I’ve been an avid AI/ML engineer and Fast.ai user for 7+ years and have always worked towards the easiest way to develop, train, and share AI implementations, so I wanted to share the following with the community:
American Data Science has just launched its Labs service which provides you with Jupyter lab workspaces. Features on sign-up include:
- Classic and familiar personal jupyter lab workspaces
- Large and growing selection of pre-built environments including Fast.ai (GPU supported)
- Persistent storage for each workspace
- Free compute with GPU availability
- User service sharing port for each workspace
Unlock a free month of Pro subscription by applying the code FASTAI23 on checkout after clicking on the Subscribe button under the Pro plan on this link.
The Pro subscription gives you access to an additional range of environments and 16GBs of GPU VRAM for each workspace. You can play with OpenAI, Pinecone, Fast.ai, and other pre-built environments ready for your experimentations. Feel free to discuss any requests or feedback in the community forum.
Share your applications securely with microservices you can develop and run in the same server workspaces you’re working in. Just run whatever service you’d like on port 5000 in your Jupyter workspace, and then immediately view it in the URL produced from running
Example service you can run in your Fast.ai workspace
import numpy as np import gradio as gr def sepia(input_img): sepia_filter = np.array([ [0.393, 0.769, 0.189], [0.349, 0.686, 0.168], [0.272, 0.534, 0.131] ]) sepia_img = input_img.dot(sepia_filter.T) sepia_img /= sepia_img.max() return sepia_img demo = gr.Interface(sepia, gr.Image(shape=(200, 200)), "image") demo.launch(server_port=5000) ## Don't forget to free the port when done # import gradio # gradio.close_all()
Additional features and information will be coming out soon. Stay updated by following AmDS on social media. Resources for GPUs may be competitive, especially during the growth phase for this service. We’ll try our best to consistently keep GPU availability up, but please be patient with us as we work on upscaling and resource limits. Also, keep an eye out for additional services that will enhance the AI engineer’s experience.
Good luck on your experiments and AI modeling!