{ "cells": [ { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "$$\n", "\\newcommand{\\mat}[1]{\\boldsymbol {#1}}\n", "\\newcommand{\\mattr}[1]{\\boldsymbol {#1}^\\top}\n", "\\newcommand{\\matinv}[1]{\\boldsymbol {#1}^{-1}}\n", "\\newcommand{\\vec}[1]{\\boldsymbol {#1}}\n", "\\newcommand{\\vectr}[1]{\\boldsymbol {#1}^\\top}\n", "\\newcommand{\\rvar}[1]{\\mathrm {#1}}\n", "\\newcommand{\\rvec}[1]{\\boldsymbol{\\mathrm{#1}}}\n", "\\newcommand{\\diag}{\\mathop{\\mathrm {diag}}}\n", "\\newcommand{\\set}[1]{\\mathbb {#1}}\n", "\\newcommand{\\norm}[1]{\\left\\lVert#1\\right\\rVert}\n", "\\newcommand{\\pderiv}[2]{\\frac{\\partial #1}{\\partial #2}}\n", "\\newcommand{\\bb}[1]{\\boldsymbol{#1}}\n", "\\newcommand{\\Tr}[0]{^\\top}\n", "\\newcommand{\\softmax}[1]{\\mathrm{softmax}\\left({#1}\\right)}\n", "$$\n", "\n", "# CS236781: Deep Learning\n", "# Tutorial 10: Transformers\n", "\n", "\n", "
" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "## Introduction\n", "\n", "In this tutorial, we will cover:\n", "\n", "- Attention is all you need\n", "- Some of the characters of sesame street\n", "- GPT\n", "\n", "\n", "
" ] }, { "cell_type": "code", "execution_count": 1, "metadata": { "execution": { "iopub.execute_input": "2022-03-24T07:25:15.707910Z", "iopub.status.busy": "2022-03-24T07:25:15.707373Z", "iopub.status.idle": "2022-03-24T07:25:16.993157Z", "shell.execute_reply": "2022-03-24T07:25:16.992767Z" }, "slideshow": { "slide_type": "subslide" } }, "outputs": [], "source": [ "# Setup\n", "%matplotlib inline\n", "import os\n", "import sys\n", "import math\n", "import time\n", "import tqdm\n", "import torch\n", "import matplotlib.pyplot as plt\n", "import numpy as np\n", "import warnings\n", "warnings.simplefilter(\"ignore\")\n", "\n", "# pytorch\n", "import torch\n", "import torch.nn as nn\n", "import torchtext\n", "import torchtext.data as data\n", "import torchtext.datasets as datasets\n", "import torch.nn.functional as F\n", "import torch.utils.data as data\n", "import torch.optim as optim" ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "execution": { "iopub.execute_input": "2022-03-24T07:25:16.995346Z", "iopub.status.busy": "2022-03-24T07:25:16.995233Z", "iopub.status.idle": "2022-03-24T07:25:17.008877Z", "shell.execute_reply": "2022-03-24T07:25:17.008540Z" }, "slideshow": { "slide_type": "fragment" } }, "outputs": [], "source": [ "plt.rcParams['font.size'] = 20\n", "data_dir = os.path.expanduser('~/.pytorch-datasets')\n", "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "## Theory reminders" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "skip" } }, "source": [ "### Attention" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "skip" } }, "source": [ "We've talked about the perpues of attention and the motivation resreach for deep learning attention mechanisem\n", "\n", "
\n" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "skip" } }, "source": [ "We also defined the scaled dot product attention as:\n", "\n", "$$\n", "\\begin{align}\n", "\\mat{B} &= \\frac{1}{\\sqrt{d}} \\mat{Q}\\mattr{K} \\ \\in\\set{R}^{m\\times n} \\\\\n", "\\mat{A} &= \\softmax{\\mat{B}}_{\\mathrm{dim}=1}, \\in\\set{R}^{m\\times n} \\\\\n", "\\mat{Y} &= \\mat{A}\\mat{V} \\ \\in\\set{R}^{m\\times d_v}.\n", "\\end{align}\n", "$$\n", "\n", "where `K`,`Q` and `V` for the self attention came as projections of the same input sequnce\n", "\n", "$$\n", "\\begin{align*}\n", "\\vec{q}_{i} &= \\mat{W}_{xq}\\vec{x}_{i} &\n", "\\vec{k}_{i} &= \\mat{W}_{xk}\\vec{x}_{i} &\n", "\\vec{v}_{i} &= \\mat{W}_{xv}\\vec{x}_{i} \n", "\\end{align*}\n", "$$" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "
\n" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "### Encoder-decoder architectures" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "A common architecture used in many tasks is the encoder-decoder pattern.\n", "\n", "- The **encoder** maps the input to some latent representation, usually of a low dimension.\n", "- The **decoder** applies a different mapping, from the latent space to some other space (sometimes back to the input space).\n", "\n", "We implemented encoder-decoder with and without attention" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "A | B\n", "- | - \n", " |" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "## All you need is love, or attention" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "### Before the tresformers\n", "\n", "Before the transformer architecture, each family of problems, was ruled by a diffrent architecture, commonly sharing only few characteristics:\n", "\n" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "
" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "This days, they all start to share a little bit more...\n", "\n", "
\n", "\n", "[Attention Is All You Need](https://arxiv.org/abs/1706.03762)" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "Recap of the attention mechanisem:\n", "\n", "
\n", "\n", "image by jay alammar\n" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "In a matrix stack:\n", "\n", "\n", "
\n", "\n", "\n", "\n", "
\n", "\n", "\n", "image by jay alammar" ] }, { "cell_type": "code", "execution_count": 3, "metadata": { "slideshow": { "slide_type": "subslide" } }, "outputs": [], "source": [ "def scaled_dot_product(q, k, v, mask=None):\n", " d_k = q.size()[-1]\n", " attn_logits = torch.matmul(q, k.transpose(-2, -1))\n", " attn_logits = attn_logits / math.sqrt(d_k)\n", " if mask is not None:\n", " attn_logits = attn_logits.masked_fill(mask == 0, -9e15)\n", " attention = F.softmax(attn_logits, dim=-1)\n", " values = torch.matmul(attention, v)\n", " return values, attention" ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "slideshow": { "slide_type": "fragment" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Q\n", " tensor([[ 0.0842, 0.6029],\n", " [ 2.2176, -0.9276],\n", " [-1.0193, 0.1904]])\n", "K\n", " tensor([[ 0.4362, 0.8059],\n", " [ 1.1931, -0.3469],\n", " [ 0.5787, 0.5497]])\n", "V\n", " tensor([[ 0.8489, 0.3791],\n", " [-0.9486, 0.4248],\n", " [ 1.4218, -2.1640]])\n", "Values\n", " tensor([[ 0.6004, -0.5132],\n", " [-0.3879, 0.0152],\n", " [ 0.6832, -0.5476]])\n", "Attention\n", " tensor([[0.3931, 0.2515, 0.3554],\n", " [0.1057, 0.7379, 0.1564],\n", " [0.4223, 0.2095, 0.3682]])\n" ] } ], "source": [ "seq_len, d_k = 3, 2\n", "q = torch.randn(seq_len, d_k)\n", "k = torch.randn(seq_len, d_k)\n", "v = torch.randn(seq_len, d_k)\n", "values, attention = scaled_dot_product(q, k, v)\n", "print(\"Q\\n\", q)\n", "print(\"K\\n\", k)\n", "print(\"V\\n\", v)\n", "print(\"Values\\n\", values)\n", "print(\"Attention\\n\", attention)" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "The paper further refined the self-attention layer by adding a mechanism called “multi-headed” attention. (We're all about childhood today..)\n", "\n", "\n", " \n", "\n", "\n", "\n" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "How do we do that?\n", "\n", " \n", "\n", "Can it improve the performance of the attention layer?" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "* It expands the model’s ability to focus on different positions. For eample `The child didn’t cross the street because it was too busy`, diffrent heads can give diffrent interpertation. once the `child` was too busy, and another where the `street` was busy. That way two heads have diffrent meaning and we want the model to capture the ambiguity in language.\n", "\n", "\n", "* It learn diffrent attention score (NxN!), as it used diffrent `Q`,`K`,`V`, when we look at meaning, sometimes we give more attention to diffrent parts of the sentence. If i ask you who didn't cross the road, the focus of the answer is on `The child`, while when i ask why, the focus is on `it was busy`. We want a mechanisem that can answer several questions.\n" ] }, { "cell_type": "code", "execution_count": 5, "metadata": { "slideshow": { "slide_type": "subslide" } }, "outputs": [], "source": [ "class MultiheadAttention(nn.Module):\n", " \n", " def __init__(self, input_dim, embed_dim, num_heads):\n", " super().__init__()\n", " assert embed_dim % num_heads == 0, \"Embedding dimension must be 0 modulo number of heads.\"\n", " #this is the q,k,v total embeding dim\n", " self.embed_dim = embed_dim\n", " self.num_heads = num_heads\n", " self.head_dim = embed_dim // num_heads\n", " \n", " # Stack all weight matrices 1...h together for efficiency\n", " self.qkv_proj = nn.Linear(input_dim, 3*embed_dim)\n", " self.o_proj = nn.Linear(embed_dim, embed_dim)\n", " \n", " self._reset_parameters()\n", "\n", " def _reset_parameters(self):\n", " # Original Transformer initialization, see PyTorch documentation of the paper if you like to\n", " nn.init.xavier_uniform_(self.qkv_proj.weight)\n", " self.qkv_proj.bias.data.fill_(0)\n", " nn.init.xavier_uniform_(self.o_proj.weight)\n", " self.o_proj.bias.data.fill_(0)\n", "\n", " def forward(self, x, mask=None, return_attention=False):\n", " #this is the input embed dim\n", " batch_size, seq_length, embed_dim = x.size()\n", " qkv = self.qkv_proj(x) #[Batch, SeqLen, 3*total_embed]\n", " \n", " # Separate Q, K, V from linear output\n", " qkv = qkv.reshape(batch_size, seq_length, self.num_heads, 3*self.head_dim)\n", " qkv = qkv.permute(0, 2, 1, 3) # [Batch, Head, SeqLen, 3*Dims]\n", " q, k, v = qkv.chunk(3, dim=-1) #[Batch, Head, SeqLen, Dims]\n", " \n", " # Determine value outputs\n", " values, attention = scaled_dot_product(q, k, v, mask=mask)\n", " values = values.permute(0, 2, 1, 3) # [Batch, SeqLen, Head, Dims]\n", " values = values.reshape(batch_size, seq_length, embed_dim) #concatination all heads\n", " o = self.o_proj(values)\n", " \n", " if return_attention:\n", " return o, attention\n", " else:\n", " return o" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [], "source": [ "seq_len,input_dim, embed_dim = 20,10,10\n", "x = torch.randn(4,seq_len,input_dim)\n", "\n", "MhA= MultiheadAttention(input_dim, embed_dim, embed_dim//4)" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "MHA output: torch.Size([4, 20, 10]), attn shape: torch.Size([4, 2, 20, 20])\n" ] } ], "source": [ "out,a = MhA(x,return_attention=True)\n", "print(f'MHA output: {out.shape}, attn shape: {a.shape}')\n" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "You can analize the attention scores in [here](https://huggingface.co/exbert/?model=gpt2&modelKind=bidirectional&sentence=The%20girl%20ran%20to%20a%20local%20pub%20to%20escape%20the%20din%20of%20her%20city.&layer=11&heads=..10&threshold=1&tokenInd=null&tokenSide=null&maskInds=..&hideClsSep=true)" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "Recall for the transformer:\n", "
\n", "\n", "\n", "feed forward is simply an MLP with GELU activision:" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "$GELU(x) = x P(X\\le x) = x \\times \\frac{1}{2} [1+ erf(x/\\sqrt(2)]$\n", "\n", "https://en.wikipedia.org/wiki/Error_function\n", "\n", "can also be aproximate by $0.5x (1+ tanh[\\sqrt{(2/\\pi)} (x+0.044715x^3)]$" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [], "source": [ "#remember?\n", "def elu_forward(z, alpha): \n", " elu_positive = z\n", " elu_negative = alpha * (torch.exp(z) - 1)\n", " elu_output = torch.where(z>0, elu_positive, elu_negative)\n", " return elu_output" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [], "source": [ "def gelu_forward(z,type=1):\n", " if type==1:\n", " return 0.5*z*(1+torch.erf(z/np.sqrt(2))) \n", " return 0.5*z*(1+torch.tanh(np.sqrt(2/np.pi)*(z+0.044715*z**3)))" ] }, { "cell_type": "code", "execution_count": 10, "metadata": { "scrolled": true }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "z = torch.linspace(-5, 5, steps=1000)\n", "plt.plot(z.numpy(), torch.relu(z).numpy(), label='ReLU(z)', linewidth=5);\n", "plt.plot(z.numpy(), elu_forward(z, alpha=.5).numpy(), label='ELU(z)', linewidth=2);\n", "plt.plot(z.numpy(), gelu_forward(z).numpy(), label='GELU(z)', linewidth=2);\n", "plt.legend(); plt.grid();" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "tensor(1.6022)" ] }, "execution_count": 11, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "z = torch.linspace(-5, 5, steps=10000)\n", "plt.plot(z.numpy(), gelu_forward(z).numpy(), label='GELU', linewidth=2);\n", "plt.plot(z.numpy(), gelu_forward(z,2).numpy(), label='GELU tanh aprox', linewidth=2);\n", "plt.legend(); plt.grid();\n", "(gelu_forward(z)- gelu_forward(z,2)).abs().sum()" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "There is one major drawback that we didn't refer to so far, can you guess what it is?" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "### Positional Encoding" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "fragment" } }, "source": [ "* Untill now, we saw a strong tool for learning, but if a token came as the first word in the sentence, or as last, we only learned the power of the relation between them, the order did not matter!\n", "\n", "* The positional encoding is an idea that reminds Spectral decomposition, like Fourier series, based just on the positions.\n", "\n", "* The basis is Sin and Cosin functions with different frequencies as coefficients (2pi, 4pi…)\n", "\n", "* The reason of that is that we want to push each embedding a bit to encode the position as well as the meaning, but we also want it to have the same norm (why?)\n", "\n", "* It's not special for text, as you saw in the lecture VIT use something similar. When we don't need to add PE?" ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [], "source": [ "class PositionalEncoding(nn.Module):\n", "\n", " def __init__(self, d_model, max_len=5000): \n", " \"\"\"\n", " Inputs\n", " d_model - Hidden dimensionality of the input.\n", " max_len - Maximum length of a sequence to expect.\n", " \"\"\"\n", " super().__init__()\n", "\n", " # Create matrix of [SeqLen, HiddenDim] representing the positional encoding for max_len inputs\n", " pe = torch.zeros(max_len, d_model)\n", " position = torch.arange(0, max_len, dtype=torch.float).unsqueeze(1)\n", " div_term = torch.exp(torch.arange(0, d_model, 2).float() * (-math.log(10000.0) / d_model))\n", " pe[:, 0::2] = torch.sin(position * div_term)\n", " pe[:, 1::2] = torch.cos(position * div_term)\n", " pe = pe.unsqueeze(0)\n", " \n", " # register_buffer => Tensor which is not a parameter, but should be part of the modules state.\n", " # Used for tensors that need to be on the same device as the module.\n", " # persistent=False tells PyTorch to not add the buffer to the state dict (e.g. when we save the model) \n", " self.register_buffer('pe', pe, persistent=False)\n", "\n", " def forward(self, x):\n", " x = x + self.pe[:, :x.size(1)]\n", " return x" ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "encod_block = PositionalEncoding(d_model=48, max_len=96)\n", "pe = encod_block.pe.squeeze().T.cpu().numpy()\n", "\n", "fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(8,3))\n", "pos = ax.imshow(pe, cmap=\"RdGy\", extent=(1,pe.shape[1]+1,pe.shape[0]+1,1))\n", "fig.colorbar(pos, ax=ax)\n", "ax.set_xlabel(\"Position in sequence\")\n", "ax.set_ylabel(\"Hidden dimension\")\n", "ax.set_title(\"Positional encoding over hidden dimensions\")\n", "ax.set_xticks([1]+[i*10 for i in range(1,1+pe.shape[1]//10)])\n", "ax.set_yticks([1]+[i*10 for i in range(1,1+pe.shape[0]//10)])\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "#!pip install seaborn #- if it's not part of the requrement file\n", "import seaborn as sns\n", "import numpy as np\n", "sns.set_theme()\n", "fig, ax = plt.subplots(2, 2, figsize=(12,4))\n", "ax = [a for a_list in ax for a in a_list]\n", "for i in range(len(ax)):\n", " ax[i].plot(np.arange(1,17), pe[i,:16], color=f'C{i}', marker=\"o\", markersize=6, markeredgecolor=\"black\")\n", " ax[i].set_title(f\"Encoding in hidden dimension {i+1}\")\n", " ax[i].set_xlabel(\"Position in sequence\", fontsize=10)\n", " ax[i].set_ylabel(\"Positional encoding\", fontsize=10)\n", " ax[i].set_xticks(np.arange(1,17))\n", " ax[i].tick_params(axis='both', which='major', labelsize=10)\n", " ax[i].tick_params(axis='both', which='minor', labelsize=8)\n", " ax[i].set_ylim(-1.2, 1.2)\n", "fig.subplots_adjust(hspace=0.8)\n", "sns.reset_orig()\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "skip" } }, "source": [ "## Ways to use it in NLP" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "* **T**ext-**T**o-**T**ext **T**ransfer **T**ransformer (T5): Translation, summerization, question answering and so on.\n", "
\n", "
\n", "
\n", "* **B**idirectional **E**ncoder **R**epresentations from **T**ransformers (BERT): **Encoder only!** mask prediction over one sequnce, can be use for error fixing, wordtune/gramerly etc.. \n", "
\n", "
\n", "
\n", "* **G**enerative **P**re**T**raining (GPT): **Decoder only!**. Generate text, stories using nothing but initial tokens." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "
\n" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "## BRRT" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "
\n", "\n", "Uses pre-training of semi (self) supervision of two tasks:\n", "\n", "1. masked prediction: mask out 15% of the tokens and predict them using bi-directional encoding of the sentense.\n", "\n", "make use of the spacial token `[MASK]` instead of some of the inputs\n", "\n", "2. NSP: Next sentence prediction: predict if the next sentence come after the current one or not.\n", "\n", "sentences of the same paragraph comes one after the other with a seperation of the special token `[SEP]`" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "## GPT" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "
\n", "\n", "GPT4: 100 Trillion PARAMS!\n" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "What's next? i call it The **moor's law of parameters**:\n", "\n", "\n", "\n", "\n", "
\n", "\n", "\n", "From a survey on LLM\n", "\n", "https://arxiv.org/pdf/2303.18223.pdf" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "Let's try to work with something that is way to big to train here, but we can play with a pre-trained version: GPT2:\n", "\n", "
\n" ] }, { "cell_type": "code", "execution_count": 17, "metadata": { "slideshow": { "slide_type": "skip" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Collecting transformers\n", " Downloading transformers-4.38.2-py3-none-any.whl.metadata (130 kB)\n", "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m130.7/130.7 kB\u001b[0m \u001b[31m1.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0ma \u001b[36m0:00:01\u001b[0m\n", "\u001b[?25hRequirement already satisfied: filelock in /Users/moshekimhi/miniconda3/envs/cs236781/lib/python3.8/site-packages (from transformers) (3.13.1)\n", "Collecting huggingface-hub<1.0,>=0.19.3 (from transformers)\n", " Downloading huggingface_hub-0.21.4-py3-none-any.whl.metadata (13 kB)\n", "Requirement already satisfied: numpy>=1.17 in /Users/moshekimhi/miniconda3/envs/cs236781/lib/python3.8/site-packages (from transformers) (1.24.4)\n", "Requirement already satisfied: packaging>=20.0 in /Users/moshekimhi/miniconda3/envs/cs236781/lib/python3.8/site-packages (from transformers) (23.2)\n", "Requirement already satisfied: pyyaml>=5.1 in /Users/moshekimhi/miniconda3/envs/cs236781/lib/python3.8/site-packages (from transformers) (6.0.1)\n", "Requirement already satisfied: regex!=2019.12.17 in /Users/moshekimhi/miniconda3/envs/cs236781/lib/python3.8/site-packages (from transformers) (2023.12.25)\n", "Requirement already satisfied: requests in /Users/moshekimhi/miniconda3/envs/cs236781/lib/python3.8/site-packages (from transformers) (2.31.0)\n", "Collecting tokenizers<0.19,>=0.14 (from transformers)\n", " Downloading tokenizers-0.15.2-cp38-cp38-macosx_11_0_arm64.whl.metadata (6.7 kB)\n", "Collecting safetensors>=0.4.1 (from transformers)\n", " Downloading safetensors-0.4.2-cp38-cp38-macosx_11_0_arm64.whl.metadata (3.8 kB)\n", "Requirement already satisfied: tqdm>=4.27 in /Users/moshekimhi/miniconda3/envs/cs236781/lib/python3.8/site-packages (from transformers) (4.66.1)\n", "Requirement already satisfied: fsspec>=2023.5.0 in /Users/moshekimhi/miniconda3/envs/cs236781/lib/python3.8/site-packages (from huggingface-hub<1.0,>=0.19.3->transformers) (2023.12.2)\n", "Requirement already satisfied: typing-extensions>=3.7.4.3 in /Users/moshekimhi/miniconda3/envs/cs236781/lib/python3.8/site-packages (from huggingface-hub<1.0,>=0.19.3->transformers) (4.9.0)\n", "Requirement already satisfied: charset-normalizer<4,>=2 in /Users/moshekimhi/miniconda3/envs/cs236781/lib/python3.8/site-packages (from requests->transformers) (3.3.2)\n", "Requirement already satisfied: idna<4,>=2.5 in /Users/moshekimhi/miniconda3/envs/cs236781/lib/python3.8/site-packages (from requests->transformers) (3.6)\n", "Requirement already satisfied: urllib3<3,>=1.21.1 in /Users/moshekimhi/miniconda3/envs/cs236781/lib/python3.8/site-packages (from requests->transformers) (2.1.0)\n", "Requirement already satisfied: certifi>=2017.4.17 in /Users/moshekimhi/miniconda3/envs/cs236781/lib/python3.8/site-packages (from requests->transformers) (2023.11.17)\n", "Downloading transformers-4.38.2-py3-none-any.whl (8.5 MB)\n", "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m8.5/8.5 MB\u001b[0m \u001b[31m24.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0ma \u001b[36m0:00:01\u001b[0m\n", "\u001b[?25hDownloading huggingface_hub-0.21.4-py3-none-any.whl (346 kB)\n", "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m346.4/346.4 kB\u001b[0m \u001b[31m6.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0mta \u001b[36m0:00:01\u001b[0m\n", "\u001b[?25hDownloading safetensors-0.4.2-cp38-cp38-macosx_11_0_arm64.whl (392 kB)\n", "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m392.9/392.9 kB\u001b[0m \u001b[31m5.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0mta \u001b[36m0:00:01\u001b[0m\n", "\u001b[?25hDownloading tokenizers-0.15.2-cp38-cp38-macosx_11_0_arm64.whl (2.4 MB)\n", "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.4/2.4 MB\u001b[0m \u001b[31m10.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0ma \u001b[36m0:00:01\u001b[0m\n", "\u001b[?25hInstalling collected packages: safetensors, huggingface-hub, tokenizers, transformers\n", "Successfully installed huggingface-hub-0.21.4 safetensors-0.4.2 tokenizers-0.15.2 transformers-4.38.2\n" ] } ], "source": [ "!pip install transformers" ] }, { "cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [ { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "4c769f5238574f1d94b4d36e8e2a97f1", "version_major": 2, "version_minor": 0 }, "text/plain": [ "config.json: 0%| | 0.00/665 [00:00\n", "To re-use, please provide attribution and link to the original.\n", "\n", "Some images in this tutorial were taken and/or adapted from the following sources:\n", "\n", "- Jay Alammar blog - https://jalammar.github.io/illustrated-transformer/\n", "- Attention is all you need paper - https://arxiv.org/abs/1706.03762\n", "- Lucas Beyer - transformer presentation" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "celltoolbar": "Slideshow", "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": "3.8.12" }, "rise": { "scroll": true }, "toc-autonumbering": false }, "nbformat": 4, "nbformat_minor": 4 }