02_finetuning_data_preprocessing.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "3e2aed51",
   "metadata": {},
   "source": [
    "# Encode the binary scores with different encoding schemes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "0072b0a9",
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "sys.path.append('../../source')\n",
    "\n",
    "from utils import *\n",
    "\n",
    "import pickle\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from tqdm import tqdm\n",
    "\n",
    "import encoders"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e0521d93",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"/home/abunn/ttmp/fragment_dataset.pkl\", \"rb\") as f:\n",
    "    x_train, y_train, x_valid, y_valid, x_test, y_test = pickle.load(f)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "9bfcb9db",
   "metadata": {},
   "source": [
    "### Sparse encodings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "14d693f0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "char_encoder\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/27999 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 27999/27999 [00:30<00:00, 904.72it/s]\n",
      "100%|██████████| 6003/6003 [00:06<00:00, 904.19it/s]\n",
      "100%|██████████| 6003/6003 [00:06<00:00, 905.10it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chroma_octave_encoder\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 27999/27999 [00:30<00:00, 906.86it/s]\n",
      "100%|██████████| 6003/6003 [00:06<00:00, 905.09it/s]\n",
      "100%|██████████| 6003/6003 [00:06<00:00, 903.18it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "interval_encoder\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 27999/27999 [00:33<00:00, 839.18it/s]\n",
      "100%|██████████| 6003/6003 [00:07<00:00, 841.63it/s]\n",
      "100%|██████████| 6003/6003 [00:07<00:00, 844.40it/s]\n"
     ]
    }
   ],
   "source": [
    "composers = np.unique(y_train).tolist()\n",
    "\n",
    "for enc_name, enc in encoders.sparse_encoders.items():\n",
    "    print(enc_name)\n",
    "    x_train_encoded = []\n",
    "    y_train_encoded = []\n",
    "    for x, y in tqdm(zip(x_train, y_train), total=len(x_train)):\n",
    "        x_train_encoded.append(enc(x))\n",
    "        y_train_encoded.append(composers.index(y))\n",
    "        \n",
    "    x_valid_encoded = []\n",
    "    y_valid_encoded = []\n",
    "    for x, y in tqdm(zip(x_valid, y_valid), total=len(x_valid)):\n",
    "        x_valid_encoded.append(enc(x))\n",
    "        y_valid_encoded.append(composers.index(y))\n",
    "        \n",
    "    x_test_encoded = []\n",
    "    y_test_encoded = []\n",
    "    for x, y in tqdm(zip(x_test, y_test), total=len(x_test)):\n",
    "        x_test_encoded.append(enc(x))\n",
    "        y_test_encoded.append(composers.index(y))\n",
    "    \n",
    "    # Data for classification finetuning\n",
    "    with open(f\"/home/abunn/ttmp/ExplorationWithLLMs/source/03_alec_temp_name/finetuning_data/{enc_name}.pkl\", \"wb\") as f:\n",
    "        pickle.dump((x_train_encoded, y_train_encoded, x_valid_encoded, y_valid_encoded, x_test_encoded, y_test_encoded), f)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "6282a145",
   "metadata": {},
   "source": [
    "### Dense encodings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "ccbde8d0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1, 1]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/27999 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 27999/27999 [01:02<00:00, 450.04it/s]\n",
      "100%|██████████| 6003/6003 [00:12<00:00, 465.54it/s]\n",
      "100%|██████████| 6003/6003 [00:08<00:00, 675.44it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1, 2]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 27999/27999 [00:37<00:00, 750.57it/s]\n",
      "100%|██████████| 6003/6003 [00:06<00:00, 888.39it/s]\n",
      "100%|██████████| 6003/6003 [00:06<00:00, 904.59it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1, 4]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 27999/27999 [00:25<00:00, 1116.94it/s]\n",
      "100%|██████████| 6003/6003 [00:05<00:00, 1130.62it/s]\n",
      "100%|██████████| 6003/6003 [00:05<00:00, 1115.71it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1, 8]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 27999/27999 [00:19<00:00, 1472.36it/s]\n",
      "100%|██████████| 6003/6003 [00:04<00:00, 1445.78it/s]\n",
      "100%|██████████| 6003/6003 [00:04<00:00, 1461.83it/s]\n"
     ]
    }
   ],
   "source": [
    "composers = np.unique(y_train).tolist()\n",
    "\n",
    "block_sizes = [\n",
    "    [1, 1],\n",
    "    [1, 2],\n",
    "    [1, 4],\n",
    "    [1, 8],\n",
    "]\n",
    "\n",
    "for block_size in block_sizes:\n",
    "    print(block_size)\n",
    "    x_train_encoded = []\n",
    "    y_train_encoded = []\n",
    "    for x, y in tqdm(zip(x_train, y_train), total=len(x_train)):\n",
    "        x_train_encoded.append(encoders.dense_encoder(x, block_size=block_size))\n",
    "        y_train_encoded.append(composers.index(y))\n",
    "        \n",
    "    x_valid_encoded = []\n",
    "    y_valid_encoded = []\n",
    "    for x, y in tqdm(zip(x_valid, y_valid), total=len(x_valid)):\n",
    "        x_valid_encoded.append(encoders.dense_encoder(x, block_size=block_size))\n",
    "        y_valid_encoded.append(composers.index(y))\n",
    "    \n",
    "    x_test_encoded = []\n",
    "    y_test_encoded = []\n",
    "    for x, y in tqdm(zip(x_test, y_test), total=len(x_test)):\n",
    "        x_test_encoded.append(encoders.dense_encoder(x, block_size=block_size))\n",
    "        y_test_encoded.append(composers.index(y))\n",
    "    \n",
    "    # Data for classification finetuning\n",
    "    with open(f\"/home/abunn/ttmp/ExplorationWithLLMs/source/03_alec_temp_name/finetuning_data/dense_{block_size[0]}_{block_size[1]}.pkl\", \"wb\") as f:\n",
    "        pickle.dump((x_train_encoded, y_train_encoded, x_valid_encoded, y_valid_encoded, x_test_encoded, y_test_encoded), f)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "554e1f5d",
   "metadata": {},
   "source": [
    "# Get an example from each encoding scheme"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "3efade06",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x1000 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "visualize_bootleg_score(merge_staff_overlaps(x_train[0]).T, lines=[13, 15, 17, 19, 21, 25, 27, 29, 31, 33])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "f7efe8aa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SPARSE\n",
      "\n",
      "Encoder: char_encoder\n",
      "\n",
      "('f | Se | V | Tf | U | Qe | N | J | TXcdf | e | l | Xg | i | Yh | Wf | Uf | h | Xg | Se | Tc | f | U | R | N | UWf | i | h | i | j | k | j | k | l | UWm | h | i | j | k | j | k | l | UWm | Yd | Rc | U | R | N | Ubd | i | l | S | i | Th | Rm | Pf | h | Sg | Ni | Oi | ej | P | M | I | QSUXb', 3)\n",
      "\n",
      "Encoder: chroma_octave_encoder\n",
      "\n",
      "('D5 | E3-C5 | A4 | F3-D5 | G3 | C3-C5 | G2 | C2 | F3-C4-A5-B5-D5 | C5 | C6 | C4-E5 | G5 | D4-F5 | B4-D5 | G3-D5 | F5 | C4-E5 | E3-C5 | F3-A5 | D5 | G3 | D3 | G2 | G3-B4-D5 | G5 | F5 | G5 | A6 | B6 | A6 | B6 | C6 | G3-B4-D6 | F5 | G5 | A6 | B6 | A6 | B6 | C6 | G3-B4-D6 | D4-B5 | D3-A5 | G3 | D3 | G2 | G3-G4-B5 | G5 | C6 | E3 | G5 | F3-F5 | D3-D6 | B3-D5 | F5 | E3-E5 | G2-G5 | A3-G5 | C5-A6 | B3 | F2 | B2 | C3-E3-G3-C4-G4', 3)\n",
      "\n",
      "Encoder: interval_encoder\n",
      "\n",
      "('D5 | E3-12 | A4 | F3-12 | G3 | C3-14 | G2 | C2 | F3-4-9-10-12 | C5 | C6 | C4-9 | G5 | D4-9 | B4-9 | G3-11 | F5 | C4-9 | E3-12 | F3-9 | D5 | G3 | D3 | G2 | G3-2-11 | G5 | F5 | G5 | A6 | B6 | A6 | B6 | C6 | G3-2-18 | F5 | G5 | A6 | B6 | A6 | B6 | C6 | G3-2-18 | D4-5 | D3-11 | G3 | D3 | G2 | G3-7-9 | G5 | C6 | E3 | G5 | F3-14 | D3-21 | B3-16 | F5 | E3-14 | G2-21 | A3-20 | C5-5 | B3 | F2 | B2 | C3-2-4-7-11', 3)\n",
      "\n",
      "DENSE\n",
      "\n",
      "Block size: [1, 1]\n",
      "\n",
      "('❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❍❌❌❌❌❍❍❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❍❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❍❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❍❌❍❌❌❍❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌', 3)\n",
      "\n",
      "Block size: [1, 2]\n",
      "\n",
      "('❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❎❌❎❌❌❏❎❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❍❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❎❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❎❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❍❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❎❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❍❍❌❌❌❎❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❍❍❌❌❌❌❌❌❌❍❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❍❍❌❌❌❌❌❌❌❍❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❍❌❎❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❎❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❍❌❌❎❎❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌ ❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❍❌❎❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❎❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❍❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❍❍❍❎❌❎❌❌❌❌❌❌❌❌❌❌❌❌', 3)\n",
      "\n",
      "Block size: [1, 4]\n",
      "\n",
      "('❌❌❌❌❌❌❌❔❌❌❌❌❌ ❌❌❌❌❐❌❌❐❌❌❌❌❌ ❌❌❌❌❌❎❌❌❌❌❌❌❌ ❌❌❌❌❔❌❌❔❌❌❌❌❌ ❌❌❌❌❌❍❌❌❌❌❌❌❌ ❌❌❌❌❍❌❌❐❌❌❌❌❌ ❌❌❌❎❌❌❌❌❌❌❌❌❌ ❌❌❎❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❔❔❌❗❌❌❌❌❌ ❌❌❌❌❌❌❌❐❌❌❌❌❌ ❌❌❌❌❌❌❌❌❌❎❌❌❌ ❌❌❌❌❌❔❌❌❍❌❌❌❌ ❌❌❌❌❌❌❌❌❐❌❌❌❌ ❌❌❌❌❌❌❍❌❎❌❌❌❌ ❌❌❌❌❌❐❌❔❌❌❌❌❌ ❌❌❌❌❌❍❌❔❌❌❌❌❌ ❌❌❌❌❌❌❌❌❎❌❌❌❌ ❌❌❌❌❌❔❌❌❍❌❌❌❌ ❌❌❌❌❐❌❌❐❌❌❌❌❌ ❌❌❌❌❔❌❌❍❌❌❌❌❌ ❌❌❌❌❌❌❌❔❌❌❌❌❌ ❌❌❌❌❌❍❌❌❌❌❌❌❌ ❌❌❌❌❎❌❌❌❌❌❌❌❌ ❌❌❌❎❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❑❌❔❌❌❌❌❌ ❌❌❌❌❌❌❌❌❐❌❌❌❌ ❌❌❌❌❌❌❌❌❎❌❌❌❌ ❌❌❌❌❌❌❌❌❐❌❌❌❌ ❌❌❌❌❌❌❌❌❔❌❌❌❌ ❌❌❌❌❌❌❌❌❌❍❌❌❌ ❌❌❌❌❌❌❌❌❔❌❌❌❌ ❌❌❌❌❌❌❌❌❌❍❌❌❌ ❌❌❌❌❌❌❌❌❌❎❌❌❌ ❌❌❌❌❌❑❌❌❌❐❌❌❌ ❌❌❌❌❌❌❌❌❎❌❌❌❌ ❌❌❌❌❌❌❌❌❐❌❌❌❌ ❌❌❌❌❌❌❌❌❔❌❌❌❌ ❌❌❌❌❌❌❌❌❌❍❌❌❌ ❌❌❌❌❌❌❌❌❔❌❌❌❌ ❌❌❌❌❌❌❌❌❌❍❌❌❌ ❌❌❌❌❌❌❌❌❌❎❌❌❌ ❌❌❌❌❌❑❌❌❌❐❌❌❌ ❌❌❌❌❌❌❍❎❌❌❌❌❌ ❌❌❌❌❎❌❌❍❌❌❌❌❌ ❌❌❌❌❌❍❌❌❌❌❌❌❌ ❌❌❌❌❎❌❌❌❌❌❌❌❌ ❌❌❌❎❌❌❌❌❌❌❌❌❌ ❌❌❌❌❌❍❔❎❌❌❌❌❌ ❌❌❌❌❌❌❌❌❐❌❌❌❌ ❌❌❌❌❌❌❌❌❌❎❌❌❌ ❌❌❌❌❐❌❌❌❌❌❌❌❌ ❌❌❌❌❌❌❌❌❐❌❌❌❌ ❌❌❌❌❔❌❌❌❎❌❌❌❌ ❌❌❌❌❎❌❌❌❌❐❌❌❌ ❌❌❌❔❌❌❌❔❌❌❌❌❌ ❌❌❌❌❌❌❌❌❎❌❌❌❌ ❌❌❌❌❐❌❌❌❍❌❌❌❌ ❌❌❌❎❌❌❌❌❐❌❌❌❌ ❌❌❌❐❌❌❌❌❐❌❌❌❌ ❌❌❌❌❌❌❌❐❔❌❌❌❌ ❌❌❌❔❌❌❌❌❌❌❌❌❌ ❌❌❌❍❌❌❌❌❌❌❌❌❌ ❌❌❍❌❌❌❌❌❌❌❌❌❌ ❌❌❌❌❑❕❔❌❌❌❌❌❌', 3)\n",
      "\n",
      "Block size: [1, 8]\n",
      "\n",
      "('❌❌❌⟌❌❌ ❌❌❐➌❌❌ ❌❌❬❌❌❌ ❌❌❔⟌❌❌ ❌❌❜❌❌❌ ❌❌❍➌❌❌ ❌❬❌❌❌❌ ❌❎❌❌❌❌ ❌❌⟔⟼❌❌ ❌❌❌➌❌❌ ❌❌❌❌❬❌ ❌❌⟌❌❍❌ ❌❌❌❌❐❌ ❌❌❌❍❎❌ ❌❌➌⟌❌❌ ❌❌❜⟌❌❌ ❌❌❌❌❎❌ ❌❌⟌❌❍❌ ❌❌❐➌❌❌ ❌❌❔❜❌❌ ❌❌❌⟌❌❌ ❌❌❜❌❌❌ ❌❌❎❌❌❌ ❌❬❌❌❌❌ ❌❌➜⟌❌❌ ❌❌❌❌❐❌ ❌❌❌❌❎❌ ❌❌❌❌❐❌ ❌❌❌❌❔❌ ❌❌❌❌❜❌ ❌❌❌❌❔❌ ❌❌❌❌❜❌ ❌❌❌❌❬❌ ❌❌➜❌➌❌ ❌❌❌❌❎❌ ❌❌❌❌❐❌ ❌❌❌❌❔❌ ❌❌❌❌❜❌ ❌❌❌❌❔❌ ❌❌❌❌❜❌ ❌❌❌❌❬❌ ❌❌➜❌➌❌ ❌❌❌❭❌❌ ❌❌❎❜❌❌ ❌❌❜❌❌❌ ❌❌❎❌❌❌ ❌❬❌❌❌❌ ❌❌❜❴❌❌ ❌❌❌❌❐❌ ❌❌❌❌❬❌ ❌❌❐❌❌❌ ❌❌❌❌❐❌ ❌❌❔❌❎❌ ❌❌❎❌➌❌ ❌⟌❌⟌❌❌ ❌❌❌❌❎❌ ❌❌❐❌❍❌ ❌❬❌❌❐❌ ❌➌❌❌❐❌ ❌❌❌➌❔❌ ❌⟌❌❌❌❌ ❌❜❌❌❌❌ ❌❍❌❌❌❌ ❌❌⟡❔❌❌', 3)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "block_sizes = [\n",
    "    [1, 1],\n",
    "    [1, 2],\n",
    "    [1, 4],\n",
    "    [1, 8],\n",
    "]\n",
    "\n",
    "print(\"SPARSE\")\n",
    "print()\n",
    "\n",
    "for enc_name in encoders.sparse_encoders.keys():\n",
    "    with open(f\"/home/abunn/ttmp/ExplorationWithLLMs/source/03_alec_temp_name/finetuning_data/{enc_name}.pkl\", \"rb\") as f:\n",
    "        x_train_encoded, y_train_encoded, x_valid_encoded, y_valid_encoded, x_test_encoded, y_test_encoded = pickle.load(f)\n",
    "    print(\"Encoder:\", enc_name)\n",
    "    print()\n",
    "    print((x_train_encoded[0], y_train_encoded[0]))\n",
    "    print()\n",
    "\n",
    "print(\"DENSE\")\n",
    "print()\n",
    "\n",
    "for block_size in block_sizes:\n",
    "    with open(f\"/home/abunn/ttmp/ExplorationWithLLMs/source/03_alec_temp_name/finetuning_data/dense_{block_size[0]}_{block_size[1]}.pkl\", \"rb\") as f:\n",
    "        x_train_encoded, y_train_encoded, x_valid_encoded, y_valid_encoded, x_test_encoded, y_test_encoded = pickle.load(f)\n",
    "    print(\"Block size:\", block_size)\n",
    "    print()\n",
    "    print((x_train_encoded[0], y_train_encoded[0]))\n",
    "    print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "20c7f2ee",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "EWLLMs",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}