Mila
Deep Neural Network Library
Loading...
Searching...
No Matches
GptTransformer.ixx File Reference

GPT-2 style transformer network (decoder-only) for autoregressive language modeling. More...

#include <string>
#include <vector>
#include <memory>
#include <sstream>
#include <type_traits>
#include <stdexcept>
#include <cstdint>
#include <cstring>
#include <ostream>
#include <iostream>
#include <filesystem>
#include <format>
#include <optional>
#include <cassert>
#include <random>
#include <chrono>
#include <algorithm>
#include <numeric>
import Serialization.Tensor;
import Serialization.PretrainedReader;
import Compute.ExecutionContextFactory;
import Compute.CpuMemoryResource;
import Logging.Logger;
import Compute.DeviceId;
import Compute.DeviceType;
import Compute.Device;
import Dnn.ActivationType;
import Serialization.Mode;
import Compute.DeviceTypeTraits;
import Dnn.RuntimeMode;
import Dnn.ModelType;
import Serialization.ModelArchive;
import Dnn.ComponentType;
import Dnn.Components.Lpe;
import Dnn.Components.GptTransformer:Presets;
import Dnn.Tensor;
import Dnn.TensorDataType;
import Dnn.ITensor;
import Dnn.Components.Linear;
import Dnn.Components.LayerNorm;
import Dnn.TensorTypes;
import Dnn.TensorDataTypeTraits;
import Dnn.LanguageNetwork;
import Compute.ExecutionContext;
import Dnn.Components.GptBlock;
import Dnn.Component;

Classes

class  Mila::Dnn::GptTransformer< TDeviceType, TPrecision >
 GPT-2 style transformer (decoder-only) for autoregressive token prediction. More...

Namespaces

namespace  Mila
 Mila main API namespace.

Detailed Description

GPT-2 style transformer network (decoder-only) for autoregressive language modeling.

Device-templated network implementing a GPT-2 style transformer decoder.