Upload 7 files

multitask_prompt_tuning.ipynb

@@ -0,0 +1,408 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "58ff91ca-ce92-43d0-ae8b-4e9e89e193f6",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "from datasets import load_dataset\n",
+    "from transformers import set_seed, AutoModelForSeq2SeqLM, AutoTokenizer\n",
+    "from peft import get_peft_model, MultitaskPromptTuningConfig, TaskType, MultitaskPromptTuningInit\n",
+    "\n",
+    "set_seed(42)\n",
+    "\n",
+    "model_name = \"google/flan-t5-base\"\n",
+    "\n",
+    "peft_config = MultitaskPromptTuningConfig(\n",
+    "    tokenizer_name_or_path=model_name,\n",
+    "    num_tasks=2,\n",
+    "    task_type=TaskType.SEQ_2_SEQ_LM,\n",
+    "    prompt_tuning_init=MultitaskPromptTuningInit.TEXT,\n",
+    "    num_virtual_tokens=50,\n",
+    "    num_transformer_submodules=1,\n",
+    "    prompt_tuning_init_text=\"classify the following into either positive or negative, or entailment, neutral or contradiction:\",\n",
+    ")\n",
+    "\n",
+    "tokenizer = AutoTokenizer.from_pretrained(model_name)\n",
+    "model = AutoModelForSeq2SeqLM.from_pretrained(model_name)\n",
+    "model = get_peft_model(model, peft_config)\n",
+    "\n",
+    "model = model.cuda()\n",
+    "\n",
+    "\n",
+    "def send_to_device(batch):\n",
+    "    for i in batch:\n",
+    "        batch[i] = batch[i].cuda()\n",
+    "    return batch"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "eb112bc1-ffaf-49fa-a216-0d601ec304ee",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "def get_sst2(split: str):\n",
+    "    examples = load_dataset(\"sst2\")[split]\n",
+    "    result_examples = []\n",
+    "    for example in examples:\n",
+    "        result_examples.append({})\n",
+    "\n",
+    "        result_examples[-1][\"input\"] = example[\"sentence\"].strip() + \"</s>\"\n",
+    "        result_examples[-1][\"output\"] = (\n",
+    "            f\"positive{tokenizer.eos_token}\" if example[\"label\"] == 1 else f\"negative{tokenizer.eos_token}\"\n",
+    "        )\n",
+    "        result_examples[-1][\"task_id\"] = 0\n",
+    "\n",
+    "    return result_examples\n",
+    "\n",
+    "\n",
+    "def get_mnli(split: str):\n",
+    "    examples = load_dataset(\"multi_nli\")[split]\n",
+    "    result_examples = []\n",
+    "    for example in examples:\n",
+    "        result_examples.append({})\n",
+    "\n",
+    "        result_examples[-1][\"input\"] = example[\"premise\"].strip() + \" \" + example[\"hypothesis\"].strip() + \"</s>\"\n",
+    "\n",
+    "        if example[\"label\"] == 0:\n",
+    "            result_examples[-1][\"output\"] = f\"entailment{tokenizer.eos_token}\"\n",
+    "        elif example[\"label\"] == 1:\n",
+    "            result_examples[-1][\"output\"] = f\"neutral{tokenizer.eos_token}\"\n",
+    "        else:\n",
+    "            result_examples[-1][\"output\"] = f\"contradiction{tokenizer.eos_token}\"\n",
+    "\n",
+    "        result_examples[-1][\"task_id\"] = 1\n",
+    "\n",
+    "    return result_examples"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "e5a16ec4-8fef-4ba9-95b6-a661eb51e50c",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "from typing import Tuple\n",
+    "from torch.utils.data import Dataset, DataLoader\n",
+    "import torch\n",
+    "\n",
+    "\n",
+    "class MyDataset(Dataset):\n",
+    "    def __init__(self, split: str, mode: str = \"source\") -> None:\n",
+    "        super().__init__()\n",
+    "\n",
+    "        if split == \"train\":\n",
+    "            if mode == \"source\":\n",
+    "                self.examples = get_sst2(split) + get_mnli(split)\n",
+    "            elif mode == \"target\":\n",
+    "                self.examples = get_sst2(split)\n",
+    "        if split == \"val\":\n",
+    "            self.examples = get_sst2(\"validation\")\n",
+    "        if split == \"test\":\n",
+    "            self.examples = get_sst2(\"validation\")\n",
+    "\n",
+    "    def __getitem__(self, index) -> dict:\n",
+    "        return self.examples[index]\n",
+    "\n",
+    "    def __len__(self) -> int:\n",
+    "        return len(self.examples)\n",
+    "\n",
+    "    def __getitem__(self, index) -> dict:\n",
+    "        return self.examples[index]\n",
+    "\n",
+    "    def __len__(self) -> int:\n",
+    "        return len(self.examples)\n",
+    "\n",
+    "\n",
+    "def collate_fn(batch: dict) -> Tuple[torch.Tensor, torch.Tensor]:\n",
+    "    input = [i[\"input\"] for i in batch]\n",
+    "    input = tokenizer(input, add_special_tokens=False, return_tensors=\"pt\", padding=True)\n",
+    "\n",
+    "    output = [i[\"output\"] for i in batch]\n",
+    "    output = tokenizer(output, add_special_tokens=False, return_tensors=\"pt\", padding=True).input_ids\n",
+    "    output[output == tokenizer.pad_token_id] = -100\n",
+    "\n",
+    "    task_ids = [i[\"task_id\"] for i in batch]\n",
+    "    task_ids = torch.tensor(task_ids)\n",
+    "\n",
+    "    return {\n",
+    "        \"input_ids\": input.input_ids,\n",
+    "        \"attention_mask\": input.attention_mask,\n",
+    "        \"labels\": output,\n",
+    "        \"task_ids\": task_ids,\n",
+    "    }\n",
+    "\n",
+    "\n",
+    "train = DataLoader(MyDataset(\"train\"), shuffle=True, batch_size=8, collate_fn=collate_fn)\n",
+    "val = DataLoader(MyDataset(\"val\"), shuffle=False, batch_size=8, collate_fn=collate_fn)\n",
+    "test = DataLoader(MyDataset(\"test\"), shuffle=False, batch_size=8, collate_fn=collate_fn)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "fe0aec7b-f61e-4b00-a90e-c1201dc1f84c",
+   "metadata": {},
+   "source": [
+    "## source training"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cceecc94-f43a-4f62-8d45-926f2f02f36d",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "from torch.optim.adamw import AdamW\n",
+    "from transformers import get_cosine_schedule_with_warmup\n",
+    "from tqdm import tqdm\n",
+    "from sklearn.metrics import f1_score"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "eae5516b-73ab-44a8-a083-4e8de6127f30",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "POSITIVE_TOKEN_ID = tokenizer(\" positive\", add_special_tokens=False)[\"input_ids\"][0]\n",
+    "NEGATIVE_TOKEN_ID = tokenizer(\" negative\", add_special_tokens=False)[\"input_ids\"][0]\n",
+    "\n",
+    "\n",
+    "def classify(batch):\n",
+    "    batch = send_to_device(batch)\n",
+    "    # we pass labels here since we need to generate and peft doesn't support generation yet.\n",
+    "    # No clue how to get around this\n",
+    "    scores = model(**batch).logits\n",
+    "    preds = []\n",
+    "    for i in range(scores.shape[0]):\n",
+    "        if scores[i, 0, POSITIVE_TOKEN_ID] > scores[i, 0, NEGATIVE_TOKEN_ID]:\n",
+    "            preds.append(POSITIVE_TOKEN_ID)\n",
+    "        else:\n",
+    "            preds.append(NEGATIVE_TOKEN_ID)\n",
+    "    return preds\n",
+    "\n",
+    "\n",
+    "@torch.inference_mode()\n",
+    "def evaluate(model, data):\n",
+    "    loss = 0\n",
+    "    preds = []\n",
+    "    golds = []\n",
+    "\n",
+    "    for batch in tqdm(data):\n",
+    "        batch = send_to_device(batch)\n",
+    "        loss += model(**batch).loss\n",
+    "        golds.extend(batch[\"labels\"][:, 0].tolist())\n",
+    "        preds.extend(classify(batch))\n",
+    "\n",
+    "    return loss / len(val), f1_score(golds, preds, pos_label=POSITIVE_TOKEN_ID)\n",
+    "\n",
+    "\n",
+    "optimizer = AdamW(model.parameters(), lr=1e-4)\n",
+    "scheduler = get_cosine_schedule_with_warmup(optimizer, 200, len(train))\n",
+    "\n",
+    "n = 1000\n",
+    "step = 0\n",
+    "train_ = tqdm(train)\n",
+    "\n",
+    "val_loss, f1 = evaluate(model, val)\n",
+    "print(\n",
+    "    f\"\"\"\n",
+    "before source training\n",
+    "val loss = {val_loss}\n",
+    "f1 = {f1}\"\"\"\n",
+    ")\n",
+    "\n",
+    "for batch in train_:\n",
+    "    if step % n == 0:\n",
+    "        val_loss, f1 = evaluate(model, val)\n",
+    "        print(\n",
+    "            f\"\"\"\n",
+    "step = {step}\n",
+    "val loss = {val_loss}\n",
+    "f1 = {f1}\"\"\"\n",
+    "        )\n",
+    "        model.save_pretrained(f\"checkpoints_source/{step}\")\n",
+    "\n",
+    "    step += 1\n",
+    "    batch = send_to_device(batch)\n",
+    "    loss = model(**batch).loss\n",
+    "    loss.backward()\n",
+    "    optimizer.step()\n",
+    "    scheduler.step()\n",
+    "    train_.set_postfix(train_loss=loss)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "74168ef3-66f3-41a7-a40b-7840b103fbf9",
+   "metadata": {},
+   "source": [
+    "## target training"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b09fd456-163e-4dc1-b24d-f2d0d349036c",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "train = DataLoader(MyDataset(\"train\", \"target\"), shuffle=True, batch_size=8, collate_fn=collate_fn)\n",
+    "val = DataLoader(MyDataset(\"val\", \"target\"), shuffle=False, batch_size=8, collate_fn=collate_fn)\n",
+    "test = DataLoader(MyDataset(\"test\", \"target\"), shuffle=False, batch_size=8, collate_fn=collate_fn)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4a539944-f16c-4c3f-bb4a-7b5d9a6042e2",
+   "metadata": {},
+   "source": [
+    "#### create a fresh model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5520d904-aa6c-4654-9335-ed4e7d76cba2",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "peft_config = MultitaskPromptTuningConfig(\n",
+    "    tokenizer_name_or_path=model_name,\n",
+    "    num_tasks=1,\n",
+    "    task_type=TaskType.SEQ_2_SEQ_LM,\n",
+    "    prompt_tuning_init=MultitaskPromptTuningInit.EXACT_SOURCE_TASK,\n",
+    "    prompt_tuning_init_state_dict_path=\"checkpoints_source/50000/adapter_model.bin\",\n",
+    "    num_virtual_tokens=50,\n",
+    "    num_transformer_submodules=1,\n",
+    ")\n",
+    "\n",
+    "tokenizer = AutoTokenizer.from_pretrained(model_name)\n",
+    "model = AutoModelForSeq2SeqLM.from_pretrained(model_name)\n",
+    "model = get_peft_model(model, peft_config)\n",
+    "\n",
+    "model = model.cuda()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "dfa39c2d-d1c5-4ed4-90f8-26e8e324371c",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "optimizer = AdamW(model.parameters(), lr=1e-4)\n",
+    "scheduler = get_cosine_schedule_with_warmup(optimizer, 200, len(train))\n",
+    "\n",
+    "n = 1000\n",
+    "step = 0\n",
+    "train_ = tqdm(train)\n",
+    "\n",
+    "val_loss, f1 = evaluate(model, val)\n",
+    "print(\n",
+    "    f\"\"\"\n",
+    "before target training\n",
+    "val loss = {val_loss}\n",
+    "f1 = {f1}\"\"\"\n",
+    ")\n",
+    "\n",
+    "for batch in train_:\n",
+    "    if step % n == 0:\n",
+    "        val_loss, f1 = evaluate(model, val)\n",
+    "        print(\n",
+    "            f\"\"\"\n",
+    "step = {step}\n",
+    "val loss = {val_loss}\n",
+    "f1 = {f1}\"\"\"\n",
+    "        )\n",
+    "        model.save_pretrained(f\"checkpoints_target/{step}\")\n",
+    "\n",
+    "    step += 1\n",
+    "    batch = send_to_device(batch)\n",
+    "    loss = model(**batch).loss\n",
+    "    loss.backward()\n",
+    "    optimizer.step()\n",
+    "    scheduler.step()\n",
+    "    train_.set_postfix(train_loss=loss)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b6a6eeda-1e09-49a6-8845-cd96c8573145",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "# load last checkpoint for now\n",
+    "from peft import set_peft_model_state_dict\n",
+    "\n",
+    "sd_6000 = torch.load(\"checkpoints_target/6000/adapter_model.bin\")\n",
+    "set_peft_model_state_dict(model, sd_6000)\n",
+    "\n",
+    "# evaluate val\n",
+    "val_loss, f1 = evaluate(model, val)\n",
+    "print(\n",
+    "    f\"\"\"\n",
+    "final\n",
+    "val loss = {val_loss}\n",
+    "f1 = {f1}\"\"\"\n",
+    ")\n",
+    "\n",
+    "# evaluate test\n",
+    "test_loss, f1 = evaluate(model, test)\n",
+    "print(\n",
+    "    f\"\"\"\n",
+    "final\n",
+    "test loss = {test_loss}\n",
+    "f1 = {f1}\"\"\"\n",
+    ")"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "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": "3.9.13"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

peft_ia3_seq2seq.ipynb

@@ -0,0 +1,2711 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "metadata": {
+    "id": "5f93b7d1"
+   },
+   "outputs": [],
+   "source": [
+    "from transformers import AutoModelForSeq2SeqLM\n",
+    "import peft\n",
+    "from peft import get_peft_config, get_peft_model, get_peft_model_state_dict, IA3Config, TaskType\n",
+    "import torch\n",
+    "from datasets import load_dataset\n",
+    "import os\n",
+    "\n",
+    "os.environ[\"TOKENIZERS_PARALLELISM\"] = \"false\"\n",
+    "from transformers import AutoTokenizer\n",
+    "from torch.utils.data import DataLoader\n",
+    "from transformers import default_data_collator, get_linear_schedule_with_warmup\n",
+    "from tqdm import tqdm\n",
+    "from datasets import load_dataset\n",
+    "\n",
+    "device = \"cuda\"\n",
+    "model_name_or_path = \"bigscience/mt0-large\"\n",
+    "tokenizer_name_or_path = \"bigscience/mt0-large\"\n",
+    "\n",
+    "checkpoint_name = \"financial_sentiment_analysis_ia3_v1.pt\"\n",
+    "text_column = \"sentence\"\n",
+    "label_column = \"text_label\"\n",
+    "max_length = 128\n",
+    "lr = 8e-3\n",
+    "num_epochs = 3\n",
+    "batch_size = 8"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "b9e6368c",
+    "outputId": "fc2888a8-4fe9-4d61-dd2d-753e751e1416"
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "<module 'peft' from '/usr/local/lib/python3.10/dist-packages/peft/__init__.py'>"
+      ]
+     },
+     "execution_count": 13,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "import importlib\n",
+    "\n",
+    "importlib.reload(peft)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {
+    "id": "8d0850ac"
+   },
+   "outputs": [],
+   "source": [
+    "# creating model\n",
+    "peft_config = IA3Config(task_type=TaskType.SEQ_2_SEQ_LM, inference_mode=False, feedforward_modules=[])\n",
+    "\n",
+    "model = AutoModelForSeq2SeqLM.from_pretrained(model_name_or_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "e10c3831",
+    "outputId": "e69c5e07-ae58-446c-8301-e99ac6b85d62"
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "MT5ForConditionalGeneration(\n",
+       "  (shared): Embedding(250112, 1024)\n",
+       "  (encoder): MT5Stack(\n",
+       "    (embed_tokens): Embedding(250112, 1024)\n",
+       "    (block): ModuleList(\n",
+       "      (0): MT5Block(\n",
+       "        (layer): ModuleList(\n",
+       "          (0): MT5LayerSelfAttention(\n",
+       "            (SelfAttention): MT5Attention(\n",
+       "              (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (relative_attention_bias): Embedding(32, 16)\n",
+       "            )\n",
+       "            (layer_norm): MT5LayerNorm()\n",
+       "            (dropout): Dropout(p=0.1, inplace=False)\n",
+       "          )\n",
+       "          (1): MT5LayerFF(\n",
+       "            (DenseReluDense): MT5DenseGatedActDense(\n",
+       "              (wi_0): Linear(in_features=1024, out_features=2816, bias=False)\n",
+       "              (wi_1): Linear(in_features=1024, out_features=2816, bias=False)\n",
+       "              (wo): Linear(in_features=2816, out_features=1024, bias=False)\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "              (act): NewGELUActivation()\n",
+       "            )\n",
+       "            (layer_norm): MT5LayerNorm()\n",
+       "            (dropout): Dropout(p=0.1, inplace=False)\n",
+       "          )\n",
+       "        )\n",
+       "      )\n",
+       "      (1-23): 23 x MT5Block(\n",
+       "        (layer): ModuleList(\n",
+       "          (0): MT5LayerSelfAttention(\n",
+       "            (SelfAttention): MT5Attention(\n",
+       "              (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "            )\n",
+       "            (layer_norm): MT5LayerNorm()\n",
+       "            (dropout): Dropout(p=0.1, inplace=False)\n",
+       "          )\n",
+       "          (1): MT5LayerFF(\n",
+       "            (DenseReluDense): MT5DenseGatedActDense(\n",
+       "              (wi_0): Linear(in_features=1024, out_features=2816, bias=False)\n",
+       "              (wi_1): Linear(in_features=1024, out_features=2816, bias=False)\n",
+       "              (wo): Linear(in_features=2816, out_features=1024, bias=False)\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "              (act): NewGELUActivation()\n",
+       "            )\n",
+       "            (layer_norm): MT5LayerNorm()\n",
+       "            (dropout): Dropout(p=0.1, inplace=False)\n",
+       "          )\n",
+       "        )\n",
+       "      )\n",
+       "    )\n",
+       "    (final_layer_norm): MT5LayerNorm()\n",
+       "    (dropout): Dropout(p=0.1, inplace=False)\n",
+       "  )\n",
+       "  (decoder): MT5Stack(\n",
+       "    (embed_tokens): Embedding(250112, 1024)\n",
+       "    (block): ModuleList(\n",
+       "      (0): MT5Block(\n",
+       "        (layer): ModuleList(\n",
+       "          (0): MT5LayerSelfAttention(\n",
+       "            (SelfAttention): MT5Attention(\n",
+       "              (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (relative_attention_bias): Embedding(32, 16)\n",
+       "            )\n",
+       "            (layer_norm): MT5LayerNorm()\n",
+       "            (dropout): Dropout(p=0.1, inplace=False)\n",
+       "          )\n",
+       "          (1): MT5LayerCrossAttention(\n",
+       "            (EncDecAttention): MT5Attention(\n",
+       "              (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "            )\n",
+       "            (layer_norm): MT5LayerNorm()\n",
+       "            (dropout): Dropout(p=0.1, inplace=False)\n",
+       "          )\n",
+       "          (2): MT5LayerFF(\n",
+       "            (DenseReluDense): MT5DenseGatedActDense(\n",
+       "              (wi_0): Linear(in_features=1024, out_features=2816, bias=False)\n",
+       "              (wi_1): Linear(in_features=1024, out_features=2816, bias=False)\n",
+       "              (wo): Linear(in_features=2816, out_features=1024, bias=False)\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "              (act): NewGELUActivation()\n",
+       "            )\n",
+       "            (layer_norm): MT5LayerNorm()\n",
+       "            (dropout): Dropout(p=0.1, inplace=False)\n",
+       "          )\n",
+       "        )\n",
+       "      )\n",
+       "      (1-23): 23 x MT5Block(\n",
+       "        (layer): ModuleList(\n",
+       "          (0): MT5LayerSelfAttention(\n",
+       "            (SelfAttention): MT5Attention(\n",
+       "              (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "            )\n",
+       "            (layer_norm): MT5LayerNorm()\n",
+       "            (dropout): Dropout(p=0.1, inplace=False)\n",
+       "          )\n",
+       "          (1): MT5LayerCrossAttention(\n",
+       "            (EncDecAttention): MT5Attention(\n",
+       "              (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "            )\n",
+       "            (layer_norm): MT5LayerNorm()\n",
+       "            (dropout): Dropout(p=0.1, inplace=False)\n",
+       "          )\n",
+       "          (2): MT5LayerFF(\n",
+       "            (DenseReluDense): MT5DenseGatedActDense(\n",
+       "              (wi_0): Linear(in_features=1024, out_features=2816, bias=False)\n",
+       "              (wi_1): Linear(in_features=1024, out_features=2816, bias=False)\n",
+       "              (wo): Linear(in_features=2816, out_features=1024, bias=False)\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "              (act): NewGELUActivation()\n",
+       "            )\n",
+       "            (layer_norm): MT5LayerNorm()\n",
+       "            (dropout): Dropout(p=0.1, inplace=False)\n",
+       "          )\n",
+       "        )\n",
+       "      )\n",
+       "    )\n",
+       "    (final_layer_norm): MT5LayerNorm()\n",
+       "    (dropout): Dropout(p=0.1, inplace=False)\n",
+       "  )\n",
+       "  (lm_head): Linear(in_features=1024, out_features=250112, bias=False)\n",
+       ")"
+      ]
+     },
+     "execution_count": 15,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "05978e96",
+    "outputId": "ea9b7d40-010f-4df0-ec64-a7146a5f8b08"
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "trainable params: 282,624 || all params: 1,229,863,936 || trainable%: 0.022980103060766553\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "PeftModelForSeq2SeqLM(\n",
+       "  (base_model): IA3Model(\n",
+       "    (model): MT5ForConditionalGeneration(\n",
+       "      (shared): Embedding(250112, 1024)\n",
+       "      (encoder): MT5Stack(\n",
+       "        (embed_tokens): Embedding(250112, 1024)\n",
+       "        (block): ModuleList(\n",
+       "          (0): MT5Block(\n",
+       "            (layer): ModuleList(\n",
+       "              (0): MT5LayerSelfAttention(\n",
+       "                (SelfAttention): MT5Attention(\n",
+       "                  (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                  (k): Linear(\n",
+       "                    in_features=1024, out_features=1024, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 1024x1])\n",
+       "                  )\n",
+       "                  (v): Linear(\n",
+       "                    in_features=1024, out_features=1024, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 1024x1])\n",
+       "                  )\n",
+       "                  (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                  (relative_attention_bias): Embedding(32, 16)\n",
+       "                )\n",
+       "                (layer_norm): MT5LayerNorm()\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "              )\n",
+       "              (1): MT5LayerFF(\n",
+       "                (DenseReluDense): MT5DenseGatedActDense(\n",
+       "                  (wi_0): Linear(in_features=1024, out_features=2816, bias=False)\n",
+       "                  (wi_1): Linear(\n",
+       "                    in_features=1024, out_features=2816, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 2816x1])\n",
+       "                  )\n",
+       "                  (wo): Linear(in_features=2816, out_features=1024, bias=False)\n",
+       "                  (dropout): Dropout(p=0.1, inplace=False)\n",
+       "                  (act): NewGELUActivation()\n",
+       "                )\n",
+       "                (layer_norm): MT5LayerNorm()\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "              )\n",
+       "            )\n",
+       "          )\n",
+       "          (1-23): 23 x MT5Block(\n",
+       "            (layer): ModuleList(\n",
+       "              (0): MT5LayerSelfAttention(\n",
+       "                (SelfAttention): MT5Attention(\n",
+       "                  (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                  (k): Linear(\n",
+       "                    in_features=1024, out_features=1024, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 1024x1])\n",
+       "                  )\n",
+       "                  (v): Linear(\n",
+       "                    in_features=1024, out_features=1024, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 1024x1])\n",
+       "                  )\n",
+       "                  (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                )\n",
+       "                (layer_norm): MT5LayerNorm()\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "              )\n",
+       "              (1): MT5LayerFF(\n",
+       "                (DenseReluDense): MT5DenseGatedActDense(\n",
+       "                  (wi_0): Linear(in_features=1024, out_features=2816, bias=False)\n",
+       "                  (wi_1): Linear(\n",
+       "                    in_features=1024, out_features=2816, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 2816x1])\n",
+       "                  )\n",
+       "                  (wo): Linear(in_features=2816, out_features=1024, bias=False)\n",
+       "                  (dropout): Dropout(p=0.1, inplace=False)\n",
+       "                  (act): NewGELUActivation()\n",
+       "                )\n",
+       "                (layer_norm): MT5LayerNorm()\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "              )\n",
+       "            )\n",
+       "          )\n",
+       "        )\n",
+       "        (final_layer_norm): MT5LayerNorm()\n",
+       "        (dropout): Dropout(p=0.1, inplace=False)\n",
+       "      )\n",
+       "      (decoder): MT5Stack(\n",
+       "        (embed_tokens): Embedding(250112, 1024)\n",
+       "        (block): ModuleList(\n",
+       "          (0): MT5Block(\n",
+       "            (layer): ModuleList(\n",
+       "              (0): MT5LayerSelfAttention(\n",
+       "                (SelfAttention): MT5Attention(\n",
+       "                  (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                  (k): Linear(\n",
+       "                    in_features=1024, out_features=1024, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 1024x1])\n",
+       "                  )\n",
+       "                  (v): Linear(\n",
+       "                    in_features=1024, out_features=1024, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 1024x1])\n",
+       "                  )\n",
+       "                  (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                  (relative_attention_bias): Embedding(32, 16)\n",
+       "                )\n",
+       "                (layer_norm): MT5LayerNorm()\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "              )\n",
+       "              (1): MT5LayerCrossAttention(\n",
+       "                (EncDecAttention): MT5Attention(\n",
+       "                  (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                  (k): Linear(\n",
+       "                    in_features=1024, out_features=1024, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 1024x1])\n",
+       "                  )\n",
+       "                  (v): Linear(\n",
+       "                    in_features=1024, out_features=1024, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 1024x1])\n",
+       "                  )\n",
+       "                  (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                )\n",
+       "                (layer_norm): MT5LayerNorm()\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "              )\n",
+       "              (2): MT5LayerFF(\n",
+       "                (DenseReluDense): MT5DenseGatedActDense(\n",
+       "                  (wi_0): Linear(in_features=1024, out_features=2816, bias=False)\n",
+       "                  (wi_1): Linear(\n",
+       "                    in_features=1024, out_features=2816, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 2816x1])\n",
+       "                  )\n",
+       "                  (wo): Linear(in_features=2816, out_features=1024, bias=False)\n",
+       "                  (dropout): Dropout(p=0.1, inplace=False)\n",
+       "                  (act): NewGELUActivation()\n",
+       "                )\n",
+       "                (layer_norm): MT5LayerNorm()\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "              )\n",
+       "            )\n",
+       "          )\n",
+       "          (1-23): 23 x MT5Block(\n",
+       "            (layer): ModuleList(\n",
+       "              (0): MT5LayerSelfAttention(\n",
+       "                (SelfAttention): MT5Attention(\n",
+       "                  (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                  (k): Linear(\n",
+       "                    in_features=1024, out_features=1024, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 1024x1])\n",
+       "                  )\n",
+       "                  (v): Linear(\n",
+       "                    in_features=1024, out_features=1024, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 1024x1])\n",
+       "                  )\n",
+       "                  (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                )\n",
+       "                (layer_norm): MT5LayerNorm()\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "              )\n",
+       "              (1): MT5LayerCrossAttention(\n",
+       "                (EncDecAttention): MT5Attention(\n",
+       "                  (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                  (k): Linear(\n",
+       "                    in_features=1024, out_features=1024, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 1024x1])\n",
+       "                  )\n",
+       "                  (v): Linear(\n",
+       "                    in_features=1024, out_features=1024, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 1024x1])\n",
+       "                  )\n",
+       "                  (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                )\n",
+       "                (layer_norm): MT5LayerNorm()\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "              )\n",
+       "              (2): MT5LayerFF(\n",
+       "                (DenseReluDense): MT5DenseGatedActDense(\n",
+       "                  (wi_0): Linear(in_features=1024, out_features=2816, bias=False)\n",
+       "                  (wi_1): Linear(\n",
+       "                    in_features=1024, out_features=2816, bias=False\n",
+       "                    (ia3_l): ParameterDict(  (default): Parameter containing: [torch.FloatTensor of size 2816x1])\n",
+       "                  )\n",
+       "                  (wo): Linear(in_features=2816, out_features=1024, bias=False)\n",
+       "                  (dropout): Dropout(p=0.1, inplace=False)\n",
+       "                  (act): NewGELUActivation()\n",
+       "                )\n",
+       "                (layer_norm): MT5LayerNorm()\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "              )\n",
+       "            )\n",
+       "          )\n",
+       "        )\n",
+       "        (final_layer_norm): MT5LayerNorm()\n",
+       "        (dropout): Dropout(p=0.1, inplace=False)\n",
+       "      )\n",
+       "      (lm_head): Linear(in_features=1024, out_features=250112, bias=False)\n",
+       "    )\n",
+       "  )\n",
+       ")"
+      ]
+     },
+     "execution_count": 16,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "model = get_peft_model(model, peft_config)\n",
+    "model.print_trainable_parameters()\n",
+    "model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/",
+     "height": 140,
+     "referenced_widgets": [
+      "bbfb7533b5ca459194e171df56b79566",
+      "c894e8237aa34c56bb250acab1466005",
+      "a5a126b229064812bf3dcb228118be50",
+      "661e1b29c59a4295b594edfa4f50ff87",
+      "1bcba805972b484d8b6aa6542c81841c",
+      "e71f5c7f1d5d4f83b58c68d2fa310d9c",
+      "6a567e0a1a5447519c5df10e777520cf",
+      "7aeca19b84904906a04c12659f84ff9e",
+      "dd4b895874ce46ceb1ad0d9bc973f98f",
+      "b138f91be7f94008806eaf0a6988bc3f",
+      "da14180f51ab44b48470cb9ea74d3864",
+      "9e12d97af6124a5a8c6627708b300c1e",
+      "faa18df899c14e9cac6721253e6c9128",
+      "79d0ede7a5b24756aa6d34fda8c29159",
+      "3b175b452f4347558aa3c4501cc90030",
+      "fc4637a1b37e4e90874c71aa4271ac74",
+      "1b8aada826a0451bb60c418b19178c8c",
+      "a91916e02e9c424e881e45b3aa978574",
+      "ca509bd409624c998e555c9a779b8aae",
+      "9c890fc422954347b86d3bde7a421caf",
+      "6f9453484ea94587a64d70f1b3a1f6e4",
+      "48770ef159f44c01be2a75c75aecd80f",
+      "0c561dab67914ea9b6e1aab803600551",
+      "1e021a1954b44d69a90101a96c360661",
+      "013e3343285f437a893bdd673fb90e22",
+      "28802da68fb04d70b1c6bc511a04676f",
+      "94174da0d6554be087d4527bea5b511a",
+      "dc8ab16a1e6c4e6893c95ccd16568f9a",
+      "72383136663448d89cf3b82b87cbb392",
+      "5b1bdaf16cbc473081e4237f839167b9",
+      "51f8fb45485540bb985b606d43ae04ea",
+      "f760cd4758334ca9a43fd15612fd808b",
+      "f60e9915d2a74ca7bc010d7684f5acf6"
+     ]
+    },
+    "id": "4ee2babf",
+    "outputId": "3c413083-247d-47da-f25c-032764be0beb"
+   },
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "WARNING:datasets.builder:Found cached dataset financial_phrasebank (/root/.cache/huggingface/datasets/financial_phrasebank/sentences_allagree/1.0.0/550bde12e6c30e2674da973a55f57edde5181d53f5a5a34c1531c53f93b7e141)\n"
+     ]
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "bbfb7533b5ca459194e171df56b79566",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "  0%|          | 0/1 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "9e12d97af6124a5a8c6627708b300c1e",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Map:   0%|          | 0/2037 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "0c561dab67914ea9b6e1aab803600551",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Map:   0%|          | 0/227 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/plain": [
+       "{'sentence': 'It will be operated by Nokia , and supported by its Nokia NetAct network and service management system .',\n",
+       " 'label': 1,\n",
+       " 'text_label': 'neutral'}"
+      ]
+     },
+     "execution_count": 17,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# loading dataset\n",
+    "dataset = load_dataset(\"financial_phrasebank\", \"sentences_allagree\")\n",
+    "dataset = dataset[\"train\"].train_test_split(test_size=0.1)\n",
+    "dataset[\"validation\"] = dataset[\"test\"]\n",
+    "del dataset[\"test\"]\n",
+    "\n",
+    "classes = dataset[\"train\"].features[\"label\"].names\n",
+    "dataset = dataset.map(\n",
+    "    lambda x: {\"text_label\": [classes[label] for label in x[\"label\"]]},\n",
+    "    batched=True,\n",
+    "    num_proc=1,\n",
+    ")\n",
+    "\n",
+    "dataset[\"train\"][0]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/",
+     "height": 17,
+     "referenced_widgets": [
+      "e1e80a68a9e7429397cafc96c3c11f80",
+      "5307864c2b1143f4b44f3f172611113e",
+      "2e2b6c3f48974ea4aca9b7710a03379e",
+      "aae78f9bd53348bda45967a38736cb78",
+      "34db17e0f28d40d6abafb8acd5dda379",
+      "8361dc2e0a834da6a0ad87f7b0cb4e1b",
+      "56f1d9d56dd44c8aa923d09a59cb0ebc",
+      "d93bfb366db14c2fa77b038752f69b38",
+      "749aaa39135841f98b344ffb840df3d4",
+      "5e5aa58adb0f48579871df33845e30b1",
+      "c25b49b7adaa48a0a3a306aa1e0661b4",
+      "21f582e1208a4a38ae3c0cdce87e5c14",
+      "d9d37b8b79f24dbf837327a250a5a346",
+      "8ba99043c350456d8623ce1d8c98f7a0",
+      "8bf37c12d5f74f7d8dbba423a9ee3ac3",
+      "f9d86ad7fa734f3a857505a542256a3c",
+      "86bf02b06ed740a88015c2b944205c1e",
+      "aef6a6be67f749908060d8038b6d3804",
+      "664c02903cb248fb9339805bccfd6c1d",
+      "82195b807b664a9585a76e0e50fe7609",
+      "8621932be14f42858d841e2ac1b173e7",
+      "71bcdb1e02144c9587879d8d815b91d4"
+     ]
+    },
+    "id": "adf9608c",
+    "outputId": "3e4bc95f-1dc4-4d34-c212-6d2374359673"
+   },
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "e1e80a68a9e7429397cafc96c3c11f80",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Running tokenizer on dataset:   0%|          | 0/2037 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "21f582e1208a4a38ae3c0cdce87e5c14",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Running tokenizer on dataset:   0%|          | 0/227 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "# data preprocessing\n",
+    "tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)\n",
+    "\n",
+    "\n",
+    "def preprocess_function(examples):\n",
+    "    inputs = examples[text_column]\n",
+    "    targets = examples[label_column]\n",
+    "    model_inputs = tokenizer(inputs, max_length=max_length, padding=\"max_length\", truncation=True, return_tensors=\"pt\")\n",
+    "    labels = tokenizer(targets, max_length=3, padding=\"max_length\", truncation=True, return_tensors=\"pt\")\n",
+    "    labels = labels[\"input_ids\"]\n",
+    "    labels[labels == tokenizer.pad_token_id] = -100\n",
+    "    model_inputs[\"labels\"] = labels\n",
+    "    return model_inputs\n",
+    "\n",
+    "\n",
+    "processed_datasets = dataset.map(\n",
+    "    preprocess_function,\n",
+    "    batched=True,\n",
+    "    num_proc=1,\n",
+    "    remove_columns=dataset[\"train\"].column_names,\n",
+    "    load_from_cache_file=False,\n",
+    "    desc=\"Running tokenizer on dataset\",\n",
+    ")\n",
+    "\n",
+    "train_dataset = processed_datasets[\"train\"]\n",
+    "eval_dataset = processed_datasets[\"validation\"]\n",
+    "\n",
+    "train_dataloader = DataLoader(\n",
+    "    train_dataset, shuffle=True, collate_fn=default_data_collator, batch_size=batch_size, pin_memory=True\n",
+    ")\n",
+    "eval_dataloader = DataLoader(eval_dataset, collate_fn=default_data_collator, batch_size=batch_size, pin_memory=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "metadata": {
+    "id": "f733a3c6"
+   },
+   "outputs": [],
+   "source": [
+    "# optimizer and lr scheduler\n",
+    "optimizer = torch.optim.AdamW(model.parameters(), lr=lr)\n",
+    "lr_scheduler = get_linear_schedule_with_warmup(\n",
+    "    optimizer=optimizer,\n",
+    "    num_warmup_steps=0,\n",
+    "    num_training_steps=(len(train_dataloader) * num_epochs),\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "6b3a4090",
+    "outputId": "369cfce9-90f2-47a1-8653-ea1168943949"
+   },
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|██████████| 255/255 [02:33<00:00,  1.67it/s]\n",
+      "100%|██████████| 29/29 [00:08<00:00,  3.48it/s]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=0: train_ppl=tensor(1.4939, device='cuda:0') train_epoch_loss=tensor(0.4014, device='cuda:0') eval_ppl=tensor(1.0514, device='cuda:0') eval_epoch_loss=tensor(0.0501, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|██████████| 255/255 [02:32<00:00,  1.67it/s]\n",
+      "100%|██████████| 29/29 [00:08<00:00,  3.43it/s]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=1: train_ppl=tensor(1.0523, device='cuda:0') train_epoch_loss=tensor(0.0510, device='cuda:0') eval_ppl=tensor(1.0383, device='cuda:0') eval_epoch_loss=tensor(0.0376, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|██████████| 255/255 [02:32<00:00,  1.68it/s]\n",
+      "100%|██████████| 29/29 [00:08<00:00,  3.44it/s]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=2: train_ppl=tensor(1.0397, device='cuda:0') train_epoch_loss=tensor(0.0389, device='cuda:0') eval_ppl=tensor(1.0392, device='cuda:0') eval_epoch_loss=tensor(0.0385, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "# training and evaluation\n",
+    "model = model.to(device)\n",
+    "\n",
+    "for epoch in range(num_epochs):\n",
+    "    model.train()\n",
+    "    total_loss = 0\n",
+    "    for step, batch in enumerate(tqdm(train_dataloader)):\n",
+    "        batch = {k: v.to(device) for k, v in batch.items()}\n",
+    "        outputs = model(**batch)\n",
+    "        loss = outputs.loss\n",
+    "        total_loss += loss.detach().float()\n",
+    "        loss.backward()\n",
+    "        optimizer.step()\n",
+    "        lr_scheduler.step()\n",
+    "        optimizer.zero_grad()\n",
+    "\n",
+    "    model.eval()\n",
+    "    eval_loss = 0\n",
+    "    eval_preds = []\n",
+    "    for step, batch in enumerate(tqdm(eval_dataloader)):\n",
+    "        batch = {k: v.to(device) for k, v in batch.items()}\n",
+    "        with torch.no_grad():\n",
+    "            outputs = model(**batch)\n",
+    "        loss = outputs.loss\n",
+    "        eval_loss += loss.detach().float()\n",
+    "        eval_preds.extend(\n",
+    "            tokenizer.batch_decode(torch.argmax(outputs.logits, -1).detach().cpu().numpy(), skip_special_tokens=True)\n",
+    "        )\n",
+    "\n",
+    "    eval_epoch_loss = eval_loss / len(eval_dataloader)\n",
+    "    eval_ppl = torch.exp(eval_epoch_loss)\n",
+    "    train_epoch_loss = total_loss / len(train_dataloader)\n",
+    "    train_ppl = torch.exp(train_epoch_loss)\n",
+    "    print(f\"{epoch=}: {train_ppl=} {train_epoch_loss=} {eval_ppl=} {eval_epoch_loss=}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "6cafa67b",
+    "outputId": "0db923d2-522c-4cb7-b694-6e2e20beae98"
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "accuracy=96.91629955947137 % on the evaluation dataset\n",
+      "eval_preds[:10]=['neutral', 'neutral', 'neutral', 'neutral', 'positive', 'neutral', 'neutral', 'neutral', 'neutral', 'neutral']\n",
+      "dataset['validation']['text_label'][:10]=['neutral', 'neutral', 'neutral', 'neutral', 'positive', 'neutral', 'neutral', 'neutral', 'neutral', 'neutral']\n"
+     ]
+    }
+   ],
+   "source": [
+    "# print accuracy\n",
+    "correct = 0\n",
+    "total = 0\n",
+    "for pred, true in zip(eval_preds, dataset[\"validation\"][\"text_label\"]):\n",
+    "    if pred.strip() == true.strip():\n",
+    "        correct += 1\n",
+    "    total += 1\n",
+    "accuracy = correct / total * 100\n",
+    "print(f\"{accuracy=} % on the evaluation dataset\")\n",
+    "print(f\"{eval_preds[:10]=}\")\n",
+    "print(f\"{dataset['validation']['text_label'][:10]=}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "metadata": {
+    "id": "a8de6005"
+   },
+   "outputs": [],
+   "source": [
+    "# saving model\n",
+    "peft_model_id = f\"{model_name_or_path}_{peft_config.peft_type}_{peft_config.task_type}\"\n",
+    "model.save_pretrained(peft_model_id)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "bd20cd4c",
+    "outputId": "0f25d837-80b1-476f-c897-92c3fef04fb2"
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "1.2M\tbigscience/mt0-large_IA3_SEQ_2_SEQ_LM/adapter_model.bin\n"
+     ]
+    }
+   ],
+   "source": [
+    "ckpt = f\"{peft_model_id}/adapter_model.bin\"\n",
+    "!du -h $ckpt"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 24,
+   "metadata": {
+    "id": "76c2fc29"
+   },
+   "outputs": [],
+   "source": [
+    "from peft import PeftModel, PeftConfig\n",
+    "\n",
+    "peft_model_id = f\"{model_name_or_path}_{peft_config.peft_type}_{peft_config.task_type}\"\n",
+    "\n",
+    "config = PeftConfig.from_pretrained(peft_model_id)\n",
+    "model = AutoModelForSeq2SeqLM.from_pretrained(config.base_model_name_or_path)\n",
+    "model = PeftModel.from_pretrained(model, peft_model_id)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 25,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "37d712ce",
+    "outputId": "4828819a-b640-4f6c-91e3-878b648e9a75"
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "25 November 2010 - Finnish paints and coatings company Tikkurila Oyj ( HEL : TIK1V ) said today that Finnish state-owned investment company Solidium Oy sold its 14.7 % stake in the company for a total of EUR98m .\n",
+      "{'input_ids': tensor([[   877,   3277,   1068,    259,    264,    515, 143136,  42068,    263,\n",
+      "            305,    259, 101264,    263,   5835,  22538,   4496,   2697,  20860,\n",
+      "            385,    274,  76347,    259,    267,    259,  93686,    353,    561,\n",
+      "            259,    271,   2426,   7883,    533,    515, 143136,   6509,    264,\n",
+      "          45815,  37624,   5835,  35133,  16558,  20860,  22026,   2476,   5006,\n",
+      "            487,   1448,    259,  96189,    281,    287,   5835,    332,    259,\n",
+      "            262,   2725,    304,   2687,   5577,    282,    259,    260,      1]]), 'attention_mask': tensor([[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
+      "         1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
+      "         1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]])}\n",
+      "tensor([[    0, 59006,     1]])\n",
+      "['neutral']\n"
+     ]
+    }
+   ],
+   "source": [
+    "model.eval()\n",
+    "i = 13\n",
+    "inputs = tokenizer(dataset[\"validation\"][text_column][i], return_tensors=\"pt\")\n",
+    "print(dataset[\"validation\"][text_column][i])\n",
+    "print(inputs)\n",
+    "\n",
+    "with torch.no_grad():\n",
+    "    outputs = model.generate(input_ids=inputs[\"input_ids\"], max_new_tokens=10)\n",
+    "    print(outputs)\n",
+    "    print(tokenizer.batch_decode(outputs.detach().cpu().numpy(), skip_special_tokens=True))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "id": "66c65ea4"
+   },
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "id": "65e71f78"
+   },
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "accelerator": "GPU",
+  "colab": {
+   "gpuType": "T4",
+   "machine_shape": "hm",
+   "provenance": []
+  },
+  "kernelspec": {
+   "display_name": "Python 3",
+   "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": "3.8.3"
+  },
+  "vscode": {
+   "interpreter": {
+    "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
+   }
+  },
+  "widgets": {
+   "application/vnd.jupyter.widget-state+json": {
+    "013e3343285f437a893bdd673fb90e22": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "FloatProgressModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "FloatProgressModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "ProgressView",
+      "bar_style": "",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_5b1bdaf16cbc473081e4237f839167b9",
+      "max": 227,
+      "min": 0,
+      "orientation": "horizontal",
+      "style": "IPY_MODEL_51f8fb45485540bb985b606d43ae04ea",
+      "value": 227
+     }
+    },
+    "0c561dab67914ea9b6e1aab803600551": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HBoxModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HBoxModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HBoxView",
+      "box_style": "",
+      "children": [
+       "IPY_MODEL_1e021a1954b44d69a90101a96c360661",
+       "IPY_MODEL_013e3343285f437a893bdd673fb90e22",
+       "IPY_MODEL_28802da68fb04d70b1c6bc511a04676f"
+      ],
+      "layout": "IPY_MODEL_94174da0d6554be087d4527bea5b511a"
+     }
+    },
+    "1b8aada826a0451bb60c418b19178c8c": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "1bcba805972b484d8b6aa6542c81841c": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "1e021a1954b44d69a90101a96c360661": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HTMLModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HTMLModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HTMLView",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_dc8ab16a1e6c4e6893c95ccd16568f9a",
+      "placeholder": "​",
+      "style": "IPY_MODEL_72383136663448d89cf3b82b87cbb392",
+      "value": "Map:   0%"
+     }
+    },
+    "21f582e1208a4a38ae3c0cdce87e5c14": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HBoxModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HBoxModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HBoxView",
+      "box_style": "",
+      "children": [
+       "IPY_MODEL_d9d37b8b79f24dbf837327a250a5a346",
+       "IPY_MODEL_8ba99043c350456d8623ce1d8c98f7a0",
+       "IPY_MODEL_8bf37c12d5f74f7d8dbba423a9ee3ac3"
+      ],
+      "layout": "IPY_MODEL_f9d86ad7fa734f3a857505a542256a3c"
+     }
+    },
+    "28802da68fb04d70b1c6bc511a04676f": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HTMLModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HTMLModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HTMLView",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_f760cd4758334ca9a43fd15612fd808b",
+      "placeholder": "​",
+      "style": "IPY_MODEL_f60e9915d2a74ca7bc010d7684f5acf6",
+      "value": " 0/227 [00:00&lt;?, ? examples/s]"
+     }
+    },
+    "2e2b6c3f48974ea4aca9b7710a03379e": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "FloatProgressModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "FloatProgressModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "ProgressView",
+      "bar_style": "",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_d93bfb366db14c2fa77b038752f69b38",
+      "max": 2037,
+      "min": 0,
+      "orientation": "horizontal",
+      "style": "IPY_MODEL_749aaa39135841f98b344ffb840df3d4",
+      "value": 2037
+     }
+    },
+    "34db17e0f28d40d6abafb8acd5dda379": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": "hidden",
+      "width": null
+     }
+    },
+    "3b175b452f4347558aa3c4501cc90030": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HTMLModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HTMLModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HTMLView",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_6f9453484ea94587a64d70f1b3a1f6e4",
+      "placeholder": "​",
+      "style": "IPY_MODEL_48770ef159f44c01be2a75c75aecd80f",
+      "value": " 0/2037 [00:00&lt;?, ? examples/s]"
+     }
+    },
+    "48770ef159f44c01be2a75c75aecd80f": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "DescriptionStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "DescriptionStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "description_width": ""
+     }
+    },
+    "51f8fb45485540bb985b606d43ae04ea": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "ProgressStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "ProgressStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "bar_color": null,
+      "description_width": ""
+     }
+    },
+    "5307864c2b1143f4b44f3f172611113e": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HTMLModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HTMLModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HTMLView",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_8361dc2e0a834da6a0ad87f7b0cb4e1b",
+      "placeholder": "​",
+      "style": "IPY_MODEL_56f1d9d56dd44c8aa923d09a59cb0ebc",
+      "value": "Running tokenizer on dataset:  98%"
+     }
+    },
+    "56f1d9d56dd44c8aa923d09a59cb0ebc": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "DescriptionStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "DescriptionStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "description_width": ""
+     }
+    },
+    "5b1bdaf16cbc473081e4237f839167b9": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "5e5aa58adb0f48579871df33845e30b1": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "661e1b29c59a4295b594edfa4f50ff87": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HTMLModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HTMLModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HTMLView",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_b138f91be7f94008806eaf0a6988bc3f",
+      "placeholder": "​",
+      "style": "IPY_MODEL_da14180f51ab44b48470cb9ea74d3864",
+      "value": " 1/1 [00:00&lt;00:00, 67.12it/s]"
+     }
+    },
+    "664c02903cb248fb9339805bccfd6c1d": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "6a567e0a1a5447519c5df10e777520cf": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "DescriptionStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "DescriptionStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "description_width": ""
+     }
+    },
+    "6f9453484ea94587a64d70f1b3a1f6e4": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "71bcdb1e02144c9587879d8d815b91d4": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "DescriptionStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "DescriptionStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "description_width": ""
+     }
+    },
+    "72383136663448d89cf3b82b87cbb392": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "DescriptionStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "DescriptionStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "description_width": ""
+     }
+    },
+    "749aaa39135841f98b344ffb840df3d4": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "ProgressStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "ProgressStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "bar_color": null,
+      "description_width": ""
+     }
+    },
+    "79d0ede7a5b24756aa6d34fda8c29159": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "FloatProgressModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "FloatProgressModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "ProgressView",
+      "bar_style": "",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_ca509bd409624c998e555c9a779b8aae",
+      "max": 2037,
+      "min": 0,
+      "orientation": "horizontal",
+      "style": "IPY_MODEL_9c890fc422954347b86d3bde7a421caf",
+      "value": 2037
+     }
+    },
+    "7aeca19b84904906a04c12659f84ff9e": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "82195b807b664a9585a76e0e50fe7609": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "ProgressStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "ProgressStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "bar_color": null,
+      "description_width": ""
+     }
+    },
+    "8361dc2e0a834da6a0ad87f7b0cb4e1b": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "8621932be14f42858d841e2ac1b173e7": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "86bf02b06ed740a88015c2b944205c1e": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "8ba99043c350456d8623ce1d8c98f7a0": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "FloatProgressModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "FloatProgressModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "ProgressView",
+      "bar_style": "",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_664c02903cb248fb9339805bccfd6c1d",
+      "max": 227,
+      "min": 0,
+      "orientation": "horizontal",
+      "style": "IPY_MODEL_82195b807b664a9585a76e0e50fe7609",
+      "value": 227
+     }
+    },
+    "8bf37c12d5f74f7d8dbba423a9ee3ac3": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HTMLModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HTMLModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HTMLView",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_8621932be14f42858d841e2ac1b173e7",
+      "placeholder": "​",
+      "style": "IPY_MODEL_71bcdb1e02144c9587879d8d815b91d4",
+      "value": " 0/227 [00:00&lt;?, ? examples/s]"
+     }
+    },
+    "94174da0d6554be087d4527bea5b511a": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": "hidden",
+      "width": null
+     }
+    },
+    "9c890fc422954347b86d3bde7a421caf": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "ProgressStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "ProgressStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "bar_color": null,
+      "description_width": ""
+     }
+    },
+    "9e12d97af6124a5a8c6627708b300c1e": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HBoxModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HBoxModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HBoxView",
+      "box_style": "",
+      "children": [
+       "IPY_MODEL_faa18df899c14e9cac6721253e6c9128",
+       "IPY_MODEL_79d0ede7a5b24756aa6d34fda8c29159",
+       "IPY_MODEL_3b175b452f4347558aa3c4501cc90030"
+      ],
+      "layout": "IPY_MODEL_fc4637a1b37e4e90874c71aa4271ac74"
+     }
+    },
+    "a5a126b229064812bf3dcb228118be50": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "FloatProgressModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "FloatProgressModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "ProgressView",
+      "bar_style": "success",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_7aeca19b84904906a04c12659f84ff9e",
+      "max": 1,
+      "min": 0,
+      "orientation": "horizontal",
+      "style": "IPY_MODEL_dd4b895874ce46ceb1ad0d9bc973f98f",
+      "value": 1
+     }
+    },
+    "a91916e02e9c424e881e45b3aa978574": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "DescriptionStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "DescriptionStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "description_width": ""
+     }
+    },
+    "aae78f9bd53348bda45967a38736cb78": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HTMLModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HTMLModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HTMLView",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_5e5aa58adb0f48579871df33845e30b1",
+      "placeholder": "​",
+      "style": "IPY_MODEL_c25b49b7adaa48a0a3a306aa1e0661b4",
+      "value": " 2000/2037 [00:00&lt;00:00, 3864.28 examples/s]"
+     }
+    },
+    "aef6a6be67f749908060d8038b6d3804": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "DescriptionStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "DescriptionStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "description_width": ""
+     }
+    },
+    "b138f91be7f94008806eaf0a6988bc3f": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "bbfb7533b5ca459194e171df56b79566": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HBoxModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HBoxModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HBoxView",
+      "box_style": "",
+      "children": [
+       "IPY_MODEL_c894e8237aa34c56bb250acab1466005",
+       "IPY_MODEL_a5a126b229064812bf3dcb228118be50",
+       "IPY_MODEL_661e1b29c59a4295b594edfa4f50ff87"
+      ],
+      "layout": "IPY_MODEL_1bcba805972b484d8b6aa6542c81841c"
+     }
+    },
+    "c25b49b7adaa48a0a3a306aa1e0661b4": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "DescriptionStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "DescriptionStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "description_width": ""
+     }
+    },
+    "c894e8237aa34c56bb250acab1466005": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HTMLModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HTMLModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HTMLView",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_e71f5c7f1d5d4f83b58c68d2fa310d9c",
+      "placeholder": "​",
+      "style": "IPY_MODEL_6a567e0a1a5447519c5df10e777520cf",
+      "value": "100%"
+     }
+    },
+    "ca509bd409624c998e555c9a779b8aae": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "d93bfb366db14c2fa77b038752f69b38": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "d9d37b8b79f24dbf837327a250a5a346": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HTMLModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HTMLModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HTMLView",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_86bf02b06ed740a88015c2b944205c1e",
+      "placeholder": "​",
+      "style": "IPY_MODEL_aef6a6be67f749908060d8038b6d3804",
+      "value": "Running tokenizer on dataset:   0%"
+     }
+    },
+    "da14180f51ab44b48470cb9ea74d3864": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "DescriptionStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "DescriptionStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "description_width": ""
+     }
+    },
+    "dc8ab16a1e6c4e6893c95ccd16568f9a": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "dd4b895874ce46ceb1ad0d9bc973f98f": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "ProgressStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "ProgressStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "bar_color": null,
+      "description_width": ""
+     }
+    },
+    "e1e80a68a9e7429397cafc96c3c11f80": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HBoxModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HBoxModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HBoxView",
+      "box_style": "",
+      "children": [
+       "IPY_MODEL_5307864c2b1143f4b44f3f172611113e",
+       "IPY_MODEL_2e2b6c3f48974ea4aca9b7710a03379e",
+       "IPY_MODEL_aae78f9bd53348bda45967a38736cb78"
+      ],
+      "layout": "IPY_MODEL_34db17e0f28d40d6abafb8acd5dda379"
+     }
+    },
+    "e71f5c7f1d5d4f83b58c68d2fa310d9c": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "f60e9915d2a74ca7bc010d7684f5acf6": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "DescriptionStyleModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "DescriptionStyleModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "StyleView",
+      "description_width": ""
+     }
+    },
+    "f760cd4758334ca9a43fd15612fd808b": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": null,
+      "width": null
+     }
+    },
+    "f9d86ad7fa734f3a857505a542256a3c": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": "hidden",
+      "width": null
+     }
+    },
+    "faa18df899c14e9cac6721253e6c9128": {
+     "model_module": "@jupyter-widgets/controls",
+     "model_module_version": "1.5.0",
+     "model_name": "HTMLModel",
+     "state": {
+      "_dom_classes": [],
+      "_model_module": "@jupyter-widgets/controls",
+      "_model_module_version": "1.5.0",
+      "_model_name": "HTMLModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/controls",
+      "_view_module_version": "1.5.0",
+      "_view_name": "HTMLView",
+      "description": "",
+      "description_tooltip": null,
+      "layout": "IPY_MODEL_1b8aada826a0451bb60c418b19178c8c",
+      "placeholder": "​",
+      "style": "IPY_MODEL_a91916e02e9c424e881e45b3aa978574",
+      "value": "Map:   0%"
+     }
+    },
+    "fc4637a1b37e4e90874c71aa4271ac74": {
+     "model_module": "@jupyter-widgets/base",
+     "model_module_version": "1.2.0",
+     "model_name": "LayoutModel",
+     "state": {
+      "_model_module": "@jupyter-widgets/base",
+      "_model_module_version": "1.2.0",
+      "_model_name": "LayoutModel",
+      "_view_count": null,
+      "_view_module": "@jupyter-widgets/base",
+      "_view_module_version": "1.2.0",
+      "_view_name": "LayoutView",
+      "align_content": null,
+      "align_items": null,
+      "align_self": null,
+      "border": null,
+      "bottom": null,
+      "display": null,
+      "flex": null,
+      "flex_flow": null,
+      "grid_area": null,
+      "grid_auto_columns": null,
+      "grid_auto_flow": null,
+      "grid_auto_rows": null,
+      "grid_column": null,
+      "grid_gap": null,
+      "grid_row": null,
+      "grid_template_areas": null,
+      "grid_template_columns": null,
+      "grid_template_rows": null,
+      "height": null,
+      "justify_content": null,
+      "justify_items": null,
+      "left": null,
+      "margin": null,
+      "max_height": null,
+      "max_width": null,
+      "min_height": null,
+      "min_width": null,
+      "object_fit": null,
+      "object_position": null,
+      "order": null,
+      "overflow": null,
+      "overflow_x": null,
+      "overflow_y": null,
+      "padding": null,
+      "right": null,
+      "top": null,
+      "visibility": "hidden",
+      "width": null
+     }
+    }
+   }
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

peft_lora_seq2seq.ipynb

@@ -0,0 +1,486 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "5f93b7d1",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "===================================BUG REPORT===================================\n",
+      "Welcome to bitsandbytes. For bug reports, please submit your error trace to: https://github.com/TimDettmers/bitsandbytes/issues\n",
+      "For effortless bug reporting copy-paste your error into this form: https://docs.google.com/forms/d/e/1FAIpQLScPB8emS3Thkp66nvqwmjTEgxp8Y9ufuWTzFyr9kJ5AoI47dQ/viewform?usp=sf_link\n",
+      "================================================================================\n",
+      "CUDA SETUP: CUDA runtime path found: /home/sourab/miniconda3/envs/ml/lib/libcudart.so\n",
+      "CUDA SETUP: Highest compute capability among GPUs detected: 7.5\n",
+      "CUDA SETUP: Detected CUDA version 117\n",
+      "CUDA SETUP: Loading binary /home/sourab/miniconda3/envs/ml/lib/python3.10/site-packages/bitsandbytes/libbitsandbytes_cuda117.so...\n"
+     ]
+    }
+   ],
+   "source": [
+    "from transformers import AutoModelForSeq2SeqLM\n",
+    "from peft import get_peft_config, get_peft_model, get_peft_model_state_dict, LoraConfig, TaskType\n",
+    "import torch\n",
+    "from datasets import load_dataset\n",
+    "import os\n",
+    "\n",
+    "os.environ[\"TOKENIZERS_PARALLELISM\"] = \"false\"\n",
+    "from transformers import AutoTokenizer\n",
+    "from torch.utils.data import DataLoader\n",
+    "from transformers import default_data_collator, get_linear_schedule_with_warmup\n",
+    "from tqdm import tqdm\n",
+    "from datasets import load_dataset\n",
+    "\n",
+    "device = \"cuda\"\n",
+    "model_name_or_path = \"bigscience/mt0-large\"\n",
+    "tokenizer_name_or_path = \"bigscience/mt0-large\"\n",
+    "\n",
+    "checkpoint_name = \"financial_sentiment_analysis_lora_v1.pt\"\n",
+    "text_column = \"sentence\"\n",
+    "label_column = \"text_label\"\n",
+    "max_length = 128\n",
+    "lr = 1e-3\n",
+    "num_epochs = 3\n",
+    "batch_size = 8"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "8d0850ac",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# creating model\n",
+    "peft_config = LoraConfig(task_type=TaskType.SEQ_2_SEQ_LM, inference_mode=False, r=8, lora_alpha=32, lora_dropout=0.1)\n",
+    "\n",
+    "model = AutoModelForSeq2SeqLM.from_pretrained(model_name_or_path)\n",
+    "model = get_peft_model(model, peft_config)\n",
+    "model.print_trainable_parameters()\n",
+    "model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "4ee2babf",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Found cached dataset financial_phrasebank (/home/sourab/.cache/huggingface/datasets/financial_phrasebank/sentences_allagree/1.0.0/550bde12e6c30e2674da973a55f57edde5181d53f5a5a34c1531c53f93b7e141)\n"
+     ]
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "3403bf3d718042018b0531848cc30209",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "  0%|          | 0/1 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "d3d5c45e3776469f9560b6eaa9346f8f",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "  0%|          | 0/3 [00:00<?, ?ba/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "e9736f26e9aa450b8d65f95c0b9c81cc",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "  0%|          | 0/1 [00:00<?, ?ba/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/plain": [
+       "{'sentence': \"The 10,000-odd square metre plot that Stockmann has bought for the Nevsky Center shopping center is located on Nevsky Prospect , St Petersburg 's high street , next to the Vosstaniya Square underground station , in the immediate vicinity of Moscow Station .\",\n",
+       " 'label': 1,\n",
+       " 'text_label': 'neutral'}"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# loading dataset\n",
+    "dataset = load_dataset(\"financial_phrasebank\", \"sentences_allagree\")\n",
+    "dataset = dataset[\"train\"].train_test_split(test_size=0.1)\n",
+    "dataset[\"validation\"] = dataset[\"test\"]\n",
+    "del dataset[\"test\"]\n",
+    "\n",
+    "classes = dataset[\"train\"].features[\"label\"].names\n",
+    "dataset = dataset.map(\n",
+    "    lambda x: {\"text_label\": [classes[label] for label in x[\"label\"]]},\n",
+    "    batched=True,\n",
+    "    num_proc=1,\n",
+    ")\n",
+    "\n",
+    "dataset[\"train\"][0]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "adf9608c",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "c460989d4ab24e3f97d81ef040b1d1b4",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Running tokenizer on dataset:   0%|          | 0/3 [00:00<?, ?ba/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "1acc389b08b94f8a87900b9fbdbccce4",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Running tokenizer on dataset:   0%|          | 0/1 [00:00<?, ?ba/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "# data preprocessing\n",
+    "tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)\n",
+    "\n",
+    "\n",
+    "def preprocess_function(examples):\n",
+    "    inputs = examples[text_column]\n",
+    "    targets = examples[label_column]\n",
+    "    model_inputs = tokenizer(inputs, max_length=max_length, padding=\"max_length\", truncation=True, return_tensors=\"pt\")\n",
+    "    labels = tokenizer(targets, max_length=3, padding=\"max_length\", truncation=True, return_tensors=\"pt\")\n",
+    "    labels = labels[\"input_ids\"]\n",
+    "    labels[labels == tokenizer.pad_token_id] = -100\n",
+    "    model_inputs[\"labels\"] = labels\n",
+    "    return model_inputs\n",
+    "\n",
+    "\n",
+    "processed_datasets = dataset.map(\n",
+    "    preprocess_function,\n",
+    "    batched=True,\n",
+    "    num_proc=1,\n",
+    "    remove_columns=dataset[\"train\"].column_names,\n",
+    "    load_from_cache_file=False,\n",
+    "    desc=\"Running tokenizer on dataset\",\n",
+    ")\n",
+    "\n",
+    "train_dataset = processed_datasets[\"train\"]\n",
+    "eval_dataset = processed_datasets[\"validation\"]\n",
+    "\n",
+    "train_dataloader = DataLoader(\n",
+    "    train_dataset, shuffle=True, collate_fn=default_data_collator, batch_size=batch_size, pin_memory=True\n",
+    ")\n",
+    "eval_dataloader = DataLoader(eval_dataset, collate_fn=default_data_collator, batch_size=batch_size, pin_memory=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "f733a3c6",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# optimizer and lr scheduler\n",
+    "optimizer = torch.optim.AdamW(model.parameters(), lr=lr)\n",
+    "lr_scheduler = get_linear_schedule_with_warmup(\n",
+    "    optimizer=optimizer,\n",
+    "    num_warmup_steps=0,\n",
+    "    num_training_steps=(len(train_dataloader) * num_epochs),\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "6b3a4090",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|████████████████████████████████████████████████████████████████████████████████████████| 255/255 [02:21<00:00,  1.81it/s]\n",
+      "100%|██████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:07<00:00,  4.13it/s]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=0: train_ppl=tensor(14.6341, device='cuda:0') train_epoch_loss=tensor(2.6834, device='cuda:0') eval_ppl=tensor(1.0057, device='cuda:0') eval_epoch_loss=tensor(0.0057, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|████████████████████████████████████████████████████████████████████████████████████████| 255/255 [02:00<00:00,  2.11it/s]\n",
+      "100%|██████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:05<00:00,  5.66it/s]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=1: train_ppl=tensor(1.7576, device='cuda:0') train_epoch_loss=tensor(0.5640, device='cuda:0') eval_ppl=tensor(1.0052, device='cuda:0') eval_epoch_loss=tensor(0.0052, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|████████████████████████████████████████████████████████████████████████████████████████| 255/255 [01:33<00:00,  2.74it/s]\n",
+      "100%|██████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:04<00:00,  6.23it/s]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=2: train_ppl=tensor(1.3830, device='cuda:0') train_epoch_loss=tensor(0.3243, device='cuda:0') eval_ppl=tensor(1.0035, device='cuda:0') eval_epoch_loss=tensor(0.0035, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "# training and evaluation\n",
+    "model = model.to(device)\n",
+    "\n",
+    "for epoch in range(num_epochs):\n",
+    "    model.train()\n",
+    "    total_loss = 0\n",
+    "    for step, batch in enumerate(tqdm(train_dataloader)):\n",
+    "        batch = {k: v.to(device) for k, v in batch.items()}\n",
+    "        outputs = model(**batch)\n",
+    "        loss = outputs.loss\n",
+    "        total_loss += loss.detach().float()\n",
+    "        loss.backward()\n",
+    "        optimizer.step()\n",
+    "        lr_scheduler.step()\n",
+    "        optimizer.zero_grad()\n",
+    "\n",
+    "    model.eval()\n",
+    "    eval_loss = 0\n",
+    "    eval_preds = []\n",
+    "    for step, batch in enumerate(tqdm(eval_dataloader)):\n",
+    "        batch = {k: v.to(device) for k, v in batch.items()}\n",
+    "        with torch.no_grad():\n",
+    "            outputs = model(**batch)\n",
+    "        loss = outputs.loss\n",
+    "        eval_loss += loss.detach().float()\n",
+    "        eval_preds.extend(\n",
+    "            tokenizer.batch_decode(torch.argmax(outputs.logits, -1).detach().cpu().numpy(), skip_special_tokens=True)\n",
+    "        )\n",
+    "\n",
+    "    eval_epoch_loss = eval_loss / len(eval_dataloader)\n",
+    "    eval_ppl = torch.exp(eval_epoch_loss)\n",
+    "    train_epoch_loss = total_loss / len(train_dataloader)\n",
+    "    train_ppl = torch.exp(train_epoch_loss)\n",
+    "    print(f\"{epoch=}: {train_ppl=} {train_epoch_loss=} {eval_ppl=} {eval_epoch_loss=}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "6cafa67b",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "accuracy=97.3568281938326 % on the evaluation dataset\n",
+      "eval_preds[:10]=['neutral', 'neutral', 'neutral', 'positive', 'neutral', 'positive', 'positive', 'neutral', 'neutral', 'neutral']\n",
+      "dataset['validation']['text_label'][:10]=['neutral', 'neutral', 'neutral', 'positive', 'neutral', 'positive', 'positive', 'neutral', 'neutral', 'neutral']\n"
+     ]
+    }
+   ],
+   "source": [
+    "# print accuracy\n",
+    "correct = 0\n",
+    "total = 0\n",
+    "for pred, true in zip(eval_preds, dataset[\"validation\"][\"text_label\"]):\n",
+    "    if pred.strip() == true.strip():\n",
+    "        correct += 1\n",
+    "    total += 1\n",
+    "accuracy = correct / total * 100\n",
+    "print(f\"{accuracy=} % on the evaluation dataset\")\n",
+    "print(f\"{eval_preds[:10]=}\")\n",
+    "print(f\"{dataset['validation']['text_label'][:10]=}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "a8de6005",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# saving model\n",
+    "peft_model_id = f\"{model_name_or_path}_{peft_config.peft_type}_{peft_config.task_type}\"\n",
+    "model.save_pretrained(peft_model_id)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "bd20cd4c",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "9,2M\tbigscience/mt0-large_LORA_SEQ_2_SEQ_LM/adapter_model.bin\r\n"
+     ]
+    }
+   ],
+   "source": [
+    "ckpt = f\"{peft_model_id}/adapter_model.bin\"\n",
+    "!du -h $ckpt"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "76c2fc29",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from peft import PeftModel, PeftConfig\n",
+    "\n",
+    "peft_model_id = f\"{model_name_or_path}_{peft_config.peft_type}_{peft_config.task_type}\"\n",
+    "\n",
+    "config = PeftConfig.from_pretrained(peft_model_id)\n",
+    "model = AutoModelForSeq2SeqLM.from_pretrained(config.base_model_name_or_path)\n",
+    "model = PeftModel.from_pretrained(model, peft_model_id)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "37d712ce",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "- Demand for fireplace products was lower than expected , especially in Germany .\n",
+      "{'input_ids': tensor([[  259,   264,   259, 82903,   332,  1090, 10040, 10371,   639,   259,\n",
+      "         19540,  2421,   259, 25505,   259,   261,   259, 21230,   281, 17052,\n",
+      "           259,   260,     1]]), 'attention_mask': tensor([[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]])}\n",
+      "tensor([[    0,   259, 32588,     1]])\n",
+      "['negative']\n"
+     ]
+    }
+   ],
+   "source": [
+    "model.eval()\n",
+    "i = 13\n",
+    "inputs = tokenizer(dataset[\"validation\"][text_column][i], return_tensors=\"pt\")\n",
+    "print(dataset[\"validation\"][text_column][i])\n",
+    "print(inputs)\n",
+    "\n",
+    "with torch.no_grad():\n",
+    "    outputs = model.generate(input_ids=inputs[\"input_ids\"], max_new_tokens=10)\n",
+    "    print(outputs)\n",
+    "    print(tokenizer.batch_decode(outputs.detach().cpu().numpy(), skip_special_tokens=True))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "66c65ea4",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "65e71f78",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "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": "3.10.5"
+  },
+  "vscode": {
+   "interpreter": {
+    "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
+   }
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

peft_lora_seq2seq_accelerate_big_model_inference.ipynb

@@ -0,0 +1,253 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "71fbfca2",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from transformers import AutoModelForSeq2SeqLM\n",
+    "from peft import PeftModel, PeftConfig\n",
+    "import torch\n",
+    "from datasets import load_dataset\n",
+    "import os\n",
+    "from transformers import AutoTokenizer\n",
+    "from torch.utils.data import DataLoader\n",
+    "from transformers import default_data_collator, get_linear_schedule_with_warmup\n",
+    "from tqdm import tqdm\n",
+    "from datasets import load_dataset\n",
+    "\n",
+    "dataset_name = \"twitter_complaints\"\n",
+    "text_column = \"Tweet text\"\n",
+    "label_column = \"text_label\"\n",
+    "batch_size = 8\n",
+    "\n",
+    "peft_model_id = \"smangrul/twitter_complaints_bigscience_T0_3B_LORA_SEQ_2_SEQ_LM\"\n",
+    "config = PeftConfig.from_pretrained(peft_model_id)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "cc55820a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "peft_model_id = \"smangrul/twitter_complaints_bigscience_T0_3B_LORA_SEQ_2_SEQ_LM\"\n",
+    "max_memory = {0: \"6GIB\", 1: \"0GIB\", 2: \"0GIB\", 3: \"0GIB\", 4: \"0GIB\", \"cpu\": \"30GB\"}\n",
+    "config = PeftConfig.from_pretrained(peft_model_id)\n",
+    "model = AutoModelForSeq2SeqLM.from_pretrained(config.base_model_name_or_path, device_map=\"auto\", max_memory=max_memory)\n",
+    "model = PeftModel.from_pretrained(model, peft_model_id, device_map=\"auto\", max_memory=max_memory)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "e1a3648b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from datasets import load_dataset\n",
+    "\n",
+    "dataset = load_dataset(\"ought/raft\", dataset_name)\n",
+    "\n",
+    "classes = [k.replace(\"_\", \" \") for k in dataset[\"train\"].features[\"Label\"].names]\n",
+    "print(classes)\n",
+    "dataset = dataset.map(\n",
+    "    lambda x: {\"text_label\": [classes[label] for label in x[\"Label\"]]},\n",
+    "    batched=True,\n",
+    "    num_proc=1,\n",
+    ")\n",
+    "print(dataset)\n",
+    "dataset[\"train\"][0]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "fe12d4d3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "tokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path)\n",
+    "target_max_length = max([len(tokenizer(class_label)[\"input_ids\"]) for class_label in classes])\n",
+    "\n",
+    "\n",
+    "def preprocess_function(examples):\n",
+    "    inputs = examples[text_column]\n",
+    "    targets = examples[label_column]\n",
+    "    model_inputs = tokenizer(inputs, truncation=True)\n",
+    "    labels = tokenizer(\n",
+    "        targets, max_length=target_max_length, padding=\"max_length\", truncation=True, return_tensors=\"pt\"\n",
+    "    )\n",
+    "    labels = labels[\"input_ids\"]\n",
+    "    labels[labels == tokenizer.pad_token_id] = -100\n",
+    "    model_inputs[\"labels\"] = labels\n",
+    "    return model_inputs\n",
+    "\n",
+    "\n",
+    "processed_datasets = dataset.map(\n",
+    "    preprocess_function,\n",
+    "    batched=True,\n",
+    "    num_proc=1,\n",
+    "    remove_columns=dataset[\"train\"].column_names,\n",
+    "    load_from_cache_file=True,\n",
+    "    desc=\"Running tokenizer on dataset\",\n",
+    ")\n",
+    "\n",
+    "train_dataset = processed_datasets[\"train\"]\n",
+    "eval_dataset = processed_datasets[\"train\"]\n",
+    "test_dataset = processed_datasets[\"test\"]\n",
+    "\n",
+    "\n",
+    "def collate_fn(examples):\n",
+    "    return tokenizer.pad(examples, padding=\"longest\", return_tensors=\"pt\")\n",
+    "\n",
+    "\n",
+    "train_dataloader = DataLoader(\n",
+    "    train_dataset, shuffle=True, collate_fn=collate_fn, batch_size=batch_size, pin_memory=True\n",
+    ")\n",
+    "eval_dataloader = DataLoader(eval_dataset, collate_fn=collate_fn, batch_size=batch_size, pin_memory=True)\n",
+    "test_dataloader = DataLoader(test_dataset, collate_fn=collate_fn, batch_size=batch_size, pin_memory=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "b33be5e6",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "@NYTsupport i have complained a dozen times & yet my papers are still thrown FAR from my door. Why is this so hard to resolve?\n",
+      "{'input_ids': tensor([[25335,  1499,     3,    10,  3320, 12056,   382, 20390,     3,    23,\n",
+      "            43, 25932,     3,     9,  9611,   648,     3,   184,  4624,   117,\n",
+      "           780,    82,  5778,    33,   341,     3, 12618,   377,  4280,    45,\n",
+      "            82,  1365,     5,  1615,    19,    48,    78,   614,    12,  7785,\n",
+      "            58, 16229,     3,    10,     3,     1]]), 'attention_mask': tensor([[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
+      "         1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]])}\n",
+      "tensor([[    0, 10394,     1]], device='cuda:0')\n",
+      "['complaint']\n"
+     ]
+    }
+   ],
+   "source": [
+    "model.eval()\n",
+    "i = 15\n",
+    "inputs = tokenizer(f'{text_column} : {dataset[\"test\"][i][\"Tweet text\"]} Label : ', return_tensors=\"pt\")\n",
+    "print(dataset[\"test\"][i][\"Tweet text\"])\n",
+    "print(inputs)\n",
+    "\n",
+    "with torch.no_grad():\n",
+    "    outputs = model.generate(input_ids=inputs[\"input_ids\"].to(\"cuda\"), max_new_tokens=10)\n",
+    "    print(outputs)\n",
+    "    print(tokenizer.batch_decode(outputs.detach().cpu().numpy(), skip_special_tokens=True))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "b6d6cd5b",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "  0%|                                                                                                    | 0/7 [00:00<?, ?it/s]You're using a T5TokenizerFast tokenizer. Please note that with a fast tokenizer, using the `__call__` method is faster than using a method to encode the text followed by a call to the `pad` method to get a padded encoding.\n",
+      "100%|████████████████████████████████████████████████████████████████████████████████████████████| 7/7 [00:10<00:00,  1.48s/it]\n"
+     ]
+    }
+   ],
+   "source": [
+    "model.eval()\n",
+    "eval_preds = []\n",
+    "for _, batch in enumerate(tqdm(eval_dataloader)):\n",
+    "    batch = {k: v.to(\"cuda\") for k, v in batch.items() if k != \"labels\"}\n",
+    "    with torch.no_grad():\n",
+    "        outputs = model.generate(**batch, max_new_tokens=10)\n",
+    "    preds = outputs.detach().cpu().numpy()\n",
+    "    eval_preds.extend(tokenizer.batch_decode(preds, skip_special_tokens=True))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "61264abe",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "accuracy=100.0\n",
+      "eval_preds[:10]=['no complaint', 'no complaint', 'complaint', 'complaint', 'no complaint', 'no complaint', 'no complaint', 'complaint', 'complaint', 'no complaint']\n",
+      "dataset['train'][label_column][:10]=['no complaint', 'no complaint', 'complaint', 'complaint', 'no complaint', 'no complaint', 'no complaint', 'complaint', 'complaint', 'no complaint']\n"
+     ]
+    }
+   ],
+   "source": [
+    "correct = 0\n",
+    "total = 0\n",
+    "for pred, true in zip(eval_preds, dataset[\"train\"][label_column]):\n",
+    "    if pred.strip() == true.strip():\n",
+    "        correct += 1\n",
+    "    total += 1\n",
+    "accuracy = correct / total * 100\n",
+    "print(f\"{accuracy=}\")\n",
+    "print(f\"{eval_preds[:10]=}\")\n",
+    "print(f\"{dataset['train'][label_column][:10]=}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "a70802a3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "model.eval()\n",
+    "test_preds = []\n",
+    "\n",
+    "for _, batch in enumerate(tqdm(test_dataloader)):\n",
+    "    batch = {k: v for k, v in batch.items() if k != \"labels\"}\n",
+    "    with torch.no_grad():\n",
+    "        outputs = model.generate(**batch, max_new_tokens=10)\n",
+    "    preds = outputs.detach().cpu().numpy()\n",
+    "    test_preds.extend(tokenizer.batch_decode(preds, skip_special_tokens=True))\n",
+    "    if len(test_preds) > 100:\n",
+    "        break\n",
+    "test_preds"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "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": "3.10.5 (v3.10.5:f377153967, Jun  6 2022, 12:36:10) [Clang 13.0.0 (clang-1300.0.29.30)]"
+  },
+  "vscode": {
+   "interpreter": {
+    "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
+   }
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

peft_prefix_tuning_seq2seq.ipynb

@@ -0,0 +1,516 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "5f93b7d1",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "===================================BUG REPORT===================================\n",
+      "Welcome to bitsandbytes. For bug reports, please submit your error trace to: https://github.com/TimDettmers/bitsandbytes/issues\n",
+      "For effortless bug reporting copy-paste your error into this form: https://docs.google.com/forms/d/e/1FAIpQLScPB8emS3Thkp66nvqwmjTEgxp8Y9ufuWTzFyr9kJ5AoI47dQ/viewform?usp=sf_link\n",
+      "================================================================================\n",
+      "CUDA SETUP: CUDA runtime path found: /home/sourab/miniconda3/envs/ml/lib/libcudart.so\n",
+      "CUDA SETUP: Highest compute capability among GPUs detected: 7.5\n",
+      "CUDA SETUP: Detected CUDA version 117\n",
+      "CUDA SETUP: Loading binary /home/sourab/miniconda3/envs/ml/lib/python3.10/site-packages/bitsandbytes/libbitsandbytes_cuda117.so...\n"
+     ]
+    }
+   ],
+   "source": [
+    "from transformers import AutoModelForSeq2SeqLM\n",
+    "from peft import get_peft_config, get_peft_model, get_peft_model_state_dict, PrefixTuningConfig, TaskType\n",
+    "import torch\n",
+    "from datasets import load_dataset\n",
+    "import os\n",
+    "\n",
+    "os.environ[\"TOKENIZERS_PARALLELISM\"] = \"false\"\n",
+    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"3\"\n",
+    "from transformers import AutoTokenizer\n",
+    "from torch.utils.data import DataLoader\n",
+    "from transformers import default_data_collator, get_linear_schedule_with_warmup\n",
+    "from tqdm import tqdm\n",
+    "from datasets import load_dataset\n",
+    "\n",
+    "device = \"cuda\"\n",
+    "model_name_or_path = \"t5-large\"\n",
+    "tokenizer_name_or_path = \"t5-large\"\n",
+    "\n",
+    "checkpoint_name = \"financial_sentiment_analysis_prefix_tuning_v1.pt\"\n",
+    "text_column = \"sentence\"\n",
+    "label_column = \"text_label\"\n",
+    "max_length = 128\n",
+    "lr = 1e-2\n",
+    "num_epochs = 5\n",
+    "batch_size = 8"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "8d0850ac",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# creating model\n",
+    "peft_config = PrefixTuningConfig(task_type=TaskType.SEQ_2_SEQ_LM, inference_mode=False, num_virtual_tokens=20)\n",
+    "\n",
+    "model = AutoModelForSeq2SeqLM.from_pretrained(model_name_or_path)\n",
+    "model = get_peft_model(model, peft_config)\n",
+    "model.print_trainable_parameters()\n",
+    "model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "4ee2babf",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Found cached dataset financial_phrasebank (/home/sourab/.cache/huggingface/datasets/financial_phrasebank/sentences_allagree/1.0.0/550bde12e6c30e2674da973a55f57edde5181d53f5a5a34c1531c53f93b7e141)\n"
+     ]
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "ec4be98991b84181bfa75f8846422b8b",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "  0%|          | 0/1 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "82a6bd694c4f4751a23c370ab51f01a4",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "  0%|          | 0/3 [00:00<?, ?ba/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "3844878631534468a1495e435563e4b0",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "  0%|          | 0/1 [00:00<?, ?ba/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/plain": [
+       "{'sentence': 'Finnish elevators and escalators maker KONE Corporation said on Tuesday ( 18 March ) that it has received a major order from Sir Robert McAlpine to supply all elevators and escalators for the Watermark Place project in the City of London .',\n",
+       " 'label': 2,\n",
+       " 'text_label': 'positive'}"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# loading dataset\n",
+    "dataset = load_dataset(\"financial_phrasebank\", \"sentences_allagree\")\n",
+    "dataset = dataset[\"train\"].train_test_split(test_size=0.1)\n",
+    "dataset[\"validation\"] = dataset[\"test\"]\n",
+    "del dataset[\"test\"]\n",
+    "\n",
+    "classes = dataset[\"train\"].features[\"label\"].names\n",
+    "dataset = dataset.map(\n",
+    "    lambda x: {\"text_label\": [classes[label] for label in x[\"label\"]]},\n",
+    "    batched=True,\n",
+    "    num_proc=1,\n",
+    ")\n",
+    "\n",
+    "dataset[\"train\"][0]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "adf9608c",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/home/sourab/transformers/src/transformers/models/t5/tokenization_t5_fast.py:155: FutureWarning: This tokenizer was incorrectly instantiated with a model max length of 512 which will be corrected in Transformers v5.\n",
+      "For now, this behavior is kept to avoid breaking backwards compatibility when padding/encoding with `truncation is True`.\n",
+      "- Be aware that you SHOULD NOT rely on t5-large automatically truncating your input to 512 when padding/encoding.\n",
+      "- If you want to encode/pad to sequences longer than 512 you can either instantiate this tokenizer with `model_max_length` or pass `max_length` when encoding/padding.\n",
+      "- To avoid this warning, please instantiate this tokenizer with `model_max_length` set to your preferred value.\n",
+      "  warnings.warn(\n"
+     ]
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "4af8c12efb5643659573347509079f3a",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Running tokenizer on dataset:   0%|          | 0/3 [00:00<?, ?ba/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "86033b6257384584afd034075af808cb",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Running tokenizer on dataset:   0%|          | 0/1 [00:00<?, ?ba/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "# data preprocessing\n",
+    "tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)\n",
+    "\n",
+    "\n",
+    "def preprocess_function(examples):\n",
+    "    inputs = examples[text_column]\n",
+    "    targets = examples[label_column]\n",
+    "    model_inputs = tokenizer(inputs, max_length=max_length, padding=\"max_length\", truncation=True, return_tensors=\"pt\")\n",
+    "    labels = tokenizer(targets, max_length=2, padding=\"max_length\", truncation=True, return_tensors=\"pt\")\n",
+    "    labels = labels[\"input_ids\"]\n",
+    "    labels[labels == tokenizer.pad_token_id] = -100\n",
+    "    model_inputs[\"labels\"] = labels\n",
+    "    return model_inputs\n",
+    "\n",
+    "\n",
+    "processed_datasets = dataset.map(\n",
+    "    preprocess_function,\n",
+    "    batched=True,\n",
+    "    num_proc=1,\n",
+    "    remove_columns=dataset[\"train\"].column_names,\n",
+    "    load_from_cache_file=False,\n",
+    "    desc=\"Running tokenizer on dataset\",\n",
+    ")\n",
+    "\n",
+    "train_dataset = processed_datasets[\"train\"]\n",
+    "eval_dataset = processed_datasets[\"validation\"]\n",
+    "\n",
+    "train_dataloader = DataLoader(\n",
+    "    train_dataset, shuffle=True, collate_fn=default_data_collator, batch_size=batch_size, pin_memory=True\n",
+    ")\n",
+    "eval_dataloader = DataLoader(eval_dataset, collate_fn=default_data_collator, batch_size=batch_size, pin_memory=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "f733a3c6",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# optimizer and lr scheduler\n",
+    "optimizer = torch.optim.AdamW(model.parameters(), lr=lr)\n",
+    "lr_scheduler = get_linear_schedule_with_warmup(\n",
+    "    optimizer=optimizer,\n",
+    "    num_warmup_steps=0,\n",
+    "    num_training_steps=(len(train_dataloader) * num_epochs),\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "6b3a4090",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|████████████████████████████████████████████████████████████████████████████████████████| 255/255 [00:49<00:00,  5.15it/s]\n",
+      "100%|██████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:03<00:00,  7.56it/s]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=0: train_ppl=tensor(2760654.5000, device='cuda:0') train_epoch_loss=tensor(14.8310, device='cuda:0') eval_ppl=tensor(1.0124, device='cuda:0') eval_epoch_loss=tensor(0.0124, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|████████████████████████████████████████████████████████████████████████████████████████| 255/255 [00:40<00:00,  6.22it/s]\n",
+      "100%|██████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:05<00:00,  5.05it/s]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=1: train_ppl=tensor(2.7329, device='cuda:0') train_epoch_loss=tensor(1.0054, device='cuda:0') eval_ppl=tensor(1.0081, device='cuda:0') eval_epoch_loss=tensor(0.0080, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|████████████████████████████████████████████████████████████████████████████████████████| 255/255 [00:58<00:00,  4.36it/s]\n",
+      "100%|██████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:05<00:00,  5.05it/s]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=2: train_ppl=tensor(2.1698, device='cuda:0') train_epoch_loss=tensor(0.7747, device='cuda:0') eval_ppl=tensor(1.0057, device='cuda:0') eval_epoch_loss=tensor(0.0057, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|████████████████████████████████████████████████████████████████████████████████████████| 255/255 [00:58<00:00,  4.35it/s]\n",
+      "100%|██████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:05<00:00,  5.06it/s]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=3: train_ppl=tensor(2.0724, device='cuda:0') train_epoch_loss=tensor(0.7287, device='cuda:0') eval_ppl=tensor(1.0051, device='cuda:0') eval_epoch_loss=tensor(0.0051, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|████████████████████████████████████████████████████████████████████████████████████████| 255/255 [01:02<00:00,  4.10it/s]\n",
+      "100%|██████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:06<00:00,  4.74it/s]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=4: train_ppl=tensor(1.7598, device='cuda:0') train_epoch_loss=tensor(0.5652, device='cuda:0') eval_ppl=tensor(1.0047, device='cuda:0') eval_epoch_loss=tensor(0.0047, device='cuda:0')\n"
+     ]
+    }
+   ],
+   "source": [
+    "# training and evaluation\n",
+    "model = model.to(device)\n",
+    "\n",
+    "for epoch in range(num_epochs):\n",
+    "    model.train()\n",
+    "    total_loss = 0\n",
+    "    for step, batch in enumerate(tqdm(train_dataloader)):\n",
+    "        batch = {k: v.to(device) for k, v in batch.items()}\n",
+    "        outputs = model(**batch)\n",
+    "        loss = outputs.loss\n",
+    "        total_loss += loss.detach().float()\n",
+    "        loss.backward()\n",
+    "        optimizer.step()\n",
+    "        lr_scheduler.step()\n",
+    "        optimizer.zero_grad()\n",
+    "\n",
+    "    model.eval()\n",
+    "    eval_loss = 0\n",
+    "    eval_preds = []\n",
+    "    for step, batch in enumerate(tqdm(eval_dataloader)):\n",
+    "        batch = {k: v.to(device) for k, v in batch.items()}\n",
+    "        with torch.no_grad():\n",
+    "            outputs = model(**batch)\n",
+    "        loss = outputs.loss\n",
+    "        eval_loss += loss.detach().float()\n",
+    "        eval_preds.extend(\n",
+    "            tokenizer.batch_decode(torch.argmax(outputs.logits, -1).detach().cpu().numpy(), skip_special_tokens=True)\n",
+    "        )\n",
+    "\n",
+    "    eval_epoch_loss = eval_loss / len(eval_dataloader)\n",
+    "    eval_ppl = torch.exp(eval_epoch_loss)\n",
+    "    train_epoch_loss = total_loss / len(train_dataloader)\n",
+    "    train_ppl = torch.exp(train_epoch_loss)\n",
+    "    print(f\"{epoch=}: {train_ppl=} {train_epoch_loss=} {eval_ppl=} {eval_epoch_loss=}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "6cafa67b",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "accuracy=96.91629955947137 % on the evaluation dataset\n",
+      "eval_preds[:10]=['negative', 'positive', 'neutral', 'neutral', 'neutral', 'neutral', 'neutral', 'neutral', 'neutral', 'neutral']\n",
+      "dataset['validation']['text_label'][:10]=['negative', 'neutral', 'neutral', 'neutral', 'neutral', 'neutral', 'neutral', 'neutral', 'neutral', 'neutral']\n"
+     ]
+    }
+   ],
+   "source": [
+    "# print accuracy\n",
+    "correct = 0\n",
+    "total = 0\n",
+    "for pred, true in zip(eval_preds, dataset[\"validation\"][\"text_label\"]):\n",
+    "    if pred.strip() == true.strip():\n",
+    "        correct += 1\n",
+    "    total += 1\n",
+    "accuracy = correct / total * 100\n",
+    "print(f\"{accuracy=} % on the evaluation dataset\")\n",
+    "print(f\"{eval_preds[:10]=}\")\n",
+    "print(f\"{dataset['validation']['text_label'][:10]=}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "a8de6005",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# saving model\n",
+    "peft_model_id = f\"{model_name_or_path}_{peft_config.peft_type}_{peft_config.task_type}\"\n",
+    "model.save_pretrained(peft_model_id)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "bd20cd4c",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "3,8M\tt5-large_PREFIX_TUNING_SEQ_2_SEQ_LM/adapter_model.bin\r\n"
+     ]
+    }
+   ],
+   "source": [
+    "ckpt = f\"{peft_model_id}/adapter_model.bin\"\n",
+    "!du -h $ckpt"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "76c2fc29",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from peft import PeftModel, PeftConfig\n",
+    "\n",
+    "peft_model_id = f\"{model_name_or_path}_{peft_config.peft_type}_{peft_config.task_type}\"\n",
+    "\n",
+    "config = PeftConfig.from_pretrained(peft_model_id)\n",
+    "model = AutoModelForSeq2SeqLM.from_pretrained(config.base_model_name_or_path)\n",
+    "model = PeftModel.from_pretrained(model, peft_model_id)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 27,
+   "id": "d997f1cc",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Acando AB ( ACANB SS ) fell 8.9 percent to 13.35 kronor , the lowest close since Dec. 11 .\n",
+      "{'input_ids': tensor([[ 4292,   232,    32,     3,  5359,    41,     3, 22029, 14972,     3,\n",
+      "          4256,     3,    61,  4728,  4848,  1298,  1093,    12,  8808,  2469,\n",
+      "             3, 22318,    29,   127,     3,     6,     8,  7402,   885,   437,\n",
+      "          4451,     5,   850,     3,     5,     1]]), 'attention_mask': tensor([[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
+      "         1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]])}\n",
+      "tensor([[   0, 2841,    1]])\n",
+      "['negative']\n"
+     ]
+    }
+   ],
+   "source": [
+    "model.eval()\n",
+    "i = 107\n",
+    "inputs = tokenizer(dataset[\"validation\"][text_column][i], return_tensors=\"pt\")\n",
+    "print(dataset[\"validation\"][text_column][i])\n",
+    "print(inputs)\n",
+    "\n",
+    "with torch.no_grad():\n",
+    "    outputs = model.generate(input_ids=inputs[\"input_ids\"], max_new_tokens=10)\n",
+    "    print(outputs)\n",
+    "    print(tokenizer.batch_decode(outputs.detach().cpu().numpy(), skip_special_tokens=True))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "fb746c1e",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "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": "3.10.5"
+  },
+  "vscode": {
+   "interpreter": {
+    "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
+   }
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

peft_prompt_tuning_seq2seq.ipynb

@@ -0,0 +1,804 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "5f93b7d1",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T08:37:58.711225Z",
+     "start_time": "2023-05-30T08:37:56.881307Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "===================================BUG REPORT===================================\n",
+      "Welcome to bitsandbytes. For bug reports, please run\n",
+      "\n",
+      "python -m bitsandbytes\n",
+      "\n",
+      " and submit this information together with your error trace to: https://github.com/TimDettmers/bitsandbytes/issues\n",
+      "================================================================================\n",
+      "bin /udir/tschilla/anaconda3/envs/peft/lib/python3.9/site-packages/bitsandbytes/libbitsandbytes_cuda117.so\n",
+      "CUDA_SETUP: WARNING! libcudart.so not found in any environmental path. Searching in backup paths...\n",
+      "CUDA SETUP: CUDA runtime path found: /usr/local/cuda/lib64/libcudart.so.11.0\n",
+      "CUDA SETUP: Highest compute capability among GPUs detected: 8.0\n",
+      "CUDA SETUP: Detected CUDA version 117\n",
+      "CUDA SETUP: Loading binary /udir/tschilla/anaconda3/envs/peft/lib/python3.9/site-packages/bitsandbytes/libbitsandbytes_cuda117.so...\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/udir/tschilla/anaconda3/envs/peft/lib/python3.9/site-packages/bitsandbytes/cuda_setup/main.py:149: UserWarning: /udir/tschilla/anaconda3 did not contain ['libcudart.so', 'libcudart.so.11.0', 'libcudart.so.12.0'] as expected! Searching further paths...\n",
+      "  warn(msg)\n",
+      "/udir/tschilla/anaconda3/envs/peft/lib/python3.9/site-packages/bitsandbytes/cuda_setup/main.py:149: UserWarning: WARNING: The following directories listed in your path were found to be non-existent: {PosixPath('Europe/Paris')}\n",
+      "  warn(msg)\n",
+      "/udir/tschilla/anaconda3/envs/peft/lib/python3.9/site-packages/bitsandbytes/cuda_setup/main.py:149: UserWarning: WARNING: The following directories listed in your path were found to be non-existent: {PosixPath('/udir/tschilla/.cache/dotnet_bundle_extract')}\n",
+      "  warn(msg)\n",
+      "/udir/tschilla/anaconda3/envs/peft/lib/python3.9/site-packages/bitsandbytes/cuda_setup/main.py:149: UserWarning: WARNING: The following directories listed in your path were found to be non-existent: {PosixPath('5002'), PosixPath('http'), PosixPath('//127.0.0.1')}\n",
+      "  warn(msg)\n",
+      "/udir/tschilla/anaconda3/envs/peft/lib/python3.9/site-packages/bitsandbytes/cuda_setup/main.py:149: UserWarning: WARNING: The following directories listed in your path were found to be non-existent: {PosixPath('() {  ( alias;\\n eval ${which_declare} ) | /usr/bin/which --tty-only --read-alias --read-functions --show-tilde --show-dot $@\\n}')}\n",
+      "  warn(msg)\n",
+      "/udir/tschilla/anaconda3/envs/peft/lib/python3.9/site-packages/bitsandbytes/cuda_setup/main.py:149: UserWarning: WARNING: The following directories listed in your path were found to be non-existent: {PosixPath('module'), PosixPath('//matplotlib_inline.backend_inline')}\n",
+      "  warn(msg)\n",
+      "/udir/tschilla/anaconda3/envs/peft/lib/python3.9/site-packages/bitsandbytes/cuda_setup/main.py:149: UserWarning: Found duplicate ['libcudart.so', 'libcudart.so.11.0', 'libcudart.so.12.0'] files: {PosixPath('/usr/local/cuda/lib64/libcudart.so.11.0'), PosixPath('/usr/local/cuda/lib64/libcudart.so')}.. We'll flip a coin and try one of these, in order to fail forward.\n",
+      "Either way, this might cause trouble in the future:\n",
+      "If you get `CUDA error: invalid device function` errors, the above might be the cause and the solution is to make sure only one ['libcudart.so', 'libcudart.so.11.0', 'libcudart.so.12.0'] in the paths that we search based on your env.\n",
+      "  warn(msg)\n"
+     ]
+    }
+   ],
+   "source": [
+    "import os\n",
+    "\n",
+    "import torch\n",
+    "from transformers import AutoModelForSeq2SeqLM, AutoTokenizer, default_data_collator, get_linear_schedule_with_warmup\n",
+    "from peft import get_peft_model, PromptTuningConfig, TaskType, PromptTuningInit\n",
+    "from torch.utils.data import DataLoader\n",
+    "from tqdm import tqdm\n",
+    "from datasets import load_dataset\n",
+    "\n",
+    "os.environ[\"TOKENIZERS_PARALLELISM\"] = \"false\"\n",
+    "\n",
+    "device = \"cuda\"\n",
+    "model_name_or_path = \"t5-large\"\n",
+    "tokenizer_name_or_path = \"t5-large\"\n",
+    "\n",
+    "checkpoint_name = \"financial_sentiment_analysis_prompt_tuning_v1.pt\"\n",
+    "text_column = \"sentence\"\n",
+    "label_column = \"text_label\"\n",
+    "max_length = 128\n",
+    "lr = 1\n",
+    "num_epochs = 5\n",
+    "batch_size = 8"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "8d0850ac",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T08:38:12.413984Z",
+     "start_time": "2023-05-30T08:38:04.601042Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "trainable params: 40960 || all params: 737709056 || trainable%: 0.005552324411210698\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/udir/tschilla/anaconda3/envs/peft/lib/python3.9/site-packages/transformers/models/t5/tokenization_t5_fast.py:155: FutureWarning: This tokenizer was incorrectly instantiated with a model max length of 512 which will be corrected in Transformers v5.\n",
+      "For now, this behavior is kept to avoid breaking backwards compatibility when padding/encoding with `truncation is True`.\n",
+      "- Be aware that you SHOULD NOT rely on t5-large automatically truncating your input to 512 when padding/encoding.\n",
+      "- If you want to encode/pad to sequences longer than 512 you can either instantiate this tokenizer with `model_max_length` or pass `max_length` when encoding/padding.\n",
+      "- To avoid this warning, please instantiate this tokenizer with `model_max_length` set to your preferred value.\n",
+      "  warnings.warn(\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "PeftModelForSeq2SeqLM(\n",
+       "  (base_model): T5ForConditionalGeneration(\n",
+       "    (shared): Embedding(32128, 1024)\n",
+       "    (encoder): T5Stack(\n",
+       "      (embed_tokens): Embedding(32128, 1024)\n",
+       "      (block): ModuleList(\n",
+       "        (0): T5Block(\n",
+       "          (layer): ModuleList(\n",
+       "            (0): T5LayerSelfAttention(\n",
+       "              (SelfAttention): T5Attention(\n",
+       "                (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (relative_attention_bias): Embedding(32, 16)\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "            (1): T5LayerFF(\n",
+       "              (DenseReluDense): T5DenseActDense(\n",
+       "                (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
+       "                (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "                (act): ReLU()\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "          )\n",
+       "        )\n",
+       "        (1-23): 23 x T5Block(\n",
+       "          (layer): ModuleList(\n",
+       "            (0): T5LayerSelfAttention(\n",
+       "              (SelfAttention): T5Attention(\n",
+       "                (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "            (1): T5LayerFF(\n",
+       "              (DenseReluDense): T5DenseActDense(\n",
+       "                (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
+       "                (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "                (act): ReLU()\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "          )\n",
+       "        )\n",
+       "      )\n",
+       "      (final_layer_norm): T5LayerNorm()\n",
+       "      (dropout): Dropout(p=0.1, inplace=False)\n",
+       "    )\n",
+       "    (decoder): T5Stack(\n",
+       "      (embed_tokens): Embedding(32128, 1024)\n",
+       "      (block): ModuleList(\n",
+       "        (0): T5Block(\n",
+       "          (layer): ModuleList(\n",
+       "            (0): T5LayerSelfAttention(\n",
+       "              (SelfAttention): T5Attention(\n",
+       "                (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (relative_attention_bias): Embedding(32, 16)\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "            (1): T5LayerCrossAttention(\n",
+       "              (EncDecAttention): T5Attention(\n",
+       "                (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "            (2): T5LayerFF(\n",
+       "              (DenseReluDense): T5DenseActDense(\n",
+       "                (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
+       "                (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "                (act): ReLU()\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "          )\n",
+       "        )\n",
+       "        (1-23): 23 x T5Block(\n",
+       "          (layer): ModuleList(\n",
+       "            (0): T5LayerSelfAttention(\n",
+       "              (SelfAttention): T5Attention(\n",
+       "                (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "            (1): T5LayerCrossAttention(\n",
+       "              (EncDecAttention): T5Attention(\n",
+       "                (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "            (2): T5LayerFF(\n",
+       "              (DenseReluDense): T5DenseActDense(\n",
+       "                (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
+       "                (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "                (act): ReLU()\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "          )\n",
+       "        )\n",
+       "      )\n",
+       "      (final_layer_norm): T5LayerNorm()\n",
+       "      (dropout): Dropout(p=0.1, inplace=False)\n",
+       "    )\n",
+       "    (lm_head): Linear(in_features=1024, out_features=32128, bias=False)\n",
+       "  )\n",
+       "  (prompt_encoder): ModuleDict(\n",
+       "    (default): PromptEmbedding(\n",
+       "      (embedding): Embedding(40, 1024)\n",
+       "    )\n",
+       "  )\n",
+       "  (word_embeddings): Embedding(32128, 1024)\n",
+       ")"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# creating model\n",
+    "peft_config = PromptTuningConfig(\n",
+    "    task_type=TaskType.SEQ_2_SEQ_LM,\n",
+    "    prompt_tuning_init=PromptTuningInit.TEXT,\n",
+    "    num_virtual_tokens=20,\n",
+    "    prompt_tuning_init_text=\"What is the sentiment of this article?\\n\",\n",
+    "    inference_mode=False,\n",
+    "    tokenizer_name_or_path=model_name_or_path,\n",
+    ")\n",
+    "\n",
+    "model = AutoModelForSeq2SeqLM.from_pretrained(model_name_or_path)\n",
+    "model = get_peft_model(model, peft_config)\n",
+    "model.print_trainable_parameters()\n",
+    "model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "4ee2babf",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T08:38:18.759143Z",
+     "start_time": "2023-05-30T08:38:17.881621Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Found cached dataset financial_phrasebank (/data/proxem/huggingface/datasets/financial_phrasebank/sentences_allagree/1.0.0/550bde12e6c30e2674da973a55f57edde5181d53f5a5a34c1531c53f93b7e141)\n"
+     ]
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "fb63f50cb7cb4f5aae10648ba74d6c4e",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "  0%|          | 0/1 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Map:   0%|          | 0/2037 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Map:   0%|          | 0/227 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/plain": [
+       "{'sentence': '`` Lining stone sales were also good in the early autumn , and order books are strong to the end of the year .',\n",
+       " 'label': 2,\n",
+       " 'text_label': 'positive'}"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# loading dataset\n",
+    "dataset = load_dataset(\"financial_phrasebank\", \"sentences_allagree\")\n",
+    "dataset = dataset[\"train\"].train_test_split(test_size=0.1)\n",
+    "dataset[\"validation\"] = dataset[\"test\"]\n",
+    "del dataset[\"test\"]\n",
+    "\n",
+    "classes = dataset[\"train\"].features[\"label\"].names\n",
+    "dataset = dataset.map(\n",
+    "    lambda x: {\"text_label\": [classes[label] for label in x[\"label\"]]},\n",
+    "    batched=True,\n",
+    "    num_proc=1,\n",
+    ")\n",
+    "\n",
+    "dataset[\"train\"][0]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "adf9608c",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T08:38:21.132266Z",
+     "start_time": "2023-05-30T08:38:20.340722Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Running tokenizer on dataset:   0%|          | 0/2037 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Running tokenizer on dataset:   0%|          | 0/227 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "# data preprocessing\n",
+    "tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)\n",
+    "target_max_length = max([len(tokenizer(class_label)[\"input_ids\"]) for class_label in classes])\n",
+    "\n",
+    "\n",
+    "def preprocess_function(examples):\n",
+    "    inputs = examples[text_column]\n",
+    "    targets = examples[label_column]\n",
+    "    model_inputs = tokenizer(inputs, max_length=max_length, padding=\"max_length\", truncation=True, return_tensors=\"pt\")\n",
+    "    labels = tokenizer(\n",
+    "        targets, max_length=target_max_length, padding=\"max_length\", truncation=True, return_tensors=\"pt\"\n",
+    "    )\n",
+    "    labels = labels[\"input_ids\"]\n",
+    "    labels[labels == tokenizer.pad_token_id] = -100\n",
+    "    model_inputs[\"labels\"] = labels\n",
+    "    return model_inputs\n",
+    "\n",
+    "\n",
+    "processed_datasets = dataset.map(\n",
+    "    preprocess_function,\n",
+    "    batched=True,\n",
+    "    num_proc=1,\n",
+    "    remove_columns=dataset[\"train\"].column_names,\n",
+    "    load_from_cache_file=False,\n",
+    "    desc=\"Running tokenizer on dataset\",\n",
+    ")\n",
+    "\n",
+    "train_dataset = processed_datasets[\"train\"]\n",
+    "eval_dataset = processed_datasets[\"validation\"]\n",
+    "\n",
+    "train_dataloader = DataLoader(\n",
+    "    train_dataset, shuffle=True, collate_fn=default_data_collator, batch_size=batch_size, pin_memory=True\n",
+    ")\n",
+    "eval_dataloader = DataLoader(eval_dataset, collate_fn=default_data_collator, batch_size=batch_size, pin_memory=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "f733a3c6",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T08:38:22.907922Z",
+     "start_time": "2023-05-30T08:38:22.901057Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "# optimizer and lr scheduler\n",
+    "optimizer = torch.optim.AdamW(model.parameters(), lr=lr)\n",
+    "lr_scheduler = get_linear_schedule_with_warmup(\n",
+    "    optimizer=optimizer,\n",
+    "    num_warmup_steps=0,\n",
+    "    num_training_steps=(len(train_dataloader) * num_epochs),\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "6b3a4090",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T08:42:29.409070Z",
+     "start_time": "2023-05-30T08:38:50.102263Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 255/255 [00:42<00:00,  6.05it/s]\n",
+      "100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:02<00:00, 14.40it/s]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=0: train_ppl=tensor(8.0846, device='cuda:0') train_epoch_loss=tensor(2.0900, device='cuda:0') eval_ppl=tensor(1.3542, device='cuda:0') eval_epoch_loss=tensor(0.3032, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 255/255 [00:41<00:00,  6.15it/s]\n",
+      "100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:02<00:00, 14.42it/s]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=1: train_ppl=tensor(1.5088, device='cuda:0') train_epoch_loss=tensor(0.4113, device='cuda:0') eval_ppl=tensor(1.2692, device='cuda:0') eval_epoch_loss=tensor(0.2384, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 255/255 [00:41<00:00,  6.18it/s]\n",
+      "100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:02<00:00, 14.45it/s]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=2: train_ppl=tensor(1.5322, device='cuda:0') train_epoch_loss=tensor(0.4267, device='cuda:0') eval_ppl=tensor(1.2065, device='cuda:0') eval_epoch_loss=tensor(0.1877, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 255/255 [00:41<00:00,  6.17it/s]\n",
+      "100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:02<00:00, 14.38it/s]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=3: train_ppl=tensor(1.4475, device='cuda:0') train_epoch_loss=tensor(0.3699, device='cuda:0') eval_ppl=tensor(1.2346, device='cuda:0') eval_epoch_loss=tensor(0.2107, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 255/255 [00:42<00:00,  5.94it/s]\n",
+      "100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 29/29 [00:02<00:00, 14.42it/s]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "epoch=4: train_ppl=tensor(1.3428, device='cuda:0') train_epoch_loss=tensor(0.2948, device='cuda:0') eval_ppl=tensor(1.2041, device='cuda:0') eval_epoch_loss=tensor(0.1857, device='cuda:0')\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "# training and evaluation\n",
+    "model = model.to(device)\n",
+    "\n",
+    "for epoch in range(num_epochs):\n",
+    "    model.train()\n",
+    "    total_loss = 0\n",
+    "    for step, batch in enumerate(tqdm(train_dataloader)):\n",
+    "        batch = {k: v.to(device) for k, v in batch.items()}\n",
+    "        outputs = model(**batch)\n",
+    "        loss = outputs.loss\n",
+    "        total_loss += loss.detach().float()\n",
+    "        loss.backward()\n",
+    "        optimizer.step()\n",
+    "        lr_scheduler.step()\n",
+    "        optimizer.zero_grad()\n",
+    "\n",
+    "    model.eval()\n",
+    "    eval_loss = 0\n",
+    "    eval_preds = []\n",
+    "    for step, batch in enumerate(tqdm(eval_dataloader)):\n",
+    "        batch = {k: v.to(device) for k, v in batch.items()}\n",
+    "        with torch.no_grad():\n",
+    "            outputs = model(**batch)\n",
+    "        loss = outputs.loss\n",
+    "        eval_loss += loss.detach().float()\n",
+    "        eval_preds.extend(\n",
+    "            tokenizer.batch_decode(torch.argmax(outputs.logits, -1).detach().cpu().numpy(), skip_special_tokens=True)\n",
+    "        )\n",
+    "\n",
+    "    eval_epoch_loss = eval_loss / len(eval_dataloader)\n",
+    "    eval_ppl = torch.exp(eval_epoch_loss)\n",
+    "    train_epoch_loss = total_loss / len(train_dataloader)\n",
+    "    train_ppl = torch.exp(train_epoch_loss)\n",
+    "    print(f\"{epoch=}: {train_ppl=} {train_epoch_loss=} {eval_ppl=} {eval_epoch_loss=}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "6cafa67b",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T08:42:42.844671Z",
+     "start_time": "2023-05-30T08:42:42.840447Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "accuracy=85.46255506607929 % on the evaluation dataset\n",
+      "eval_preds[:10]=['neutral', 'neutral', 'neutral', 'neutral', 'neutral', 'positive', 'neutral', 'negative', 'neutral', 'positive']\n",
+      "dataset['validation']['text_label'][:10]=['neutral', 'neutral', 'neutral', 'neutral', 'neutral', 'positive', 'neutral', 'negative', 'positive', 'neutral']\n"
+     ]
+    }
+   ],
+   "source": [
+    "# print accuracy\n",
+    "correct = 0\n",
+    "total = 0\n",
+    "for pred, true in zip(eval_preds, dataset[\"validation\"][\"text_label\"]):\n",
+    "    if pred.strip() == true.strip():\n",
+    "        correct += 1\n",
+    "    total += 1\n",
+    "accuracy = correct / total * 100\n",
+    "print(f\"{accuracy=} % on the evaluation dataset\")\n",
+    "print(f\"{eval_preds[:10]=}\")\n",
+    "print(f\"{dataset['validation']['text_label'][:10]=}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "a8de6005",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T08:42:45.752765Z",
+     "start_time": "2023-05-30T08:42:45.742397Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "# saving model\n",
+    "peft_model_id = f\"{model_name_or_path}_{peft_config.peft_type}_{peft_config.task_type}\"\n",
+    "model.save_pretrained(peft_model_id)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "bd20cd4c",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T08:42:47.660873Z",
+     "start_time": "2023-05-30T08:42:47.488293Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "164K\tt5-large_PROMPT_TUNING_SEQ_2_SEQ_LM/adapter_model.bin\r\n"
+     ]
+    }
+   ],
+   "source": [
+    "ckpt = f\"{peft_model_id}/adapter_model.bin\"\n",
+    "!du -h $ckpt"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "76c2fc29",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T08:42:56.721990Z",
+     "start_time": "2023-05-30T08:42:49.060700Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "from peft import PeftModel, PeftConfig\n",
+    "\n",
+    "peft_model_id = f\"{model_name_or_path}_{peft_config.peft_type}_{peft_config.task_type}\"\n",
+    "\n",
+    "config = PeftConfig.from_pretrained(peft_model_id)\n",
+    "model = AutoModelForSeq2SeqLM.from_pretrained(config.base_model_name_or_path)\n",
+    "model = PeftModel.from_pretrained(model, peft_model_id)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "d997f1cc",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T08:42:59.600916Z",
+     "start_time": "2023-05-30T08:42:58.961468Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Danske Bank is Denmark 's largest bank with 3.5 million customers .\n",
+      "tensor([[ 3039,  1050,  1925,    19, 18001,     3,    31,     7,  2015,  2137,\n",
+      "            28,     3,  9285,   770,   722,     3,     5,     1]])\n",
+      "tensor([[   0, 7163,    1]])\n",
+      "['neutral']\n"
+     ]
+    }
+   ],
+   "source": [
+    "model.eval()\n",
+    "i = 107\n",
+    "input_ids = tokenizer(dataset[\"validation\"][text_column][i], return_tensors=\"pt\").input_ids\n",
+    "print(dataset[\"validation\"][text_column][i])\n",
+    "print(input_ids)\n",
+    "\n",
+    "with torch.no_grad():\n",
+    "    outputs = model.generate(input_ids=input_ids, max_new_tokens=10)\n",
+    "    print(outputs)\n",
+    "    print(tokenizer.batch_decode(outputs.detach().cpu().numpy(), skip_special_tokens=True))"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "peft",
+   "language": "python",
+   "name": "peft"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.16"
+  },
+  "toc": {
+   "base_numbering": 1,
+   "nav_menu": {},
+   "number_sections": true,
+   "sideBar": true,
+   "skip_h1_title": false,
+   "title_cell": "Table of Contents",
+   "title_sidebar": "Contents",
+   "toc_cell": false,
+   "toc_position": {},
+   "toc_section_display": true,
+   "toc_window_display": false
+  },
+  "varInspector": {
+   "cols": {
+    "lenName": 16,
+    "lenType": 16,
+    "lenVar": 40
+   },
+   "kernels_config": {
+    "python": {
+     "delete_cmd_postfix": "",
+     "delete_cmd_prefix": "del ",
+     "library": "var_list.py",
+     "varRefreshCmd": "print(var_dic_list())"
+    },
+    "r": {
+     "delete_cmd_postfix": ") ",
+     "delete_cmd_prefix": "rm(",
+     "library": "var_list.r",
+     "varRefreshCmd": "cat(var_dic_list()) "
+    }
+   },
+   "types_to_exclude": [
+    "module",
+    "function",
+    "builtin_function_or_method",
+    "instance",
+    "_Feature"
+   ],
+   "window_display": false
+  },
+  "vscode": {
+   "interpreter": {
+    "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
+   }
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

peft_prompt_tuning_seq2seq_with_generate.ipynb

@@ -0,0 +1,757 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "5f93b7d1",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T09:49:56.334329Z",
+     "start_time": "2023-05-30T09:49:54.494916Z"
+    }
+   },
+   "outputs": [
+    {
+     "ename": "KeyboardInterrupt",
+     "evalue": "",
+     "output_type": "error",
+     "traceback": [
+      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
+      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
+      "Cell \u001b[0;32mIn[1], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtransformers\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m AutoModelForSeq2SeqLM, Seq2SeqTrainingArguments, Seq2SeqTrainer, GenerationConfig\n\u001b[1;32m      2\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mpeft\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m get_peft_model, PromptTuningInit, PromptTuningConfig, TaskType\n\u001b[1;32m      3\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\n",
+      "File \u001b[0;32m<frozen importlib._bootstrap>:1055\u001b[0m, in \u001b[0;36m_handle_fromlist\u001b[0;34m(module, fromlist, import_, recursive)\u001b[0m\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/transformers/utils/import_utils.py:1076\u001b[0m, in \u001b[0;36m_LazyModule.__getattr__\u001b[0;34m(self, name)\u001b[0m\n\u001b[1;32m   1074\u001b[0m     value \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_get_module(name)\n\u001b[1;32m   1075\u001b[0m \u001b[38;5;28;01melif\u001b[39;00m name \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_class_to_module\u001b[38;5;241m.\u001b[39mkeys():\n\u001b[0;32m-> 1076\u001b[0m     module \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_get_module\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_class_to_module\u001b[49m\u001b[43m[\u001b[49m\u001b[43mname\u001b[49m\u001b[43m]\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1077\u001b[0m     value \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mgetattr\u001b[39m(module, name)\n\u001b[1;32m   1078\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/transformers/utils/import_utils.py:1086\u001b[0m, in \u001b[0;36m_LazyModule._get_module\u001b[0;34m(self, module_name)\u001b[0m\n\u001b[1;32m   1084\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21m_get_module\u001b[39m(\u001b[38;5;28mself\u001b[39m, module_name: \u001b[38;5;28mstr\u001b[39m):\n\u001b[1;32m   1085\u001b[0m     \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[0;32m-> 1086\u001b[0m         \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mimportlib\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mimport_module\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43m.\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m \u001b[49m\u001b[38;5;241;43m+\u001b[39;49m\u001b[43m \u001b[49m\u001b[43mmodule_name\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[38;5;18;43m__name__\u001b[39;49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1087\u001b[0m     \u001b[38;5;28;01mexcept\u001b[39;00m \u001b[38;5;167;01mException\u001b[39;00m \u001b[38;5;28;01mas\u001b[39;00m e:\n\u001b[1;32m   1088\u001b[0m         \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mRuntimeError\u001b[39;00m(\n\u001b[1;32m   1089\u001b[0m             \u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mFailed to import \u001b[39m\u001b[38;5;132;01m{\u001b[39;00m\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m\u001b[38;5;18m__name__\u001b[39m\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m.\u001b[39m\u001b[38;5;132;01m{\u001b[39;00mmodule_name\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m because of the following error (look up to see its\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m   1090\u001b[0m             \u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124m traceback):\u001b[39m\u001b[38;5;130;01m\\n\u001b[39;00m\u001b[38;5;132;01m{\u001b[39;00me\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m   1091\u001b[0m         ) \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01me\u001b[39;00m\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/importlib/__init__.py:127\u001b[0m, in \u001b[0;36mimport_module\u001b[0;34m(name, package)\u001b[0m\n\u001b[1;32m    125\u001b[0m             \u001b[38;5;28;01mbreak\u001b[39;00m\n\u001b[1;32m    126\u001b[0m         level \u001b[38;5;241m+\u001b[39m\u001b[38;5;241m=\u001b[39m \u001b[38;5;241m1\u001b[39m\n\u001b[0;32m--> 127\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43m_bootstrap\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_gcd_import\u001b[49m\u001b[43m(\u001b[49m\u001b[43mname\u001b[49m\u001b[43m[\u001b[49m\u001b[43mlevel\u001b[49m\u001b[43m:\u001b[49m\u001b[43m]\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mpackage\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mlevel\u001b[49m\u001b[43m)\u001b[49m\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/transformers/training_args_seq2seq.py:21\u001b[0m\n\u001b[1;32m     18\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtyping\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m Optional, Union\n\u001b[1;32m     20\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mgeneration\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mconfiguration_utils\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m GenerationConfig\n\u001b[0;32m---> 21\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mtraining_args\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m TrainingArguments\n\u001b[1;32m     22\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mutils\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m add_start_docstrings\n\u001b[1;32m     25\u001b[0m logger \u001b[38;5;241m=\u001b[39m logging\u001b[38;5;241m.\u001b[39mgetLogger(\u001b[38;5;18m__name__\u001b[39m)\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/transformers/training_args.py:29\u001b[0m\n\u001b[1;32m     25\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtyping\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m Any, Dict, List, Optional, Union\n\u001b[1;32m     27\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mpackaging\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m version\n\u001b[0;32m---> 29\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mdebug_utils\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m DebugOption\n\u001b[1;32m     30\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mtrainer_utils\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (\n\u001b[1;32m     31\u001b[0m     EvaluationStrategy,\n\u001b[1;32m     32\u001b[0m     FSDPOption,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     36\u001b[0m     ShardedDDPOption,\n\u001b[1;32m     37\u001b[0m )\n\u001b[1;32m     38\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mutils\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (\n\u001b[1;32m     39\u001b[0m     ExplicitEnum,\n\u001b[1;32m     40\u001b[0m     cached_property,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     53\u001b[0m     requires_backends,\n\u001b[1;32m     54\u001b[0m )\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/transformers/debug_utils.py:21\u001b[0m\n\u001b[1;32m     17\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mutils\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m ExplicitEnum, is_torch_available, logging\n\u001b[1;32m     20\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m is_torch_available():\n\u001b[0;32m---> 21\u001b[0m     \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\n\u001b[1;32m     24\u001b[0m logger \u001b[38;5;241m=\u001b[39m logging\u001b[38;5;241m.\u001b[39mget_logger(\u001b[38;5;18m__name__\u001b[39m)\n\u001b[1;32m     27\u001b[0m \u001b[38;5;28;01mclass\u001b[39;00m \u001b[38;5;21;01mDebugUnderflowOverflow\u001b[39;00m:\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/torch/__init__.py:1465\u001b[0m\n\u001b[1;32m   1463\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m library\n\u001b[1;32m   1464\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m TYPE_CHECKING:\n\u001b[0;32m-> 1465\u001b[0m     \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m _meta_registrations\n\u001b[1;32m   1467\u001b[0m \u001b[38;5;66;03m# Enable CUDA Sanitizer\u001b[39;00m\n\u001b[1;32m   1468\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mTORCH_CUDA_SANITIZER\u001b[39m\u001b[38;5;124m'\u001b[39m \u001b[38;5;129;01min\u001b[39;00m os\u001b[38;5;241m.\u001b[39menviron:\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/torch/_meta_registrations.py:7\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_prims_common\u001b[39;00m \u001b[38;5;28;01mas\u001b[39;00m \u001b[38;5;21;01mutils\u001b[39;00m\n\u001b[1;32m      6\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m Tensor\n\u001b[0;32m----> 7\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_decomp\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m _add_op_to_registry, global_decomposition_table, meta_table\n\u001b[1;32m      8\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_ops\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m OpOverload\n\u001b[1;32m      9\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_prims\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m _elementwise_meta, ELEMENTWISE_PRIM_TYPE_PROMOTION_KIND\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/torch/_decomp/__init__.py:169\u001b[0m\n\u001b[1;32m    165\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m decompositions\n\u001b[1;32m    168\u001b[0m \u001b[38;5;66;03m# populate the table\u001b[39;00m\n\u001b[0;32m--> 169\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_decomp\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mdecompositions\u001b[39;00m\n\u001b[1;32m    170\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_refs\u001b[39;00m\n\u001b[1;32m    172\u001b[0m \u001b[38;5;66;03m# This list was copied from torch/_inductor/decomposition.py\u001b[39;00m\n\u001b[1;32m    173\u001b[0m \u001b[38;5;66;03m# excluding decompositions that results in prim ops\u001b[39;00m\n\u001b[1;32m    174\u001b[0m \u001b[38;5;66;03m# Resulting opset of decomposition is core aten ops\u001b[39;00m\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/torch/_decomp/decompositions.py:10\u001b[0m\n\u001b[1;32m      7\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtyping\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m Callable, cast, Iterable, List, Optional, Tuple, Union\n\u001b[1;32m      9\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\n\u001b[0;32m---> 10\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_prims\u001b[39;00m \u001b[38;5;28;01mas\u001b[39;00m \u001b[38;5;21;01mprims\u001b[39;00m\n\u001b[1;32m     11\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_prims_common\u001b[39;00m \u001b[38;5;28;01mas\u001b[39;00m \u001b[38;5;21;01mutils\u001b[39;00m\n\u001b[1;32m     12\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mnn\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mfunctional\u001b[39;00m \u001b[38;5;28;01mas\u001b[39;00m \u001b[38;5;21;01mF\u001b[39;00m\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/torch/_prims/__init__.py:33\u001b[0m\n\u001b[1;32m     17\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_prims_common\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (\n\u001b[1;32m     18\u001b[0m     check,\n\u001b[1;32m     19\u001b[0m     Dim,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     30\u001b[0m     type_to_dtype,\n\u001b[1;32m     31\u001b[0m )\n\u001b[1;32m     32\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_prims_common\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mwrappers\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m backwards_not_supported\n\u001b[0;32m---> 33\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_subclasses\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mfake_tensor\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m FakeTensor, FakeTensorMode\n\u001b[1;32m     34\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01moverrides\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m handle_torch_function, has_torch_function\n\u001b[1;32m     35\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mutils\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_pytree\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m tree_flatten, tree_map, tree_unflatten\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/torch/_subclasses/__init__.py:3\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\n\u001b[0;32m----> 3\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_subclasses\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mfake_tensor\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (\n\u001b[1;32m      4\u001b[0m     DynamicOutputShapeException,\n\u001b[1;32m      5\u001b[0m     FakeTensor,\n\u001b[1;32m      6\u001b[0m     FakeTensorMode,\n\u001b[1;32m      7\u001b[0m     UnsupportedFakeTensorException,\n\u001b[1;32m      8\u001b[0m )\n\u001b[1;32m     10\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_subclasses\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mfake_utils\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m CrossRefFakeMode\n\u001b[1;32m     12\u001b[0m __all__ \u001b[38;5;241m=\u001b[39m [\n\u001b[1;32m     13\u001b[0m     \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mFakeTensor\u001b[39m\u001b[38;5;124m\"\u001b[39m,\n\u001b[1;32m     14\u001b[0m     \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mFakeTensorMode\u001b[39m\u001b[38;5;124m\"\u001b[39m,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     17\u001b[0m     \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mCrossRefFakeMode\u001b[39m\u001b[38;5;124m\"\u001b[39m,\n\u001b[1;32m     18\u001b[0m ]\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/torch/_subclasses/fake_tensor.py:13\u001b[0m\n\u001b[1;32m     10\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mweakref\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m ReferenceType\n\u001b[1;32m     12\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\n\u001b[0;32m---> 13\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_guards\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m Source\n\u001b[1;32m     14\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_ops\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m OpOverload\n\u001b[1;32m     15\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mtorch\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_prims_common\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (\n\u001b[1;32m     16\u001b[0m     elementwise_dtypes,\n\u001b[1;32m     17\u001b[0m     ELEMENTWISE_TYPE_PROMOTION_KIND,\n\u001b[1;32m     18\u001b[0m     is_float_dtype,\n\u001b[1;32m     19\u001b[0m     is_integer_dtype,\n\u001b[1;32m     20\u001b[0m )\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/torch/_guards.py:14\u001b[0m\n\u001b[1;32m     11\u001b[0m \u001b[38;5;66;03m# TODO(voz): Stolen pattern, not sure why this is the case,\u001b[39;00m\n\u001b[1;32m     12\u001b[0m \u001b[38;5;66;03m# but mypy complains.\u001b[39;00m\n\u001b[1;32m     13\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[0;32m---> 14\u001b[0m     \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m  \u001b[38;5;66;03m# type: ignore[import]\u001b[39;00m\n\u001b[1;32m     15\u001b[0m \u001b[38;5;28;01mexcept\u001b[39;00m \u001b[38;5;167;01mImportError\u001b[39;00m:\n\u001b[1;32m     16\u001b[0m     log\u001b[38;5;241m.\u001b[39mwarning(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mNo sympy found\u001b[39m\u001b[38;5;124m\"\u001b[39m)\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/sympy/__init__.py:74\u001b[0m\n\u001b[1;32m     67\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mlogic\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (to_cnf, to_dnf, to_nnf, And, Or, Not, Xor, Nand, Nor,\n\u001b[1;32m     68\u001b[0m         Implies, Equivalent, ITE, POSform, SOPform, simplify_logic, bool_map,\n\u001b[1;32m     69\u001b[0m         true, false, satisfiable)\n\u001b[1;32m     71\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01massumptions\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (AppliedPredicate, Predicate, AssumptionsContext,\n\u001b[1;32m     72\u001b[0m         assuming, Q, ask, register_handler, remove_handler, refine)\n\u001b[0;32m---> 74\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolys\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (Poly, PurePoly, poly_from_expr, parallel_poly_from_expr,\n\u001b[1;32m     75\u001b[0m         degree, total_degree, degree_list, LC, LM, LT, pdiv, prem, pquo,\n\u001b[1;32m     76\u001b[0m         pexquo, div, rem, quo, exquo, half_gcdex, gcdex, invert,\n\u001b[1;32m     77\u001b[0m         subresultants, resultant, discriminant, cofactors, gcd_list, gcd,\n\u001b[1;32m     78\u001b[0m         lcm_list, lcm, terms_gcd, trunc, monic, content, primitive, compose,\n\u001b[1;32m     79\u001b[0m         decompose, sturm, gff_list, gff, sqf_norm, sqf_part, sqf_list, sqf,\n\u001b[1;32m     80\u001b[0m         factor_list, factor, intervals, refine_root, count_roots, real_roots,\n\u001b[1;32m     81\u001b[0m         nroots, ground_roots, nth_power_roots_poly, cancel, reduced, groebner,\n\u001b[1;32m     82\u001b[0m         is_zero_dimensional, GroebnerBasis, poly, symmetrize, horner,\n\u001b[1;32m     83\u001b[0m         interpolate, rational_interpolate, viete, together,\n\u001b[1;32m     84\u001b[0m         BasePolynomialError, ExactQuotientFailed, PolynomialDivisionFailed,\n\u001b[1;32m     85\u001b[0m         OperationNotSupported, HeuristicGCDFailed, HomomorphismFailed,\n\u001b[1;32m     86\u001b[0m         IsomorphismFailed, ExtraneousFactors, EvaluationFailed,\n\u001b[1;32m     87\u001b[0m         RefinementFailed, CoercionFailed, NotInvertible, NotReversible,\n\u001b[1;32m     88\u001b[0m         NotAlgebraic, DomainError, PolynomialError, UnificationFailed,\n\u001b[1;32m     89\u001b[0m         GeneratorsError, GeneratorsNeeded, ComputationFailed,\n\u001b[1;32m     90\u001b[0m         UnivariatePolynomialError, MultivariatePolynomialError,\n\u001b[1;32m     91\u001b[0m         PolificationFailed, OptionError, FlagError, minpoly,\n\u001b[1;32m     92\u001b[0m         minimal_polynomial, primitive_element, field_isomorphism,\n\u001b[1;32m     93\u001b[0m         to_number_field, isolate, round_two, prime_decomp, prime_valuation,\n\u001b[1;32m     94\u001b[0m         galois_group, itermonomials, Monomial, lex, grlex,\n\u001b[1;32m     95\u001b[0m         grevlex, ilex, igrlex, igrevlex, CRootOf, rootof, RootOf,\n\u001b[1;32m     96\u001b[0m         ComplexRootOf, RootSum, roots, Domain, FiniteField, IntegerRing,\n\u001b[1;32m     97\u001b[0m         RationalField, RealField, ComplexField, PythonFiniteField,\n\u001b[1;32m     98\u001b[0m         GMPYFiniteField, PythonIntegerRing, GMPYIntegerRing, PythonRational,\n\u001b[1;32m     99\u001b[0m         GMPYRationalField, AlgebraicField, PolynomialRing, FractionField,\n\u001b[1;32m    100\u001b[0m         ExpressionDomain, FF_python, FF_gmpy, ZZ_python, ZZ_gmpy, QQ_python,\n\u001b[1;32m    101\u001b[0m         QQ_gmpy, GF, FF, ZZ, QQ, ZZ_I, QQ_I, RR, CC, EX, EXRAW,\n\u001b[1;32m    102\u001b[0m         construct_domain, swinnerton_dyer_poly, cyclotomic_poly,\n\u001b[1;32m    103\u001b[0m         symmetric_poly, random_poly, interpolating_poly, jacobi_poly,\n\u001b[1;32m    104\u001b[0m         chebyshevt_poly, chebyshevu_poly, hermite_poly, hermite_prob_poly,\n\u001b[1;32m    105\u001b[0m         legendre_poly, laguerre_poly, apart, apart_list, assemble_partfrac_list,\n\u001b[1;32m    106\u001b[0m         Options, ring, xring, vring, sring, field, xfield, vfield, sfield)\n\u001b[1;32m    108\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mseries\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (Order, O, limit, Limit, gruntz, series, approximants,\n\u001b[1;32m    109\u001b[0m         residue, EmptySequence, SeqPer, SeqFormula, sequence, SeqAdd, SeqMul,\n\u001b[1;32m    110\u001b[0m         fourier_series, fps, difference_delta, limit_seq)\n\u001b[1;32m    112\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mfunctions\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (factorial, factorial2, rf, ff, binomial,\n\u001b[1;32m    113\u001b[0m         RisingFactorial, FallingFactorial, subfactorial, carmichael,\n\u001b[1;32m    114\u001b[0m         fibonacci, lucas, motzkin, tribonacci, harmonic, bernoulli, bell, euler,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m    133\u001b[0m         Znm, elliptic_k, elliptic_f, elliptic_e, elliptic_pi, beta, mathieus,\n\u001b[1;32m    134\u001b[0m         mathieuc, mathieusprime, mathieucprime, riemann_xi, betainc, betainc_regularized)\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/sympy/polys/__init__.py:78\u001b[0m\n\u001b[1;32m      3\u001b[0m __all__ \u001b[38;5;241m=\u001b[39m [\n\u001b[1;32m      4\u001b[0m     \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mPoly\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mPurePoly\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mpoly_from_expr\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mparallel_poly_from_expr\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mdegree\u001b[39m\u001b[38;5;124m'\u001b[39m,\n\u001b[1;32m      5\u001b[0m     \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mtotal_degree\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mdegree_list\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mLC\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mLM\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mLT\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mpdiv\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mprem\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mpquo\u001b[39m\u001b[38;5;124m'\u001b[39m,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     65\u001b[0m     \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mfield\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mxfield\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mvfield\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124msfield\u001b[39m\u001b[38;5;124m'\u001b[39m\n\u001b[1;32m     66\u001b[0m ]\n\u001b[1;32m     68\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolytools\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (Poly, PurePoly, poly_from_expr,\n\u001b[1;32m     69\u001b[0m         parallel_poly_from_expr, degree, total_degree, degree_list, LC, LM,\n\u001b[1;32m     70\u001b[0m         LT, pdiv, prem, pquo, pexquo, div, rem, quo, exquo, half_gcdex, gcdex,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     75\u001b[0m         count_roots, real_roots, nroots, ground_roots, nth_power_roots_poly,\n\u001b[1;32m     76\u001b[0m         cancel, reduced, groebner, is_zero_dimensional, GroebnerBasis, poly)\n\u001b[0;32m---> 78\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolyfuncs\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (symmetrize, horner, interpolate,\n\u001b[1;32m     79\u001b[0m         rational_interpolate, viete)\n\u001b[1;32m     81\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mrationaltools\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m together\n\u001b[1;32m     83\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolyerrors\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (BasePolynomialError, ExactQuotientFailed,\n\u001b[1;32m     84\u001b[0m         PolynomialDivisionFailed, OperationNotSupported, HeuristicGCDFailed,\n\u001b[1;32m     85\u001b[0m         HomomorphismFailed, IsomorphismFailed, ExtraneousFactors,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m     90\u001b[0m         MultivariatePolynomialError, PolificationFailed, OptionError,\n\u001b[1;32m     91\u001b[0m         FlagError)\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/sympy/polys/polyfuncs.py:10\u001b[0m\n\u001b[1;32m      8\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolys\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolyoptions\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m allowed_flags, build_options\n\u001b[1;32m      9\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolys\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolytools\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m poly_from_expr, Poly\n\u001b[0;32m---> 10\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolys\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mspecialpolys\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (\n\u001b[1;32m     11\u001b[0m     symmetric_poly, interpolating_poly)\n\u001b[1;32m     12\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolys\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mrings\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m sring\n\u001b[1;32m     13\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mutilities\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m numbered_symbols, take, public\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/sympy/polys/specialpolys.py:298\u001b[0m\n\u001b[1;32m    294\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m dmp_mul(f, h, n, K), dmp_mul(g, h, n, K), h\n\u001b[1;32m    296\u001b[0m \u001b[38;5;66;03m# A few useful polynomials from Wang's paper ('78).\u001b[39;00m\n\u001b[0;32m--> 298\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolys\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mrings\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m ring\n\u001b[1;32m    300\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21m_f_0\u001b[39m():\n\u001b[1;32m    301\u001b[0m     R, x, y, z \u001b[38;5;241m=\u001b[39m ring(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mx,y,z\u001b[39m\u001b[38;5;124m\"\u001b[39m, ZZ)\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/sympy/polys/rings.py:30\u001b[0m\n\u001b[1;32m     26\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolys\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolyoptions\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (Domain \u001b[38;5;28;01mas\u001b[39;00m DomainOpt,\n\u001b[1;32m     27\u001b[0m                                      Order \u001b[38;5;28;01mas\u001b[39;00m OrderOpt, build_options)\n\u001b[1;32m     28\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolys\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpolyutils\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (expr_from_dict, _dict_reorder,\n\u001b[1;32m     29\u001b[0m                                    _parallel_dict_from_expr)\n\u001b[0;32m---> 30\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mprinting\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mdefaults\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m DefaultPrinting\n\u001b[1;32m     31\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mutilities\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m public, subsets\n\u001b[1;32m     32\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mutilities\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01miterables\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m is_sequence\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/sympy/printing/__init__.py:5\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[38;5;124;03m\"\"\"Printing subsystem\"\"\"\u001b[39;00m\n\u001b[1;32m      3\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpretty\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m pager_print, pretty, pretty_print, pprint, pprint_use_unicode, pprint_try_use_unicode\n\u001b[0;32m----> 5\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mlatex\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m latex, print_latex, multiline_latex\n\u001b[1;32m      7\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mmathml\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m mathml, print_mathml\n\u001b[1;32m      9\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpython\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m python, print_python\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/sympy/printing/latex.py:18\u001b[0m\n\u001b[1;32m     16\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mcore\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01msympify\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m SympifyError\n\u001b[1;32m     17\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mlogic\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mboolalg\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m true, BooleanTrue, BooleanFalse\n\u001b[0;32m---> 18\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mtensor\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01marray\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m NDimArray\n\u001b[1;32m     20\u001b[0m \u001b[38;5;66;03m# sympy.printing imports\u001b[39;00m\n\u001b[1;32m     21\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mprinting\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mprecedence\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m precedence_traditional\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/sympy/tensor/__init__.py:4\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[38;5;124;03m\"\"\"A module to manipulate symbolic objects with indices including tensors\u001b[39;00m\n\u001b[1;32m      2\u001b[0m \n\u001b[1;32m      3\u001b[0m \u001b[38;5;124;03m\"\"\"\u001b[39;00m\n\u001b[0;32m----> 4\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mindexed\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m IndexedBase, Idx, Indexed\n\u001b[1;32m      5\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mindex_methods\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m get_contraction_structure, get_indices\n\u001b[1;32m      6\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mfunctions\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m shape\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/sympy/tensor/indexed.py:114\u001b[0m\n\u001b[1;32m    112\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mcore\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mlogic\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m fuzzy_bool, fuzzy_not\n\u001b[1;32m    113\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mcore\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01msympify\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m _sympify\n\u001b[0;32m--> 114\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mfunctions\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mspecial\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mtensor_functions\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m KroneckerDelta\n\u001b[1;32m    115\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mmultipledispatch\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m dispatch\n\u001b[1;32m    116\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mutilities\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01miterables\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m is_sequence, NotIterable\n",
+      "File \u001b[0;32m~/anaconda3/envs/peft/lib/python3.9/site-packages/sympy/functions/__init__.py:21\u001b[0m\n\u001b[1;32m     17\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mfunctions\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01melementary\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mtrigonometric\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (sin, cos, tan,\n\u001b[1;32m     18\u001b[0m         sec, csc, cot, sinc, asin, acos, atan, asec, acsc, acot, atan2)\n\u001b[1;32m     19\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mfunctions\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01melementary\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mexponential\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (exp_polar, exp, log,\n\u001b[1;32m     20\u001b[0m         LambertW)\n\u001b[0;32m---> 21\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mfunctions\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01melementary\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mhyperbolic\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (sinh, cosh, tanh, coth,\n\u001b[1;32m     22\u001b[0m         sech, csch, asinh, acosh, atanh, acoth, asech, acsch)\n\u001b[1;32m     23\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mfunctions\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01melementary\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mintegers\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m floor, ceiling, frac\n\u001b[1;32m     24\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01msympy\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mfunctions\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01melementary\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mpiecewise\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m (Piecewise, piecewise_fold,\n\u001b[1;32m     25\u001b[0m                                                   piecewise_exclusive)\n",
+      "File \u001b[0;32m<frozen importlib._bootstrap>:1007\u001b[0m, in \u001b[0;36m_find_and_load\u001b[0;34m(name, import_)\u001b[0m\n",
+      "File \u001b[0;32m<frozen importlib._bootstrap>:986\u001b[0m, in \u001b[0;36m_find_and_load_unlocked\u001b[0;34m(name, import_)\u001b[0m\n",
+      "File \u001b[0;32m<frozen importlib._bootstrap>:680\u001b[0m, in \u001b[0;36m_load_unlocked\u001b[0;34m(spec)\u001b[0m\n",
+      "File \u001b[0;32m<frozen importlib._bootstrap_external>:846\u001b[0m, in \u001b[0;36mexec_module\u001b[0;34m(self, module)\u001b[0m\n",
+      "File \u001b[0;32m<frozen importlib._bootstrap_external>:978\u001b[0m, in \u001b[0;36mget_code\u001b[0;34m(self, fullname)\u001b[0m\n",
+      "File \u001b[0;32m<frozen importlib._bootstrap_external>:647\u001b[0m, in \u001b[0;36m_compile_bytecode\u001b[0;34m(data, name, bytecode_path, source_path)\u001b[0m\n",
+      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
+     ]
+    }
+   ],
+   "source": [
+    "import os\n",
+    "\n",
+    "import torch\n",
+    "from transformers import (\n",
+    "    AutoTokenizer,\n",
+    "    default_data_collator,\n",
+    "    AutoModelForSeq2SeqLM,\n",
+    "    Seq2SeqTrainingArguments,\n",
+    "    Seq2SeqTrainer,\n",
+    "    GenerationConfig,\n",
+    ")\n",
+    "from peft import get_peft_model, PromptTuningInit, PromptTuningConfig, TaskType\n",
+    "from datasets import load_dataset\n",
+    "\n",
+    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"0\"\n",
+    "os.environ[\"TOKENIZERS_PARALLELISM\"] = \"false\"\n",
+    "\n",
+    "device = \"cuda\"\n",
+    "model_name_or_path = \"t5-large\"\n",
+    "tokenizer_name_or_path = \"t5-large\"\n",
+    "\n",
+    "checkpoint_name = \"financial_sentiment_analysis_prefix_tuning_v1.pt\"\n",
+    "text_column = \"sentence\"\n",
+    "label_column = \"text_label\"\n",
+    "max_length = 8\n",
+    "lr = 1e0\n",
+    "num_epochs = 5\n",
+    "batch_size = 8"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "8d0850ac",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T09:50:04.808527Z",
+     "start_time": "2023-05-30T09:49:56.953075Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "trainable params: 40960 || all params: 737709056 || trainable%: 0.005552324411210698\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "PeftModelForSeq2SeqLM(\n",
+       "  (base_model): T5ForConditionalGeneration(\n",
+       "    (shared): Embedding(32128, 1024)\n",
+       "    (encoder): T5Stack(\n",
+       "      (embed_tokens): Embedding(32128, 1024)\n",
+       "      (block): ModuleList(\n",
+       "        (0): T5Block(\n",
+       "          (layer): ModuleList(\n",
+       "            (0): T5LayerSelfAttention(\n",
+       "              (SelfAttention): T5Attention(\n",
+       "                (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (relative_attention_bias): Embedding(32, 16)\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "            (1): T5LayerFF(\n",
+       "              (DenseReluDense): T5DenseActDense(\n",
+       "                (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
+       "                (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "                (act): ReLU()\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "          )\n",
+       "        )\n",
+       "        (1-23): 23 x T5Block(\n",
+       "          (layer): ModuleList(\n",
+       "            (0): T5LayerSelfAttention(\n",
+       "              (SelfAttention): T5Attention(\n",
+       "                (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "            (1): T5LayerFF(\n",
+       "              (DenseReluDense): T5DenseActDense(\n",
+       "                (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
+       "                (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "                (act): ReLU()\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "          )\n",
+       "        )\n",
+       "      )\n",
+       "      (final_layer_norm): T5LayerNorm()\n",
+       "      (dropout): Dropout(p=0.1, inplace=False)\n",
+       "    )\n",
+       "    (decoder): T5Stack(\n",
+       "      (embed_tokens): Embedding(32128, 1024)\n",
+       "      (block): ModuleList(\n",
+       "        (0): T5Block(\n",
+       "          (layer): ModuleList(\n",
+       "            (0): T5LayerSelfAttention(\n",
+       "              (SelfAttention): T5Attention(\n",
+       "                (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (relative_attention_bias): Embedding(32, 16)\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "            (1): T5LayerCrossAttention(\n",
+       "              (EncDecAttention): T5Attention(\n",
+       "                (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "            (2): T5LayerFF(\n",
+       "              (DenseReluDense): T5DenseActDense(\n",
+       "                (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
+       "                (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "                (act): ReLU()\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "          )\n",
+       "        )\n",
+       "        (1-23): 23 x T5Block(\n",
+       "          (layer): ModuleList(\n",
+       "            (0): T5LayerSelfAttention(\n",
+       "              (SelfAttention): T5Attention(\n",
+       "                (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "            (1): T5LayerCrossAttention(\n",
+       "              (EncDecAttention): T5Attention(\n",
+       "                (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "                (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "            (2): T5LayerFF(\n",
+       "              (DenseReluDense): T5DenseActDense(\n",
+       "                (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
+       "                (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
+       "                (dropout): Dropout(p=0.1, inplace=False)\n",
+       "                (act): ReLU()\n",
+       "              )\n",
+       "              (layer_norm): T5LayerNorm()\n",
+       "              (dropout): Dropout(p=0.1, inplace=False)\n",
+       "            )\n",
+       "          )\n",
+       "        )\n",
+       "      )\n",
+       "      (final_layer_norm): T5LayerNorm()\n",
+       "      (dropout): Dropout(p=0.1, inplace=False)\n",
+       "    )\n",
+       "    (lm_head): Linear(in_features=1024, out_features=32128, bias=False)\n",
+       "  )\n",
+       "  (prompt_encoder): ModuleDict(\n",
+       "    (default): PromptEmbedding(\n",
+       "      (embedding): Embedding(40, 1024)\n",
+       "    )\n",
+       "  )\n",
+       "  (word_embeddings): Embedding(32128, 1024)\n",
+       ")"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# creating model\n",
+    "peft_config = peft_config = PromptTuningConfig(\n",
+    "    task_type=TaskType.SEQ_2_SEQ_LM,\n",
+    "    prompt_tuning_init=PromptTuningInit.TEXT,\n",
+    "    num_virtual_tokens=20,\n",
+    "    prompt_tuning_init_text=\"What is the sentiment of this article?\\n\",\n",
+    "    inference_mode=False,\n",
+    "    tokenizer_name_or_path=model_name_or_path,\n",
+    ")\n",
+    "\n",
+    "model = AutoModelForSeq2SeqLM.from_pretrained(model_name_or_path)\n",
+    "model = get_peft_model(model, peft_config)\n",
+    "model.print_trainable_parameters()\n",
+    "model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "4ee2babf",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T09:50:09.224782Z",
+     "start_time": "2023-05-30T09:50:08.172611Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Found cached dataset financial_phrasebank (/data/proxem/huggingface/datasets/financial_phrasebank/sentences_allagree/1.0.0/550bde12e6c30e2674da973a55f57edde5181d53f5a5a34c1531c53f93b7e141)\n"
+     ]
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "d3a799c64a2c43258dc6166c90e2e49f",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "  0%|          | 0/1 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Map:   0%|          | 0/2037 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Map:   0%|          | 0/227 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/plain": [
+       "{'sentence': 'The price of the 10,000 kroon par value bonds was 9663,51 kroons in the primary issue .',\n",
+       " 'label': 1,\n",
+       " 'text_label': 'neutral'}"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# loading dataset\n",
+    "dataset = load_dataset(\"financial_phrasebank\", \"sentences_allagree\")\n",
+    "dataset = dataset[\"train\"].train_test_split(test_size=0.1)\n",
+    "dataset[\"validation\"] = dataset[\"test\"]\n",
+    "del dataset[\"test\"]\n",
+    "\n",
+    "classes = dataset[\"train\"].features[\"label\"].names\n",
+    "dataset = dataset.map(\n",
+    "    lambda x: {\"text_label\": [classes[label] for label in x[\"label\"]]},\n",
+    "    batched=True,\n",
+    "    num_proc=1,\n",
+    ")\n",
+    "\n",
+    "dataset[\"train\"][0]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "adf9608c",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T09:50:12.176663Z",
+     "start_time": "2023-05-30T09:50:11.421273Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/udir/tschilla/anaconda3/envs/peft/lib/python3.9/site-packages/transformers/models/t5/tokenization_t5_fast.py:155: FutureWarning: This tokenizer was incorrectly instantiated with a model max length of 512 which will be corrected in Transformers v5.\n",
+      "For now, this behavior is kept to avoid breaking backwards compatibility when padding/encoding with `truncation is True`.\n",
+      "- Be aware that you SHOULD NOT rely on t5-large automatically truncating your input to 512 when padding/encoding.\n",
+      "- If you want to encode/pad to sequences longer than 512 you can either instantiate this tokenizer with `model_max_length` or pass `max_length` when encoding/padding.\n",
+      "- To avoid this warning, please instantiate this tokenizer with `model_max_length` set to your preferred value.\n",
+      "  warnings.warn(\n"
+     ]
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Running tokenizer on dataset:   0%|          | 0/2037 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Running tokenizer on dataset:   0%|          | 0/227 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "# data preprocessing\n",
+    "tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)\n",
+    "\n",
+    "\n",
+    "def preprocess_function(examples):\n",
+    "    inputs = examples[text_column]\n",
+    "    targets = examples[label_column]\n",
+    "    model_inputs = tokenizer(inputs, max_length=max_length, padding=\"max_length\", truncation=True, return_tensors=\"pt\")\n",
+    "    labels = tokenizer(targets, max_length=2, padding=\"max_length\", truncation=True, return_tensors=\"pt\")\n",
+    "    labels = labels[\"input_ids\"]\n",
+    "    labels[labels == tokenizer.pad_token_id] = -100\n",
+    "    model_inputs[\"labels\"] = labels\n",
+    "    return model_inputs\n",
+    "\n",
+    "\n",
+    "processed_datasets = dataset.map(\n",
+    "    preprocess_function,\n",
+    "    batched=True,\n",
+    "    num_proc=1,\n",
+    "    remove_columns=dataset[\"train\"].column_names,\n",
+    "    load_from_cache_file=False,\n",
+    "    desc=\"Running tokenizer on dataset\",\n",
+    ")\n",
+    "\n",
+    "train_dataset = processed_datasets[\"train\"].shuffle()\n",
+    "eval_dataset = processed_datasets[\"validation\"]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "6b3a4090",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T09:53:10.336984Z",
+     "start_time": "2023-05-30T09:50:14.780995Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/udir/tschilla/anaconda3/envs/peft/lib/python3.9/site-packages/transformers/optimization.py:407: FutureWarning: This implementation of AdamW is deprecated and will be removed in a future version. Use the PyTorch implementation torch.optim.AdamW instead, or set `no_deprecation_warning=True` to disable this warning\n",
+      "  warnings.warn(\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "\n",
+       "    <div>\n",
+       "      \n",
+       "      <progress value='1275' max='1275' style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
+       "      [1275/1275 02:52, Epoch 5/5]\n",
+       "    </div>\n",
+       "    <table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       " <tr style=\"text-align: left;\">\n",
+       "      <th>Epoch</th>\n",
+       "      <th>Training Loss</th>\n",
+       "      <th>Validation Loss</th>\n",
+       "      <th>Accuracy</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <td>1</td>\n",
+       "      <td>4.784800</td>\n",
+       "      <td>0.576933</td>\n",
+       "      <td>0.559471</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <td>2</td>\n",
+       "      <td>0.648200</td>\n",
+       "      <td>0.437575</td>\n",
+       "      <td>0.577093</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <td>3</td>\n",
+       "      <td>0.536200</td>\n",
+       "      <td>0.397857</td>\n",
+       "      <td>0.625551</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <td>4</td>\n",
+       "      <td>0.472200</td>\n",
+       "      <td>0.373160</td>\n",
+       "      <td>0.643172</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <td>5</td>\n",
+       "      <td>0.452500</td>\n",
+       "      <td>0.370234</td>\n",
+       "      <td>0.656388</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table><p>"
+      ],
+      "text/plain": [
+       "<IPython.core.display.HTML object>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/plain": [
+       "TrainOutput(global_step=1275, training_loss=1.3787811279296875, metrics={'train_runtime': 173.3699, 'train_samples_per_second': 58.747, 'train_steps_per_second': 7.354, 'total_flos': 344546979840000.0, 'train_loss': 1.3787811279296875, 'epoch': 5.0})"
+      ]
+     },
+     "execution_count": 5,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# training and evaluation\n",
+    "\n",
+    "\n",
+    "def compute_metrics(eval_preds):\n",
+    "    preds, labels = eval_preds\n",
+    "    preds = tokenizer.batch_decode(preds, skip_special_tokens=True)\n",
+    "    labels = tokenizer.batch_decode(labels, skip_special_tokens=True)\n",
+    "\n",
+    "    correct = 0\n",
+    "    total = 0\n",
+    "    for pred, true in zip(preds, labels):\n",
+    "        if pred.strip() == true.strip():\n",
+    "            correct += 1\n",
+    "        total += 1\n",
+    "    accuracy = correct / total\n",
+    "    return {\"accuracy\": accuracy}\n",
+    "\n",
+    "\n",
+    "training_args = Seq2SeqTrainingArguments(\n",
+    "    \"out\",\n",
+    "    per_device_train_batch_size=batch_size,\n",
+    "    learning_rate=lr,\n",
+    "    num_train_epochs=num_epochs,\n",
+    "    evaluation_strategy=\"epoch\",\n",
+    "    logging_strategy=\"epoch\",\n",
+    "    save_strategy=\"no\",\n",
+    "    report_to=[],\n",
+    "    predict_with_generate=True,\n",
+    "    generation_config=GenerationConfig(max_length=max_length),\n",
+    ")\n",
+    "trainer = Seq2SeqTrainer(\n",
+    "    model=model,\n",
+    "    tokenizer=tokenizer,\n",
+    "    args=training_args,\n",
+    "    train_dataset=train_dataset,\n",
+    "    eval_dataset=eval_dataset,\n",
+    "    data_collator=default_data_collator,\n",
+    "    compute_metrics=compute_metrics,\n",
+    ")\n",
+    "trainer.train()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "a8de6005",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T09:53:13.045146Z",
+     "start_time": "2023-05-30T09:53:13.035612Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "# saving model\n",
+    "peft_model_id = f\"{model_name_or_path}_{peft_config.peft_type}_{peft_config.task_type}\"\n",
+    "model.save_pretrained(peft_model_id)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "bd20cd4c",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T09:53:15.240763Z",
+     "start_time": "2023-05-30T09:53:15.059304Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "164K\tt5-large_PROMPT_TUNING_SEQ_2_SEQ_LM/adapter_model.bin\r\n"
+     ]
+    }
+   ],
+   "source": [
+    "ckpt = f\"{peft_model_id}/adapter_model.bin\"\n",
+    "!du -h $ckpt"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "76c2fc29",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T09:53:25.055105Z",
+     "start_time": "2023-05-30T09:53:17.797989Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "from peft import PeftModel, PeftConfig\n",
+    "\n",
+    "peft_model_id = f\"{model_name_or_path}_{peft_config.peft_type}_{peft_config.task_type}\"\n",
+    "\n",
+    "config = PeftConfig.from_pretrained(peft_model_id)\n",
+    "model = AutoModelForSeq2SeqLM.from_pretrained(config.base_model_name_or_path)\n",
+    "model = PeftModel.from_pretrained(model, peft_model_id)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "d997f1cc",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2023-05-30T09:53:26.777030Z",
+     "start_time": "2023-05-30T09:53:26.013697Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Aspocomp Group , headquartered in Helsinki , Finland , develops interconnection solutions for the electronics industry .\n",
+      "{'input_ids': tensor([[   71,  7990,  7699,  1531,     3,     6,     3, 27630,    16, 29763,\n",
+      "             3,     6, 16458,     3,     6,  1344,     7,  1413, 28102,  1275,\n",
+      "            21,     8, 12800,   681,     3,     5,     1]]), 'attention_mask': tensor([[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
+      "         1, 1, 1]])}\n",
+      "tensor([[   0, 7163,    1]])\n",
+      "['neutral']\n"
+     ]
+    }
+   ],
+   "source": [
+    "model.eval()\n",
+    "i = 107\n",
+    "inputs = tokenizer(dataset[\"validation\"][text_column][i], return_tensors=\"pt\")\n",
+    "print(dataset[\"validation\"][text_column][i])\n",
+    "print(inputs)\n",
+    "\n",
+    "with torch.no_grad():\n",
+    "    outputs = model.generate(input_ids=inputs[\"input_ids\"], max_new_tokens=10)\n",
+    "    print(outputs)\n",
+    "    print(tokenizer.batch_decode(outputs.detach().cpu().numpy(), skip_special_tokens=True))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "fb746c1e",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "peft",
+   "language": "python",
+   "name": "peft"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.16"
+  },
+  "toc": {
+   "base_numbering": 1,
+   "nav_menu": {},
+   "number_sections": true,
+   "sideBar": true,
+   "skip_h1_title": false,
+   "title_cell": "Table of Contents",
+   "title_sidebar": "Contents",
+   "toc_cell": false,
+   "toc_position": {},
+   "toc_section_display": true,
+   "toc_window_display": false
+  },
+  "varInspector": {
+   "cols": {
+    "lenName": 16,
+    "lenType": 16,
+    "lenVar": 40
+   },
+   "kernels_config": {
+    "python": {
+     "delete_cmd_postfix": "",
+     "delete_cmd_prefix": "del ",
+     "library": "var_list.py",
+     "varRefreshCmd": "print(var_dic_list())"
+    },
+    "r": {
+     "delete_cmd_postfix": ") ",
+     "delete_cmd_prefix": "rm(",
+     "library": "var_list.r",
+     "varRefreshCmd": "cat(var_dic_list()) "
+    }
+   },
+   "types_to_exclude": [
+    "module",
+    "function",
+    "builtin_function_or_method",
+    "instance",
+    "_Feature"
+   ],
+   "window_display": false
+  },
+  "vscode": {
+   "interpreter": {
+    "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
+   }
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}