,6 +31,7 @@ RUN_PARAM_KEYS = {
     "skip_existing",
     "template_selection",
     "ground_truth_conformer_selection",
+    "cyclic_chains",
 }
 """Run parameters that should be passed to engine.run(), not __init__."""
 
@@ -159,8 +160,8 @@ def compare_structures(
 @pytest.mark.parametrize(
     "example_id,rmsd_tolerance,csv_tolerance",
     [
-        ("5vht_from_file", 0.1, 0.003),
-        ("8vkf_from_file", 0.1, 0.003),
+        ("5vht_from_file", 0.1, 0.01),
+        ("8vkf_from_file", 0.1, 0.01),
     ],
 )
 def test_inference_regression(example_id, rmsd_tolerance, csv_tolerance):
@@ -241,8 +242,8 @@ def test_inference_regression(example_id, rmsd_tolerance, csv_tolerance):
 @pytest.mark.parametrize(
     "example_id,rmsd_tolerance,csv_tolerance",
     [
-        ("5vht_from_file", 0.1, 0.003),
-        ("8vkf_from_file", 0.1, 0.003),
+        ("5vht_from_file", 0.1, 0.01),
+        ("8vkf_from_file", 0.1, 0.01),
     ],
 )
 def test_inference_regression_in_memory(example_id, rmsd_tolerance, csv_tolerance):
@@ -325,3 +326,6 @@ def test_inference_regression_in_memory(example_id, rmsd_tolerance, csv_toleranc
         assert (
             rmsd_difference < rmsd_tolerance
         ), f"Mean RMSD difference {rmsd_difference:.4f}Å exceeds {rmsd_tolerance}Å tolerance for {example_id}"
+
+if __name__ == "__main__":
+    pytest.main([__file__])

models/rfd3/configs/datasets/train/af2db_interdomain_distillation.yaml

@@ -4,7 +4,7 @@ defaults:
 
 interdomain_distillation:
   dataset:
-    _target_: atomworks.ml.datasets.datasets.StructuralDatasetWrapper
+    _target_: atomworks.ml.datasets.StructuralDatasetWrapper
 
     cif_parser_args:
       cache_dir: null
@@ -19,7 +19,7 @@ interdomain_distillation:
       path_colname: path
 
     dataset:
-      _target_: atomworks.ml.datasets.datasets.PandasDataset
+      _target_: atomworks.ml.datasets.PandasDataset
       name: af2db_interdomain_distillation
       data: /projects/ml/datahub/dfs/af2db_interdomain_dset.parquet
       id_column: example_id

models/rfd3/configs/datasets/train/bcov_ppi_distillation.yaml

@@ -5,7 +5,7 @@ defaults:
 
 bcov_ppi_distillation:
   dataset:
-    _target_: atomworks.ml.datasets.datasets.StructuralDatasetWrapper
+    _target_: atomworks.ml.datasets.StructuralDatasetWrapper
 
     cif_parser_args:
       cache_dir: null
@@ -20,7 +20,7 @@ bcov_ppi_distillation:
       path_colname: path
 
     dataset:
-      _target_: atomworks.ml.datasets.datasets.PandasDataset
+      _target_: atomworks.ml.datasets.PandasDataset
       name: bcov_ppi_distillation
       data: /projects/ml/datahub/dfs/bcov_ppi_dset_af3_preds.parquet
       id_column: example_id

models/rfd3/configs/datasets/train/pdb/base_no_weights.yaml

@@ -3,14 +3,14 @@ defaults:
   - _self_
 
 dataset:
-  _target_: atomworks.ml.datasets.datasets.StructuralDatasetWrapper
+  _target_: atomworks.ml.datasets.StructuralDatasetWrapper
   save_failed_examples_to_dir: ${paths.data.failed_examples_dir}
   cif_parser_args:
     cache_dir: null
     load_from_cache: false
     save_to_cache: false
   dataset:
-    _target_: atomworks.ml.datasets.datasets.PandasDataset
+    _target_: atomworks.ml.datasets.PandasDataset
     # we will use the example_id as the unique column
     id_column: example_id
   transform:

models/rfd3/configs/datasets/train/rfd3_monomer_distillation.yaml

@@ -4,7 +4,7 @@ defaults:
 
 monomer_distillation:
   dataset:
-    _target_: atomworks.ml.datasets.datasets.StructuralDatasetWrapper
+    _target_: atomworks.ml.datasets.StructuralDatasetWrapper
     save_failed_examples_to_dir: ${paths.data.failed_examples_dir}
     # Explicitly do not load from cache.
     # Dataset too big, and structures are small
@@ -15,7 +15,7 @@ monomer_distillation:
 
     # metadata dataset
     dataset:
-      _target_: atomworks.ml.datasets.datasets.PandasDataset
+      _target_: atomworks.ml.datasets.PandasDataset
       name: af2fb_distillation
       id_column: example_id
       data: ${paths.data.monomer_distillation_parquet_dir}/af2_distillation_facebook.parquet

models/rfd3/src/rfd3/inference/datasets.py

@@ -12,8 +12,7 @@ from typing import Any, List
 import yaml
 from atomworks.io.parser import parse_atom_array
 
-# from atomworks.ml.datasets.datasets import BaseDataset
-from atomworks.ml.datasets.datasets import MolecularDataset
+from atomworks.ml.datasets import MolecularDataset
 from atomworks.ml.transforms.base import Compose, Transform, TransformedDict
 from biotite.structure import BondList
 from omegaconf import DictConfig, OmegaConf
@@ -41,7 +40,7 @@ logger = RankedLogger(__name__, rank_zero_only=True)
 all_ranks_logger = RankedLogger(__name__, rank_zero_only=False)
 
 
-class ContigJsonDataset(MolecularDataset):  # atomworks.ml.datasets.datasets
+class ContigJsonDataset(MolecularDataset):
     """
     Enables loading of JSON files containing contig data for benchmark design tasks,
     or the passing of examples through analogously-structured hydra configs.

src/modelhub/trainers/fabric.py

@@ -118,8 +118,12 @@ class FabricTrainer(ABC):
             (3) Fabric Loggers (https://lightning.ai/docs/fabric/2.4.0/api/loggers.html)
             (4) Efficient Gradient Accumulation (https://lightning.ai/docs/fabric/2.4.0/advanced/gradient_accumulation.html)
         """
-        # DDP strategy requires a manual timeout higher than the default
-        if strategy == "ddp" and not is_interactive_environment():
+        # Use custom DDP strategy only for multi-device, non-interactive environments
+        if (
+            strategy == "ddp"
+            and not is_interactive_environment()
+            and not (num_nodes == 1 and devices_per_node == 1)
+        ):
             strategy = DDPStrategy(
                 timeout=timedelta(seconds=nccl_timeout),
                 find_unused_parameters=find_unused_parameters,

src/modelhub/utils/datasets.py

@@ -25,14 +25,14 @@ from modelhub.utils.ddp import RankedLogger
 
 ranked_logger = RankedLogger(__name__, rank_zero_only=True)
 try:
-    from atomworks.ml.datasets.datasets import (
+    from atomworks.ml.datasets import (
         ConcatDatasetWithID,
         FallbackDatasetWrapper,
         get_row_and_index_by_example_id,
     )
 except Exception as e:
     ranked_logger.warning(
-        f"Failed to import atomworks.ml.datasets.datasets: {type(e).__name__}: {e}. "
+        f"Failed to import atomworks.ml.datasets: {type(e).__name__}: {e}. "
         "If training networks, the PDB_MIRROR environment variable must be set."
     )