UniRec/examples/training/multi_gpu_train_ml100k.sh at main · microsoft/UniRec · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
# Copyright (c) Microsoft Corporation.
# Licensed under the MIT license.

#!/bin/bash

###############################################################################################
### Please modify the following variables according to your device and mission requirements ###
###############################################################################################
# ============DDP Configuration===========
GPU_INDICES="0" # e.g. "0,1"

# how many nodes to use? (one node may have multiple GPUs)
NUM_NODES=1

# number of process in each node. (the number should be equal to the number of GPU_INDICES)
NPROC_PER_NODE=1

# ========================================

LOCAL_ROOT="$HOME/workspace/UniRec"  # path to UniRec
###############################################################################################


# default parameters for local run
MY_DIR=$LOCAL_ROOT
ALL_DATA_ROOT="$LOCAL_ROOT/data"
OUTPUT_ROOT="$LOCAL_ROOT/output"
MODEL_NAME='SASRec' # [AvgHist, AttHist, MF, SVDPlusPlus, GRU, SASRec, ConvFormer, FASTConvFormer]
loss_type='bce' # [bce, bpr, softmax]
DATASET_NAME="ml-100k"
max_seq_len=10
verbose=2

cd $MY_DIR
export PYTHONPATH=$PWD

# overall config
DATA_TYPE='SeqRecDataset'  # BaseDataset SeqRecDataset

# train
learning_rate=0.001
test_protocol='one_vs_all'  # 'one_vs_k' 'one_vs_all' 'session_aware'

exp_name="SASRec-ml100k"

ALL_RESULTS_ROOT="$OUTPUT_ROOT/$DATASET_NAME/$MODEL_NAME"
mkdir -p $ALL_RESULTS_ROOT
### train ###################################

CUDA_VISIBLE_DEVICES=$GPU_INDICES torchrun --nnodes=$NUM_NODES --nproc_per_node=$NPROC_PER_NODE --rdzv_backend=c10d --rdzv_endpoint=127.0.0.1:29400 unirec/main/main.py \
    --config_dir="unirec/config" \
    --model=$MODEL_NAME \
    --dataloader=$DATA_TYPE \
    --dataset=$DATASET_NAME \
    --dataset_path=$ALL_DATA_ROOT"/"$DATASET_NAME \
    --output_path=$ALL_RESULTS_ROOT"/train" \
    --learning_rate=$learning_rate \
    --dropout_prob=0.0 \
    --embedding_size=64 \
    --hidden_size=64 \
    --use_pre_item_emb=0 \
    --loss_type=$loss_type \
    --max_seq_len=$max_seq_len \
    --has_user_bias=0 \
    --has_item_bias=0 \
    --epochs=100  \
    --early_stop=10 \
    --batch_size=512 \
    --n_sample_neg_train=5 \
    --neg_by_pop_alpha=1.0 \
    --valid_protocol=$test_protocol \
    --test_protocol=$test_protocol \
    --grad_clip_value=-1 \
    --weight_decay=1e-6 \
    --history_mask_mode='autoregressive' \
    --user_history_filename="user_history" \
    --user_history_file_format="user-item_seq" \
    --metrics="['hit@10;20', 'ndcg@10;20']" \
    --key_metric="ndcg@10" \
    --num_workers=4 \
    --num_workers_test=0 \
    --verbose=$verbose \
    --exp_name=$exp_name \
    --use_wandb=0 \
    --gpu_id=-1
# done