-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathgptq_deepseek_nohup.sh
108 lines (82 loc) · 3.57 KB
/
gptq_deepseek_nohup.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
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
#!/bin/bash
export PYTHONPATH=/home/LeiFeng/xiaolong/moe_quantize/optimum/:$PYTHONPATH:/home/LeiFeng/xiaolong/moe_quantize/auto_gptq/:$PYTHONPATH
export DEBUG=0
# Command 1
export CUDA_VISIBLE_DEVICES=3
nohup python quantize_gptq_deepseek.py \
--model_name deepseek-ai/deepseek-moe-16b-base \
--nsamples 512 \
--group_size 64 \
--bits all_4 > log_cuda_3_all_4.out &
# Command 2
export CUDA_VISIBLE_DEVICES=4
nohup python quantize_gptq_deepseek.py \
--model_name deepseek-ai/deepseek-moe-16b-base \
--nsamples 512 \
--group_size 64 \
--bits all_2 > log_cuda_4_all_2.out &
# Command 3
export CUDA_VISIBLE_DEVICES=5
nohup python quantize_gptq_deepseek.py \
--model_name deepseek-ai/deepseek-moe-16b-base \
--nsamples 512 \
--group_size 64 \
--bits moe.all_mlp.2+other_block.4 > log_cuda_5_moe.all_mlp.2+other_block.4.out &
# Command 4
export CUDA_VISIBLE_DEVICES=6
nohup python quantize_gptq_deepseek.py \
--model_name deepseek-ai/deepseek-moe-16b-base \
--nsamples 512 \
--group_size 64 \
--bits moe.shared_4.other.2+other_block_4 > log_cuda_6_moe.shared_4.other.2+other_block_4.out &
# Command 5
export CUDA_VISIBLE_DEVICES=7
nohup python quantize_gptq_deepseek.py \
--model_name deepseek-ai/deepseek-moe-16b-base \
--nsamples 512 \
--group_size 64 \
--bits moe.shared_2.other.4+other_block_4 > log_cuda_7_moe.shared_2.other.4+other_block_4.out &
# Command 6
export CUDA_VISIBLE_DEVICES=7
nohup python quantize_gptq_deepseek.py \
--model_name deepseek-ai/deepseek-moe-16b-base \
--nsamples 512 \
--group_size 64 \
--bits all_8 > log_cuda_7_all_8.out &
# Command 7
export CUDA_VISIBLE_DEVICES=6
nohup python quantize_gptq_deepseek.py \
--model_name deepseek-ai/deepseek-moe-16b-base \
--nsamples 512 \
--group_size 64 \
--bits moe.all_mlp.4+other_block.8 > log_cuda_6_moe.all_mlp.4+other_block.8.out &
# Command 8
export CUDA_VISIBLE_DEVICES=5
nohup python quantize_gptq_deepseek.py \
--model_name deepseek-ai/deepseek-moe-16b-base \
--nsamples 512 \
--group_size 64 \
--bits moe.shared_4.other.2+other_block.8 > log_cuda_5_moe.shared_4.other.2+other_block.8.out &
# Command 9
export CUDA_VISIBLE_DEVICES=4
nohup python quantize_gptq_deepseek.py \
--model_name deepseek-ai/deepseek-moe-16b-base \
--nsamples 512 \
--group_size 64 \
--bits moe.shared_2.other.4+other_block.8 > log_cuda_4_moe.shared_2.other.4+other_block.8.out &
# export PYTHONPATH=/home/LeiFeng/xiaolong/moe_quantize/optimum/:$PYTHONPATH:/home/LeiFeng/xiaolong/moe_quantize/awq/:$PYTHONPATH
# export CUDA_VISIBLE_DEVICES=5
# nohup python quantize_awq_deepseek.py --bits 4 > quantize_awq_deepseek_bit.4.log &
# export PYTHONPATH=/home/LeiFeng/xiaolong/moe_quantize/optimum/:$PYTHONPATH:/home/LeiFeng/xiaolong/moe_quantize/awq/:$PYTHONPATH
# export CUDA_VISIBLE_DEVICES=6
# nohup python quantize_awq_deepseek.py --bits 2 > quantize_awq_deepseek_bit.2.log &
# export PYTHONPATH=/home/LeiFeng/xiaolong/moe_quantize/optimum/:$PYTHONPATH:/home/LeiFeng/xiaolong/moe_quantize/awq/:$PYTHONPATH
# export CUDA_VISIBLE_DEVICES=7
# nohup python quantize_awq_deepseek.py --bits 8 > quantize_awq_deepseek_bit.8.log &
export PYTHONPATH=/home/LeiFeng/xiaolong/moe_quantize/optimum/:$PYTHONPATH:/home/LeiFeng/xiaolong/moe_quantize/auto_gptq/:$PYTHONPATH
export CUDA_VISIBLE_DEVICES=4
nohup python quantize_gptq_deepseek.py \
--model_name deepseek-ai/deepseek-moe-16b-base \
--nsamples 512 \
--group_size 64 \
--bits moe.shared_8.top1_8.other_2+other_block.4 > run_log/log_cuda_4_moe.shared_8.top1_8.other_2+other_block.4.out &