code/04.Finetuning/Phi_3_Inference_Finetuning.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PnA661gpWtFy"
      },
      "source": [
        "## This Phi-3 Fine Tuning Notebook provides instructions on how to:\n",
        "\n",
        "- Fine-tune the Phi-3 mini model using QLoRA and LoRA techniques\n",
        "- Quantize the Phi-3 mini model using BitsandBytes and GPTQ for efficient memory usage\n",
        "- Execute the Phi-3 mini model using Hugging Face's Transformers library\n",
        "- Each section in this notebook is designed to be run independently, allowing you to focus on specific tasks as needed.\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## Fine-Tuning Phi-3\n",
        "Welcome to the guide on fine-tuning the Phi-3 model. Phi-3 is a powerful language model developed by Microsoft, designed to generate human-like text based on the input it receives. Fine-tuning is a process that involves training a pre-trained model (like Phi-3) on a specific task, allowing it to adapt its pre-learned knowledge to the new task.\n",
        "\n",
        "In this guide, we will walk you through the steps of fine-tuning the Phi-3 model. This process can help improve the model's performance on specific tasks or domains that were not covered in its original training data.\n",
        "\n",
        "The fine-tuning process involves several steps, including setting up the environment, loading the pre-trained model, preparing the training data, and finally, training the model on the new data.\n",
        "\n",
        "By the end of this guide, you should have a good understanding of how to fine-tune the Phi-3 model for your specific needs. Let's get started!"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VKMVLej2Fx6T"
      },
      "source": [
        "# Inference"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0BJnJ48qOYSP"
      },
      "source": [
        "This section demonstrates how to run inference using the Phi-3 mini model with Hugging Face's Transformers library, specifically the 16-bit version."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ydzzo_AwOhIg",
        "outputId": "9367c36e-ef45-4188-d99b-b9925bbb5fa0"
      },
      "outputs": [],
      "source": [
        "# This command is run in a bash shell due to '%%bash' at the beginning.\n",
        "# 'pip install -qqq' is used to install Python packages with pip, Python's package installer, in a less verbose mode.\n",
        "# 'accelerate', 'transformers', 'auto-gptq', and 'optimum' are the packages being installed.\n",
        "# These packages are necessary for the fine-tuning and inference of the Phi-3 model.\n",
        "%%bash\n",
        "pip install -qqq accelerate transformers auto-gptq optimum"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-doiRu9FL4o5"
      },
      "source": [
        "Using the original model (16-bit version)\n",
        "\n",
        "It requires 7.4 GB of GPU RAM"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 156,
          "referenced_widgets": [
            "d7d3fc8c16a844a7bdad297bc7a76546",
            "6a08690bc594415e91f0a5bc0148d093",
            "780bca400b83462f84fa965fc5aa21a6",
            "967d69ea7e61481281db28cbd45b897a",
            "f1bb4cba99c94580893910bea8fe3d40",
            "ef4c83fa2850481eb2c8bfc522a63401",
            "ed8e4ab7c0ad4e91a9dda4f493f85d4d",
            "8a7e6cafd50b4d88a3a85019af2b1561",
            "2e51fa700c1944edbb5fbce97de9092b",
            "348b7399643c433b900d92808b7de19a",
            "09c513ba74794a57a82a7f9ccb0042fa"
          ]
        },
        "id": "xBD2kd0wL4LL",
        "outputId": "3a246aa0-1c42-485a-bbd0-163b2064be0a"
      },
      "outputs": [],
      "source": [
        "# Import necessary libraries\n",
        "import torch\n",
        "from transformers import AutoTokenizer, AutoModelForCausalLM, set_seed\n",
        "\n",
        "# Set a seed for reproducibility\n",
        "set_seed(2024)\n",
        "\n",
        "# Define the prompt for the model\n",
        "prompt = \"insert your prompt here\"\n",
        "\n",
        "# Define the model checkpoint simply replace with Phi-3 Model Required\n",
        "model_checkpoint = \"microsoft/Phi-3-mini-4k-instruct\"\n",
        "\n",
        "# Load the tokenizer from the model checkpoint\n",
        "# trust_remote_code=True allows the execution of code from the model files\n",
        "tokenizer = AutoTokenizer.from_pretrained(model_checkpoint,trust_remote_code=True)\n",
        "\n",
        "# Load the model from the model checkpoint\n",
        "# trust_remote_code=True allows the execution of code from the model files\n",
        "# torch_dtype=\"auto\" automatically determines the appropriate torch.dtype\n",
        "# device_map=\"cuda\" specifies that the model should be loaded to the GPU\n",
        "model = AutoModelForCausalLM.from_pretrained(model_checkpoint,\n",
        "                                             trust_remote_code=True,\n",
        "                                             torch_dtype=\"auto\",\n",
        "                                             device_map=\"cuda\")\n",
        "\n",
        "# Tokenize the prompt and move the tensors to the GPU\n",
        "inputs = tokenizer(prompt,\n",
        "                   return_tensors=\"pt\").to(\"cuda\")\n",
        "\n",
        "# Generate a response from the model\n",
        "# do_sample=True means the model will generate text by sampling from the distribution of possible outputs\n",
        "# max_new_tokens=120 limits the length of the generated text to 120 tokens\n",
        "outputs = model.generate(**inputs,\n",
        "                         do_sample=True, max_new_tokens=120)\n",
        "\n",
        "# Decode the generated tokens and remove any special tokens\n",
        "response = tokenizer.decode(outputs[0], skip_special_tokens=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "q7IKK-0xHq6P",
        "outputId": "7b2b1bf1-1c79-48dd-ed8a-71502a1ff2f1"
      },
      "outputs": [],
      "source": [
        "# Print the generated response from the model\n",
        "print(response)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ikXZjUiyKSSb"
      },
      "source": [
        "## Code Generation"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121,
          "referenced_widgets": [
            "b1c02b6a844748d896c68d231e7b49ab",
            "1c91b47f9df94852972be0a7d0899233",
            "2cd72050e4a64e3d8a50af99463e9864",
            "2e2740c3ffc04be18f464b9a347edc7b",
            "72d86577268341dfa656fbabb7807e86",
            "7b659540e031480bad94e2ae0b709663",
            "2e8531f7923e40fea430a46f15da9020",
            "d18d5bea3fb246eaa8960bc14bd52ff8",
            "d9460fe3db8f400a9bdb24454516e1e1",
            "090c57feb7a1451fb14ebb4e3f7fbe22",
            "bc5dc14387824eafad12b9d8e64dde12"
          ]
        },
        "id": "p0kNIpmAKUBT",
        "outputId": "c049cbb2-4228-4c14-e8f8-86db949676ca"
      },
      "outputs": [],
      "source": [
        "# Import necessary libraries\n",
        "import torch\n",
        "from transformers import AutoTokenizer, AutoModelForCausalLM, set_seed\n",
        "\n",
        "# Set a seed for reproducibility\n",
        "set_seed(2024)\n",
        "\n",
        "# Define the prompt for the model. In this case, the prompt is a request for C# code.\n",
        "prompt = \"Write a C# code that reads the content of multiple text files and save the result as CSV\"\n",
        "\n",
        "# Define the model checkpoint and Phi-3 Model Required\n",
        "model_checkpoint = \"microsoft/Phi-3-mini-4k-instruct\"\n",
        "\n",
        "# Load the tokenizer from the model checkpoint\n",
        "# trust_remote_code=True allows the execution of code from the model files\n",
        "tokenizer = AutoTokenizer.from_pretrained(model_checkpoint,trust_remote_code=True)\n",
        "\n",
        "# Load the model from the model checkpoint\n",
        "# trust_remote_code=True allows the execution of code from the model files\n",
        "# torch_dtype=\"auto\" automatically determines the appropriate torch.dtype\n",
        "# device_map=\"cuda\" specifies that the model should be loaded to the GPU\n",
        "model = AutoModelForCausalLM.from_pretrained(model_checkpoint,\n",
        "                                             trust_remote_code=True,\n",
        "                                             torch_dtype=\"auto\",\n",
        "                                             device_map=\"cuda\")\n",
        "\n",
        "# Tokenize the prompt and move the tensors to the GPU\n",
        "inputs = tokenizer(prompt,\n",
        "                   return_tensors=\"pt\").to(\"cuda\")\n",
        "\n",
        "# Generate a response from the model\n",
        "# do_sample=True means the model will generate text by sampling from the distribution of possible outputs\n",
        "# max_new_tokens=200 limits the length of the generated text to 200 tokens\n",
        "outputs = model.generate(**inputs,\n",
        "                         do_sample=True, max_new_tokens=200)\n",
        "\n",
        "# Decode the generated tokens and remove any special tokens\n",
        "response = tokenizer.decode(outputs[0], skip_special_tokens=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Xi-o38J0KhKn",
        "outputId": "2d232ea7-f9d4-470f-c72c-2beb91f55062"
      },
      "outputs": [],
      "source": [
        "# Print the generated response from the model\n",
        "print(response)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eQE0EZNnF5BI"
      },
      "source": [
        "# Quantization"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7JeieNRaOR-p"
      },
      "source": [
        "When the Phi-3 model is fine-tuned using Hugging Face's Transformers and subsequently quantized using 4-bit GPTQ, it necessitates 2.7 GB of GPU RAM.4\n",
        "\n",
        "\"Bitsandbytes NF4\" is a specific configuration or method within the Bitsandbytes library, which is used for quantization. Quantization is a process that reduces the numerical precision of the weights in a model to make it smaller and faster. "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {},
      "outputs": [],
      "source": [
        "Bitsandbytes NF4"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TEOhsfGVSkRQ",
        "outputId": "f1ea90a2-7754-49d2-a565-774e4ffd58dc"
      },
      "outputs": [],
      "source": [
        "# This command is used to install and upgrade necessary Python packages using pip, Python's package installer.\n",
        "# The '!' at the beginning allows you to run shell commands in the notebook.\n",
        "# '-qqq' is used to make the installation process less verbose.\n",
        "# '--upgrade' ensures that if the packages are already installed, they are upgraded to the latest version.\n",
        "# 'transformers', 'bitsandbytes', 'accelerate', and 'datasets' are the packages being installed/upgraded.\n",
        "!pip install -qqq --upgrade transformers bitsandbytes accelerate datasets"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TcoqOxjgFKhH"
      },
      "source": [
        "# Phi-3 Fine-tuning"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "k6_NXV4VbLpL",
        "outputId": "16d79b66-dbcb-4f0f-cbf9-e4cd52d98947"
      },
      "outputs": [],
      "source": [
        "# This command is run in a bash shell due to '%%bash' at the beginning.\n",
        "# 'pip -q install' is used to install Python packages with pip, Python's package installer, in a quiet mode which reduces the output verbosity.\n",
        "# 'huggingface_hub', 'transformers', 'peft', and 'bitsandbytes' are the packages being installed by the first command.\n",
        "# These packages are necessary for the fine-tuning and inference of the Phi-3 model.\n",
        "# 'trl' and 'xformers' are additional packages being installed by the second command.\n",
        "# 'datasets' is a package for providing access to a vast range of datasets, installed by the third command.\n",
        "# The last command ensures that 'torch' version is at least 1.10. If it's already installed but the version is lower, it will be upgraded.\n",
        "%%bash\n",
        "pip -q install huggingface_hub transformers peft bitsandbytes\n",
        "pip -q install trl xformers\n",
        "pip -q install datasets\n",
        "pip install torch>=1.10"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "3EAOk-wZbB3G"
      },
      "outputs": [],
      "source": [
        "# Import necessary modules from the transformers library\n",
        "# AutoModelForCausalLM: This is a class for causal language models. It's used for tasks like text generation.\n",
        "# AutoTokenizer: This class is used for tokenizing input data, a necessary step before feeding data into a model.\n",
        "# TrainingArguments: This class is used for defining the parameters for model training, like learning rate, batch size, etc.\n",
        "# BitsAndBytesConfig: This class is used for configuring the BitsAndBytes quantization process.\n",
        "from transformers import AutoModelForCausalLM, AutoTokenizer, TrainingArguments, BitsAndBytesConfig\n",
        "\n",
        "# Import necessary modules from the huggingface_hub library\n",
        "# ModelCard: This class is used for creating a model card, which provides information about a model.\n",
        "# ModelCardData: This class is used for defining the data of a model card.\n",
        "# HfApi: This class provides an interface to the Hugging Face API, allowing you to interact with the Hugging Face Model Hub.\n",
        "from huggingface_hub import ModelCard, ModelCardData, HfApi\n",
        "\n",
        "# Import the load_dataset function from the datasets library. This function is used for loading datasets.\n",
        "from datasets import load_dataset\n",
        "\n",
        "# Import the Template class from the jinja2 library. This class is used for creating dynamic HTML templates.\n",
        "from jinja2 import Template\n",
        "\n",
        "# Import the SFTTrainer class from the trl library. This class is used for training models.\n",
        "from trl import SFTTrainer\n",
        "\n",
        "# Import the yaml module. This module is used for working with YAML files.\n",
        "import yaml\n",
        "\n",
        "# Import the torch library. This library provides tools for training and running deep learning models.\n",
        "import torch"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "5keZmmhJbBzJ"
      },
      "outputs": [],
      "source": [
        "# MODEL_ID is a string that specifies the identifier of the pre-trained model that will be fine-tuned. \n",
        "# In this case, the model is 'Phi-3-mini-4k-instruct' from Microsoft.\n",
        "MODEL_ID = \"microsoft/Phi-3-mini-4k-instruct\"\n",
        "\n",
        "# NEW_MODEL_NAME is a string that specifies the name of the new model after fine-tuning.\n",
        "# Here, the new model will be named 'opus-samantha-phi-3-mini-4k'.\n",
        "NEW_MODEL_NAME = \"New-Model-phi-3-mini-4k\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "TtO_Q4OwbBvr"
      },
      "outputs": [],
      "source": [
        "# DATASET_NAME is a string that specifies the name of the dataset to be used for fine-tuning.\n",
        "# Replace \"replace with your dataset\" with the actual name of your dataset.\n",
        "DATASET_NAME = \"replace with your dataset\"\n",
        "\n",
        "# SPLIT specifies the portion of the dataset to be used. In this case, the 'train' split of the dataset will be used.\n",
        "SPLIT = \"train\"\n",
        "\n",
        "# MAX_SEQ_LENGTH is an integer that specifies the maximum length of the sequences that the model will handle.\n",
        "MAX_SEQ_LENGTH = 2048\n",
        "\n",
        "# num_train_epochs is an integer that specifies the number of times the training process will go through the entire dataset.\n",
        "num_train_epochs = 1\n",
        "\n",
        "# license is a string that specifies the license under which the model is distributed. In this case, it's Apache License 2.0.\n",
        "license = \"apache-2.0\"\n",
        "\n",
        "# username is a string that specifies the GitHub username of the person who is fine-tuning the model.\n",
        "username = \"GitHubUsername\"\n",
        "\n",
        "# learning_rate is a float that specifies the learning rate to be used during training.\n",
        "learning_rate = 1.41e-5\n",
        "\n",
        "# per_device_train_batch_size is an integer that specifies the number of samples to work through before updating the internal model parameters.\n",
        "per_device_train_batch_size = 4\n",
        "\n",
        "# gradient_accumulation_steps is an integer that specifies the number of steps to accumulate gradients before performing a backward/update pass.\n",
        "gradient_accumulation_steps = 1"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "R2tmmOeVdGez"
      },
      "outputs": [],
      "source": [
        "# This code checks if the current CUDA device supports bfloat16 (Brain Floating Point) computations.\n",
        "# If bfloat16 is supported, it sets the compute_dtype to torch.bfloat16.\n",
        "# If not, it sets the compute_dtype to torch.float16.\n",
        "# bfloat16 and float16 are both half-precision floating-point formats, but bfloat16 provides better performance on some hardware.\n",
        "if torch.cuda.is_bf16_supported():\n",
        "  compute_dtype = torch.bfloat16\n",
        "else:\n",
        "  compute_dtype = torch.float16"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 260,
          "referenced_widgets": [
            "3bfe80fa3d8744fb90fe08d5416b1bbc",
            "ae383e2094594370849cf3409ae56fb5",
            "1d4e2797a9964ed08c8e237c462e90ac",
            "a29197379b264fe7a9d644ab090d2f6c",
            "f631c0450ce14cdeb99d37bdb8c1a3c6",
            "f5402736c73142caaa323cd70aec8ae1",
            "029b6e32a4164189a09381268e8dc57f",
            "11f85ad9fe0b442282d468b1da8c0437",
            "15ec765469404365ae541c926ae917f8",
            "624eeb8629284a10bd98740121076553",
            "0e4e003ca04744dd87bc29a5c77c13a6"
          ]
        },
        "id": "D4KIQBMocYkS",
        "outputId": "98b75683-3a44-4642-c15b-b26f2a563b56"
      },
      "outputs": [],
      "source": [
        "# Load the pre-trained model specified by MODEL_ID using the AutoModelForCausalLM class.\n",
        "# The 'trust_remote_code=True' argument allows the execution of code from the model card (if any).\n",
        "model = AutoModelForCausalLM.from_pretrained(MODEL_ID, trust_remote_code=True)\n",
        "\n",
        "# Load the tokenizer associated with the pre-trained model specified by MODEL_ID using the AutoTokenizer class.\n",
        "# The 'trust_remote_code=True' argument allows the execution of code from the model card (if any).\n",
        "tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, trust_remote_code=True)\n",
        "\n",
        "# Load the dataset specified by DATASET_NAME using the load_dataset function.\n",
        "# The 'split=\"train\"' argument specifies that we want to load the training split of the dataset.\n",
        "dataset = load_dataset(DATASET_NAME, split=\"train\")\n",
        "\n",
        "# Get the ID of the end-of-sentence (EOS) token from the tokenizer and store it in EOS_TOKEN.\n",
        "# This token is used to mark the end of a sentence in the input data.\n",
        "EOS_TOKEN=tokenizer.eos_token_id"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "83DkqVQ5cipS",
        "outputId": "14f46138-6407-44c7-9787-e8d270f5c116"
      },
      "outputs": [],
      "source": [
        "# This line simply prints the contents of the 'dataset' variable.\n",
        "# 'dataset' is expected to be a Dataset object loaded from the 'datasets' library.\n",
        "# Printing it will display information about the dataset such as the number of samples, the features, and a few example data points.\n",
        "dataset"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vU6t6xlJckts"
      },
      "outputs": [],
      "source": [
        "# Select a subset of the data for faster processing\n",
        "dataset = dataset.select(range(100))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iyOxewwpcnlc",
        "outputId": "7eac88bf-4a5e-417c-ca16-696408e9a7c1"
      },
      "outputs": [],
      "source": [
        "# This line simply prints the contents of the 'dataset' variable.\n",
        "# 'dataset' is expected to be a Dataset object loaded from the 'datasets' library.\n",
        "# Printing it will display information about the dataset such as the number of samples, the features, and a few example data points.\n",
        "dataset"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mLYxbQ7RcpUY",
        "outputId": "4c3c302b-607b-49d5-ff73-d57cb683a20e"
      },
      "outputs": [],
      "source": [
        "# Define a function to format the prompts in the dataset.\n",
        "# This function takes a batch of examples and returns a dictionary with the key 'text' and the value being a list of formatted texts.\n",
        "def formatting_prompts_func(examples):\n",
        "    # Extract the conversations from the examples.\n",
        "    convos = examples[\"conversations\"]\n",
        "    # Initialize an empty list to store the formatted texts.\n",
        "    texts = []\n",
        "    # Define a dictionary to map the 'from' field in the conversation to a prefix.\n",
        "    mapper = {\"system\": \"system\\n\", \"human\": \"\\nuser\\n\", \"gpt\": \"\\nassistant\\n\"}\n",
        "    # Define a dictionary to map the 'from' field in the conversation to a suffix.\n",
        "    end_mapper = {\"system\": \"\", \"human\": \"\", \"gpt\": \"\"}\n",
        "    # Iterate over each conversation.\n",
        "    for convo in convos:\n",
        "        # Format the conversation by joining each turn with its corresponding prefix and suffix.\n",
        "        # Append the EOS token to the end of the conversation.\n",
        "        text = \"\".join(f\"{mapper[(turn := x['from'])]} {x['value']}\\n{end_mapper[turn]}\" for x in convo)\n",
        "        texts.append(f\"{text}{EOS_TOKEN}\")\n",
        "    # Return the formatted texts.\n",
        "    return {\"text\": texts}\n",
        "\n",
        "# Apply the formatting function to the dataset using the map method.\n",
        "# The 'batched=True' argument means that the function is applied to batches of examples.\n",
        "dataset = dataset.map(formatting_prompts_func, batched=True)\n",
        "\n",
        "# Print the 9th example from the 'text' field of the dataset to check the result.\n",
        "print(dataset['text'][8])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "124j2HGTctc7"
      },
      "outputs": [],
      "source": [
        "# Create a TrainingArguments object, which is used to define the parameters for model training.\n",
        "\n",
        "args = TrainingArguments(\n",
        "    # 'evaluation_strategy' is set to \"steps\", which means evaluation is done at each logging step.\n",
        "    evaluation_strategy=\"steps\",\n",
        "\n",
        "    # 'per_device_train_batch_size' is set to 7, which means each training batch will contain 7 samples per device.\n",
        "    per_device_train_batch_size=7,\n",
        "\n",
        "    # 'gradient_accumulation_steps' is set to 4, which means gradients are accumulated for 4 steps before performing a backward/update pass.\n",
        "    gradient_accumulation_steps=4,\n",
        "\n",
        "    # 'gradient_checkpointing' is set to True, which means model gradients are stored in memory during training to reduce memory usage.\n",
        "    gradient_checkpointing=True,\n",
        "\n",
        "    # 'learning_rate' is set to 1e-4, which is the learning rate for the optimizer.\n",
        "    learning_rate=1e-4,\n",
        "\n",
        "    # 'fp16' is set to True if bfloat16 is not supported, which means the model will use 16-bit floating point precision for training if possible.\n",
        "    fp16 = not torch.cuda.is_bf16_supported(),\n",
        "\n",
        "    # 'bf16' is set to True if bfloat16 is supported, which means the model will use bfloat16 precision for training if possible.\n",
        "    bf16 = torch.cuda.is_bf16_supported(),\n",
        "\n",
        "    # 'max_steps' is set to -1, which means there is no maximum number of training steps.\n",
        "    max_steps=-1,\n",
        "\n",
        "    # 'num_train_epochs' is set to 3, which means the training process will go through the entire dataset 3 times.\n",
        "    num_train_epochs=3,\n",
        "\n",
        "    # 'save_strategy' is set to \"epoch\", which means the model is saved at the end of each epoch.\n",
        "    save_strategy=\"epoch\",\n",
        "\n",
        "    # 'logging_steps' is set to 10, which means logging is done every 10 steps.\n",
        "    logging_steps=10,\n",
        "\n",
        "    # 'output_dir' is set to NEW_MODEL_NAME, which is the directory where the model and its configuration will be saved.\n",
        "    output_dir=NEW_MODEL_NAME,\n",
        "\n",
        "    # 'optim' is set to \"paged_adamw_32bit\", which is the optimizer to be used for training.\n",
        "    optim=\"paged_adamw_32bit\",\n",
        "\n",
        "    # 'lr_scheduler_type' is set to \"linear\", which means the learning rate scheduler type is linear.\n",
        "    lr_scheduler_type=\"linear\"\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sNn4_hJDCUxd",
        "outputId": "9fa5170f-0e86-4676-d80c-a56886872112"
      },
      "outputs": [],
      "source": [
        "# Create an instance of the SFTTrainer class, which is used to fine-tune the model.\n",
        "\n",
        "trainer = SFTTrainer(\n",
        "    # 'model' is the pre-trained model that will be fine-tuned.\n",
        "    model=model,\n",
        "\n",
        "    # 'args' are the training arguments that specify the training parameters.\n",
        "    args=args,\n",
        "\n",
        "    # 'train_dataset' is the dataset that will be used for training.\n",
        "    train_dataset=dataset,\n",
        "\n",
        "    # 'dataset_text_field' is the key in the dataset that contains the text data.\n",
        "    dataset_text_field=\"text\",\n",
        "\n",
        "    # 'max_seq_length' is the maximum length of the sequences that the model will handle.\n",
        "    max_seq_length=128,\n",
        "\n",
        "    # 'formatting_func' is the function that will be used to format the prompts in the dataset.\n",
        "    formatting_func=formatting_prompts_func\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "_bMqB8QQEdJa"
      },
      "outputs": [],
      "source": [
        "# 'device' is set to 'cuda', which means the CUDA device will be used for computations if available.\n",
        "device = 'cuda'\n",
        "\n",
        "# Import the 'gc' module, which provides an interface to the garbage collector.\n",
        "import gc\n",
        "\n",
        "# Import the 'os' module, which provides a way of using operating system dependent functionality.\n",
        "import os\n",
        "\n",
        "# Call the 'collect' method of the 'gc' module to start a garbage collection, which can help free up memory.\n",
        "gc.collect()\n",
        "\n",
        "# Call the 'empty_cache' method of 'torch.cuda' to release all unused cached memory from PyTorch so that it can be used by other GPU applications.\n",
        "torch.cuda.empty_cache()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        },
        "id": "inuNQbzKcwiB",
        "outputId": "3f5db482-9ba1-4e3f-b52c-ffe1e61667dc"
      },
      "outputs": [],
      "source": [
        "# Call the 'train' method of the 'trainer' object to start the training process.\n",
        "# This method will fine-tune the model on the training dataset according to the parameters specified in the 'args' object.\n",
        "trainer.train()"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "gpuType": "A100",
      "machine_shape": "hm",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "029b6e32a4164189a09381268e8dc57f": {
          "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": ""
          }
        },
        "090c57feb7a1451fb14ebb4e3f7fbe22": {
          "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
          }
        },
        "09c513ba74794a57a82a7f9ccb0042fa": {
          "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": ""
          }
        },
        "0e4e003ca04744dd87bc29a5c77c13a6": {
          "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": ""
          }
        },
        "11f85ad9fe0b442282d468b1da8c0437": {
          "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
          }
        },
        "15ec765469404365ae541c926ae917f8": {
          "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": ""
          }
        },
        "1c91b47f9df94852972be0a7d0899233": {
          "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_7b659540e031480bad94e2ae0b709663",
            "placeholder": "​",
            "style": "IPY_MODEL_2e8531f7923e40fea430a46f15da9020",
            "value": "Loading checkpoint shards: 100%"
          }
        },
        "1d4e2797a9964ed08c8e237c462e90ac": {
          "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_11f85ad9fe0b442282d468b1da8c0437",
            "max": 2,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_15ec765469404365ae541c926ae917f8",
            "value": 2
          }
        },
        "2cd72050e4a64e3d8a50af99463e9864": {
          "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_d18d5bea3fb246eaa8960bc14bd52ff8",
            "max": 2,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_d9460fe3db8f400a9bdb24454516e1e1",
            "value": 2
          }
        },
        "2e2740c3ffc04be18f464b9a347edc7b": {
          "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_090c57feb7a1451fb14ebb4e3f7fbe22",
            "placeholder": "​",
            "style": "IPY_MODEL_bc5dc14387824eafad12b9d8e64dde12",
            "value": " 2/2 [00:03&lt;00:00,  1.43s/it]"
          }
        },
        "2e51fa700c1944edbb5fbce97de9092b": {
          "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": ""
          }
        },
        "2e8531f7923e40fea430a46f15da9020": {
          "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": ""
          }
        },
        "348b7399643c433b900d92808b7de19a": {
          "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
          }
        },
        "3bfe80fa3d8744fb90fe08d5416b1bbc": {
          "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_ae383e2094594370849cf3409ae56fb5",
              "IPY_MODEL_1d4e2797a9964ed08c8e237c462e90ac",
              "IPY_MODEL_a29197379b264fe7a9d644ab090d2f6c"
            ],
            "layout": "IPY_MODEL_f631c0450ce14cdeb99d37bdb8c1a3c6"
          }
        },
        "624eeb8629284a10bd98740121076553": {
          "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
          }
        },
        "6a08690bc594415e91f0a5bc0148d093": {
          "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_ef4c83fa2850481eb2c8bfc522a63401",
            "placeholder": "​",
            "style": "IPY_MODEL_ed8e4ab7c0ad4e91a9dda4f493f85d4d",
            "value": "Loading checkpoint shards: 100%"
          }
        },
        "72d86577268341dfa656fbabb7807e86": {
          "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
          }
        },
        "780bca400b83462f84fa965fc5aa21a6": {
          "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_8a7e6cafd50b4d88a3a85019af2b1561",
            "max": 2,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_2e51fa700c1944edbb5fbce97de9092b",
            "value": 2
          }
        },
        "7b659540e031480bad94e2ae0b709663": {
          "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
          }
        },
        "8a7e6cafd50b4d88a3a85019af2b1561": {
          "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
          }
        },
        "967d69ea7e61481281db28cbd45b897a": {
          "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_348b7399643c433b900d92808b7de19a",
            "placeholder": "​",
            "style": "IPY_MODEL_09c513ba74794a57a82a7f9ccb0042fa",
            "value": " 2/2 [00:03&lt;00:00,  1.42s/it]"
          }
        },
        "a29197379b264fe7a9d644ab090d2f6c": {
          "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_624eeb8629284a10bd98740121076553",
            "placeholder": "​",
            "style": "IPY_MODEL_0e4e003ca04744dd87bc29a5c77c13a6",
            "value": " 2/2 [00:37&lt;00:00, 18.01s/it]"
          }
        },
        "ae383e2094594370849cf3409ae56fb5": {
          "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_f5402736c73142caaa323cd70aec8ae1",
            "placeholder": "​",
            "style": "IPY_MODEL_029b6e32a4164189a09381268e8dc57f",
            "value": "Loading checkpoint shards: 100%"
          }
        },
        "b1c02b6a844748d896c68d231e7b49ab": {
          "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_1c91b47f9df94852972be0a7d0899233",
              "IPY_MODEL_2cd72050e4a64e3d8a50af99463e9864",
              "IPY_MODEL_2e2740c3ffc04be18f464b9a347edc7b"
            ],
            "layout": "IPY_MODEL_72d86577268341dfa656fbabb7807e86"
          }
        },
        "bc5dc14387824eafad12b9d8e64dde12": {
          "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": ""
          }
        },
        "d18d5bea3fb246eaa8960bc14bd52ff8": {
          "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
          }
        },
        "d7d3fc8c16a844a7bdad297bc7a76546": {
          "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_6a08690bc594415e91f0a5bc0148d093",
              "IPY_MODEL_780bca400b83462f84fa965fc5aa21a6",
              "IPY_MODEL_967d69ea7e61481281db28cbd45b897a"
            ],
            "layout": "IPY_MODEL_f1bb4cba99c94580893910bea8fe3d40"
          }
        },
        "d9460fe3db8f400a9bdb24454516e1e1": {
          "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": ""
          }
        },
        "ed8e4ab7c0ad4e91a9dda4f493f85d4d": {
          "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": ""
          }
        },
        "ef4c83fa2850481eb2c8bfc522a63401": {
          "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
          }
        },
        "f1bb4cba99c94580893910bea8fe3d40": {
          "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
          }
        },
        "f5402736c73142caaa323cd70aec8ae1": {
          "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
          }
        },
        "f631c0450ce14cdeb99d37bdb8c1a3c6": {
          "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
          }
        }
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}