brucewayne0459
commited on
Update README.md
Browse files
README.md
CHANGED
@@ -1,187 +1,101 @@
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
---
|
4 |
-
|
5 |
-
# Model Card for Model ID
|
6 |
-
|
7 |
-
<!-- Provide a quick summary of what the model is/does. -->
|
8 |
-
|
9 |
-
This modelcard aims to be a base template for new models. It has been generated using [this raw template](https://github.com/huggingface/huggingface_hub/blob/main/src/huggingface_hub/templates/modelcard_template.md?plain=1).
|
10 |
|
11 |
## Model Details
|
12 |
|
13 |
### Model Description
|
14 |
|
15 |
-
|
16 |
-
|
17 |
|
18 |
|
19 |
- **Developed by:** Bruce_Wayne
|
20 |
- **Funded by [optional]:** Jhonny and koti
|
21 |
- **Model type:** vision model
|
22 |
- **Finetuned from model [optional]:** https://huggingface.co/google/paligemma-3b-pt-224
|
23 |
-
|
|
|
|
|
24 |
## Uses
|
25 |
-
|
26 |
-
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
27 |
-
|
28 |
### Direct Use
|
29 |
|
30 |
-
|
31 |
|
32 |
-
[More Information Needed]
|
33 |
-
|
34 |
-
### Downstream Use [optional]
|
35 |
-
|
36 |
-
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
37 |
-
|
38 |
-
[More Information Needed]
|
39 |
-
|
40 |
-
### Out-of-Scope Use
|
41 |
-
|
42 |
-
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
43 |
-
|
44 |
-
[More Information Needed]
|
45 |
|
46 |
## Bias, Risks, and Limitations
|
47 |
|
48 |
-
|
49 |
-
|
50 |
-
[More Information Needed]
|
51 |
-
|
52 |
-
### Recommendations
|
53 |
-
|
54 |
-
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
55 |
-
|
56 |
-
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
57 |
|
58 |
## How to Get Started with the Model
|
59 |
|
60 |
-
|
61 |
|
62 |
-
|
63 |
|
|
|
|
|
|
|
|
|
64 |
## Training Details
|
65 |
|
66 |
### Training Data
|
67 |
|
68 |
-
|
69 |
-
|
70 |
-
[More Information Needed]
|
71 |
|
72 |
### Training Procedure
|
73 |
|
74 |
-
|
75 |
-
|
76 |
-
#### Preprocessing [optional]
|
77 |
-
|
78 |
-
[More Information Needed]
|
79 |
-
|
80 |
|
81 |
#### Training Hyperparameters
|
82 |
|
83 |
-
- **Training regime:**
|
|
|
|
|
|
|
|
|
84 |
|
85 |
-
#### Speeds, Sizes, Times [optional]
|
86 |
-
|
87 |
-
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
88 |
-
|
89 |
-
[More Information Needed]
|
90 |
|
91 |
## Evaluation
|
92 |
|
93 |
-
<!-- This section describes the evaluation protocols and provides the results. -->
|
94 |
-
|
95 |
### Testing Data, Factors & Metrics
|
96 |
|
97 |
#### Testing Data
|
98 |
|
99 |
-
|
100 |
-
|
101 |
-
[More Information Needed]
|
102 |
-
|
103 |
-
#### Factors
|
104 |
-
|
105 |
-
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
106 |
-
|
107 |
-
[More Information Needed]
|
108 |
|
109 |
#### Metrics
|
110 |
-
|
111 |
-
|
112 |
-
|
113 |
-
[More Information Needed]
|
114 |
-
|
115 |
### Results
|
116 |
|
117 |
-
|
118 |
|
119 |
#### Summary
|
120 |
|
121 |
|
122 |
-
|
123 |
-
## Model Examination [optional]
|
124 |
-
|
125 |
-
<!-- Relevant interpretability work for the model goes here -->
|
126 |
-
|
127 |
-
[More Information Needed]
|
128 |
-
|
129 |
## Environmental Impact
|
130 |
|
131 |
-
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
132 |
-
|
133 |
-
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
134 |
|
135 |
-
- **Hardware Type:**
|
136 |
-
- **Hours used:**
|
137 |
-
- **Cloud Provider:**
|
138 |
-
- **Compute Region:**
|
139 |
-
- **Carbon Emitted:**
|
140 |
|
141 |
-
## Technical Specifications
|
142 |
|
143 |
### Model Architecture and Objective
|
144 |
|
145 |
-
|
|
|
146 |
|
147 |
### Compute Infrastructure
|
148 |
|
149 |
-
[More Information Needed]
|
150 |
-
|
151 |
#### Hardware
|
152 |
|
153 |
-
|
154 |
-
|
155 |
-
|
156 |
-
|
157 |
-
[More Information Needed]
|
158 |
-
|
159 |
-
## Citation [optional]
|
160 |
-
|
161 |
-
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
162 |
-
|
163 |
-
**BibTeX:**
|
164 |
-
|
165 |
-
[More Information Needed]
|
166 |
-
|
167 |
-
**APA:**
|
168 |
-
|
169 |
-
[More Information Needed]
|
170 |
-
|
171 |
-
## Glossary [optional]
|
172 |
-
|
173 |
-
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
174 |
-
|
175 |
-
[More Information Needed]
|
176 |
-
|
177 |
-
## More Information [optional]
|
178 |
-
|
179 |
-
[More Information Needed]
|
180 |
-
|
181 |
-
## Model Card Authors [optional]
|
182 |
-
|
183 |
-
[More Information Needed]
|
184 |
-
|
185 |
-
## Model Card Contact
|
186 |
-
|
187 |
-
[More Information Needed]
|
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
---
|
4 |
+
# Model Card for PaliGemma Dermatology Model
|
|
|
|
|
|
|
|
|
|
|
5 |
|
6 |
## Model Details
|
7 |
|
8 |
### Model Description
|
9 |
|
10 |
+
This model, based on the PaliGemma-3B architecture, has been fine-tuned for dermatology-related image and text processing tasks. The model is designed to assist in the identification of various skin conditions using a combination of image analysis and natural language processing.
|
|
|
11 |
|
12 |
|
13 |
- **Developed by:** Bruce_Wayne
|
14 |
- **Funded by [optional]:** Jhonny and koti
|
15 |
- **Model type:** vision model
|
16 |
- **Finetuned from model [optional]:** https://huggingface.co/google/paligemma-3b-pt-224
|
17 |
+
- **LoRa Adaptors used:** Yes
|
18 |
+
- **Intended use:** Medical image analysis, specifically for dermatology
|
19 |
+
**
|
20 |
## Uses
|
|
|
|
|
|
|
21 |
### Direct Use
|
22 |
|
23 |
+
The model can be directly used for analyzing dermatology images, providing insights into potential skin conditions.
|
24 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
25 |
|
26 |
## Bias, Risks, and Limitations
|
27 |
|
28 |
+
**Skin Tone Bias:** The model may have been trained on a dataset that does not adequately represent all skin tones, potentially leading to biased results.
|
29 |
+
**Geographic Bias:** The model's performance may vary depending on the prevalence of certain conditions in different geographic regions.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
30 |
|
31 |
## How to Get Started with the Model
|
32 |
|
33 |
+
** python
|
34 |
|
35 |
+
from transformers import AutoProcessor, PaliGemmaForConditionalGeneration
|
36 |
|
37 |
+
model_id = "brucewayne0459/paligemma_derm"
|
38 |
+
processor = AutoProcessor.from_pretrained(model_id)
|
39 |
+
model = PaliGemmaForConditionalGeneration.from_pretrained(model_id)
|
40 |
+
**
|
41 |
## Training Details
|
42 |
|
43 |
### Training Data
|
44 |
|
45 |
+
The model was fine-tuned on a dataset of dermatological images combined with disease names
|
|
|
|
|
46 |
|
47 |
### Training Procedure
|
48 |
|
49 |
+
The model was fine-tuned using LoRA (Low-Rank Adaptation) for more efficient training. Mixed precision (bfloat16) was used to speed up training and reduce memory usage.
|
|
|
|
|
|
|
|
|
|
|
50 |
|
51 |
#### Training Hyperparameters
|
52 |
|
53 |
+
- **Training regime:** Mixed precision (bfloat16)
|
54 |
+
- **Epochs:** 10
|
55 |
+
- **Learning rate:** 2e-5
|
56 |
+
- **Batch size:** 6
|
57 |
+
- **Gradient accumulation steps:** 4
|
58 |
|
|
|
|
|
|
|
|
|
|
|
59 |
|
60 |
## Evaluation
|
61 |
|
|
|
|
|
62 |
### Testing Data, Factors & Metrics
|
63 |
|
64 |
#### Testing Data
|
65 |
|
66 |
+
The model was evaluated on a separate validation set of dermatological images and Disease Names, distinct from the training data.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
67 |
|
68 |
#### Metrics
|
69 |
+
- **Validation Loss:** The loss was tracked throughout the training process to evaluate model performance.
|
70 |
+
- **Accuracy:** The primary metric for assessing model predictions.
|
|
|
|
|
|
|
71 |
### Results
|
72 |
|
73 |
+
The model achieved a final validation loss of approximately 0.2214, indicating reasonable performance in predicting skin conditions based on the dataset used.
|
74 |
|
75 |
#### Summary
|
76 |
|
77 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
78 |
## Environmental Impact
|
79 |
|
|
|
|
|
|
|
80 |
|
81 |
+
- **Hardware Type:** 1 x L4 GPU
|
82 |
+
- **Hours used:** ~22 HOURS
|
83 |
+
- **Cloud Provider:** LIGHTNING AI
|
84 |
+
- **Compute Region:** USA
|
85 |
+
- **Carbon Emitted:** 0.9 kg eq. CO2
|
86 |
|
87 |
+
## Technical Specifications
|
88 |
|
89 |
### Model Architecture and Objective
|
90 |
|
91 |
+
- **Architecture:** Vision-Language model based on PaliGemma-3B
|
92 |
+
- **Objective:** To classify and diagnose dermatological conditions from images and text
|
93 |
|
94 |
### Compute Infrastructure
|
95 |
|
|
|
|
|
96 |
#### Hardware
|
97 |
|
98 |
+
- **GPU:** 1xL4 GPU
|
99 |
+
-
|
100 |
+
## Model Card Authors
|
101 |
+
Bruce_Wayne
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|