From 97dc77aacf4afb33e13b0bdbe3f79128bd32078f Mon Sep 17 00:00:00 2001
From: Arthur BATEL <arthur.batel@insa-lyon.fr>
Date: Thu, 15 Feb 2024 17:38:05 +0100
Subject: [PATCH] Explainations
---
Experiments.ipynb | 97 ++++++++++++++++++++++-------------------------
1 file changed, 45 insertions(+), 52 deletions(-)
diff --git a/Experiments.ipynb b/Experiments.ipynb
index 37bf511..8e52a6f 100644
--- a/Experiments.ipynb
+++ b/Experiments.ipynb
@@ -22,9 +22,9 @@
"source": [
"## Table 2 and 3: compute ACC, AUC and RMSE and DOA\n",
"\n",
- "Warning : can be long to compute\n",
- "Prerequisite : Running the makefile\n",
- "Output : stored in `../../results/table_2`"
+ "**Warning** : can be long to compute\n",
+ "**Prerequisite** : Running the makefile\n",
+ "**Output** : stored in `../../results/table_2`"
],
"metadata": {
"collapsed": false
@@ -48,7 +48,15 @@
{
"cell_type": "markdown",
"source": [
- "## Table 3: compute DOA"
+ "## Table 3: compute DOA\n",
+ "\n",
+ "**Warning** : can be long to compute\n",
+ "**Prerequisite** : Running the makefile\n",
+ "**Output** : printed in the notebook\n",
+ "\n",
+ "The doa metrics are computed for the dataset whose index is specified by the variable `i`, for CD-BPR and for all competitors simultaneously, and for each data splitting of the cross validation.\n",
+ "\n",
+ "The embeddings are given pre-computed in the `result` directory. However, for the sake of reproducibility, they will be overwritten by the ones computed in the previous cell. "
],
"metadata": {
"collapsed": false
@@ -62,35 +70,31 @@
"name": "stdout",
"output_type": "stream",
"text": [
- "_create_directory() - handled error when creating a directory at ../experiments_logs/3: [Errno 17] File exists: '../experiments_logs/3'\r\n",
"dataset math_1\r\n",
"fold 0\r\n",
"model DINA\r\n",
+ "num_kc 11\r\n",
+ "model MCD\r\n",
+ "num_kc 11\r\n",
+ "model NCDM\r\n",
+ "num_kc 11\r\n",
+ "model MIRT\r\n",
+ "num_kc 11\r\n",
+ "model BPR\r\n",
+ "num_kc 11\r\n",
+ "fold 1\r\n",
+ "model DINA\r\n",
+ "num_kc 11\r\n",
+ "model MCD\r\n",
+ "num_kc 11\r\n",
+ "model NCDM\r\n",
+ "num_kc 11\r\n",
"^C\r\n",
"Traceback (most recent call last):\r\n",
- " File \"/home/arthurb/Programmation/cd-bpr/code/./DOA.py\", line 161, in <module>\r\n",
- " exp.train_test_split(dataset_name, test_proportion=0.2, valid_proportion=0.2, n_folds=5, i_fold=i_fold)\r\n",
- " File \"/home/arthurb/Programmation/cd-bpr/code/utility/DatasetProcessor.py\", line 938, in train_test_split\r\n",
- " self.test_steps[dataset_name] = self.get_test_dataset(dataset_name=dataset_name)['train_valid_log_nb'].to_numpy().astype(int)\r\n",
- " ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/Programmation/cd-bpr/code/utility/DatasetProcessor.py\", line 1416, in get_test_dataset\r\n",
- " return self.datasets[dataset_name][self.datasets[dataset_name]['split']=='test']\r\n",
- " ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/pandas/core/ops/common.py\", line 76, in new_method\r\n",
- " return method(self, other)\r\n",
- " ^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/pandas/core/arraylike.py\", line 40, in __eq__\r\n",
- " return self._cmp_method(other, operator.eq)\r\n",
- " ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/pandas/core/series.py\", line 5803, in _cmp_method\r\n",
- " res_values = ops.comparison_op(lvalues, rvalues, op)\r\n",
- " ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/pandas/core/ops/array_ops.py\", line 346, in comparison_op\r\n",
- " res_values = comp_method_OBJECT_ARRAY(op, lvalues, rvalues)\r\n",
- " ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/pandas/core/ops/array_ops.py\", line 131, in comp_method_OBJECT_ARRAY\r\n",
- " result = libops.scalar_compare(x.ravel(), y, op)\r\n",
- " ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n",
+ " File \"/home/arthurb/Programmation/cd-bpr/code/./DOA.py\", line 185, in <module>\r\n",
+ " r_train = dao_creuse(F, kc_user, kc_user_val, dico_u)\r\n",
+ " ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n",
+ " File \"/home/arthurb/Programmation/cd-bpr/code/./DOA.py\", line -1, in dao_creuse\r\n",
"KeyboardInterrupt\r\n"
]
}
@@ -107,12 +111,12 @@
"metadata": {
"collapsed": false,
"ExecuteTime": {
- "end_time": "2024-02-15T15:25:00.814018170Z",
- "start_time": "2024-02-15T15:24:53.134103941Z"
+ "end_time": "2024-02-15T16:21:51.656146352Z",
+ "start_time": "2024-02-15T15:49:46.331839222Z"
}
},
"id": "68a2dc1287b5cac",
- "execution_count": 7
+ "execution_count": 18
},
{
"cell_type": "markdown",
@@ -124,12 +128,12 @@
},
{
"cell_type": "code",
- "execution_count": 3,
+ "execution_count": 19,
"id": "a9e32954",
"metadata": {
"ExecuteTime": {
- "end_time": "2024-02-15T15:24:35.092972342Z",
- "start_time": "2024-02-15T15:24:06.236900058Z"
+ "end_time": "2024-02-15T16:22:26.443494416Z",
+ "start_time": "2024-02-15T16:22:01.974704182Z"
}
},
"outputs": [
@@ -140,26 +144,15 @@
"Ablation (0 no ablation, 1 ablation L2, 2 ablation init, 3 both) 0\r\n",
"assist0910_tkde\r\n",
"NB (question-answer) in the data 17283\r\n",
- "Epoch 1, Loss: 1.37166428565979\r\n",
+ "Epoch 1, Loss: 1.3824448585510254\r\n",
"^C\r\n",
"Traceback (most recent call last):\r\n",
" File \"/home/arthurb/Programmation/cd-bpr/code/binary_bpr_ablation/main.py\", line 219, in <module>\r\n",
" acc = bpr_model.train(train, len(dico_kc), epochs, batch_size, y_train, ablation)\r\n",
" ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/Programmation/cd-bpr/code/binary_bpr_ablation/BPR_model.py\", line 105, in train\r\n",
- " self.optimizer.step()\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/torch/optim/optimizer.py\", line 385, in wrapper\r\n",
- " out = func(*args, **kwargs)\r\n",
- " ^^^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/torch/optim/optimizer.py\", line 76, in _use_grad\r\n",
- " ret = func(self, *args, **kwargs)\r\n",
- " ^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/torch/optim/adam.py\", line 166, in step\r\n",
- " adam(\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/torch/optim/adam.py\", line 316, in adam\r\n",
- " func(params,\r\n",
- " File \"/home/arthurb/anaconda3/envs/cdbpr-env/lib/python3.11/site-packages/torch/optim/adam.py\", line 391, in _single_tensor_adam\r\n",
- " exp_avg.lerp_(grad, 1 - beta1)\r\n",
+ " File \"/home/arthurb/Programmation/cd-bpr/code/binary_bpr_ablation/BPR_model.py\", line 99, in train\r\n",
+ " comp = comp.cpu()\r\n",
+ " ^^^^^^^^^^\r\n",
"KeyboardInterrupt\r\n"
]
}
@@ -180,12 +173,12 @@
},
{
"cell_type": "code",
- "execution_count": 8,
+ "execution_count": 20,
"id": "e46ff5d4",
"metadata": {
"ExecuteTime": {
- "end_time": "2024-02-15T15:25:07.559037819Z",
- "start_time": "2024-02-15T15:25:07.001579876Z"
+ "end_time": "2024-02-15T16:22:29.824107304Z",
+ "start_time": "2024-02-15T16:22:29.286318951Z"
}
},
"outputs": [
--
GitLab