# Week 15 Discussion Thread **URL:** \ **Category:** Fastbook Reading Group\ **Created:** [September 22, 2021, 1:42pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740 "2021-09-22T13:42:15Z")\ **Posts on this page:** 20\ **Page:** 1
**Author:** ![lessing](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/lessing/32/200_2.png) [@lessing](https://community.wandb.ai/u/lessing)\ **Post date:** [September 22, 2021, 1:42pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/1 "2021-09-22T13:42:16Z")
[![](https://us1.discourse-cdn.com/flex020/uploads/wandb/original/2X/8/854a54ffa069d30b07e5bcad14a5cb972942ece1.jpeg "W&B Fastbook Reading Group — 15. Application Architectures Deep Dive") ](https://www.youtube.com/watch?v=NXiM8X0O7E4) ---
**Author:** ![lessing](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/lessing/32/200_2.png) [@lessing](https://community.wandb.ai/u/lessing)\ **Post date:** [September 22, 2021, 1:55pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/2 "2021-09-22T13:55:36Z")
**Resources:** **Blog posts from 2 weeks ago** - @vinayak_nayak’s blog post “[Understanding activations with color dimension](https://elisonsherton.github.io//fastbook/deep%20learning/2021/09/08/understanding-activations.html)” and “[Building a resnet34 architecture from scratch](https://elisonsherton.github.io//fastbook/deep%20learning/2021/09/22/resnets.html#building-a-resnet34-architecture-from-scratch)” - @ravimashru’s blog posts on “[Understanding Super-Convergence](https://ravimashru.dev/blog/2021-09-09-superconvergence/)” and “[Understanding FixRes - Fixing the train-test resolution discrepancy](https://ravimashru.dev/blog/2021-09-22-fix-res/)” ---
**Author:** ![bhutanisanyam1](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/bhutanisanyam1/32/18_2.png) [@bhutanisanyam1](https://community.wandb.ai/u/bhutanisanyam1)\ **Post date:** [September 22, 2021, 5:02pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/3 "2021-09-22T17:02:37Z")
The group is ending? 😱 😱 😢 ---
**Author:** ![amanarora](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/amanarora/32/59_2.png) [@amanarora](https://community.wandb.ai/u/amanarora)\ **Post date:** [September 22, 2021, 5:05pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/4 "2021-09-22T17:05:07Z")
This is the blog post on Optimizers - [“Adam” and friends | Committed towards better future](https://amaarora.github.io/2021/03/13/optimizers.html) ---
**Author:** ![bhutanisanyam1](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/bhutanisanyam1/32/18_2.png) [@bhutanisanyam1](https://community.wandb.ai/u/bhutanisanyam1)\ **Post date:** [September 22, 2021, 5:05pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/5 "2021-09-22T17:05:29Z")
Edit: Aman mentioned it too. I spoke too soon 🙂 We also have a PyTorch study group too where the PyTorch topics will be covered soon for anyone interested 🍵 > **[PyTorch Book Reading Group](https://community.wandb.ai/c/community-events/pytorch-book/32)** > > First Special Session with Thomas Viehmann Announced here ---
**Author:** ![bhutanisanyam1](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/bhutanisanyam1/32/18_2.png) [@bhutanisanyam1](https://community.wandb.ai/u/bhutanisanyam1)\ **Post date:** [September 22, 2021, 5:09pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/6 "2021-09-22T17:09:22Z")
Hopefully others will have a better idea, but my “big project” from that break between fastai versions, which took over my life was Chai Time Data Science 😃 ---
**Author:** ![ramesh](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/ramesh/32/98_2.png) [@ramesh](https://community.wandb.ai/u/ramesh)\ **Post date:** [September 22, 2021, 5:16pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/7 "2021-09-22T17:16:46Z")
Thanks Aman for doing these sessions. You have given us enough momentum to continue the learnings. Looking forward to stay connected in the Paper Reading Group. ---
**Author:** ![skj092](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/skj092/32/228_2.png) [@skj092](https://community.wandb.ai/u/skj092)\ **Post date:** [September 22, 2021, 5:21pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/8 "2021-09-22T17:21:46Z")
How to get source code in vscode? ---
**Author:** ![ramesh](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/ramesh/32/98_2.png) [@ramesh](https://community.wandb.ai/u/ramesh)\ **Post date:** [September 22, 2021, 5:22pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/9 "2021-09-22T17:22:12Z")
I have a question on Head. For Classification, we could use certain architectures (VGG, ResNet etc). But for Segmentation or Object Detection, we usually hv different base architectures like UNet / RetinaNet etc. Most of the work is really in the Data Prep. Can you speak a bit about Data Prep step in Object Detection. In it, we have location of objects, but the architecture might use Anchors or other techniques. So we no longer have the 1:1 we used to have in Classification or even Segmentation. How do you go about building Dataset for Object Detection tasks where each architecture might have different losses / outputs (unlike Classification or Segmentation tasks). ---
**Author:** ![bhutanisanyam1](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/bhutanisanyam1/32/18_2.png) [@bhutanisanyam1](https://community.wandb.ai/u/bhutanisanyam1)\ **Post date:** [September 22, 2021, 5:22pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/10 "2021-09-22T17:22:55Z")
One neat trick you can try-\> Head over to fastai repo on GH. And press dot on your keyboard 🙂 (It launches a VSCode instance, not powerful enough for DL but p great for reading source code) ---
**Author:** ![darkhorizon](https://avatars.discourse-cdn.com/v4/letter/d/b5a626/32.png) [@darkhorizon](https://community.wandb.ai/u/darkhorizon)\ **Post date:** [September 22, 2021, 5:24pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/11 "2021-09-22T17:24:57Z")
If that doesn’t work change github to github1s ---
**Author:** ![bhutanisanyam1](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/bhutanisanyam1/32/18_2.png) [@bhutanisanyam1](https://community.wandb.ai/u/bhutanisanyam1)\ **Post date:** [September 22, 2021, 5:41pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/12 "2021-09-22T17:41:23Z")
We ran a hugging face study group too 🍵 Link: [W&B Study Group: fastai x Hugging Face - YouTube](https://www.youtube.com/playlist?list=PLD80i8An1OEF8UOb9N9uSoidOGIMKW96t) ---
**Author:** ![amanarora](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/amanarora/32/59_2.png) [@amanarora](https://community.wandb.ai/u/amanarora)\ **Post date:** [September 22, 2021, 5:42pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/13 "2021-09-22T17:42:57Z")
NLP course by Rachel - [new fast.ai course: A Code-First Introduction to Natural Language Processing · fast.ai](https://www.fast.ai/2019/07/08/fastai-nlp/) ---
**Author:** ![bhutanisanyam1](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/bhutanisanyam1/32/18_2.png) [@bhutanisanyam1](https://community.wandb.ai/u/bhutanisanyam1)\ **Post date:** [September 22, 2021, 5:43pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/14 "2021-09-22T17:43:35Z")
Maybe I could start a fastai (NLP) study group next 🤔 ---
**Author:** ![amanarora](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/amanarora/32/59_2.png) [@amanarora](https://community.wandb.ai/u/amanarora)\ **Post date:** [September 22, 2021, 5:44pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/15 "2021-09-22T17:44:14Z")
Once you get to transformers - here is another resource to help with your journeys! [The Annotated GPT-2 | Committed towards better future](https://amaarora.github.io/2020/02/18/annotatedGPT2.html) ---
**Author:** ![bhutanisanyam1](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/bhutanisanyam1/32/18_2.png) [@bhutanisanyam1](https://community.wandb.ai/u/bhutanisanyam1)\ **Post date:** [September 22, 2021, 5:45pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/16 "2021-09-22T17:45:09Z")
Professor Zach had created one ---
**Author:** ![amanarora](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/amanarora/32/59_2.png) [@amanarora](https://community.wandb.ai/u/amanarora)\ **Post date:** [September 22, 2021, 5:45pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/17 "2021-09-22T17:45:27Z")
For tabular - [Using AutoEncoders with Tabular Data (Intermediate) | walkwithfastai](https://walkwithfastai.com/tab.ae) ---
**Author:** ![amanarora](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/amanarora/32/59_2.png) [@amanarora](https://community.wandb.ai/u/amanarora)\ **Post date:** [September 22, 2021, 5:46pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/18 "2021-09-22T17:46:10Z")
Two more resources on CV: > **[GitHub - fastai/timmdocs: Documentation for Ross Wightman's timm image model...](https://github.com/fastai/timmdocs)** > > Documentation for Ross Wightman's timm image model library - GitHub - fastai/timmdocs: Documentation for Ross Wightman's timm image model library > **[GitHub - rwightman/pytorch-image-models: PyTorch image models, scripts,...](https://github.com/rwightman/pytorch-image-models)** > > PyTorch image models, scripts, pretrained weights -- ResNet, ResNeXT, EfficientNet, EfficientNetV2, NFNet, Vision Transformer, MixNet, MobileNet-V3/V2, RegNet, DPN, CSPNet, and more - GitHub - rwig... ---
**Author:** ![amanarora](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/amanarora/32/59_2.png) [@amanarora](https://community.wandb.ai/u/amanarora)\ **Post date:** [September 22, 2021, 5:47pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/19 "2021-09-22T17:47:41Z")
Segmentation models pytorch is another great repository - [GitHub - qubvel/segmentation\_models.pytorch: Segmentation models with pretrained backbones. PyTorch.](https://github.com/qubvel/segmentation_models.pytorch) ---
**Author:** ![bhutanisanyam1](https://sea2.discourse-cdn.com/flex020/user_avatar/community.wandb.ai/bhutanisanyam1/32/18_2.png) [@bhutanisanyam1](https://community.wandb.ai/u/bhutanisanyam1)\ **Post date:** [September 22, 2021, 5:49pm UTC](https://community.wandb.ai/t/week-15-discussion-thread/740/20 "2021-09-22T17:49:02Z")
This will always be remembered as our favourite (unofficial) fastai 2021 course. Thank you for hosting it! 🍵 [Next page](https://community.wandb.ai/t/week-15-discussion-thread/740.md?page=2)