-
Notifications
You must be signed in to change notification settings - Fork 1
/
run.sh
executable file
·72 lines (53 loc) · 2.62 KB
/
run.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
69
70
71
CUDA_VISIBLE_DEVICES=6 torchrun --nproc_per_node 1 --master_port=1234 \
train_custom.py \
--cfg-path instructiontuning_configs/instructtuning_imgsingle_resumecheck.yaml
CUDA_VISIBLE_DEVICES=3,4,5 torchrun --nproc_per_node 4 --master_port=1734 \
evaluate.py \
--cfg-path eval_configs/qa/okvqa_eval_isosrcnew.yaml
torchrun --nproc_per_node 1 --master_port=1434 \
train_custom.py \
--cfg-path instructiontuning_configs/instructtuning_all_newweight_womm.yaml
torchrun --nproc_per_node 8 --master_port=1234 \
train_custom.py \
--cfg-path train_configs/minigpt4_stage1_fromtwomodal_doubledata_newpt.yaml
torchrun --nproc_per_node 8 --master_port=1231 \
evaluate.py \
--cfg-path eval_configs/caption/clothocap_eval.yaml \
--options model.ckpt=/mnt/bn/vlpopt/zhaozijia/LLM/minigpt4/minigpt4_stage1_pretrain_fromaudioonly_ftqformer_audio/20230511000/checkpoint_9.pth \
run.output_dir="$output_dir$type";
torchrun --nproc_per_node 8 --master_port=1231 \
evaluate.py \
--cfg-path eval_configs/qa/okvqa_eval.yaml \
--options model.ckpt=/mnt/bn/vlpopt/zhaozijia/LLM/minigpt4/minigpt4_stage1_fromblipflan_cclaion/20230510202/checkpoint_7.pth
template="###Human: Given following image: <query>, Answer the following question with a one-word answer based on the content of the given image, <question>###Assistant:"
echo $template
torchrun --nproc_per_node 8 --master_port=1231 \
evaluate.py \
--cfg-path eval_configs/qa/vqa_eval.yaml \
--options model.ckpt=/mnt/bn/vlpopt/zhaozijia/LLM/minigpt4/minigpt4_stage1_fromblipflan_cclaion/20230510202/checkpoint_7.pth \
torchrun --nproc_per_node 8 --master_port=1231 \
evaluate.py \
--cfg-path eval_configs/caption/asvd/avsd_eval_video.yaml
CUDA_VISIBLE_DEVICES=2 torchrun --nproc_per_node 1 --master_port=1325 \
evaluate.py \
--cfg-path eval_configs/qa/msrvtt_avqa_eval_4f.yaml
CUDA_VISIBLE_DEVICES=4,5,6,7 torchrun --nproc_per_node 4 --master_port=1239 \
evaluate.py \
--cfg-path eval_configs/caption/clothocap_eval.yaml
CUDA_VISIBLE_DEVICES=2 torchrun --nproc_per_node 1 --master_port=1236 \
train_custom.py \
--cfg-path instructiontuning_configs/instructtuning_imgsingle.yaml
CUDA_VISIBLE_DEVICES=3 torchrun --nproc_per_node 1 --master_port=1235 \
train_custom.py \
--cfg-path instructiontuning_configs/instructtuning_imgsingle_resumecheck.yaml
type=msvdqa
torchrun --nproc_per_node 8 --master_port=1231 \
evaluate.py \
--cfg-path eval_configs/qa_new/msvd_videoqa_eval_4f.yaml \
--options model.ckpt=$path \
run.output_dir="$output_dir$type";
torchrun --nproc_per_node 1 --master_port=1231 \
evaluate.py \
--cfg-path eval_configs/caption/valorcaps_eval.yaml \
--options model.ckpt=$path \
run.output_dir="$output_dir$type";