patcdaniel
commited on
Commit
•
35cfd3b
1
Parent(s):
4f9d5a1
Added model card template.md
Browse files
README.md
CHANGED
@@ -1,3 +1,128 @@
|
|
1 |
---
|
2 |
license: cc-by-sa-4.0
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: cc-by-sa-4.0
|
3 |
+
language:
|
4 |
+
- en
|
5 |
---
|
6 |
+
# phytoClassUCSC
|
7 |
+
|
8 |
+
Sections and prompts from the [model cards paper](https://arxiv.org/abs/1810.03993), v2.
|
9 |
+
|
10 |
+
Jump to section:
|
11 |
+
|
12 |
+
- [Model details](#model-details)
|
13 |
+
- [Intended use](#intended-use)
|
14 |
+
- [Factors](#factors)
|
15 |
+
- [Metrics](#metrics)
|
16 |
+
- [Evaluation data](#evaluation-data)
|
17 |
+
- [Training data](#training-data)
|
18 |
+
- [Quantitative analyses](#quantitative-analyses)
|
19 |
+
- [Ethical considerations](#ethical-considerations)
|
20 |
+
- [Caveats and recommendations](#caveats-and-recommendations)
|
21 |
+
|
22 |
+
## Model details
|
23 |
+
|
24 |
+
Review section 4.1 of the [model cards paper](https://arxiv.org/abs/1810.03993).
|
25 |
+
|
26 |
+
- Developed by the Kudela Lab from the Ocean Sciences Department at University of California, Santa Cruz.
|
27 |
+
- Current version trained in February, 2023.
|
28 |
+
- phytoClassUCSC-SoftNone02162023
|
29 |
+
- phytoClassUCSC is a depthwise- CNN based on the Xception architecture [Chollet, F., 2017](https://arxiv.org/abs/1610.02357) with 134 layers using weights pretrained on ImageNet.
|
30 |
+
- An average pooling layer is used.
|
31 |
+
- Paper or other resource for more information
|
32 |
+
- Citation details
|
33 |
+
- License
|
34 |
+
- Email Patrick Daniel ([pcdaniel@ucsc.edu](pcdaniel@ucsc.edu)) for questions
|
35 |
+
|
36 |
+
## Intended use
|
37 |
+
|
38 |
+
_Use cases that were envisioned during development._
|
39 |
+
|
40 |
+
Review section 4.2 of the [model cards paper](https://arxiv.org/abs/1810.03993).
|
41 |
+
|
42 |
+
### Primary intended uses
|
43 |
+
|
44 |
+
### Primary intended users
|
45 |
+
|
46 |
+
### Out-of-scope use cases
|
47 |
+
|
48 |
+
## Factors
|
49 |
+
|
50 |
+
_Factors could include demographic or phenotypic groups, environmental conditions, technical
|
51 |
+
attributes, or others listed in Section 4.3._
|
52 |
+
|
53 |
+
Review section 4.3 of the [model cards paper](https://arxiv.org/abs/1810.03993).
|
54 |
+
|
55 |
+
### Relevant factors
|
56 |
+
|
57 |
+
### Evaluation factors
|
58 |
+
|
59 |
+
## Metrics
|
60 |
+
|
61 |
+
_The appropriate metrics to feature in a model card depend on the type of model that is being tested.
|
62 |
+
For example, classification systems in which the primary output is a class label differ significantly
|
63 |
+
from systems whose primary output is a score. In all cases, the reported metrics should be determined
|
64 |
+
based on the model’s structure and intended use._
|
65 |
+
|
66 |
+
Review section 4.4 of the [model cards paper](https://arxiv.org/abs/1810.03993).
|
67 |
+
|
68 |
+
### Model performance measures
|
69 |
+
|
70 |
+
### Decision thresholds
|
71 |
+
|
72 |
+
### Approaches to uncertainty and variability
|
73 |
+
|
74 |
+
## Evaluation data
|
75 |
+
|
76 |
+
_All referenced datasets would ideally point to any set of documents that provide visibility into the
|
77 |
+
source and composition of the dataset. Evaluation datasets should include datasets that are publicly
|
78 |
+
available for third-party use. These could be existing datasets or new ones provided alongside the model
|
79 |
+
card analyses to enable further benchmarking._
|
80 |
+
|
81 |
+
Review section 4.5 of the [model cards paper](https://arxiv.org/abs/1810.03993).
|
82 |
+
|
83 |
+
### Datasets
|
84 |
+
|
85 |
+
### Motivation
|
86 |
+
|
87 |
+
### Preprocessing
|
88 |
+
|
89 |
+
## Training data
|
90 |
+
|
91 |
+
Review section 4.6 of the [model cards paper](https://arxiv.org/abs/1810.03993).
|
92 |
+
|
93 |
+
## Quantitative analyses
|
94 |
+
|
95 |
+
_Quantitative analyses should be disaggregated, that is, broken down by the chosen factors. Quantitative
|
96 |
+
analyses should provide the results of evaluating the model according to the chosen metrics, providing
|
97 |
+
confidence interval values when possible._
|
98 |
+
|
99 |
+
Review section 4.7 of the [model cards paper](https://arxiv.org/abs/1810.03993).
|
100 |
+
|
101 |
+
### Unitary results
|
102 |
+
|
103 |
+
### Intersectional result
|
104 |
+
|
105 |
+
## Ethical considerations
|
106 |
+
|
107 |
+
_This section is intended to demonstrate the ethical considerations that went into model development,
|
108 |
+
surfacing ethical challenges and solutions to stakeholders. Ethical analysis does not always lead to
|
109 |
+
precise solutions, but the process of ethical contemplation is worthwhile to inform on responsible
|
110 |
+
practices and next steps in future work._
|
111 |
+
|
112 |
+
Review section 4.8 of the [model cards paper](https://arxiv.org/abs/1810.03993).
|
113 |
+
|
114 |
+
### Data
|
115 |
+
|
116 |
+
### Human life
|
117 |
+
|
118 |
+
### Mitigations
|
119 |
+
|
120 |
+
### Risks and harms
|
121 |
+
|
122 |
+
### Use cases
|
123 |
+
|
124 |
+
## Caveats and recommendations
|
125 |
+
|
126 |
+
_This section should list additional concerns that were not covered in the previous sections._
|
127 |
+
|
128 |
+
Review section 4.9 of the [model cards paper](https://arxiv.org/abs/1810.03993).
|