Compute
Elastic Cloud Server
Huawei Cloud Flexus
Bare Metal Server
Auto Scaling
Image Management Service
Dedicated Host
FunctionGraph
Cloud Phone Host
Huawei Cloud EulerOS
Networking
Virtual Private Cloud
Elastic IP
Elastic Load Balance
NAT Gateway
Direct Connect
Virtual Private Network
VPC Endpoint
Cloud Connect
Enterprise Router
Enterprise Switch
Global Accelerator
Management & Governance
Cloud Eye
Identity and Access Management
Cloud Trace Service
Resource Formation Service
Tag Management Service
Log Tank Service
Config
OneAccess
Resource Access Manager
Simple Message Notification
Application Performance Management
Application Operations Management
Organizations
Optimization Advisor
IAM Identity Center
Cloud Operations Center
Resource Governance Center
Migration
Server Migration Service
Object Storage Migration Service
Cloud Data Migration
Migration Center
Cloud Ecosystem
KooGallery
Partner Center
User Support
My Account
Billing Center
Cost Center
Resource Center
Enterprise Management
Service Tickets
HUAWEI CLOUD (International) FAQs
ICP Filing
Support Plans
My Credentials
Customer Operation Capabilities
Partner Support Plans
Professional Services
Analytics
MapReduce Service
Data Lake Insight
CloudTable Service
Cloud Search Service
Data Lake Visualization
Data Ingestion Service
GaussDB(DWS)
DataArts Studio
Data Lake Factory
DataArts Lake Formation
IoT
IoT Device Access
Others
Product Pricing Details
System Permissions
Console Quick Start
Common FAQs
Instructions for Associating with a HUAWEI CLOUD Partner
Message Center
Security & Compliance
Security Technologies and Applications
Web Application Firewall
Host Security Service
Cloud Firewall
SecMaster
Anti-DDoS Service
Data Encryption Workshop
Database Security Service
Cloud Bastion Host
Data Security Center
Cloud Certificate Manager
Edge Security
Situation Awareness
Managed Threat Detection
Blockchain
Blockchain Service
Web3 Node Engine Service
Media Services
Media Processing Center
Video On Demand
Live
SparkRTC
MetaStudio
Storage
Object Storage Service
Elastic Volume Service
Cloud Backup and Recovery
Storage Disaster Recovery Service
Scalable File Service Turbo
Scalable File Service
Volume Backup Service
Cloud Server Backup Service
Data Express Service
Dedicated Distributed Storage Service
Containers
Cloud Container Engine
SoftWare Repository for Container
Application Service Mesh
Ubiquitous Cloud Native Service
Cloud Container Instance
Databases
Relational Database Service
Document Database Service
Data Admin Service
Data Replication Service
GeminiDB
GaussDB
Distributed Database Middleware
Database and Application Migration UGO
TaurusDB
Middleware
Distributed Cache Service
API Gateway
Distributed Message Service for Kafka
Distributed Message Service for RabbitMQ
Distributed Message Service for RocketMQ
Cloud Service Engine
Multi-Site High Availability Service
EventGrid
Dedicated Cloud
Dedicated Computing Cluster
Business Applications
Workspace
ROMA Connect
Message & SMS
Domain Name Service
Edge Data Center Management
Meeting
AI
Face Recognition Service
Graph Engine Service
Content Moderation
Image Recognition
Optical Character Recognition
ModelArts
ImageSearch
Conversational Bot Service
Speech Interaction Service
Huawei HiLens
Video Intelligent Analysis Service
Developer Tools
SDK Developer Guide
API Request Signing Guide
Terraform
Koo Command Line Interface
Content Delivery & Edge Computing
Content Delivery Network
Intelligent EdgeFabric
CloudPond
Intelligent EdgeCloud
Solutions
SAP Cloud
High Performance Computing
Developer Services
ServiceStage
CodeArts
CodeArts PerfTest
CodeArts Req
CodeArts Pipeline
CodeArts Build
CodeArts Deploy
CodeArts Artifact
CodeArts TestPlan
CodeArts Check
CodeArts Repo
Cloud Application Engine
MacroVerse aPaaS
KooMessage
KooPhone
KooDrive
Help Center/ ModelArts/ ModelArts User Guide (Lite Server)/ Using Lite Server Resources/ PyTorch GPU Training and Inference Guide for GPT-2

PyTorch GPU Training and Inference Guide for GPT-2

Updated on 2025-02-13 GMT+08:00

Scenario

This section describes how to use the DeepSpeed framework to train GPT-2 in a GP Ant8 BMS (single-node single-card and single-node multi-card). After the training, the automatically generated content and interactive dialog box mode are provided.

Background

  • Megatron-DeepSpeed

    DeepSpeed is a PyTorch-based deep learning model training framework. It combines Megatron-LM and DeepSpeed to train on systems with distributed computing, and takes full advantage of the parallel processing capabilities of multiple GPUs and deep learning accelerators. Large-scale language models can be efficiently trained.

    Megatron-LM is a model for large-scale language modeling. Based on the Generative Pre-trained Transformer (GPT) architecture, it is a neural network model based on the self-attention mechanism and is widely used in natural language processing tasks, such as text generation, machine translation, and dialog systems.

    DeepSpeed is an open-source library for accelerating deep learning training. It is optimized for large-scale model and distributed training, which can significantly improve the training speed and efficiency. DeepSpeed provides various technologies and optimization policies, including distributed gradient descent, model parallelization, gradient accumulation, and dynamic precision scaling. It also supports optimizing the memory usage and compute resource allocation of foundation models.

  • GPT2

    GPT2 is a new pre-trained model released by OpenAI based on the GPT model in 2018. It is a large language model based on Transformer. It is trained on a large number of datasets and directly runs a pre-trained GPT-2 model. Given a preset start word or sentence, it can randomly generate subsequent text.

Preparing the Environment

Purchase a GPU-powered BMS on ModelArts and select a general-purpose image for AIGC to train a GPT2 model. The following images and specifications are used as example:

  • Image: Ubuntu 20.04 x86 64bit SDI3 for Ant8 BareMetal with RoCE and NVIDIA-525 CUDA-12.0
  • BMS specifications: GP Ant8, including eight GPU cards and eight RoCE NICs

To purchase an Ant8 BMS, submit a service ticket on the Huawei Cloud official website.

Step 1 Installing a Model

  1. Install the Megatron-DeepSpeed framework.

    1. Log in to the GPU BMS using SSH as user root. For details, see Logging In to a BMS Using an SSH Key Pair.
    2. Pull the PyTorch image. You can download a common image source.
      docker pull nvcr.io/nvidia/pytorch:21.10-py3
    3. Start the container.
      docker run -d -t --network=host --gpus all --privileged --ipc=host --ulimit memlock=-1 --ulimit stack=67108864 --name megatron-deepspeed -v /etc/localtime:/etc/localtime -v /root/.ssh:/root/.ssh nvcr.io/nvidia/pytorch:21.10-py3
    4. Access the container terminal.
      docker exec -it megatron-deepspeed bash
    5. Download the Megatron-DeepSpeed framework.
      git clone https://github.com/bigscience-workshop/Megatron-DeepSpeed
      NOTE:

      If the git clone fails to be executed, download the model to the local host, copy it to the server, and run the docker cp command to copy it to the container.

    6. Install the Megatron-DeepSpeed framework.
      cd Megatron-DeepSpeed
      pip install -r requirements.txt -i http://mirrors.myhuaweicloud.com/pypi/web/simple --trusted-host mirrors.myhuaweicloud.com
      pip install mpi4py -i http://mirrors.myhuaweicloud.com/pypi/web/simple --trusted-host mirrors.myhuaweicloud.com
    7. Modify the test code and comment out the line where the assertion is located.
      vim /workspace/Megatron-DeepSpeed/megatron/model/fused_softmax.py +191

      Add # before assert mask is None, "Mask is silently ignored due to the use of a custom kernel".

      # assert mask is None, "Mask is silently ignored due to the use of a custom kernel"

  2. Download and preprocess the dataset.

    The OSCAR dataset in JSON format with 1 GB 79K-record is used.

    1. Download the dataset.
      wget https://huggingface.co/bigscience/misc-test-data/resolve/main/stas/oscar-1GB.jsonl.xz 
      wget https://s3.amazonaws.com/models.huggingface.co/bert/gpt2-vocab.json
      wget https://s3.amazonaws.com/models.huggingface.co/bert/gpt2-merges.txt
    2. Decompress the dataset.
      xz -d oscar-1GB.jsonl.xz
    3. Preprocess data.
      python3 tools/preprocess_data.py \
          --input oscar-1GB.jsonl \
          --output-prefix meg-gpt2 \
          --vocab gpt2-vocab.json \
          --dataset-impl mmap \
          --tokenizer-type GPT2BPETokenizer \
          --merge-file gpt2-merges.txt \
          --append-eod \
          --workers 8
      NOTE:
      If the error message "np.float" is displayed, change it to "float" as prompted.
      Figure 1 Data preprocessing error
    4. Complete the preprocessing.
      Figure 2 Data preprocessed
    5. Create the data directory and move the processed data.
      mkdir data 
      mv meg-gpt2* ./data 
      mv gpt2* ./data

Step 2 Single-Node Single-Card Training

This section describes how to use a GP Ant8 BMS to train the GPT-2 MEDIUM model on a single node.

  1. Create the pre-training script file.

    1. Create the script.
      vim pretrain_gpt2.sh
    2. Add the following information to the script:
      #! /bin/bash
      
      # Runs the "345M" parameter model
      
      GPUS_PER_NODE=1
      # Change for multinode config
      MASTER_ADDR=localhost
      MASTER_PORT=6000
      NNODES=1
      NODE_RANK=0
      WORLD_SIZE=$(($GPUS_PER_NODE*$NNODES))
      
      DATA_PATH=data/meg-gpt2_text_document
      CHECKPOINT_PATH=checkpoints/gpt2
      
      DISTRIBUTED_ARGS="--nproc_per_node $GPUS_PER_NODE --nnodes $NNODES --node_rank $NODE_RANK --master_addr $MASTER_ADDR --master_port $MASTER_PORT"
      
      python -m torch.distributed.launch $DISTRIBUTED_ARGS \
             pretrain_gpt.py \
             --tensor-model-parallel-size 1 \
             --pipeline-model-parallel-size 1 \
             --num-layers 24 \
             --hidden-size 1024 \
             --num-attention-heads 16 \
             --micro-batch-size 4 \
             --global-batch-size 8 \
             --seq-length 1024 \
             --max-position-embeddings 1024 \
             --train-iters 5000 \
             --lr-decay-iters 320000 \
             --save $CHECKPOINT_PATH \
             --load $CHECKPOINT_PATH \
             --data-path $DATA_PATH \
             --vocab-file data/gpt2-vocab.json \
             --merge-file data/gpt2-merges.txt \
             --data-impl mmap \
             --split 949,50,1 \
             --distributed-backend nccl \
             --lr 0.00015 \
             --lr-decay-style cosine \
             --min-lr 1.0e-5 \
             --weight-decay 1e-2 \
             --clip-grad 1.0 \
             --lr-warmup-fraction .01 \
             --checkpoint-activations \
             --log-interval 10 \
             --save-interval 500 \
             --eval-interval 100 \
             --eval-iters 10 \
             --fp16

  2. Start training.

    Configure the parameters in the pre-training script for the single-node single-card training.

    GPUS_PER_NODE=1
    NNODES=1
    NODE_RANK=0
    1. Start the pre-training.
      nohup sh ./pretrain_gpt2.sh &
      Figure 3 Starting pre-training
    2. View training logs and monitor programs in real time.
      tail -f nohup.out

      If the following information is displayed, the model is trained.

      Figure 4 Model training completed

      Observe GPU usage during training.

      Figure 5 GPU usage

  3. View the checkpoint of the generated model.

    The checkpoint of the model used in this case is saved in /workspace/Megatron-DeepSpeed/checkpoints/gpt2.

    ll ./checkpoints/gpt2
    Figure 6 Model checkpoint

Step 3 Single-Node Multi-Card Training

Compared with single-node single-card training, single-node multi-card training only needs to set multi-card parameters in the pre-training script. Other steps are the same as those of single-node single-card training.

  1. Currently, eight GPUs are selected for BMS. Adjust the following parameter in the pre-training script:

    GPUS_PER_NODE=8

  2. Configure the global batch size, micro batch size, and data_parallel_size parameters. The value of global_batch_size can be exactly divided by the value of micro_batch_size * data_parallel_size.

    Configure the parameters.

    global_batch_size = 64 
    micro_batch_size = 4 
    data_parallel_size = 8

  3. The content of the pre-training script is as follows:

    #! /bin/bash
    
    # Runs the "345M" parameter model
    
    GPUS_PER_NODE=8
    # Change for multinode config
    MASTER_ADDR=localhost
    MASTER_PORT=6000
    NNODES=1
    NODE_RANK=0
    WORLD_SIZE=$(($GPUS_PER_NODE*$NNODES))
    
    DATA_PATH=data/meg-gpt2_text_document
    CHECKPOINT_PATH=checkpoints/gpt2
    
    DISTRIBUTED_ARGS="--nproc_per_node $GPUS_PER_NODE --nnodes $NNODES --node_rank $NODE_RANK --master_addr $MASTER_ADDR --master_port $MASTER_PORT"
    
    python -m torch.distributed.launch $DISTRIBUTED_ARGS \
           pretrain_gpt.py \
           --tensor-model-parallel-size 1 \
           --pipeline-model-parallel-size 1 \
           --num-layers 24 \
           --hidden-size 1024 \
           --num-attention-heads 16 \
           --micro-batch-size 4 \
           --global-batch-size 64 \
           --seq-length 1024 \
           --max-position-embeddings 1024 \
           --train-iters 5000 \
           --lr-decay-iters 320000 \
           --save $CHECKPOINT_PATH \
           --load $CHECKPOINT_PATH \
           --data-path $DATA_PATH \
           --vocab-file data/gpt2-vocab.json \
           --merge-file data/gpt2-merges.txt \
           --data-impl mmap \
           --split 949,50,1 \
           --distributed-backend nccl \
           --lr 0.00015 \
           --lr-decay-style cosine \
           --min-lr 1.0e-5 \
           --weight-decay 1e-2 \
           --clip-grad 1.0 \
           --lr-warmup-fraction .01 \
           --checkpoint-activations \
           --log-interval 10 \
           --save-interval 500 \
           --eval-interval 100 \
           --eval-iters 10 \
           --fp16

    The following figure shows the GPU usage.

    Figure 7 GPU usage

Step4 Using the GPT-2 Model to Generate Text

  1. Generate text automatically.

    1. Create a text generation script.
      vim generate_text.sh

      Add the following content:

      #!/bin/bash
      
      CHECKPOINT_PATH=checkpoints/gpt2
      VOCAB_FILE=data/gpt2-vocab.json
      MERGE_FILE=data/gpt2-merges.txt
      
      python tools/generate_samples_gpt.py \
             --tensor-model-parallel-size 1 \
             --num-layers 24 \
             --hidden-size 1024 \
             --load $CHECKPOINT_PATH \
             --num-attention-heads 16 \
             --max-position-embeddings 1024 \
             --tokenizer-type GPT2BPETokenizer \
             --fp16 \
             --micro-batch-size 2 \
             --seq-length 1024 \
             --out-seq-length 1024 \
             --temperature 1.0 \
             --vocab-file $VOCAB_FILE \
             --merge-file $MERGE_FILE \
             --genfile unconditional_samples.json \
             --num-samples 2 \
             --top_p 0.9 \
             --recompute
    2. Generate text.
      sh ./generate_text.sh

      If the following information is displayed, the text is generated.

      Figure 8 Text generated
    3. View the generated text.
      cat unconditional_samples.json

      The following figure shows the output.

      Figure 9 File information

  2. Enable interactive dialog.

    1. Create a text generation script.
      vim interactive_text.sh

      Enter the following information:

      #!/bin/bash
      
      CHECKPOINT_PATH=/workspace/Megatron-DeepSpeed/checkpoints/gpt2_345m
      VOCAB_FILE=/workspace/Megatron-DeepSpeed/data/gpt2-vocab.json
      MERGE_FILE=/workspace/Megatron-DeepSpeed/data/gpt2-merges.txt
      
      deepspeed /workspace/Megatron-DeepSpeed/tools/generate_samples_gpt.py \
             --tensor-model-parallel-size 1 \
             --num-layers 24 \
             --hidden-size 1024 \
             --load $CHECKPOINT_PATH \
             --num-attention-heads 16 \
             --max-position-embeddings 1024 \
             --tokenizer-type GPT2BPETokenizer \
             --fp16 \
             --micro-batch-size 2 \
             --seq-length 1024 \
             --out-seq-length 1024 \
             --temperature 1.0 \
             --vocab-file $VOCAB_FILE \
             --merge-file $MERGE_FILE \
             --genfile unconditional_samples.json \
             --num-samples 0 \
             --top_p 0.9 \
             --recompute
    2. Start interactive dialog.
      bash interactive_text.sh

      The following information is displayed. Enter huawei and press Enter.

      Context prompt (stop to exit) >>> huawei

      The text is automatically generated. The output depends on the model training and dataset.

      Figure 10 Model output

We use cookies to improve our site and your experience. By continuing to browse our site you accept our cookie policy. Find out more

Feedback

Feedback

Feedback

0/500

Selected Content

Submit selected content with the feedback