{ "cells": [ { "cell_type": "markdown", "id": "35d79007", "metadata": {}, "source": [ "# Training SchNet on rMD17 (ethanol)\n", "\n", "This trains the `xnn` **SchNet** (Schütt *et al.*, NIPS 2017) from scratch on\n", "the paper's own benchmark task: energies **and forces** of a molecular-dynamics\n", "trajectory of a small organic molecule (the paper's MD17, Table 2 — here the\n", "noise-free revision [rMD17](https://doi.org/10.1088/2632-2153/abba6f),\n", "downloaded in seconds through the `xnn` hub), and reproduces the paper's\n", "headline demonstrations at small scale:\n", "\n", "* joint **energy + force training** with the paper's loss weighting\n", " (eq. 5, $\\rho = 0.01$) — forces come for free via autograd through the\n", " `ForceStressOutput` wrapper the `Trainer` adds (the model is\n", " energy-conserving by construction, eq. 4);\n", "* an **energy/force parity** plot vs DFT on held-out conformations;\n", "* a **smooth 1-D potential-energy scan**, the property that continuous\n", " filters + shifted softplus buy over discretized approaches (paper Fig. 1).\n", "\n", "The model is the paper architecture ($F = 64$, $T = 3$, Gaussian RBFs every\n", "0.1 Å with $\\gamma = 10$ Å$^{-2}$); the implementation is verified against\n", "the manuscripts' equations block by block in\n", "`examples/fidelity_checks/schnet_verification.ipynb`.\n", "\n", "Run with the **`xnn`** kernel (uses the GPU when available)." ] }, { "cell_type": "markdown", "id": "fdeabc4a", "metadata": {}, "source": [ "## 0. Setup and data" ] }, { "cell_type": "code", "execution_count": 1, "id": "806082e0", "metadata": { "execution": { "iopub.execute_input": "2026-07-20T04:18:18.176205Z", "iopub.status.busy": "2026-07-20T04:18:18.176031Z", "iopub.status.idle": "2026-07-20T04:18:21.342354Z", "shell.execute_reply": "2026-07-20T04:18:21.341655Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "train 1000 test 500 atoms/molecule 9\n" ] } ], "source": [ "import time, warnings\n", "warnings.filterwarnings(\"ignore\")\n", "import numpy as np\n", "import torch\n", "import matplotlib.pyplot as plt\n", "\n", "torch.set_default_dtype(torch.float64)\n", "torch.manual_seed(0)\n", "\n", "from xnn.common.data import load_dataset\n", "\n", "# The paper trains without a distance cutoff -- its RBF grid just ends beyond\n", "# every pair distance. Ethanol spans < 5 A, so a 10 A radius reproduces that\n", "# no-cutoff setting (complete graph) with a matching 0.1 A RBF grid; the DTNN\n", "# paper likewise picks the grid end \"depending on the range of distances in\n", "# the data\".\n", "CUTOFF, N_RBF = 10.0, 101\n", "\n", "splits = load_dataset(\"rmd17\", molecule=\"ethanol\",\n", " n_train=1000, n_test=500, quiet=True)\n", "train_structs, test_structs = splits[\"train\"], splits[\"test\"]\n", "N_ATOMS = len(train_structs[0][\"atomic_numbers\"])\n", "print(f\"train {len(train_structs)} test {len(test_structs)} \"\n", " f\"atoms/molecule {N_ATOMS}\")" ] }, { "cell_type": "markdown", "id": "c4363c0a", "metadata": {}, "source": [ "## 1. Per-atom energy standardization (DTNN Methods, step 4)\n", "\n", "SchNet inherits the DTNN output convention: the network predicts a\n", "*standardized* per-atom energy $\\hat E_i$, and the final contribution is\n", "$E_i = E_\\sigma \\hat E_i + E_\\mu$ with $E_\\mu$/$E_\\sigma$ the mean and\n", "standard deviation of the **energy per atom** over the training set. With the\n", "zero-initialized output head, the model's first prediction is exactly the\n", "training-set mean — \"a good starting point for the training\"." ] }, { "cell_type": "code", "execution_count": 2, "id": "bf794d09", "metadata": { "execution": { "iopub.execute_input": "2026-07-20T04:18:21.344112Z", "iopub.status.busy": "2026-07-20T04:18:21.343843Z", "iopub.status.idle": "2026-07-20T04:18:21.348367Z", "shell.execute_reply": "2026-07-20T04:18:21.347562Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "E_mu = -467.7363 eV/atom E_sigma = 0.01981 eV/atom\n", "(total-energy spread of the trajectory: 0.178 eV)\n" ] } ], "source": [ "E_train = np.array([s[\"energy\"] for s in train_structs])\n", "E_SHIFT = float(E_train.mean() / N_ATOMS) # E_mu (eV / atom)\n", "E_SCALE = float(E_train.std() / N_ATOMS) # E_sigma (eV / atom)\n", "print(f\"E_mu = {E_SHIFT:.4f} eV/atom E_sigma = {E_SCALE:.5f} eV/atom\")\n", "print(f\"(total-energy spread of the trajectory: {E_train.std():.3f} eV)\")" ] }, { "cell_type": "markdown", "id": "17d1bec6", "metadata": {}, "source": [ "## 2. Build the paper's SchNet and train on energies + forces\n", "\n", "The paper's combined loss (eq. 5) is\n", "$\\ell = \\rho\\,\\|E - \\hat E\\|^2 + \\frac{1}{n}\\sum_i \\|F_i + \\partial\\hat\n", "E/\\partial R_i\\|^2$ with $\\rho = 0.01$. The `xnn` trainer normalizes the\n", "energy term per atom (`energy_weight * ((E - Ê)/n)²`), so `energy_weight=1,\n", "force_weight=1` gives an effective energy:force ratio of $1/n^2 = 1/81\n", "\\approx 0.012$ — the paper's $\\rho$." ] }, { "cell_type": "code", "execution_count": 3, "id": "9ce2c350", "metadata": { "execution": { "iopub.execute_input": "2026-07-20T04:18:21.349739Z", "iopub.status.busy": "2026-07-20T04:18:21.349612Z", "iopub.status.idle": "2026-07-20T04:22:13.625476Z", "shell.execute_reply": "2026-07-20T04:22:13.624492Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "SchNet parameters: 78,117 | device: cuda\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "test loss 7.6798e-03\n", "trained 300 epochs in 231s | test loss 7.6798e-03\n" ] } ], "source": [ "from torch.utils.data import Subset\n", "from xnn.common.data import AtomicDataset\n", "from xnn.common.config import Config, ModelConfig, DataConfig, OptimConfig\n", "from xnn.common.train import Trainer\n", "\n", "EPOCHS, BS, LR = 300, 32, 1e-3\n", "\n", "train_ds = AtomicDataset(train_structs, CUTOFF)\n", "test_ds = AtomicDataset(test_structs, CUTOFF)\n", "n_val = 100\n", "val_idx = list(range(n_val))\n", "tr_idx = list(range(n_val, len(train_ds)))\n", "\n", "cfg = Config(\n", " model=ModelConfig(name=\"schnet\", cutoff=CUTOFF, n_features=64,\n", " n_interactions=3, n_rbf=N_RBF,\n", " extra={\"gamma\": 10.0,\n", " \"energy_shift\": E_SHIFT,\n", " \"energy_scale\": E_SCALE}),\n", " data=DataConfig(cutoff=CUTOFF, batch_size=BS),\n", " optim=OptimConfig(lr=LR, epochs=EPOCHS, energy_weight=1.0,\n", " force_weight=1.0, scheduler=\"plateau\"),\n", " output_dir=\"runs/schnet_rmd17\",\n", ")\n", "\n", "trainer = Trainer(cfg, Subset(train_ds, tr_idx), Subset(train_ds, val_idx),\n", " test_ds)\n", "n_params = sum(p.numel() for p in trainer.module.parameters())\n", "print(f\"SchNet parameters: {n_params:,} | device: {trainer.device}\")\n", "\n", "hist = {\"train\": [], \"val\": []}\n", "trainer._log = lambda ep, tr, va: (hist[\"train\"].append(tr.get(\"loss\")),\n", " hist[\"val\"].append(va.get(\"loss\")))\n", "t0 = time.time()\n", "metrics = trainer.fit()\n", "print(f\"trained {EPOCHS} epochs in {time.time()-t0:.0f}s | \"\n", " f\"test loss {metrics['test']['loss']:.4e}\")" ] }, { "cell_type": "markdown", "id": "5c344e22", "metadata": {}, "source": [ "## 3. Learning curves" ] }, { "cell_type": "code", "execution_count": 4, "id": "064da43f", "metadata": { "execution": { "iopub.execute_input": "2026-07-20T04:22:13.627513Z", "iopub.status.busy": "2026-07-20T04:22:13.627194Z", "iopub.status.idle": "2026-07-20T04:22:14.611405Z", "shell.execute_reply": "2026-07-20T04:22:14.610342Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, ax = plt.subplots(figsize=(5, 3.2))\n", "ax.plot(hist[\"train\"], label=\"train\")\n", "ax.plot(hist[\"val\"], label=\"val\")\n", "ax.set_yscale(\"log\"); ax.set_xlabel(\"epoch\"); ax.set_ylabel(\"loss\")\n", "ax.set_title(\"SchNet on rMD17 ethanol (energies + forces)\")\n", "ax.legend(); fig.tight_layout()\n", "fig.savefig(\"schnet_loss_curves.png\", dpi=110); plt.show()" ] }, { "cell_type": "markdown", "id": "ce415b24", "metadata": {}, "source": [ "## 4. Energy & force parity vs DFT (the paper's Table 2 task)\n", "\n", "Predict on the 500 held-out conformations and compare to the reference DFT.\n", "We report MAEs in kcal/mol — the paper's unit. For scale: SchNet trained on\n", "$N = 1000$ MD17 ethanol conformations reaches 0.08 kcal/mol (energy) and\n", "0.39 kcal/mol/Å (forces) in Table 2, with ~10x longer training and learning-\n", "rate decay; this 300-epoch run lands within a few times those numbers\n", "(sub-kcal/mol energies)." ] }, { "cell_type": "code", "execution_count": 5, "id": "7cdd2d72", "metadata": { "execution": { "iopub.execute_input": "2026-07-20T04:22:14.615430Z", "iopub.status.busy": "2026-07-20T04:22:14.615292Z", "iopub.status.idle": "2026-07-20T04:22:15.461723Z", "shell.execute_reply": "2026-07-20T04:22:15.450525Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "energy MAE : 25.87 meV (0.597 kcal/mol)\n", "force MAE : 62.58 meV/A (1.443 kcal/mol/A)\n", "energy RMSE: 33.49 meV (0.772 kcal/mol)\n", "force RMSE: 87.76 meV/A (2.024 kcal/mol/A)\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from torch.utils.data import DataLoader\n", "from xnn.common.data import collate\n", "\n", "model = trainer.model.eval()\n", "device = trainer.device\n", "EV2KCAL = 23.060541945329334\n", "\n", "E_pred, E_ref, F_pred, F_ref = [], [], [], []\n", "for batch in DataLoader(test_ds, batch_size=50, collate_fn=collate):\n", " batch = batch.to(device)\n", " out = model(batch)\n", " E_pred.append(out[\"energy\"].detach().cpu().numpy())\n", " E_ref.append(batch.energy.detach().cpu().numpy())\n", " F_pred.append(out[\"forces\"].detach().cpu().numpy())\n", " F_ref.append(batch.forces.detach().cpu().numpy())\n", "E_pred, E_ref = np.concatenate(E_pred), np.concatenate(E_ref)\n", "F_pred, F_ref = np.concatenate(F_pred), np.concatenate(F_ref)\n", "\n", "e_mae = np.abs(E_pred - E_ref).mean()\n", "f_mae = np.abs(F_pred - F_ref).mean()\n", "e_rmse = np.sqrt(np.mean((E_pred - E_ref) ** 2))\n", "f_rmse = np.sqrt(np.mean((F_pred - F_ref) ** 2))\n", "print(f\"energy MAE : {e_mae*1e3:6.2f} meV ({e_mae*EV2KCAL:.3f} kcal/mol)\")\n", "print(f\"force MAE : {f_mae*1e3:6.2f} meV/A ({f_mae*EV2KCAL:.3f} kcal/mol/A)\")\n", "print(f\"energy RMSE: {e_rmse*1e3:6.2f} meV ({e_rmse*EV2KCAL:.3f} kcal/mol)\")\n", "print(f\"force RMSE: {f_rmse*1e3:6.2f} meV/A ({f_rmse*EV2KCAL:.3f} kcal/mol/A)\")\n", "\n", "fig, (a1, a2) = plt.subplots(1, 2, figsize=(9, 4))\n", "a1.scatter(E_ref - E_ref.mean(), E_pred - E_ref.mean(), s=8, alpha=0.5)\n", "lim = np.array([(E_ref - E_ref.mean()).min(), (E_ref - E_ref.mean()).max()])\n", "a1.plot(lim, lim, \"k--\", lw=1)\n", "a1.set_xlabel(\"DFT energy (eV, centered)\")\n", "a1.set_ylabel(\"SchNet energy (eV, centered)\")\n", "a1.set_title(f\"energy MAE {e_mae*EV2KCAL:.3f} kcal/mol\")\n", "a2.scatter(F_ref.ravel(), F_pred.ravel(), s=3, alpha=0.2)\n", "lim = np.array([F_ref.min(), F_ref.max()])\n", "a2.plot(lim, lim, \"k--\", lw=1)\n", "a2.set_xlabel(\"DFT force (eV/A)\"); a2.set_ylabel(\"SchNet force (eV/A)\")\n", "a2.set_title(f\"forces MAE {f_mae*EV2KCAL:.3f} kcal/mol/A\")\n", "fig.tight_layout()\n", "fig.savefig(\"schnet_parity.png\", dpi=110); plt.show()" ] }, { "cell_type": "markdown", "id": "45441870", "metadata": {}, "source": [ "## 5. A smooth 1-D potential-energy scan (paper Fig. 1)\n", "\n", "The point of *continuous* filters is a smooth, differentiable PES (the\n", "paper's Fig. 1 contrasts this with the jagged surface a discretized filter\n", "produces). We stretch the C–O bond of a test geometry and evaluate the\n", "trained SchNet along the scan: a smooth single-well curve is what enables\n", "stable geometry optimization and MD." ] }, { "cell_type": "code", "execution_count": 6, "id": "8bded005", "metadata": { "execution": { "iopub.execute_input": "2026-07-20T04:22:15.464063Z", "iopub.status.busy": "2026-07-20T04:22:15.463395Z", "iopub.status.idle": "2026-07-20T04:22:16.654883Z", "shell.execute_reply": "2026-07-20T04:22:16.648241Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from xnn.common.data.neighborlist import build_neighbor_list\n", "from xnn.common.data import AtomicGraph\n", "\n", "s = test_structs[0]\n", "Z = torch.tensor(np.asarray(s[\"atomic_numbers\"]), dtype=torch.long)\n", "pos0 = torch.tensor(s[\"pos\"], dtype=torch.float64)\n", "\n", "# ethanol: stretch the C-O bond (C index nearest the O)\n", "O = int((Z == 8).nonzero()[0])\n", "Cs = (Z == 6).nonzero().flatten()\n", "d = torch.linalg.norm(pos0[Cs] - pos0[O], dim=1)\n", "C = int(Cs[d.argmin()])\n", "axis = pos0[O] - pos0[C]; axis = axis / axis.norm()\n", "r0 = float(torch.linalg.norm(pos0[O] - pos0[C]))\n", "\n", "scan = np.linspace(-0.35, 0.6, 60)\n", "energies = []\n", "for dr in scan:\n", " pos = pos0.clone(); pos[O] = pos0[O] + dr * axis\n", " ei, cs = build_neighbor_list(pos, CUTOFF)\n", " g = AtomicGraph(pos=pos.to(device), atomic_numbers=Z.to(device),\n", " edge_index=ei.to(device), cell_shifts=cs.to(device),\n", " batch=torch.zeros(len(Z), dtype=torch.long, device=device),\n", " n_atoms=torch.tensor([len(Z)], device=device))\n", " with torch.no_grad():\n", " energies.append(float(trainer.module.model(g)[\"energy\"]))\n", "energies = np.array(energies) - min(energies)\n", "\n", "fig, ax = plt.subplots(figsize=(5, 3.2))\n", "ax.plot(r0 + scan, energies, \"o-\", ms=3)\n", "ax.set_xlabel(\"C-O distance (A)\"); ax.set_ylabel(\"relative energy (eV)\")\n", "ax.set_title(\"SchNet C-O bond scan (smooth PES)\")\n", "fig.tight_layout()\n", "fig.savefig(\"schnet_bond_scan.png\", dpi=110); plt.show()" ] }, { "cell_type": "markdown", "id": "e4910380", "metadata": {}, "source": [ "## Summary\n", "\n", "From scratch on 900 ethanol conformations, the paper-architecture `xnn`\n", "SchNet trained jointly on energies and forces reaches kcal/mol-scale accuracy\n", "on held-out DFT data and yields a smooth potential-energy surface — the two\n", "properties the NIPS paper demonstrates on MD17. The trained checkpoint is\n", "saved under `runs/schnet_rmd17/`; `schnet_ethanol_md.ipynb` picks it up to\n", "run NVE molecular dynamics and verify that the autograd forces conserve the\n", "total energy (the paper's energy-conservation-by-construction claim, eq. 4).\n", "The implementation itself is verified against the manuscripts' equations in\n", "`examples/fidelity_checks/schnet_verification.ipynb`." ] } ], "metadata": { "kernelspec": { "display_name": "xnn (.venv)", "language": "python", "name": "xnn" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.13.12" } }, "nbformat": 4, "nbformat_minor": 5 }