mirror of
https://github.com/NVIDIA/TensorRT-LLM.git
synced 2026-01-14 06:27:45 +08:00
415 lines
46 KiB
HTML
415 lines
46 KiB
HTML
<!DOCTYPE html>
|
|
<html class="writer-html5" lang="en" data-content_root="../../../">
|
|
<head>
|
|
<meta charset="utf-8" />
|
|
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
|
<title>tensorrt_llm.runtime.session — tensorrt_llm documentation</title>
|
|
<link rel="stylesheet" type="text/css" href="../../../_static/pygments.css?v=80d5e7a1" />
|
|
<link rel="stylesheet" type="text/css" href="../../../_static/css/theme.css?v=19f00094" />
|
|
|
|
|
|
<!--[if lt IE 9]>
|
|
<script src="../../../_static/js/html5shiv.min.js"></script>
|
|
<![endif]-->
|
|
|
|
<script src="../../../_static/jquery.js?v=5d32c60e"></script>
|
|
<script src="../../../_static/_sphinx_javascript_frameworks_compat.js?v=2cd50e6c"></script>
|
|
<script src="../../../_static/documentation_options.js?v=5929fcd5"></script>
|
|
<script src="../../../_static/doctools.js?v=888ff710"></script>
|
|
<script src="../../../_static/sphinx_highlight.js?v=dc90522c"></script>
|
|
<script src="../../../_static/js/theme.js"></script>
|
|
<link rel="index" title="Index" href="../../../genindex.html" />
|
|
<link rel="search" title="Search" href="../../../search.html" />
|
|
</head>
|
|
|
|
<body class="wy-body-for-nav">
|
|
<div class="wy-grid-for-nav">
|
|
<nav data-toggle="wy-nav-shift" class="wy-nav-side">
|
|
<div class="wy-side-scroll">
|
|
<div class="wy-side-nav-search" >
|
|
|
|
|
|
|
|
<a href="../../../index.html" class="icon icon-home">
|
|
tensorrt_llm
|
|
</a>
|
|
<div role="search">
|
|
<form id="rtd-search-form" class="wy-form" action="../../../search.html" method="get">
|
|
<input type="text" name="q" placeholder="Search docs" aria-label="Search docs" />
|
|
<input type="hidden" name="check_keywords" value="yes" />
|
|
<input type="hidden" name="area" value="default" />
|
|
</form>
|
|
</div>
|
|
</div><div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="Navigation menu">
|
|
<p class="caption" role="heading"><span class="caption-text">Contents:</span></p>
|
|
<ul>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../architecture.html">TensorRT-LLM Architecture</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../gpt_runtime.html">C++ GPT Runtime</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../batch_manager.html">The Batch Manager in TensorRT-LLM</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../gpt_attention.html">Multi-head, Multi-query and Group-query Attention</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../precision.html">Numerical Precision</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../installation.html">Build TensorRT-LLM</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../performance.html">Performance of TensorRT-LLM</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../2023-05-19-how-to-debug.html">How to debug</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../2023-05-17-how-to-add-a-new-model.html">How to add a new model</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../graph-rewriting.html">Graph Rewriting Module</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../memory.html">Memory Usage of TensorRT-LLM</a></li>
|
|
</ul>
|
|
<p class="caption" role="heading"><span class="caption-text">Python API</span></p>
|
|
<ul>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../python-api/tensorrt_llm.layers.html">Layers</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../python-api/tensorrt_llm.functional.html">Functionals</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../python-api/tensorrt_llm.models.html">Models</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../python-api/tensorrt_llm.plugin.html">Plugin</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../python-api/tensorrt_llm.quantization.html">Quantization</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../python-api/tensorrt_llm.runtime.html">Runtime</a></li>
|
|
</ul>
|
|
<p class="caption" role="heading"><span class="caption-text">C++ API</span></p>
|
|
<ul>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../_cpp_gen/runtime.html">Runtime</a></li>
|
|
</ul>
|
|
<p class="caption" role="heading"><span class="caption-text">Blogs</span></p>
|
|
<ul>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../blogs/H100vsA100.html">H100 has 4.6x A100 Performance in TensorRT-LLM, achieving 10,000 tok/s at 100ms to first token</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../../../blogs/H200launch.html">H200 achieves nearly 12,000 tokens/sec on Llama2-13B with TensorRT-LLM</a></li>
|
|
</ul>
|
|
|
|
</div>
|
|
</div>
|
|
</nav>
|
|
|
|
<section data-toggle="wy-nav-shift" class="wy-nav-content-wrap"><nav class="wy-nav-top" aria-label="Mobile navigation menu" >
|
|
<i data-toggle="wy-nav-top" class="fa fa-bars"></i>
|
|
<a href="../../../index.html">tensorrt_llm</a>
|
|
</nav>
|
|
|
|
<div class="wy-nav-content">
|
|
<div class="rst-content">
|
|
<div role="navigation" aria-label="Page navigation">
|
|
<ul class="wy-breadcrumbs">
|
|
<li><a href="../../../index.html" class="icon icon-home" aria-label="Home"></a></li>
|
|
<li class="breadcrumb-item"><a href="../../index.html">Module code</a></li>
|
|
<li class="breadcrumb-item active">tensorrt_llm.runtime.session</li>
|
|
<li class="wy-breadcrumbs-aside">
|
|
</li>
|
|
</ul>
|
|
<hr/>
|
|
</div>
|
|
<div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
|
|
<div itemprop="articleBody">
|
|
|
|
<h1>Source code for tensorrt_llm.runtime.session</h1><div class="highlight"><pre>
|
|
<span></span><span class="c1"># SPDX-FileCopyrightText: Copyright (c) 2022-2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved.</span>
|
|
<span class="c1"># SPDX-License-Identifier: Apache-2.0</span>
|
|
<span class="c1">#</span>
|
|
<span class="c1"># Licensed under the Apache License, Version 2.0 (the "License");</span>
|
|
<span class="c1"># you may not use this file except in compliance with the License.</span>
|
|
<span class="c1"># You may obtain a copy of the License at</span>
|
|
<span class="c1">#</span>
|
|
<span class="c1"># http://www.apache.org/licenses/LICENSE-2.0</span>
|
|
<span class="c1">#</span>
|
|
<span class="c1"># Unless required by applicable law or agreed to in writing, software</span>
|
|
<span class="c1"># distributed under the License is distributed on an "AS IS" BASIS,</span>
|
|
<span class="c1"># WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.</span>
|
|
<span class="c1"># See the License for the specific language governing permissions and</span>
|
|
<span class="c1"># limitations under the License.</span>
|
|
<span class="kn">from</span> <span class="nn">__future__</span> <span class="kn">import</span> <span class="n">annotations</span>
|
|
|
|
<span class="kn">import</span> <span class="nn">contextlib</span>
|
|
<span class="kn">from</span> <span class="nn">dataclasses</span> <span class="kn">import</span> <span class="n">dataclass</span>
|
|
<span class="kn">from</span> <span class="nn">typing</span> <span class="kn">import</span> <span class="n">Any</span><span class="p">,</span> <span class="n">Dict</span><span class="p">,</span> <span class="n">List</span><span class="p">,</span> <span class="n">Optional</span>
|
|
|
|
<span class="c1"># isort: off</span>
|
|
<span class="kn">import</span> <span class="nn">torch</span>
|
|
<span class="kn">import</span> <span class="nn">tensorrt</span> <span class="k">as</span> <span class="nn">trt</span>
|
|
<span class="c1"># isort: on</span>
|
|
|
|
<span class="kn">from</span> <span class="nn">.._utils</span> <span class="kn">import</span> <span class="n">trt_dtype_to_torch</span>
|
|
<span class="kn">from</span> <span class="nn">..logger</span> <span class="kn">import</span> <span class="n">logger</span>
|
|
|
|
|
|
<span class="nd">@contextlib</span><span class="o">.</span><span class="n">contextmanager</span>
|
|
<span class="k">def</span> <span class="nf">_scoped_stream</span><span class="p">():</span>
|
|
<span class="w"> </span><span class="sd">'''Create a scoped cuda stream, and synchronize it when the context is destroyed</span>
|
|
<span class="sd"> '''</span>
|
|
<span class="c1">#TODO: delete torch, use cuda native python bindings</span>
|
|
<span class="kn">import</span> <span class="nn">torch</span>
|
|
<span class="n">stream</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">cuda</span><span class="o">.</span><span class="n">current_stream</span><span class="p">()</span>
|
|
<span class="k">try</span><span class="p">:</span>
|
|
<span class="c1"># return a handle, trt and other lib does not recognize torch.cuda.Stream</span>
|
|
<span class="k">yield</span> <span class="n">stream</span><span class="o">.</span><span class="n">cuda_stream</span>
|
|
<span class="k">finally</span><span class="p">:</span>
|
|
<span class="n">stream</span><span class="o">.</span><span class="n">synchronize</span><span class="p">()</span>
|
|
|
|
|
|
<div class="viewcode-block" id="TensorInfo">
|
|
<a class="viewcode-back" href="../../../python-api/tensorrt_llm.runtime.html#tensorrt_llm.runtime.TensorInfo">[docs]</a>
|
|
<span class="nd">@dataclass</span>
|
|
<span class="k">class</span> <span class="nc">TensorInfo</span><span class="p">:</span>
|
|
<span class="n">name</span><span class="p">:</span> <span class="nb">str</span>
|
|
<span class="n">dtype</span><span class="p">:</span> <span class="n">trt</span><span class="o">.</span><span class="n">DataType</span>
|
|
<span class="n">shape</span><span class="p">:</span> <span class="nb">tuple</span></div>
|
|
|
|
<span class="c1"># add more info like strides, formats if needed</span>
|
|
|
|
|
|
<div class="viewcode-block" id="Session">
|
|
<a class="viewcode-back" href="../../../python-api/tensorrt_llm.runtime.html#tensorrt_llm.runtime.Session">[docs]</a>
|
|
<span class="k">class</span> <span class="nc">Session</span><span class="p">(</span><span class="nb">object</span><span class="p">):</span>
|
|
<span class="w"> </span><span class="sd">''' Session is a managed TensorRT runtime. '''</span>
|
|
|
|
<span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
|
|
<span class="c1"># use Session.from_serialized_engine to create a session</span>
|
|
<span class="k">pass</span>
|
|
|
|
<span class="k">def</span> <span class="nf">_init</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">engine_buffer</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
|
|
<span class="w"> </span><span class="sd">'''</span>
|
|
<span class="sd"> @brief: Setup TensorRT engines and context from a serialized engine file</span>
|
|
<span class="sd"> @param engine_buffer: a buffer holds the serialized TRT engine</span>
|
|
<span class="sd"> '''</span>
|
|
<span class="bp">self</span><span class="o">.</span><span class="n">_runtime</span> <span class="o">=</span> <span class="n">trt</span><span class="o">.</span><span class="n">Runtime</span><span class="p">(</span><span class="n">logger</span><span class="o">.</span><span class="n">trt_logger</span><span class="p">)</span>
|
|
<span class="k">if</span> <span class="n">engine_buffer</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
|
|
<span class="bp">self</span><span class="o">.</span><span class="n">_engine</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">runtime</span><span class="o">.</span><span class="n">deserialize_cuda_engine</span><span class="p">(</span><span class="n">engine_buffer</span><span class="p">)</span>
|
|
<span class="bp">self</span><span class="o">.</span><span class="n">_context</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">create_execution_context</span><span class="p">()</span>
|
|
<span class="k">with</span> <span class="n">_scoped_stream</span><span class="p">()</span> <span class="k">as</span> <span class="n">stream</span><span class="p">:</span>
|
|
<span class="bp">self</span><span class="o">.</span><span class="n">_context</span><span class="o">.</span><span class="n">set_optimization_profile_async</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">stream</span><span class="p">)</span>
|
|
<span class="k">return</span> <span class="bp">self</span>
|
|
|
|
<div class="viewcode-block" id="Session.from_serialized_engine">
|
|
<a class="viewcode-back" href="../../../python-api/tensorrt_llm.runtime.html#tensorrt_llm.runtime.Session.from_serialized_engine">[docs]</a>
|
|
<span class="nd">@staticmethod</span>
|
|
<span class="k">def</span> <span class="nf">from_serialized_engine</span><span class="p">(</span><span class="n">engine</span><span class="p">)</span> <span class="o">-></span> <span class="n">Session</span><span class="p">:</span>
|
|
<span class="w"> </span><span class="sd">'''</span>
|
|
<span class="sd"> @brief: Create a session from a serialized engine</span>
|
|
<span class="sd"> @param engine: a serialized engine</span>
|
|
<span class="sd"> @return: a Session object</span>
|
|
<span class="sd"> '''</span>
|
|
<span class="n">session</span> <span class="o">=</span> <span class="n">Session</span><span class="p">()</span>
|
|
<span class="k">return</span> <span class="n">session</span><span class="o">.</span><span class="n">_init</span><span class="p">(</span><span class="n">engine</span><span class="p">)</span></div>
|
|
|
|
|
|
<div class="viewcode-block" id="Session.from_engine">
|
|
<a class="viewcode-back" href="../../../python-api/tensorrt_llm.runtime.html#tensorrt_llm.runtime.Session.from_engine">[docs]</a>
|
|
<span class="nd">@staticmethod</span>
|
|
<span class="k">def</span> <span class="nf">from_engine</span><span class="p">(</span><span class="n">engine</span><span class="p">)</span> <span class="o">-></span> <span class="n">Session</span><span class="p">:</span>
|
|
<span class="w"> </span><span class="sd">'''</span>
|
|
<span class="sd"> @brief: Create a session from an existing ICudaEngine engine</span>
|
|
<span class="sd"> @param engine: an ICudaEngine</span>
|
|
<span class="sd"> @return: a Session object</span>
|
|
<span class="sd"> '''</span>
|
|
<span class="n">session</span> <span class="o">=</span> <span class="n">Session</span><span class="p">()</span>
|
|
<span class="n">session</span><span class="o">.</span><span class="n">engine</span> <span class="o">=</span> <span class="n">engine</span>
|
|
<span class="k">return</span> <span class="n">session</span><span class="o">.</span><span class="n">_init</span><span class="p">()</span></div>
|
|
|
|
|
|
<span class="nd">@property</span>
|
|
<span class="k">def</span> <span class="nf">runtime</span><span class="p">(</span><span class="bp">self</span><span class="p">)</span> <span class="o">-></span> <span class="n">trt</span><span class="o">.</span><span class="n">Runtime</span><span class="p">:</span>
|
|
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_runtime</span>
|
|
|
|
<span class="nd">@property</span>
|
|
<span class="k">def</span> <span class="nf">engine</span><span class="p">(</span><span class="bp">self</span><span class="p">)</span> <span class="o">-></span> <span class="n">trt</span><span class="o">.</span><span class="n">ICudaEngine</span><span class="p">:</span>
|
|
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_engine</span>
|
|
|
|
<span class="nd">@engine</span><span class="o">.</span><span class="n">setter</span>
|
|
<span class="k">def</span> <span class="nf">engine</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">engine</span><span class="p">:</span> <span class="n">trt</span><span class="o">.</span><span class="n">ICudaEngine</span><span class="p">):</span>
|
|
<span class="bp">self</span><span class="o">.</span><span class="n">_engine</span> <span class="o">=</span> <span class="n">engine</span>
|
|
|
|
<span class="nd">@property</span>
|
|
<span class="k">def</span> <span class="nf">context</span><span class="p">(</span><span class="bp">self</span><span class="p">)</span> <span class="o">-></span> <span class="n">trt</span><span class="o">.</span><span class="n">IExecutionContext</span><span class="p">:</span>
|
|
<span class="w"> </span><span class="sd">'''</span>
|
|
<span class="sd"> @brief: Get the default TensorRT execution context,</span>
|
|
<span class="sd"> use self.engine.create_execution_context() to create a new context if needed</span>
|
|
<span class="sd"> @return: one TensorRT execution context object</span>
|
|
<span class="sd"> '''</span>
|
|
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_context</span>
|
|
|
|
<span class="k">def</span> <span class="nf">_print_engine_info</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
|
|
<span class="w"> </span><span class="sd">'''print engine info for debug purpose, internal use only.</span>
|
|
<span class="sd"> '''</span>
|
|
<span class="n">refitable</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">refittable</span>
|
|
<span class="n">num_layers</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">num_layers</span>
|
|
<span class="n">device_memory_size</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">device_memory_size</span>
|
|
<span class="n">name</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">name</span>
|
|
<span class="n">nb_profiles</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">num_optimization_profiles</span>
|
|
<span class="n">logger</span><span class="o">.</span><span class="n">info</span><span class="p">(</span>
|
|
<span class="sa">f</span><span class="s2">"Engine:</span><span class="si">{</span><span class="n">name</span><span class="si">=:}</span><span class="s2">, </span><span class="si">{</span><span class="n">refitable</span><span class="si">=:}</span><span class="s2">, </span><span class="si">{</span><span class="n">num_layers</span><span class="si">=:}</span><span class="s2">, </span><span class="si">{</span><span class="n">device_memory_size</span><span class="si">=:}</span><span class="s2">, </span><span class="si">{</span><span class="n">nb_profiles</span><span class="si">=:}</span><span class="s2">"</span>
|
|
<span class="p">)</span>
|
|
<span class="bp">self</span><span class="o">.</span><span class="n">_print_io_info</span><span class="p">()</span>
|
|
|
|
<span class="k">def</span> <span class="nf">_print_io_info</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
|
|
<span class="w"> </span><span class="sd">'''print engine i/o info for debug purpose, internal use only.</span>
|
|
<span class="sd"> '''</span>
|
|
<span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">num_bindings</span><span class="p">):</span>
|
|
<span class="n">name</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_binding_name</span><span class="p">(</span><span class="n">i</span><span class="p">)</span>
|
|
<span class="n">dtype</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_binding_dtype</span><span class="p">(</span><span class="n">i</span><span class="p">)</span>
|
|
<span class="n">shape</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_binding_shape</span><span class="p">(</span><span class="n">i</span><span class="p">)</span>
|
|
<span class="n">is_input</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">binding_is_input</span><span class="p">(</span><span class="n">i</span><span class="p">)</span>
|
|
<span class="n">logger</span><span class="o">.</span><span class="n">info</span><span class="p">(</span>
|
|
<span class="sa">f</span><span class="s2">"Binding:</span><span class="si">{</span><span class="n">i</span><span class="si">=:}</span><span class="s2">, </span><span class="si">{</span><span class="n">name</span><span class="si">=:}</span><span class="s2">, </span><span class="si">{</span><span class="n">dtype</span><span class="si">=:}</span><span class="s2">, </span><span class="si">{</span><span class="n">shape</span><span class="si">=:}</span><span class="s2">, </span><span class="si">{</span><span class="n">is_input</span><span class="si">=:}</span><span class="s2">"</span><span class="p">)</span>
|
|
|
|
<span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">num_io_tensors</span><span class="p">):</span>
|
|
<span class="n">name</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_tensor_name</span><span class="p">(</span><span class="n">i</span><span class="p">)</span>
|
|
<span class="n">mode</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_tensor_mode</span><span class="p">(</span><span class="n">name</span><span class="p">)</span>
|
|
<span class="n">shape</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_tensor_shape</span><span class="p">(</span><span class="n">name</span><span class="p">)</span>
|
|
<span class="n">dtype</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_tensor_dtype</span><span class="p">(</span><span class="n">name</span><span class="p">)</span>
|
|
<span class="n">tformat</span> <span class="o">=</span> <span class="s2">";"</span><span class="o">.</span><span class="n">join</span><span class="p">([</span>
|
|
<span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_tensor_format_desc</span><span class="p">(</span><span class="n">name</span><span class="p">,</span> <span class="n">p</span><span class="p">)</span>
|
|
<span class="k">for</span> <span class="n">p</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">num_optimization_profiles</span><span class="p">)</span>
|
|
<span class="p">])</span>
|
|
<span class="n">logger</span><span class="o">.</span><span class="n">info</span><span class="p">(</span>
|
|
<span class="sa">f</span><span class="s2">"Tensor:</span><span class="si">{</span><span class="n">name</span><span class="si">=:}</span><span class="s2">, </span><span class="si">{</span><span class="n">mode</span><span class="si">=:}</span><span class="s2">, </span><span class="si">{</span><span class="n">shape</span><span class="si">=:}</span><span class="s2">, </span><span class="si">{</span><span class="n">dtype</span><span class="si">=:}</span><span class="s2">, </span><span class="si">{</span><span class="n">tformat</span><span class="si">=:}</span><span class="s2">"</span><span class="p">)</span>
|
|
|
|
<div class="viewcode-block" id="Session.set_shapes">
|
|
<a class="viewcode-back" href="../../../python-api/tensorrt_llm.runtime.html#tensorrt_llm.runtime.Session.set_shapes">[docs]</a>
|
|
<span class="k">def</span> <span class="nf">set_shapes</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span>
|
|
<span class="n">tensor_dict</span><span class="p">:</span> <span class="n">Dict</span><span class="p">[</span><span class="nb">str</span><span class="p">,</span> <span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">],</span>
|
|
<span class="n">context</span><span class="p">:</span> <span class="n">Optional</span><span class="p">[</span><span class="n">trt</span><span class="o">.</span><span class="n">IExecutionContext</span><span class="p">]</span> <span class="o">=</span> <span class="kc">None</span><span class="p">):</span>
|
|
<span class="k">if</span> <span class="n">context</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
|
|
<span class="n">context</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">context</span>
|
|
|
|
<span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">num_io_tensors</span><span class="p">):</span>
|
|
<span class="n">name</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_tensor_name</span><span class="p">(</span><span class="n">i</span><span class="p">)</span>
|
|
<span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_tensor_mode</span><span class="p">(</span><span class="n">name</span><span class="p">)</span> <span class="o">==</span> <span class="n">trt</span><span class="o">.</span><span class="n">TensorIOMode</span><span class="o">.</span><span class="n">INPUT</span><span class="p">:</span>
|
|
<span class="n">ok</span> <span class="o">=</span> <span class="n">context</span><span class="o">.</span><span class="n">set_input_shape</span><span class="p">(</span><span class="n">name</span><span class="p">,</span> <span class="n">tensor_dict</span><span class="p">[</span><span class="n">name</span><span class="p">]</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span>
|
|
<span class="n">logger</span><span class="o">.</span><span class="n">debug</span><span class="p">(</span>
|
|
<span class="sa">f</span><span class="s2">"setting input tensor </span><span class="si">{</span><span class="n">name</span><span class="si">}</span><span class="s2"> with shape </span><span class="si">{</span><span class="n">tensor_dict</span><span class="p">[</span><span class="n">name</span><span class="p">]</span><span class="o">.</span><span class="n">shape</span><span class="si">}</span><span class="s2">"</span>
|
|
<span class="p">)</span>
|
|
<span class="k">if</span> <span class="ow">not</span> <span class="n">ok</span><span class="p">:</span>
|
|
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span>
|
|
<span class="sa">f</span><span class="s2">"Couldn't assign </span><span class="si">{</span><span class="n">name</span><span class="si">}</span><span class="s2"> with shape </span><span class="si">{</span><span class="n">tensor_dict</span><span class="p">[</span><span class="n">name</span><span class="p">]</span><span class="o">.</span><span class="n">shape</span><span class="si">}</span><span class="s2">, "</span>
|
|
<span class="sa">f</span><span class="s2">"engine supports [min, opt, max] = </span><span class="si">{</span><span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_profile_shape</span><span class="p">(</span><span class="n">context</span><span class="o">.</span><span class="n">active_optimization_profile</span><span class="p">,</span><span class="w"> </span><span class="n">name</span><span class="p">)</span><span class="si">}</span><span class="s2">"</span>
|
|
<span class="p">)</span></div>
|
|
|
|
|
|
<div class="viewcode-block" id="Session.infer_shapes">
|
|
<a class="viewcode-back" href="../../../python-api/tensorrt_llm.runtime.html#tensorrt_llm.runtime.Session.infer_shapes">[docs]</a>
|
|
<span class="k">def</span> <span class="nf">infer_shapes</span><span class="p">(</span>
|
|
<span class="bp">self</span><span class="p">,</span>
|
|
<span class="n">inputs</span><span class="p">:</span> <span class="n">List</span><span class="p">[</span><span class="n">TensorInfo</span><span class="p">],</span>
|
|
<span class="n">context</span><span class="p">:</span> <span class="n">Optional</span><span class="p">[</span><span class="n">trt</span><span class="o">.</span><span class="n">IExecutionContext</span><span class="p">]</span> <span class="o">=</span> <span class="kc">None</span>
|
|
<span class="p">)</span> <span class="o">-></span> <span class="n">List</span><span class="p">[</span><span class="n">TensorInfo</span><span class="p">]:</span>
|
|
<span class="w"> </span><span class="sd">'''</span>
|
|
<span class="sd"> @brief: Set input shapes to given context, and infer the output shapes from the given input shapes.</span>
|
|
<span class="sd"> This function should be called every time when the input shapes are changed before calling run().</span>
|
|
<span class="sd"> Or call the context.set_input_shape on all dynamic shaped input tensors manually.</span>
|
|
<span class="sd"> @param inputs: list of TensorInfo object, each item represents an input tensor</span>
|
|
<span class="sd"> @param context: TensorRT execution context, if None, use the default context</span>
|
|
<span class="sd"> @return: list of TensorInfo object, each item represents an output tensor, returns None if failed</span>
|
|
<span class="sd"> '''</span>
|
|
<span class="c1"># set shape to the default context if context is not specified</span>
|
|
<span class="k">if</span> <span class="n">context</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
|
|
<span class="n">context</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">context</span>
|
|
<span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="n">inputs</span><span class="p">:</span>
|
|
<span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_tensor_mode</span><span class="p">(</span><span class="n">i</span><span class="o">.</span><span class="n">name</span><span class="p">)</span> <span class="o">!=</span> <span class="n">trt</span><span class="o">.</span><span class="n">TensorIOMode</span><span class="o">.</span><span class="n">INPUT</span><span class="p">:</span>
|
|
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Tensor:</span><span class="si">{</span><span class="n">i</span><span class="o">.</span><span class="n">name</span><span class="si">}</span><span class="s2"> is not an input tensor"</span><span class="p">)</span>
|
|
<span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_tensor_dtype</span><span class="p">(</span><span class="n">i</span><span class="o">.</span><span class="n">name</span><span class="p">)</span> <span class="o">!=</span> <span class="n">i</span><span class="o">.</span><span class="n">dtype</span><span class="p">:</span>
|
|
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">"Tensor:</span><span class="si">{</span><span class="n">i</span><span class="o">.</span><span class="n">name</span><span class="si">}</span><span class="s2"> has wrong dtype"</span><span class="p">)</span>
|
|
<span class="k">if</span> <span class="ow">not</span> <span class="n">context</span><span class="o">.</span><span class="n">set_input_shape</span><span class="p">(</span><span class="n">i</span><span class="o">.</span><span class="n">name</span><span class="p">,</span> <span class="n">i</span><span class="o">.</span><span class="n">shape</span><span class="p">):</span>
|
|
<span class="k">raise</span> <span class="ne">RuntimeError</span><span class="p">(</span>
|
|
<span class="sa">f</span><span class="s2">"Could not set shape </span><span class="si">{</span><span class="n">i</span><span class="o">.</span><span class="n">shape</span><span class="si">}</span><span class="s2"> for tensor </span><span class="si">{</span><span class="n">i</span><span class="o">.</span><span class="n">name</span><span class="si">}</span><span class="s2">. Please check the profile range for which your model was build."</span>
|
|
<span class="p">)</span>
|
|
|
|
<span class="n">outputs</span> <span class="o">=</span> <span class="p">[]</span>
|
|
<span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">num_io_tensors</span><span class="p">):</span>
|
|
<span class="n">name</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_tensor_name</span><span class="p">(</span><span class="n">i</span><span class="p">)</span>
|
|
<span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_tensor_mode</span><span class="p">(</span><span class="n">name</span><span class="p">)</span> <span class="o">==</span> <span class="n">trt</span><span class="o">.</span><span class="n">TensorIOMode</span><span class="o">.</span><span class="n">OUTPUT</span><span class="p">:</span>
|
|
<span class="n">shape</span> <span class="o">=</span> <span class="n">context</span><span class="o">.</span><span class="n">get_tensor_shape</span><span class="p">(</span><span class="n">name</span><span class="p">)</span>
|
|
<span class="n">dtype</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">engine</span><span class="o">.</span><span class="n">get_tensor_dtype</span><span class="p">(</span><span class="n">name</span><span class="p">)</span>
|
|
<span class="n">outputs</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">TensorInfo</span><span class="p">(</span><span class="n">name</span><span class="p">,</span> <span class="n">dtype</span><span class="p">,</span> <span class="n">shape</span><span class="p">))</span>
|
|
<span class="k">return</span> <span class="n">outputs</span></div>
|
|
|
|
|
|
<div class="viewcode-block" id="Session.run">
|
|
<a class="viewcode-back" href="../../../python-api/tensorrt_llm.runtime.html#tensorrt_llm.runtime.Session.run">[docs]</a>
|
|
<span class="k">def</span> <span class="nf">run</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span>
|
|
<span class="n">inputs</span><span class="p">:</span> <span class="n">Dict</span><span class="p">[</span><span class="nb">str</span><span class="p">,</span> <span class="n">Any</span><span class="p">],</span>
|
|
<span class="n">outputs</span><span class="p">:</span> <span class="n">Dict</span><span class="p">[</span><span class="nb">str</span><span class="p">,</span> <span class="n">Any</span><span class="p">],</span>
|
|
<span class="n">stream</span><span class="p">,</span>
|
|
<span class="n">context</span><span class="o">=</span><span class="kc">None</span><span class="p">)</span> <span class="o">-></span> <span class="nb">bool</span><span class="p">:</span>
|
|
<span class="w"> </span><span class="sd">'''</span>
|
|
<span class="sd"> @brief: Run the TensorRT engine with the given inputs and outputs</span>
|
|
<span class="sd"> @param inputs: dict of input tensors, key is tensor name, value is tensor pointer or torch tensor</span>
|
|
<span class="sd"> @param outputs: dict of output tensors, key is tensor name, value is tensor pointer or torch tensor</span>
|
|
<span class="sd"> @param stream: cuda stream to enqueue the TensorRT engine on</span>
|
|
<span class="sd"> @param context: TensorRT execution context, if None, use the default context</span>
|
|
<span class="sd"> @return: True if enqueue succeeded, note the enqueue is an async call,</span>
|
|
<span class="sd"> returning True does not mean the execution is finished</span>
|
|
<span class="sd"> '''</span>
|
|
<span class="c1"># enqueue to the default context if context is not specified</span>
|
|
<span class="k">if</span> <span class="n">context</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
|
|
<span class="n">context</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">context</span>
|
|
|
|
<span class="kn">import</span> <span class="nn">torch</span>
|
|
<span class="k">for</span> <span class="n">tensor_name</span> <span class="ow">in</span> <span class="n">inputs</span><span class="p">:</span>
|
|
<span class="n">tensor</span> <span class="o">=</span> <span class="n">inputs</span><span class="p">[</span><span class="n">tensor_name</span><span class="p">]</span>
|
|
<span class="n">ptr</span> <span class="o">=</span> <span class="n">tensor</span><span class="o">.</span><span class="n">data_ptr</span><span class="p">()</span> <span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">tensor</span><span class="p">,</span>
|
|
<span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">)</span> <span class="k">else</span> <span class="n">tensor</span>
|
|
<span class="n">context</span><span class="o">.</span><span class="n">set_tensor_address</span><span class="p">(</span><span class="n">tensor_name</span><span class="p">,</span> <span class="n">ptr</span><span class="p">)</span>
|
|
<span class="k">for</span> <span class="n">tensor_name</span> <span class="ow">in</span> <span class="n">outputs</span><span class="p">:</span>
|
|
<span class="n">tensor</span> <span class="o">=</span> <span class="n">outputs</span><span class="p">[</span><span class="n">tensor_name</span><span class="p">]</span>
|
|
<span class="n">ptr</span> <span class="o">=</span> <span class="n">tensor</span><span class="o">.</span><span class="n">data_ptr</span><span class="p">()</span> <span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">tensor</span><span class="p">,</span>
|
|
<span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">)</span> <span class="k">else</span> <span class="n">tensor</span>
|
|
<span class="n">context</span><span class="o">.</span><span class="n">set_tensor_address</span><span class="p">(</span><span class="n">tensor_name</span><span class="p">,</span> <span class="n">ptr</span><span class="p">)</span>
|
|
<span class="n">ok</span> <span class="o">=</span> <span class="n">context</span><span class="o">.</span><span class="n">execute_async_v3</span><span class="p">(</span><span class="n">stream</span><span class="p">)</span>
|
|
<span class="k">return</span> <span class="n">ok</span></div>
|
|
|
|
|
|
<span class="k">def</span> <span class="nf">_debug_run</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span>
|
|
<span class="n">inputs</span><span class="p">:</span> <span class="n">Dict</span><span class="p">[</span><span class="nb">str</span><span class="p">,</span> <span class="s2">"torch.Tensor"</span><span class="p">],</span>
|
|
<span class="n">context</span><span class="o">=</span><span class="kc">None</span><span class="p">)</span> <span class="o">-></span> <span class="n">Dict</span><span class="p">[</span><span class="nb">str</span><span class="p">,</span> <span class="s2">"torch.Tensor"</span><span class="p">]:</span>
|
|
<span class="w"> </span><span class="sd">'''Run the engine enqueue with allocated output tensors, for debug purpose, since it is a sync call and slower than run</span>
|
|
<span class="sd"> '''</span>
|
|
<span class="kn">import</span> <span class="nn">torch</span>
|
|
<span class="n">torch_dtype_to_trt</span> <span class="o">=</span> <span class="p">{</span>
|
|
<span class="n">torch</span><span class="o">.</span><span class="n">float16</span><span class="p">:</span> <span class="n">trt</span><span class="o">.</span><span class="n">float16</span><span class="p">,</span>
|
|
<span class="n">torch</span><span class="o">.</span><span class="n">float32</span><span class="p">:</span> <span class="n">trt</span><span class="o">.</span><span class="n">float32</span><span class="p">,</span>
|
|
<span class="n">torch</span><span class="o">.</span><span class="n">int32</span><span class="p">:</span> <span class="n">trt</span><span class="o">.</span><span class="n">int32</span>
|
|
<span class="p">}</span>
|
|
<span class="n">inputs_info</span> <span class="o">=</span> <span class="p">[</span>
|
|
<span class="n">TensorInfo</span><span class="p">(</span><span class="n">name</span><span class="p">,</span> <span class="n">torch_dtype_to_trt</span><span class="p">[</span><span class="n">tensor</span><span class="o">.</span><span class="n">dtype</span><span class="p">],</span> <span class="n">tensor</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span>
|
|
<span class="k">for</span> <span class="n">name</span><span class="p">,</span> <span class="n">tensor</span> <span class="ow">in</span> <span class="n">inputs</span><span class="o">.</span><span class="n">items</span><span class="p">()</span>
|
|
<span class="p">]</span>
|
|
<span class="n">outputs_info</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">infer_shapes</span><span class="p">(</span><span class="n">inputs_info</span><span class="p">)</span>
|
|
<span class="n">outputs</span> <span class="o">=</span> <span class="p">{</span>
|
|
<span class="n">t</span><span class="o">.</span><span class="n">name</span><span class="p">:</span> <span class="n">torch</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="nb">tuple</span><span class="p">(</span><span class="n">t</span><span class="o">.</span><span class="n">shape</span><span class="p">),</span>
|
|
<span class="n">dtype</span><span class="o">=</span><span class="n">trt_dtype_to_torch</span><span class="p">(</span><span class="n">t</span><span class="o">.</span><span class="n">dtype</span><span class="p">),</span>
|
|
<span class="n">device</span><span class="o">=</span><span class="s1">'cuda'</span><span class="p">)</span>
|
|
<span class="k">for</span> <span class="n">t</span> <span class="ow">in</span> <span class="n">outputs_info</span>
|
|
<span class="p">}</span>
|
|
<span class="k">with</span> <span class="n">_scoped_stream</span><span class="p">()</span> <span class="k">as</span> <span class="n">stream</span><span class="p">:</span>
|
|
<span class="bp">self</span><span class="o">.</span><span class="n">run</span><span class="p">(</span><span class="n">inputs</span><span class="o">=</span><span class="n">inputs</span><span class="p">,</span>
|
|
<span class="n">outputs</span><span class="o">=</span><span class="n">outputs</span><span class="p">,</span>
|
|
<span class="n">stream</span><span class="o">=</span><span class="n">stream</span><span class="p">,</span>
|
|
<span class="n">context</span><span class="o">=</span><span class="n">context</span><span class="p">)</span>
|
|
<span class="k">return</span> <span class="n">outputs</span></div>
|
|
|
|
</pre></div>
|
|
|
|
</div>
|
|
</div>
|
|
<footer>
|
|
|
|
<hr/>
|
|
|
|
<div role="contentinfo">
|
|
<p>© Copyright 2023, NVidia.</p>
|
|
</div>
|
|
|
|
Built with <a href="https://www.sphinx-doc.org/">Sphinx</a> using a
|
|
<a href="https://github.com/readthedocs/sphinx_rtd_theme">theme</a>
|
|
provided by <a href="https://readthedocs.org">Read the Docs</a>.
|
|
|
|
|
|
</footer>
|
|
</div>
|
|
</div>
|
|
</section>
|
|
</div>
|
|
<script>
|
|
jQuery(function () {
|
|
SphinxRtdTheme.Navigation.enable(true);
|
|
});
|
|
</script>
|
|
|
|
</body>
|
|
</html> |