forked from NVIDIA/Megatron-LM
-
Notifications
You must be signed in to change notification settings - Fork 344
/
pretrain_gpt_distributed.sh
executable file
·68 lines (59 loc) · 1.44 KB
/
pretrain_gpt_distributed.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
# Runs the "345M" parameter model
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>/gpt2-vocab.json
MERGE_FILE=<Specify path to file>/gpt2-merges.txt
DATA_PATH=<Specify path and file prefix>_text_document
DISTRIBUTED_ARGS="
--nproc_per_node $GPUS_PER_NODE \
--nnodes $NNODES \
--node_rank $NODE_RANK \
--master_addr $MASTER_ADDR \
--master_port $MASTER_PORT
"
GPT_ARGS="
--num-layers 24 \
--hidden-size 1024 \
--num-attention-heads 16 \
--seq-length 1024 \
--max-position-embeddings 1024 \
--micro-batch-size 8 \
--global-batch-size 64 \
--lr 0.00015 \
--train-iters 500000 \
--lr-decay-iters 320000 \
--lr-decay-style cosine \
--min-lr 1.0e-5 \
--weight-decay 1e-2 \
--lr-warmup-fraction .01 \
--clip-grad 1.0 \
--fp16
"
DATA_ARGS="
--data-path $DATA_PATH \
--vocab-file $VOCAB_FILE \
--merge-file $MERGE_FILE \
--data-impl mmap \
--split 949,50,1
"
OUTPUT_ARGS="
--log-interval 100 \
--save-interval 10000 \
--eval-interval 1000 \
--eval-iters 10
"
torchrun $DISTRIBUTED_ARGS pretrain_gpt.py \
$GPT_ARGS \
$DATA_ARGS \
$OUTPUT_ARGS \
--distributed-backend nccl \
--save $CHECKPOINT_PATH \
--load $CHECKPOINT_PATH