-
Notifications
You must be signed in to change notification settings - Fork 1
/
pretrain_t5_distributed_with_mp.sh
68 lines (60 loc) · 1.46 KB
/
pretrain_t5_distributed_with_mp.sh
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
#!/bin/bash
export CUDA_DEVICE_MAX_CONNECTIONS=1
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))
CHECKPOINT_PATH=<Specify path>
VOCAB_FILE=<Specify path to file>/t5-vocab.txt
DATA_PATH=<Specify path and file prefix>_text_sentence
DISTRIBUTED_ARGS="
--nproc_per_node $GPUS_PER_NODE \
--nnodes $NNODES \
--node_rank $NODE_RANK \
--master_addr $MASTER_ADDR \
--master_port $MASTER_PORT
"
T5_ARGS="
--tensor-model-parallel-size 2 \
--num-layers 12 \
--hidden-size 768 \
--num-attention-heads 12 \
--kv-channels 64 \
--ffn-hidden-size 3072 \
--encoder-seq-length 512 \
--decoder-seq-length 128 \
--max-position-embeddings 512 \
--micro-batch-size 16 \
--global-batch-size 128 \
--lr 0.0001 \
--train-iters 1000000 \
--lr-decay-iters 1000000 \
--lr-decay-style linear \
--min-lr 0.00001 \
--weight-decay 1e-2 \
--lr-warmup-fraction .01 \
--clip-grad 1.0 \
--fp16 \
--vocab-extra-ids 100
"
DATA_ARGS="
--data-path $DATA_PATH \
--vocab-file $VOCAB_FILE \
--split 949,50,1
"
OUTPUT_ARGS="
--log-interval 100 \
--save-interval 10000 \
--eval-interval 1000 \
--eval-iters 10
"
torchrun $DISTRIBUTED_ARGS pretrain_t5.py \
$T5_ARGS \
$DATA_ARGS \
$OUTPUT_ARGS \
--distributed-backend nccl \
--save $CHECKPOINT_PATH \
--load $CHECKPOINT_PATH