-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathresults_ab1.log
41 lines (41 loc) · 7.43 KB
/
results_ab1.log
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
[2024-02-14 20:40:57,766] [INFO] [real_accelerator.py:161:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Namespace(model='JackFram/llama-68m', target='meta-llama/Llama-2-13b-hf', dataset='cnn', growmap='/home/zhuominc/workspace/Sequoia/68m_13b/growmaps/L40-CNN-68m-13b-stochastic-S4.pt', start=0, end=200, T=0.6, P=1.0, DP=0.99, D=1, B=10, W=32, M=800, Mode='greedy', decay=0.85, negative=False, static=False, offloading=False)
total time :152.59179s, latency :0.02342s, decoding step: 6516, large model step: 3332, 1.9555822328931574
[2024-02-14 20:43:59,048] [INFO] [real_accelerator.py:161:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Namespace(model='JackFram/llama-68m', target='meta-llama/Llama-2-13b-hf', dataset='cnn', growmap='/home/zhuominc/workspace/Sequoia/68m_13b/growmaps/L40-CNN-68m-13b-stochastic-S8.pt', start=0, end=200, T=0.6, P=1.0, DP=0.99, D=1, B=10, W=32, M=800, Mode='greedy', decay=0.85, negative=False, static=False, offloading=False)
total time :136.77507s, latency :0.02087s, decoding step: 6553, large model step: 2930, 2.236518771331058
[2024-02-14 20:46:44,410] [INFO] [real_accelerator.py:161:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Namespace(model='JackFram/llama-68m', target='meta-llama/Llama-2-13b-hf', dataset='cnn', growmap='/home/zhuominc/workspace/Sequoia/68m_13b/growmaps/L40-CNN-68m-13b-stochastic-S16.pt', start=0, end=200, T=0.6, P=1.0, DP=0.99, D=1, B=10, W=32, M=800, Mode='greedy', decay=0.85, negative=False, static=False, offloading=False)
total time :126.98852s, latency :0.01933s, decoding step: 6570, large model step: 2640, 2.4886363636363638
[2024-02-14 20:49:20,874] [INFO] [real_accelerator.py:161:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Namespace(model='JackFram/llama-68m', target='meta-llama/Llama-2-13b-hf', dataset='cnn', growmap='/home/zhuominc/workspace/Sequoia/68m_13b/growmaps/L40-CNN-68m-13b-stochastic-S32.pt', start=0, end=200, T=0.6, P=1.0, DP=0.99, D=1, B=10, W=32, M=800, Mode='greedy', decay=0.85, negative=False, static=False, offloading=False)
total time :125.63275s, latency :0.01923s, decoding step: 6534, large model step: 2340, 2.792307692307692
[2024-02-14 20:51:58,169] [INFO] [real_accelerator.py:161:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Namespace(model='JackFram/llama-68m', target='meta-llama/Llama-2-13b-hf', dataset='cnn', growmap='/home/zhuominc/workspace/Sequoia/68m_13b/growmaps/L40-CNN-68m-13b-stochastic-S64.pt', start=0, end=200, T=0.6, P=1.0, DP=0.99, D=1, B=10, W=32, M=800, Mode='greedy', decay=0.85, negative=False, static=False, offloading=False)
total time :124.98124s, latency :0.01880s, decoding step: 6647, large model step: 2199, 3.022737608003638
[2024-02-14 20:54:39,387] [INFO] [real_accelerator.py:161:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Namespace(model='JackFram/llama-68m', target='meta-llama/Llama-2-13b-hf', dataset='cnn', growmap='/home/zhuominc/workspace/Sequoia/68m_13b/growmaps/L40-CNN-68m-13b-stochastic-S128.pt', start=0, end=200, T=0.6, P=1.0, DP=0.99, D=1, B=10, W=32, M=800, Mode='greedy', decay=0.85, negative=False, static=False, offloading=False)
[2024-02-14 20:55:46,676] [INFO] [real_accelerator.py:161:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Namespace(model='JackFram/llama-68m', target='meta-llama/Llama-2-13b-hf', dataset='cnn', growmap='/home/zhuominc/workspace/Sequoia/68m_13b/growmaps/L40-CNN-68m-13b-stochastic-S4.pt', start=0, end=200, T=0.6, P=1.0, DP=0.99, D=1, B=10, W=32, M=800, Mode='greedy', decay=0.85, negative=False, static=False, offloading=False)
total time :568.53068s, latency :0.02262s, decoding step: 25137, large model step: 12634, 1.9896311540288112
[2024-02-14 21:05:44,258] [INFO] [real_accelerator.py:161:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Namespace(model='JackFram/llama-68m', target='meta-llama/Llama-2-13b-hf', dataset='cnn', growmap='/home/zhuominc/workspace/Sequoia/68m_13b/growmaps/L40-CNN-68m-13b-stochastic-S8.pt', start=0, end=200, T=0.6, P=1.0, DP=0.99, D=1, B=10, W=32, M=800, Mode='greedy', decay=0.85, negative=False, static=False, offloading=False)
total time :498.62317s, latency :0.01999s, decoding step: 24938, large model step: 10888, 2.2904114621601765
[2024-02-14 21:14:38,851] [INFO] [real_accelerator.py:161:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Namespace(model='JackFram/llama-68m', target='meta-llama/Llama-2-13b-hf', dataset='cnn', growmap='/home/zhuominc/workspace/Sequoia/68m_13b/growmaps/L40-CNN-68m-13b-stochastic-S16.pt', start=0, end=200, T=0.6, P=1.0, DP=0.99, D=1, B=10, W=32, M=800, Mode='greedy', decay=0.85, negative=False, static=False, offloading=False)
total time :455.64840s, latency :0.01809s, decoding step: 25182, large model step: 9621, 2.617399438727783
[2024-02-14 21:22:49,403] [INFO] [real_accelerator.py:161:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Namespace(model='JackFram/llama-68m', target='meta-llama/Llama-2-13b-hf', dataset='cnn', growmap='/home/zhuominc/workspace/Sequoia/68m_13b/growmaps/L40-CNN-68m-13b-stochastic-S32.pt', start=0, end=200, T=0.6, P=1.0, DP=0.99, D=1, B=10, W=32, M=800, Mode='greedy', decay=0.85, negative=False, static=False, offloading=False)
total time :458.34797s, latency :0.01831s, decoding step: 25039, large model step: 8685, 2.8830166954519285
[2024-02-14 21:30:59,538] [INFO] [real_accelerator.py:161:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Namespace(model='JackFram/llama-68m', target='meta-llama/Llama-2-13b-hf', dataset='cnn', growmap='/home/zhuominc/workspace/Sequoia/68m_13b/growmaps/L40-CNN-68m-13b-stochastic-S64.pt', start=0, end=200, T=0.6, P=1.0, DP=0.99, D=1, B=10, W=32, M=800, Mode='greedy', decay=0.85, negative=False, static=False, offloading=False)
total time :439.14907s, latency :0.01745s, decoding step: 25163, large model step: 7907, 3.182370051852789
[2024-02-14 21:38:54,533] [INFO] [real_accelerator.py:161:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Namespace(model='JackFram/llama-68m', target='meta-llama/Llama-2-13b-hf', dataset='cnn', growmap='/home/zhuominc/workspace/Sequoia/68m_13b/growmaps/L40-CNN-68m-13b-stochastic-S128.pt', start=0, end=200, T=0.6, P=1.0, DP=0.99, D=1, B=10, W=32, M=800, Mode='greedy', decay=0.85, negative=False, static=False, offloading=False)
total time :449.04854s, latency :0.01801s, decoding step: 24933, large model step: 7207, 3.4595532121548493
[2024-02-14 21:47:08,678] [INFO] [real_accelerator.py:161:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Namespace(model='JackFram/llama-68m', target='meta-llama/Llama-2-13b-hf', dataset='cnn', growmap='/home/zhuominc/workspace/Sequoia/68m_13b/growmaps/L40-CNN-68m-13b-stochastic-S256.pt', start=0, end=200, T=0.6, P=1.0, DP=0.99, D=1, B=10, W=32, M=800, Mode='greedy', decay=0.85, negative=False, static=False, offloading=False)
total time :561.24852s, latency :0.02205s, decoding step: 25450, large model step: 6829, 3.726753551032362
[2024-02-14 21:57:32,258] [INFO] [real_accelerator.py:161:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Namespace(model='JackFram/llama-68m', target='meta-llama/Llama-2-13b-hf', dataset='cnn', growmap='/home/zhuominc/workspace/Sequoia/68m_13b/growmaps/L40-CNN-68m-13b-stochastic-S512.pt', start=0, end=200, T=0.6, P=1.0, DP=0.99, D=1, B=10, W=32, M=800, Mode='greedy', decay=0.85, negative=False, static=False, offloading=False)
total time :930.74772s, latency :0.03685s, decoding step: 25255, large model step: 6324, 3.993516761543327