diff --git a/examples/quickstart.ipynb b/examples/quickstart.ipynb index 45a8534..c165b05 100644 --- a/examples/quickstart.ipynb +++ b/examples/quickstart.ipynb @@ -2,91 +2,124 @@ "cells": [ { "cell_type": "markdown", - "id": "5881277c", + "id": "c6bfb327", "metadata": {}, "source": [ "# scmidas quickstart\n", "\n", "Five lines from `pip install scmidas` to a UMAP of mosaic-integrated single-cell multi-omic data.\n", "\n", - "This notebook trains MIDAS on a **toy 1600-cell PBMC mosaic dataset** (RNA + ADT, 4 batches with paired and unpaired modality combinations) bundled inside the package. On a free Colab T4 GPU end-to-end takes about 2 minutes; on a local mid-range GPU about 1 minute.\n", + "**On a free Colab T4 GPU** the entire notebook (install + train + UMAP) runs in roughly 2 minutes. **On a local mid-range GPU** the training step alone is ~1 minute.\n", "\n", - "> ⚠️ **The default `scmidas.integrate(...)` hyperparameters are tuned for this toy dataset**, not for your own data. For a real analysis, see [demo2 / demo3](https://scmidas.readthedocs.io/en/latest/) and tune `max_epochs`, `batch_size` accordingly." + "This notebook trains MIDAS on a 1600-cell PBMC RNA+ADT mosaic dataset bundled inside the package — no extra downloads, no config files, no data prep.\n", + "\n", + "> ⚠️ **The defaults in `scmidas.integrate()` are tuned for this toy dataset.** They are *not* general-purpose. For a real analysis, override `max_epochs` (1000-2000) and `batch_size` (256). See [demo2 / demo3](https://scmidas.readthedocs.io/en/latest/) for full pipelines." ] }, { "cell_type": "markdown", - "id": "8792bc14", + "id": "884fe652", "metadata": {}, "source": [ - "## 1. Install (Colab only)" + "## 1. Install\n", + "\n", + "Skipped if `scmidas` is already importable (e.g. local Jupyter). On Colab the `pip install` may print a benign \"dependency resolver\" warning about Colab's pre-pinned `pandas` — this is harmless and doesn't affect the demo." ] }, { "cell_type": "code", "execution_count": 1, - "id": "c46a8c24", + "id": "2652e5a8", "metadata": { "execution": { - "iopub.execute_input": "2026-05-03T05:44:21.109192Z", - "iopub.status.busy": "2026-05-03T05:44:21.108326Z", - "iopub.status.idle": "2026-05-03T05:44:21.110400Z", - "shell.execute_reply": "2026-05-03T05:44:21.109789Z" + "iopub.execute_input": "2026-05-03T07:17:54.442360Z", + "iopub.status.busy": "2026-05-03T07:17:54.441760Z", + "iopub.status.idle": "2026-05-03T07:17:54.445939Z", + "shell.execute_reply": "2026-05-03T07:17:54.446382Z" } }, "outputs": [], "source": [ - "# Uncomment the next line on Colab; skip if scmidas is already installed.\n", - "# !pip install scmidas" + "import importlib.util\n", + "if importlib.util.find_spec('scmidas') is None:\n", + " !pip install --quiet scmidas" ] }, { "cell_type": "markdown", - "id": "023f43b7", + "id": "5d181140", "metadata": {}, "source": [ - "## 2. Load the bundled quickstart dataset\n", + "## 2. Confirm a GPU is available\n", "\n", - "A 1600-cell PBMC mosaic MuData (RNA + ADT). 4 batches with the typical mosaic structure: one RNA-only, one ADT-only, two with both modalities." + "If you see a warning here, the notebook is running on CPU and training will take ~5-10 minutes instead of ~1 minute. **To enable a GPU on Colab**: `Runtime → Change runtime type → Hardware accelerator → T4 GPU`, then `Runtime → Restart and run all`." ] }, { "cell_type": "code", "execution_count": 2, - "id": "ee9385e5", + "id": "ba29b01f", "metadata": { "execution": { - "iopub.execute_input": "2026-05-03T05:44:21.116404Z", - "iopub.status.busy": "2026-05-03T05:44:21.115804Z", - "iopub.status.idle": "2026-05-03T05:44:25.934178Z", - "shell.execute_reply": "2026-05-03T05:44:25.935356Z" + "iopub.execute_input": "2026-05-03T07:17:54.450879Z", + "iopub.status.busy": "2026-05-03T07:17:54.450252Z", + "iopub.status.idle": "2026-05-03T07:17:56.003269Z", + "shell.execute_reply": "2026-05-03T07:17:56.004440Z" } }, "outputs": [ { - "data": { - "text/html": [ - "
MuData object with n_obs × n_vars = 1600 × 724\n",
-       "  obs:\t'batch', 'celltype'\n",
-       "  2 modalities\n",
-       "    rna:\t1200 × 500\n",
-       "      obs:\t'batch', 'celltype'\n",
-       "    adt:\t1200 × 224\n",
-       "      obs:\t'batch', 'celltype'
" - ], - "text/plain": [ - "MuData object with n_obs × n_vars = 1600 × 724\n", - " obs:\t'batch', 'celltype'\n", - " 2 modalities\n", - " rna:\t1200 × 500\n", - " obs:\t'batch', 'celltype'\n", - " adt:\t1200 × 224\n", - " obs:\t'batch', 'celltype'" - ] - }, - "execution_count": 1, - "metadata": {}, - "output_type": "execute_result" + "name": "stdout", + "output_type": "stream", + "text": [ + "✓ GPU detected: Tesla V100S-PCIE-32GB\n" + ] + } + ], + "source": [ + "import torch\n", + "if torch.cuda.is_available():\n", + " print(f'✓ GPU detected: {torch.cuda.get_device_name(0)}')\n", + "else:\n", + " print('⚠️ No GPU detected. Training will fall back to CPU (~5-10 min).')\n", + " print(' Colab: Runtime → Change runtime type → T4 GPU, then Run all.')" + ] + }, + { + "cell_type": "markdown", + "id": "12d309f9", + "metadata": {}, + "source": [ + "## 3. Load and inspect the bundled quickstart dataset\n", + "\n", + "The bundled MuData has 4 batches arranged in the **classic mosaic structure** that MIDAS is built to integrate: one RNA-only batch, one ADT-only batch, and two batches with both modalities." + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "id": "f7a8798e", + "metadata": { + "execution": { + "iopub.execute_input": "2026-05-03T07:17:56.011366Z", + "iopub.status.busy": "2026-05-03T07:17:56.010732Z", + "iopub.status.idle": "2026-05-03T07:17:59.017212Z", + "shell.execute_reply": "2026-05-03T07:17:59.018372Z" + } + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Per-batch breakdown:\n", + " p1_0: 400 cells modalities: RNA \n", + " p2_0: 400 cells modalities: ADT \n", + " p3_0: 400 cells modalities: RNA+ADT\n", + " p4_0: 400 cells modalities: RNA+ADT\n", + "\n", + "Total: 1600 cells across 4 batches | RNA features: 500 | ADT features: 224\n" + ] } ], "source": [ @@ -98,29 +131,37 @@ "import scmidas\n", "\n", "mdata = scmidas.datasets.quickstart()\n", - "mdata" + "\n", + "print('Per-batch breakdown:')\n", + "for b in sorted(set(mdata['rna'].obs['batch']) | set(mdata['adt'].obs['batch'])):\n", + " n_rna = int((mdata['rna'].obs['batch'] == b).sum())\n", + " n_adt = int((mdata['adt'].obs['batch'] == b).sum())\n", + " mods = ('RNA' if n_rna else '') + ('+' if n_rna and n_adt else '') + ('ADT' if n_adt else '')\n", + " n_cells = max(n_rna, n_adt)\n", + " print(f' {b}: {n_cells:>4} cells modalities: {mods:<7}')\n", + "print(f'\\nTotal: {mdata.n_obs} cells across 4 batches | RNA features: {mdata[\"rna\"].n_vars} | ADT features: {mdata[\"adt\"].n_vars}')" ] }, { "cell_type": "markdown", - "id": "602dc63b", + "id": "3407bd94", "metadata": {}, "source": [ - "## 3. Train\n", + "## 4. Train\n", "\n", - "One call. `scmidas.integrate(...)` infers feature dimensions from the MuData and uses defaults that produce a clean UMAP for this toy dataset in about a minute." + "One call. `scmidas.integrate(...)` infers feature dimensions from the MuData and uses defaults that produce a clean UMAP for this toy dataset in about a minute. (The full `MIDAS` class is still available for users who want explicit control.)" ] }, { "cell_type": "code", - "execution_count": 3, - "id": "e91fadfd", + "execution_count": 4, + "id": "f7c86538", "metadata": { "execution": { - "iopub.execute_input": "2026-05-03T05:44:25.940862Z", - "iopub.status.busy": "2026-05-03T05:44:25.940266Z", - "iopub.status.idle": "2026-05-03T05:45:29.365469Z", - "shell.execute_reply": "2026-05-03T05:45:29.364139Z" + "iopub.execute_input": "2026-05-03T07:17:59.023328Z", + "iopub.status.busy": "2026-05-03T07:17:59.022581Z", + "iopub.status.idle": "2026-05-03T07:19:00.766054Z", + "shell.execute_reply": "2026-05-03T07:19:00.767379Z" } }, "outputs": [ @@ -178,7 +219,7 @@ { "data": { "application/vnd.jupyter.widget-view+json": { - "model_id": "30186375b1a0421e92b5cf6e4c5dfda4", + "model_id": "7118de400bb1442eafa5752cc6c5bbcb", "version_major": 2, "version_minor": 0 }, @@ -195,7 +236,7 @@ "text": [ "INFO: `Trainer.fit` stopped: `max_epochs=65` reached.\n", "INFO:lightning.pytorch.utilities.rank_zero:`Trainer.fit` stopped: `max_epochs=65` reached.\n", - "INFO:scmidas.model:Checkpoint successfully saved to \"./saved_models/scmidas/model_epoch65_20260503-054529.pt\".\n" + "INFO:scmidas.model:Checkpoint successfully saved to \"./saved_models/scmidas/model_epoch65_20260503-071900.pt\".\n" ] } ], @@ -205,22 +246,24 @@ }, { "cell_type": "markdown", - "id": "2550404a", + "id": "ca3dc944", "metadata": {}, "source": [ - "## 4. Compute latent embeddings" + "## 5. Compute joint latent embeddings\n", + "\n", + "Stitch the per-batch latent blocks into one (cells × 32) matrix aligned with the MuData." ] }, { "cell_type": "code", - "execution_count": 4, - "id": "90864802", + "execution_count": 5, + "id": "ab5b338c", "metadata": { "execution": { - "iopub.execute_input": "2026-05-03T05:45:29.373813Z", - "iopub.status.busy": "2026-05-03T05:45:29.373168Z", - "iopub.status.idle": "2026-05-03T05:45:31.692584Z", - "shell.execute_reply": "2026-05-03T05:45:31.693556Z" + "iopub.execute_input": "2026-05-03T07:19:00.775726Z", + "iopub.status.busy": "2026-05-03T07:19:00.774939Z", + "iopub.status.idle": "2026-05-03T07:19:03.418845Z", + "shell.execute_reply": "2026-05-03T07:19:03.419798Z" } }, "outputs": [ @@ -230,13 +273,13 @@ "text": [ "INFO:scmidas.model:Predicting using device: cuda\n", "INFO:scmidas.model:Processing batch p1_0: ['rna']\n", - "predict:p1_0: 100%|██████████| 4/4 [00:00<00:00, 7.04it/s]\n", + "predict:p1_0: 100%|██████████| 4/4 [00:00<00:00, 6.48it/s]\n", "INFO:scmidas.model:Processing batch p2_0: ['adt']\n", - "predict:p2_0: 100%|██████████| 4/4 [00:00<00:00, 6.92it/s]\n", + "predict:p2_0: 100%|██████████| 4/4 [00:00<00:00, 6.68it/s]\n", "INFO:scmidas.model:Processing batch p3_0: ['rna', 'adt']\n", - "predict:p3_0: 100%|██████████| 4/4 [00:00<00:00, 7.30it/s]\n", + "predict:p3_0: 100%|██████████| 4/4 [00:00<00:00, 5.18it/s]\n", "INFO:scmidas.model:Processing batch p4_0: ['rna', 'adt']\n", - "predict:p4_0: 100%|██████████| 4/4 [00:00<00:00, 7.04it/s]" + "predict:p4_0: 100%|██████████| 4/4 [00:00<00:00, 6.58it/s]" ] }, { @@ -259,7 +302,6 @@ "\n", "out = model.predict(joint_latent=True)\n", "\n", - "# stitch per-batch z_c blocks into one (cells × 32) matrix aligned with mdata\n", "z_c_blocks, all_ids = [], []\n", "for b in ['p1_0', 'p2_0', 'p3_0', 'p4_0']:\n", " z_c_blocks.append(out[b]['z_c']['joint'])\n", @@ -270,29 +312,29 @@ " all_ids.extend(rna_ids + [c for c in adt_ids if c not in set(rna_ids)])\n", "z_c = np.vstack(z_c_blocks)\n", "\n", - "print('latent shape:', z_c.shape)" + "print(f'latent shape: {z_c.shape}')" ] }, { "cell_type": "markdown", - "id": "398454cf", + "id": "df5c3594", "metadata": {}, "source": [ - "## 5. UMAP\n", + "## 6. UMAP\n", "\n", - "Color by cell type (left) and source batch (right). MIDAS should produce **lineage-separated clusters** that **mix freely across batches** — the central promise of mosaic integration." + "Color by cell type (left) and source batch (right). MIDAS should produce **lineage-separated clusters** (Mono / CD4 T / CD8 T / NK / B / DC) that **mix freely across batches** — the central promise of mosaic integration." ] }, { "cell_type": "code", - "execution_count": 5, - "id": "02880ec2", + "execution_count": 6, + "id": "c82e6023", "metadata": { "execution": { - "iopub.execute_input": "2026-05-03T05:45:31.748236Z", - "iopub.status.busy": "2026-05-03T05:45:31.733683Z", - "iopub.status.idle": "2026-05-03T05:45:39.671502Z", - "shell.execute_reply": "2026-05-03T05:45:39.672634Z" + "iopub.execute_input": "2026-05-03T07:19:03.476071Z", + "iopub.status.busy": "2026-05-03T07:19:03.466191Z", + "iopub.status.idle": "2026-05-03T07:19:12.361318Z", + "shell.execute_reply": "2026-05-03T07:19:12.362617Z" } }, "outputs": [ @@ -345,17 +387,22 @@ }, { "cell_type": "markdown", - "id": "be94ec90", + "id": "0e2f9dc1", "metadata": {}, "source": [ "## Next steps\n", "\n", - "- See `docs/source/tutorials/basics/` for the full demos (with real-sized data, multi-GPU training, prediction tasks like imputation, batch correction, modality translation).\n", + "- See `docs/source/tutorials/basics/` for the full demos (real-sized data, multi-GPU training, prediction tasks like imputation, batch correction, modality translation).\n", "- For your own MuData: `scmidas.integrate(your_mdata, max_epochs=2000, batch_size=256)` (paper-default hyperparameters)." ] } ], "metadata": { + "accelerator": "GPU", + "colab": { + "gpuType": "T4", + "provenance": [] + }, "kernelspec": { "display_name": "Python 3", "language": "python", @@ -376,31 +423,25 @@ "widgets": { "application/vnd.jupyter.widget-state+json": { "state": { - "30186375b1a0421e92b5cf6e4c5dfda4": { + "1a800a568376411ea0d807398580bb58": { "model_module": "@jupyter-widgets/controls", "model_module_version": "2.0.0", - "model_name": "HBoxModel", + "model_name": "HTMLStyleModel", "state": { - "_dom_classes": [], "_model_module": "@jupyter-widgets/controls", "_model_module_version": "2.0.0", - "_model_name": "HBoxModel", + "_model_name": "HTMLStyleModel", "_view_count": null, - "_view_module": "@jupyter-widgets/controls", + "_view_module": "@jupyter-widgets/base", "_view_module_version": "2.0.0", - "_view_name": "HBoxView", - "box_style": "", - "children": [ - "IPY_MODEL_dd59645be4d9462ea87eeb3280ee9463", - "IPY_MODEL_c04fd7c0d62f4db1ac61bce7e3fc8987", - "IPY_MODEL_90c165cf45a04c5292db0b7eea9db213" - ], - "layout": "IPY_MODEL_4ee8a0e03b0c4957a25a128c2d50f5d4", - "tabbable": null, - "tooltip": null + "_view_name": "StyleView", + "background": null, + "description_width": "", + "font_size": null, + "text_color": null } }, - "39ac3c9b2c9248e58188bde88e5d542a": { + "3a9fe25d23e24cbeb45c0c92a582182a": { "model_module": "@jupyter-widgets/base", "model_module_version": "2.0.0", "model_name": "LayoutModel", @@ -453,7 +494,80 @@ "width": null } }, - "4ee8a0e03b0c4957a25a128c2d50f5d4": { + "4e3b240aefe841f8a0248084158f993d": { + "model_module": "@jupyter-widgets/controls", + "model_module_version": "2.0.0", + "model_name": "FloatProgressModel", + "state": { + "_dom_classes": [], + "_model_module": "@jupyter-widgets/controls", + "_model_module_version": "2.0.0", + "_model_name": "FloatProgressModel", + "_view_count": null, + "_view_module": "@jupyter-widgets/controls", + "_view_module_version": "2.0.0", + "_view_name": "ProgressView", + "bar_style": "success", + "description": "", + "description_allow_html": false, + "layout": "IPY_MODEL_3a9fe25d23e24cbeb45c0c92a582182a", + "max": 16.0, + "min": 0.0, + "orientation": "horizontal", + "style": "IPY_MODEL_b4b562fe298d47f3b4d56e3846f5eacb", + "tabbable": null, + "tooltip": null, + "value": 16.0 + } + }, + "7118de400bb1442eafa5752cc6c5bbcb": { + "model_module": "@jupyter-widgets/controls", + "model_module_version": "2.0.0", + "model_name": "HBoxModel", + "state": { + "_dom_classes": [], + "_model_module": "@jupyter-widgets/controls", + "_model_module_version": "2.0.0", + "_model_name": "HBoxModel", + "_view_count": null, + "_view_module": "@jupyter-widgets/controls", + "_view_module_version": "2.0.0", + "_view_name": "HBoxView", + "box_style": "", + "children": [ + "IPY_MODEL_7329104c37c542b293ccfd0611046139", + "IPY_MODEL_4e3b240aefe841f8a0248084158f993d", + "IPY_MODEL_80513bd1cec3464e8cadbfa72c52df33" + ], + "layout": "IPY_MODEL_7426bd7df0a441728b2b306dd4fce118", + "tabbable": null, + "tooltip": null + } + }, + "7329104c37c542b293ccfd0611046139": { + "model_module": "@jupyter-widgets/controls", + "model_module_version": "2.0.0", + "model_name": "HTMLModel", + "state": { + "_dom_classes": [], + "_model_module": "@jupyter-widgets/controls", + "_model_module_version": "2.0.0", + "_model_name": "HTMLModel", + "_view_count": null, + "_view_module": "@jupyter-widgets/controls", + "_view_module_version": "2.0.0", + "_view_name": "HTMLView", + "description": "", + "description_allow_html": false, + "layout": "IPY_MODEL_a1a9045e48c84d919c015d37188c6284", + "placeholder": "​", + "style": "IPY_MODEL_1a800a568376411ea0d807398580bb58", + "tabbable": null, + "tooltip": null, + "value": "Epoch 64: 100%" + } + }, + "7426bd7df0a441728b2b306dd4fce118": { "model_module": "@jupyter-widgets/base", "model_module_version": "2.0.0", "model_name": "LayoutModel", @@ -506,23 +620,30 @@ "width": "100%" } }, - "54721dddcd6445aab2ee8a974bc36b54": { + "80513bd1cec3464e8cadbfa72c52df33": { "model_module": "@jupyter-widgets/controls", "model_module_version": "2.0.0", - "model_name": "ProgressStyleModel", + "model_name": "HTMLModel", "state": { + "_dom_classes": [], "_model_module": "@jupyter-widgets/controls", "_model_module_version": "2.0.0", - "_model_name": "ProgressStyleModel", + "_model_name": "HTMLModel", "_view_count": null, - "_view_module": "@jupyter-widgets/base", + "_view_module": "@jupyter-widgets/controls", "_view_module_version": "2.0.0", - "_view_name": "StyleView", - "bar_color": null, - "description_width": "" + "_view_name": "HTMLView", + "description": "", + "description_allow_html": false, + "layout": "IPY_MODEL_f15f94f2c30b49c19984b19e5fa9f386", + "placeholder": "​", + "style": "IPY_MODEL_e104da1ca6b84b08a2ab6dd117788448", + "tabbable": null, + "tooltip": null, + "value": " 16/16 [00:00<00:00, 16.88it/s, v_num=1, loss_/recon_loss_step=414.0, loss_/kld_loss_step=61.10, loss_/consistency_loss_step=0.000, loss/net_step=411.0, loss/dsc_step=63.20, loss_/recon_loss_epoch=1.51e+3, loss_/kld_loss_epoch=79.20, loss_/consistency_loss_epoch=58.20, loss/net_epoch=1.57e+3, loss/dsc_epoch=78.00]" } }, - "737974cb592f48dd96dc726efe5b61c6": { + "a1a9045e48c84d919c015d37188c6284": { "model_module": "@jupyter-widgets/base", "model_module_version": "2.0.0", "model_name": "LayoutModel", @@ -575,74 +696,23 @@ "width": null } }, - "90c165cf45a04c5292db0b7eea9db213": { - "model_module": "@jupyter-widgets/controls", - "model_module_version": "2.0.0", - "model_name": "HTMLModel", - "state": { - "_dom_classes": [], - "_model_module": "@jupyter-widgets/controls", - "_model_module_version": "2.0.0", - "_model_name": "HTMLModel", - "_view_count": null, - "_view_module": "@jupyter-widgets/controls", - "_view_module_version": "2.0.0", - "_view_name": "HTMLView", - "description": "", - "description_allow_html": false, - "layout": "IPY_MODEL_e288a669839d41f39581cb43cdc38b85", - "placeholder": "​", - "style": "IPY_MODEL_bbab4f6701f74e20b55d21f3119e0bc2", - "tabbable": null, - "tooltip": null, - "value": " 16/16 [00:00<00:00, 17.14it/s, v_num=0, loss_/recon_loss_step=414.0, loss_/kld_loss_step=61.10, loss_/consistency_loss_step=0.000, loss/net_step=411.0, loss/dsc_step=63.20, loss_/recon_loss_epoch=1.51e+3, loss_/kld_loss_epoch=79.20, loss_/consistency_loss_epoch=58.20, loss/net_epoch=1.57e+3, loss/dsc_epoch=78.00]" - } - }, - "bbab4f6701f74e20b55d21f3119e0bc2": { + "b4b562fe298d47f3b4d56e3846f5eacb": { "model_module": "@jupyter-widgets/controls", "model_module_version": "2.0.0", - "model_name": "HTMLStyleModel", + "model_name": "ProgressStyleModel", "state": { "_model_module": "@jupyter-widgets/controls", "_model_module_version": "2.0.0", - "_model_name": "HTMLStyleModel", + "_model_name": "ProgressStyleModel", "_view_count": null, "_view_module": "@jupyter-widgets/base", "_view_module_version": "2.0.0", "_view_name": "StyleView", - "background": null, - "description_width": "", - "font_size": null, - "text_color": null - } - }, - "c04fd7c0d62f4db1ac61bce7e3fc8987": { - "model_module": "@jupyter-widgets/controls", - "model_module_version": "2.0.0", - "model_name": "FloatProgressModel", - "state": { - "_dom_classes": [], - "_model_module": "@jupyter-widgets/controls", - "_model_module_version": "2.0.0", - "_model_name": "FloatProgressModel", - "_view_count": null, - "_view_module": "@jupyter-widgets/controls", - "_view_module_version": "2.0.0", - "_view_name": "ProgressView", - "bar_style": "success", - "description": "", - "description_allow_html": false, - "layout": "IPY_MODEL_39ac3c9b2c9248e58188bde88e5d542a", - "max": 16.0, - "min": 0.0, - "orientation": "horizontal", - "style": "IPY_MODEL_54721dddcd6445aab2ee8a974bc36b54", - "tabbable": null, - "tooltip": null, - "value": 16.0 + "bar_color": null, + "description_width": "" } }, - "c2d1dd3030f7426bb2485c04c3d7a500": { + "e104da1ca6b84b08a2ab6dd117788448": { "model_module": "@jupyter-widgets/controls", "model_module_version": "2.0.0", "model_name": "HTMLStyleModel", @@ -660,30 +730,7 @@ "text_color": null } }, - "dd59645be4d9462ea87eeb3280ee9463": { - "model_module": "@jupyter-widgets/controls", - "model_module_version": "2.0.0", - "model_name": "HTMLModel", - "state": { - "_dom_classes": [], - "_model_module": "@jupyter-widgets/controls", - "_model_module_version": "2.0.0", - "_model_name": "HTMLModel", - "_view_count": null, - "_view_module": "@jupyter-widgets/controls", - "_view_module_version": "2.0.0", - "_view_name": "HTMLView", - "description": "", - "description_allow_html": false, - "layout": "IPY_MODEL_737974cb592f48dd96dc726efe5b61c6", - "placeholder": "​", - "style": "IPY_MODEL_c2d1dd3030f7426bb2485c04c3d7a500", - "tabbable": null, - "tooltip": null, - "value": "Epoch 64: 100%" - } - }, - "e288a669839d41f39581cb43cdc38b85": { + "f15f94f2c30b49c19984b19e5fa9f386": { "model_module": "@jupyter-widgets/base", "model_module_version": "2.0.0", "model_name": "LayoutModel",