Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,178 @@
|
|
1 |
---
|
2 |
license: wtfpl
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: wtfpl
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
pipeline_tag: text-generation
|
6 |
+
tags:
|
7 |
+
- llama
|
8 |
+
- w++
|
9 |
+
- meme++
|
10 |
+
- tiny
|
11 |
---
|
12 |
+
|
13 |
+
# Model Card for Model ID
|
14 |
+
|
15 |
+
Meme++ generator.
|
16 |
+
|
17 |
+
## Model Details
|
18 |
+
|
19 |
+
### Model Description
|
20 |
+
|
21 |
+
This is a tiny LLaMA model trained from scratch for 31000 steps (253952000 tokens) out of `i forgor :skull:`.
|
22 |
+
|
23 |
+
- **Developed by:** mrsteyk
|
24 |
+
- **Model type:** LLaMA
|
25 |
+
- **Language(s) (NLP):** English
|
26 |
+
- **License:** WTFPL
|
27 |
+
|
28 |
+
### Model Sources [optional]
|
29 |
+
|
30 |
+
- **Repository:** maybe someday
|
31 |
+
|
32 |
+
## Uses
|
33 |
+
|
34 |
+
This was intended for Meme++ character chard generation, trained a small demo.
|
35 |
+
|
36 |
+
### Direct Use
|
37 |
+
|
38 |
+
Random Meme++ card generation.
|
39 |
+
|
40 |
+
### Out-of-Scope Use
|
41 |
+
|
42 |
+
CSAM related stuff.
|
43 |
+
|
44 |
+
## Bias, Risks, and Limitations
|
45 |
+
|
46 |
+
This model was trained on a randomly scraped DataSet, I tried filtering
|
47 |
+
|
48 |
+
### Recommendations
|
49 |
+
|
50 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
51 |
+
|
52 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
53 |
+
|
54 |
+
## How to Get Started with the Model
|
55 |
+
|
56 |
+
Use the code below to get started with the model.
|
57 |
+
|
58 |
+
[More Information Needed]
|
59 |
+
|
60 |
+
## Training Details
|
61 |
+
|
62 |
+
### Training Data
|
63 |
+
|
64 |
+
Meme++ character definition taken off the internet.
|
65 |
+
|
66 |
+
### Training Procedure
|
67 |
+
|
68 |
+
This was trained using `lit-llama` based model code and `pytorch-lightning` CLI based trainer code.
|
69 |
+
|
70 |
+
|
71 |
+
#### Training Hyperparameters
|
72 |
+
|
73 |
+
- **Training regime:** fp32
|
74 |
+
- **Optimizer and LR:** DeepSpeed FusedAdamW with 1e-5
|
75 |
+
|
76 |
+
#### Speeds, Sizes, Times [optional]
|
77 |
+
|
78 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
79 |
+
|
80 |
+
[More Information Needed]
|
81 |
+
|
82 |
+
## Evaluation
|
83 |
+
|
84 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
85 |
+
|
86 |
+
### Testing Data, Factors & Metrics
|
87 |
+
|
88 |
+
#### Testing Data
|
89 |
+
|
90 |
+
<!-- This should link to a Data Card if possible. -->
|
91 |
+
|
92 |
+
[More Information Needed]
|
93 |
+
|
94 |
+
#### Factors
|
95 |
+
|
96 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
97 |
+
|
98 |
+
[More Information Needed]
|
99 |
+
|
100 |
+
#### Metrics
|
101 |
+
|
102 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
103 |
+
|
104 |
+
[More Information Needed]
|
105 |
+
|
106 |
+
### Results
|
107 |
+
|
108 |
+
[More Information Needed]
|
109 |
+
|
110 |
+
#### Summary
|
111 |
+
|
112 |
+
[W&B run](https://wandb.ai/mrsteyk/memepp-llama/runs/44e3aut4)
|
113 |
+
|
114 |
+
## Model Examination [optional]
|
115 |
+
|
116 |
+
<!-- Relevant interpretability work for the model goes here -->
|
117 |
+
|
118 |
+
[More Information Needed]
|
119 |
+
|
120 |
+
## Environmental Impact
|
121 |
+
|
122 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
123 |
+
|
124 |
+
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).
|
125 |
+
|
126 |
+
- **Hardware Type:** 1050 Ti Mobile
|
127 |
+
- **Hours used:** ~6
|
128 |
+
- **Cloud Provider:** Local Machine(C)(TM)
|
129 |
+
- **Compute Region:** RU
|
130 |
+
- **Carbon Emitted:** ~~450kg~~
|
131 |
+
|
132 |
+
## Technical Specifications [optional]
|
133 |
+
|
134 |
+
### Model Architecture and Objective
|
135 |
+
|
136 |
+
[More Information Needed]
|
137 |
+
|
138 |
+
### Compute Infrastructure
|
139 |
+
|
140 |
+
[More Information Needed]
|
141 |
+
|
142 |
+
#### Hardware
|
143 |
+
|
144 |
+
[More Information Needed]
|
145 |
+
|
146 |
+
#### Software
|
147 |
+
|
148 |
+
[More Information Needed]
|
149 |
+
|
150 |
+
## Citation [optional]
|
151 |
+
|
152 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
153 |
+
|
154 |
+
**BibTeX:**
|
155 |
+
|
156 |
+
[More Information Needed]
|
157 |
+
|
158 |
+
**APA:**
|
159 |
+
|
160 |
+
[More Information Needed]
|
161 |
+
|
162 |
+
## Glossary [optional]
|
163 |
+
|
164 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
165 |
+
|
166 |
+
[More Information Needed]
|
167 |
+
|
168 |
+
## More Information [optional]
|
169 |
+
|
170 |
+
[More Information Needed]
|
171 |
+
|
172 |
+
## Model Card Authors [optional]
|
173 |
+
|
174 |
+
[More Information Needed]
|
175 |
+
|
176 |
+
## Model Card Contact
|
177 |
+
|
178 |
+
[More Information Needed]
|