-
Notifications
You must be signed in to change notification settings - Fork 1
/
valohai.yaml
100 lines (95 loc) · 3.76 KB
/
valohai.yaml
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
---
- step:
name: build-cpu-gpu-uberjar
image: neomatrix369/dl4j-nlp-cuda:v0.5
inputs:
- name: src-main-resources
default: https://github.com/neomatrix369/dl4j-nlp-cuda-example/releases/download/dl4j-nlp-src-main-resources-v0.1/dl4j-nlp-src-main-resources.tgz
description: NLP data for training, prediction, evaluation
command:
- ./buildUberJar.sh
environment: aws-eu-west-1-g2-2xlarge
- step:
name: train-cpu-linux
image: neomatrix369/dl4j-nlp-cuda:v0.5
inputs:
- name: cpu-linux-uberjar
default: datum://
description: dl4j nlp cpu linux uberjar
- name: imdb-reviews
default: http://ai.stanford.edu/~amaas/data/sentiment/aclImdb_v1.tar.gz
description: IMDB Review dataset for sentiment analysis
- name: google-word2vec
default: https://s3.amazonaws.com/dl4j-distribution/GoogleNews-vectors-negative300.bin.gz
description: word2vec pre-trained Google News corpus
command:
- cd ${VH_REPOSITORY_DIR}
- export BACKEND=cpu
- export ACTION=train
- time ./runUberJar.sh --action ${ACTION} --output-model-dir .
environment: aws-eu-west-1-g3-4xlarge
- step:
name: train-gpu-linux
image: neomatrix369/dl4j-nlp-cuda:v0.5
inputs:
- name: gpu-linux-uberjar
default: datum://
description: dl4j nlp gpu linux uberjar
- name: imdb-reviews
default: http://ai.stanford.edu/~amaas/data/sentiment/aclImdb_v1.tar.gz
description: IMDB Review dataset for sentiment analysis
- name: google-word2vec
default: https://s3.amazonaws.com/dl4j-distribution/GoogleNews-vectors-negative300.bin.gz
description: word2vec pre-trained Google News corpus
command:
- cd ${VH_REPOSITORY_DIR}
- export BACKEND=gpu
- export ACTION=train
- time ./runUberJar.sh --action ${ACTION} --output-model-dir .
environment: aws-eu-west-1-g3-4xlarge
- step:
name: evaluate-model-linux
image: neomatrix369/dl4j-nlp-cuda:v0.5
inputs:
- name: linux-uberjar
default: datum://
description: dl4j nlp linux uberjar
- name: model
default: datum://
description: nlp model trained on Google news corpus
- name: imdb-reviews
default: http://ai.stanford.edu/~amaas/data/sentiment/aclImdb_v1.tar.gz
description: IMDB Review dataset for sentiment analysis
- name: google-word2vec
default: https://s3.amazonaws.com/dl4j-distribution/GoogleNews-vectors-negative300.bin.gz
description: word2vec pre-trained Google News corpus
command:
- cd ${VH_REPOSITORY_DIR}
- export ACTION=evaluate
- echo "~~~ Copying jar and model into ${VH_REPOSITORY_DIR}"
- cp ${VH_INPUTS_DIR}/linux-uberjar/*.jar ${VH_REPOSITORY_DIR}
- cp ${VH_INPUTS_DIR}/model/*.pb .
- time ./runUberJar.sh --action ${ACTION} --input-model-file $(ls *.pb)
environment: aws-eu-west-1-g3-4xlarge
###
### aws-eu-west-1-g2-2xlarge
### GPU Specification | 1x GPU - NVIDIA GRID K520 (4GB)
### Description | 8 cores, 15GB memory, 500GB storage
### Per-Hour USD$ | 0.90700
### aws-eu-west-1-g3-4xlarge
### GPU Specification | 1x GPU - NVIDIA Tesla M60 (8GB)
### Description | 16 cores, 122GB memory, 500GB storage
### Per-Hour USD$ | 1.21000
### aws-eu-west-1-g2-8xlarge
### GPU Specification | 4x GPU - NVIDIA GRID K520 (16GB)
### Description | 32 cores, 60GB memory, 500GB storage
### Per-Hour USD$ | 2.80800
###
- step:
name: know-your-gpus
image: neomatrix369/dl4j-nlp-cuda:v0.5
command:
- cd ${VH_REPOSITORY_DIR}
- ./know-your-gpus.sh &> "${VH_OUTPUTS_DIR}/know-your-gpus.logs"
- cat "${VH_OUTPUTS_DIR}/know-your-gpus.logs"
environment: aws-eu-west-1-g3-4xlarge