Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
|
@@ -42,8 +42,10 @@ demo = gr.Blocks()
|
|
| 42 |
with demo:
|
| 43 |
gr.Markdown("# CO2 Inference Demo")
|
| 44 |
gr.Markdown("## TL;DR - We ran a series of experiments to measure the energy efficiency and carbon emissions of different\
|
| 45 |
-
models from the HuggingFace Hub, and to see how different tasks and models compare
|
| 46 |
-
|
|
|
|
|
|
|
| 47 |
with gr.Accordion("More details about our methodology:", open=False):
|
| 48 |
gr.Markdown("We chose ten ML tasks: text classification, token classification, question answering, \
|
| 49 |
), masked language modeling, text generation, summarization, image classification, object detection, \
|
|
|
|
| 42 |
with demo:
|
| 43 |
gr.Markdown("# CO2 Inference Demo")
|
| 44 |
gr.Markdown("## TL;DR - We ran a series of experiments to measure the energy efficiency and carbon emissions of different\
|
| 45 |
+
models from the HuggingFace Hub, and to see how different tasks and models compare. \
|
| 46 |
+
We found that multi-purpose, generative models are orders of magnitude more energy-intensive than task-specific systems\
|
| 47 |
+
for a variety of tasks, even for models with a similar number of parameters")
|
| 48 |
+
gr.Markdown("### Explore the plots below to get more insights about the different models and tasks from our study.")
|
| 49 |
with gr.Accordion("More details about our methodology:", open=False):
|
| 50 |
gr.Markdown("We chose ten ML tasks: text classification, token classification, question answering, \
|
| 51 |
), masked language modeling, text generation, summarization, image classification, object detection, \
|