notebooks/benchmark_pipelines.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cloning into 'onnx_transformers'...\n",
      "remote: Enumerating objects: 105, done.\u001b[K\n",
      "remote: Counting objects: 100% (105/105), done.\u001b[K\n",
      "remote: Compressing objects: 100% (65/65), done.\u001b[K\n",
      "remote: Total 105 (delta 48), reused 92 (delta 37), pack-reused 0\u001b[K\n",
      "Receiving objects: 100% (105/105), 34.52 KiB | 415.00 KiB/s, done.\n",
      "Resolving deltas: 100% (48/48), done.\n"
     ]
    }
   ],
   "source": [
    "!pip install -qqq --upgrade torch==1.6.0+cpu torchvision==0.7.0+cpu -f https://download.pytorch.org/whl/torch_stable.html\n",
    "!git clone https://github.com/patil-suraj/onnx_transformers.git\n",
    "!pip install -U -qqq -e ./onnx_transformers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Architecture:        x86_64\n",
      "CPU op-mode(s):      32-bit, 64-bit\n",
      "Byte Order:          Little Endian\n",
      "CPU(s):              4\n",
      "On-line CPU(s) list: 0-3\n",
      "Thread(s) per core:  2\n",
      "Core(s) per socket:  2\n",
      "Socket(s):           1\n",
      "NUMA node(s):        1\n",
      "Vendor ID:           AuthenticAMD\n",
      "CPU family:          23\n",
      "Model:               1\n",
      "Model name:          AMD EPYC 7571\n",
      "Stepping:            2\n",
      "CPU MHz:             2541.395\n",
      "BogoMIPS:            4399.98\n",
      "Hypervisor vendor:   KVM\n",
      "Virtualization type: full\n",
      "L1d cache:           32K\n",
      "L1i cache:           64K\n",
      "L2 cache:            512K\n",
      "L3 cache:            8192K\n",
      "NUMA node0 CPU(s):   0-3\n",
      "Flags:               fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc cpuid extd_apicid aperfmperf tsc_known_freq pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy cr8_legacy abm sse4a misalignsse 3dnowprefetch topoext vmmcall fsgsbase bmi1 avx2 smep bmi2 rdseed adx smap clflushopt sha_ni xsaveopt xsavec xgetbv1 clzero xsaveerptr arat npt nrip_save\n"
     ]
    }
   ],
   "source": [
    "!lscpu"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Restart notebook after installation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from onnx_transformers import pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/notebooks/onnx_tests/onnx_transformers/.onnx/distilbert-base-cased/distilbert-base-cased.onnx.input_names.json\n",
      "Creating folder /notebooks/onnx_tests/onnx_transformers/.onnx/distilbert-base-cased\n",
      "Using framework PyTorch: 1.6.0+cpu\n",
      "Found input input_ids with shape: {0: 'batch', 1: 'sequence'}\n",
      "Found input attention_mask with shape: {0: 'batch', 1: 'sequence'}\n",
      "Found output output_0 with shape: {0: 'batch', 1: 'sequence'}\n",
      "Ensuring inputs are in correct order\n",
      "head_mask is not present in the generated input list.\n",
      "Generated inputs order: ['input_ids', 'attention_mask']\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.7/dist-packages/transformers/modeling_utils.py:1570: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!\n",
      "  input_tensor.shape == tensor_shape for input_tensor in input_tensors\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found input input_ids with shape: {0: 'batch', 1: 'sequence'}\n",
      "Found input attention_mask with shape: {0: 'batch', 1: 'sequence'}\n",
      "Found output output_0 with shape: {0: 'batch', 1: 'sequence'}\n"
     ]
    }
   ],
   "source": [
    "# load onnx pipeline\n",
    "nlp_onnx = pipeline(\"feature-extraction\",  onnx=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load torch pipeline\n",
    "nlp_torch = pipeline(\"feature-extraction\", onnx=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from contextlib import contextmanager\n",
    "from dataclasses import dataclass\n",
    "from time import time\n",
    "from tqdm import trange\n",
    "\n",
    "@contextmanager\n",
    "def track_infer_time(buffer: [int]):\n",
    "    start = time()\n",
    "    yield\n",
    "    end = time()\n",
    "\n",
    "    buffer.append(end - start)\n",
    "\n",
    "\n",
    "@dataclass\n",
    "class OnnxInferenceResult:\n",
    "  model_inference_time: [int]  \n",
    "  optimized_model_path: str"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Tracking inference time for Pytorch CPU: 100%|██████████| 100/100 [00:04<00:00, 22.69it/s]\n",
      "Tracking inference time for ONNX CPU: 100%|██████████| 100/100 [00:02<00:00, 37.85it/s]\n"
     ]
    }
   ],
   "source": [
    "pipelines = [(\"Pytorch CPU\", nlp_torch), (\"ONNX CPU\", nlp_onnx)]\n",
    "results = {}\n",
    "for label, pipeline_ in pipelines:\n",
    "  # Compute \n",
    "  time_buffer = []\n",
    "  for _ in trange(100, desc=f\"Tracking inference time for {label}\"):\n",
    "    with track_infer_time(time_buffer):\n",
    "      pipeline_(\"My name is BERT\")\n",
    "\n",
    "  # Store the result\n",
    "  results[label] = OnnxInferenceResult(\n",
    "      time_buffer, \n",
    "      None\n",
    "  )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "sns.set()\n",
    "import numpy as np\n",
    "import os\n",
    "\n",
    "\n",
    "def plot_benchmark(results):\n",
    "    # Compute average inference time + std\n",
    "    time_results = {k: np.mean(v.model_inference_time) * 1e3 for k, v in results.items()}\n",
    "    time_results_std = np.std([v.model_inference_time for v in results.values()]) * 1000\n",
    "\n",
    "    plt.rcdefaults()\n",
    "    fig, ax = plt.subplots(figsize=(16, 12))\n",
    "    ax.set_ylabel(\"Avg Inference time (ms)\")\n",
    "    ax.set_title(\"Average inference time (ms) for each provider\")\n",
    "    ax.bar(time_results.keys(), time_results.values(), yerr=time_results_std)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1600x1200 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_benchmark(results)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, let's benchmark anothe pipline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/notebooks/onnx_tests/onnx_transformers/.onnx/distilbert-base-cased-distilled-squad/distilbert-base-cased-distilled-squad.onnx.input_names.json\n"
     ]
    }
   ],
   "source": [
    "# load onnx pipeline\n",
    "nlp_onnx = pipeline(\"question-answering\",  onnx=True)\n",
    "# load torch pipeline\n",
    "nlp_torch = pipeline(\"question-answering\", onnx=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Tracking inference time for Pytorch CPU: 100%|██████████| 100/100 [00:25<00:00,  3.93it/s]\n",
      "Tracking inference time for ONNX CPU: 100%|██████████| 100/100 [00:20<00:00,  4.82it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1600x1200 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pipelines = [(\"Pytorch CPU\", nlp_torch), (\"ONNX CPU\", nlp_onnx)]\n",
    "results = {}\n",
    "    \n",
    "with warnings.catch_warnings():\n",
    "    warnings.simplefilter(\"ignore\")\n",
    "    for label, pipeline_ in pipelines:\n",
    "      # Compute \n",
    "      time_buffer = []\n",
    "      for _ in trange(100, desc=f\"Tracking inference time for {label}\"):\n",
    "        with track_infer_time(time_buffer):\n",
    "          pipeline_({\"question\": \"What is ONNX runtime ?\", \"context\": \"ONNX Runtime is a highly performant single inference engine for multiple platforms and hardware\"})\n",
    "\n",
    "      # Store the result\n",
    "      results[label] = OnnxInferenceResult(\n",
    "          time_buffer, \n",
    "          None\n",
    "      )\n",
    "\n",
    "plot_benchmark(results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}