wake up babe it's time to build pytorch from source again
December 15, 2025 at 7:23 PM
wake up babe it's time to build pytorch from source again
Here is Claude Opus's attempt at one-shot refactoring the PyTorch implementation of ProteinMPNN. Impressive work, but it could be better
github.com/delalamo/Pro...
github.com/delalamo/Pro...
GitHub - delalamo/ProteinMPNN: Code for the ProteinMPNN paper
Code for the ProteinMPNN paper. Contribute to delalamo/ProteinMPNN development by creating an account on GitHub.
github.com
December 14, 2025 at 6:30 PM
Here is Claude Opus's attempt at one-shot refactoring the PyTorch implementation of ProteinMPNN. Impressive work, but it could be better
github.com/delalamo/Pro...
github.com/delalamo/Pro...
90% of deep learning projects start with PyTorch.
But most beginners get stuck because:
> they skip the basics
> jump straight into building models
Master the PyTorch fundamentals step by step:
> with clear explanations
> hands-on code
But most beginners get stuck because:
> they skip the basics
> jump straight into building models
Master the PyTorch fundamentals step by step:
> with clear explanations
> hands-on code
December 16, 2025 at 11:40 AM
90% of deep learning projects start with PyTorch.
But most beginners get stuck because:
> they skip the basics
> jump straight into building models
Master the PyTorch fundamentals step by step:
> with clear explanations
> hands-on code
But most beginners get stuck because:
> they skip the basics
> jump straight into building models
Master the PyTorch fundamentals step by step:
> with clear explanations
> hands-on code
Can Chapel be used to write transformers? How does the resulting performance compare against C++ or PyTorch? Check out the second and final article in Thitrin Sastarasadhit’s “Transformers from Scratch in Chapel and C++” series on the Chapel blog: chapel-lang.org/blog/posts/t...
December 15, 2025 at 9:30 PM
Can Chapel be used to write transformers? How does the resulting performance compare against C++ or PyTorch? Check out the second and final article in Thitrin Sastarasadhit’s “Transformers from Scratch in Chapel and C++” series on the Chapel blog: chapel-lang.org/blog/posts/t...
Sneak peek of the GeoAI QGIS plugin! It supports Moondream AI vision language models, pytorch segmentation models, and many more. Training deep learning models in QGIS without writing a single line of code!
Stay tuned for more demos and tutorials!
GitHub: github.com/opengeos/geo...
Stay tuned for more demos and tutorials!
GitHub: github.com/opengeos/geo...
December 10, 2025 at 11:49 PM
Sneak peek of the GeoAI QGIS plugin! It supports Moondream AI vision language models, pytorch segmentation models, and many more. Training deep learning models in QGIS without writing a single line of code!
Stay tuned for more demos and tutorials!
GitHub: github.com/opengeos/geo...
Stay tuned for more demos and tutorials!
GitHub: github.com/opengeos/geo...
🧩 Academic software fragmentation is real. Your model is in Python, your collaborator needs MATLAB, the bifurcation analysis tool requires Fortran, and someone wants to run parameter sweeps on GPU via PyTorch...
PyRates solves this: define your model once, auto-generate code for any backend.
PyRates solves this: define your model once, auto-generate code for any backend.
December 16, 2025 at 12:28 PM
🧩 Academic software fragmentation is real. Your model is in Python, your collaborator needs MATLAB, the bifurcation analysis tool requires Fortran, and someone wants to run parameter sweeps on GPU via PyTorch...
PyRates solves this: define your model once, auto-generate code for any backend.
PyRates solves this: define your model once, auto-generate code for any backend.
Sometimes, when you combine the right tools, a nice repeatable pattern emerges.
Enter the "marimo-pydantic-pytorch-skypilot" trick!
youtu.be/JCVzYz8PrRQ
Enter the "marimo-pydantic-pytorch-skypilot" trick!
youtu.be/JCVzYz8PrRQ
Ah yes, the "marimo-pydantic-pytorch-skypilot" trick!
SkyPilot let's you configure jobs on a general suite of compute providers. It let's you search for spot instances across all cloud providers. And if you're k...
www.youtube.com
December 9, 2025 at 7:37 PM
Sometimes, when you combine the right tools, a nice repeatable pattern emerges.
Enter the "marimo-pydantic-pytorch-skypilot" trick!
youtu.be/JCVzYz8PrRQ
Enter the "marimo-pydantic-pytorch-skypilot" trick!
youtu.be/JCVzYz8PrRQ
🧰 Practical GeoAI Tools
A PyTorch library built for remote sensing and geospatial data, making it easy to work with satellite imagery, and ML-ready datasets. 🚀
🔗 Explore: github.com/torchgeo/tor...
#GeoTech #GIStools #RemoteSensing #SpatialAI #TorchGeo #GeospatialTech
A PyTorch library built for remote sensing and geospatial data, making it easy to work with satellite imagery, and ML-ready datasets. 🚀
🔗 Explore: github.com/torchgeo/tor...
#GeoTech #GIStools #RemoteSensing #SpatialAI #TorchGeo #GeospatialTech
GitHub - torchgeo/torchgeo: TorchGeo: datasets, samplers, transforms, and pre-trained models for geospatial data
TorchGeo: datasets, samplers, transforms, and pre-trained models for geospatial data - torchgeo/torchgeo
github.com
December 10, 2025 at 8:52 AM
🧰 Practical GeoAI Tools
A PyTorch library built for remote sensing and geospatial data, making it easy to work with satellite imagery, and ML-ready datasets. 🚀
🔗 Explore: github.com/torchgeo/tor...
#GeoTech #GIStools #RemoteSensing #SpatialAI #TorchGeo #GeospatialTech
A PyTorch library built for remote sensing and geospatial data, making it easy to work with satellite imagery, and ML-ready datasets. 🚀
🔗 Explore: github.com/torchgeo/tor...
#GeoTech #GIStools #RemoteSensing #SpatialAI #TorchGeo #GeospatialTech
Case in point, the Lil Transformy project we did together. Built a 2-layer, 2-head, 2-expert and trained it to tell terrible children's stories. That was Alph showing me how to PyTorch, basically.
github.com/jefferyharre...
github.com/jefferyharre...
Workshop/projects/Lil_Transformy at main · jefferyharrell/Workshop
Contribute to jefferyharrell/Workshop development by creating an account on GitHub.
github.com
December 11, 2025 at 2:31 AM
Case in point, the Lil Transformy project we did together. Built a 2-layer, 2-head, 2-expert and trained it to tell terrible children's stories. That was Alph showing me how to PyTorch, basically.
github.com/jefferyharre...
github.com/jefferyharre...
🚀 Hugging Face is hiring a Cloud Machine Learning Engineer (Remote US)!
Build the future of open-source AI at scale.
Key skills: Python, PyTorch/TensorFlow, Kubernetes, AWS/GCP, ML Ops
Apply fast 👉 tr.ee/4VabYT
RTs appreciated! ❤️
#AIJobs #MachineLearning #RemoteJobs
Build the future of open-source AI at scale.
Key skills: Python, PyTorch/TensorFlow, Kubernetes, AWS/GCP, ML Ops
Apply fast 👉 tr.ee/4VabYT
RTs appreciated! ❤️
#AIJobs #MachineLearning #RemoteJobs
December 10, 2025 at 5:29 AM
🚀 Hugging Face is hiring a Cloud Machine Learning Engineer (Remote US)!
Build the future of open-source AI at scale.
Key skills: Python, PyTorch/TensorFlow, Kubernetes, AWS/GCP, ML Ops
Apply fast 👉 tr.ee/4VabYT
RTs appreciated! ❤️
#AIJobs #MachineLearning #RemoteJobs
Build the future of open-source AI at scale.
Key skills: Python, PyTorch/TensorFlow, Kubernetes, AWS/GCP, ML Ops
Apply fast 👉 tr.ee/4VabYT
RTs appreciated! ❤️
#AIJobs #MachineLearning #RemoteJobs
🧰 Practical GeoAI Tools
Have you tried TorchGeo? 🌍✨
It’s a PyTorch library built for remote sensing and geospatial data, basically torchvision for Earth observation. 🚀
With ready-to-use datasets, spatial transforms, and pre-trained models, TorchGeo makes it easy to train deep learning models […]
Have you tried TorchGeo? 🌍✨
It’s a PyTorch library built for remote sensing and geospatial data, basically torchvision for Earth observation. 🚀
With ready-to-use datasets, spatial transforms, and pre-trained models, TorchGeo makes it easy to train deep learning models […]
Original post on mastodon.social
mastodon.social
December 10, 2025 at 8:50 AM
🧰 Practical GeoAI Tools
Have you tried TorchGeo? 🌍✨
It’s a PyTorch library built for remote sensing and geospatial data, basically torchvision for Earth observation. 🚀
With ready-to-use datasets, spatial transforms, and pre-trained models, TorchGeo makes it easy to train deep learning models […]
Have you tried TorchGeo? 🌍✨
It’s a PyTorch library built for remote sensing and geospatial data, basically torchvision for Earth observation. 🚀
With ready-to-use datasets, spatial transforms, and pre-trained models, TorchGeo makes it easy to train deep learning models […]
currently taking bets on which happens first
the heat death of the universe
or
my pytorch compilation finishing
the heat death of the universe
or
my pytorch compilation finishing
December 6, 2025 at 1:47 AM
currently taking bets on which happens first
the heat death of the universe
or
my pytorch compilation finishing
the heat death of the universe
or
my pytorch compilation finishing
still mad that pytorch deprecated my gtx 960m when it's only a decade old
December 5, 2025 at 9:03 PM
still mad that pytorch deprecated my gtx 960m when it's only a decade old
Note that they are currently not pytorch compatible as far as I can tell (github.com/colmap/colma...)
pycolmap-cuda12 and pytorch are incompatible · Issue #3796 · colmap/colmap
Running this setup uv init --bare uv add torch pycolmap-cuda12 will yield the following error: × No solution found when resolving dependencies: ╰─▶ Because only the following versions of torch are ...
github.com
December 4, 2025 at 7:56 AM
Note that they are currently not pytorch compatible as far as I can tell (github.com/colmap/colma...)
Picklescan Bugs Allow Malicious PyTorch Models to Evade Scans and Execute Code
Picklescan Bugs Allow Malicious PyTorch Models to Evade Scans and Execute Code
thehackernews.com
December 3, 2025 at 10:05 AM
Picklescan Bugs Allow Malicious PyTorch Models to Evade Scans and Execute Code
Your container does not have GPU drivers installed
So, how does PyTorch inside it actually use the host's GPU?
Let me explain 🧵
So, how does PyTorch inside it actually use the host's GPU?
Let me explain 🧵
December 1, 2025 at 12:46 PM
Your container does not have GPU drivers installed
So, how does PyTorch inside it actually use the host's GPU?
Let me explain 🧵
So, how does PyTorch inside it actually use the host's GPU?
Let me explain 🧵
📢 Whoop is #hiring a Staff Machine Learning Engineer (Foundation AI)!
💰 $125k - $190k
📍 Boston, MA
🔗 http://jbs.ink/XUTASgchl17i
#jobalert #jobsearch #JobSeekers #sportstechjobs #python #softwareengineer #dataengineer #machinelearning #pytorch #design
💰 $125k - $190k
📍 Boston, MA
🔗 http://jbs.ink/XUTASgchl17i
#jobalert #jobsearch #JobSeekers #sportstechjobs #python #softwareengineer #dataengineer #machinelearning #pytorch #design
December 3, 2025 at 2:46 AM
📢 Whoop is #hiring a Staff Machine Learning Engineer (Foundation AI)!
💰 $125k - $190k
📍 Boston, MA
🔗 http://jbs.ink/XUTASgchl17i
#jobalert #jobsearch #JobSeekers #sportstechjobs #python #softwareengineer #dataengineer #machinelearning #pytorch #design
💰 $125k - $190k
📍 Boston, MA
🔗 http://jbs.ink/XUTASgchl17i
#jobalert #jobsearch #JobSeekers #sportstechjobs #python #softwareengineer #dataengineer #machinelearning #pytorch #design
AI/ML for Biology & Healthcare: A Learning Path
The learning path author have been using to upskill self and be ready to work on biology and healthcare problems with Machine Learning.
Mastering PyTorch: From Linear Regression to Computer Vision: www.iamtk.co/mastering-py...
The learning path author have been using to upskill self and be ready to work on biology and healthcare problems with Machine Learning.
Mastering PyTorch: From Linear Regression to Computer Vision: www.iamtk.co/mastering-py...
Mastering PyTorch: From Linear Regression to Computer Vision
Learning PyTorch: tensors, operations, linear regression, datasets, dataloaders, and computer vision
www.iamtk.co
December 1, 2025 at 1:04 AM
AI/ML for Biology & Healthcare: A Learning Path
The learning path author have been using to upskill self and be ready to work on biology and healthcare problems with Machine Learning.
Mastering PyTorch: From Linear Regression to Computer Vision: www.iamtk.co/mastering-py...
The learning path author have been using to upskill self and be ready to work on biology and healthcare problems with Machine Learning.
Mastering PyTorch: From Linear Regression to Computer Vision: www.iamtk.co/mastering-py...
🚀 SAM3 is now available on conda-forge!
anaconda.org/conda-forge/...
You can install it directly with:
conda install -c conda-forge sam3 "pytorch=*=cuda*"
anaconda.org/conda-forge/...
You can install it directly with:
conda install -c conda-forge sam3 "pytorch=*=cuda*"
December 2, 2025 at 3:40 AM
🚀 SAM3 is now available on conda-forge!
anaconda.org/conda-forge/...
You can install it directly with:
conda install -c conda-forge sam3 "pytorch=*=cuda*"
anaconda.org/conda-forge/...
You can install it directly with:
conda install -c conda-forge sam3 "pytorch=*=cuda*"
There are errors everywhere in the paper, incomplete formulas, unnecessary repetitions. No legitimate ML researcher would train a deep learning model in PyTorch
- on CUDA...
- on 120 samples...
- with 115 features...
- then visualize results in Excel.
As a reviewer, 5-min quick scan -> nope.
- on CUDA...
- on 120 samples...
- with 115 features...
- then visualize results in Excel.
As a reviewer, 5-min quick scan -> nope.
November 27, 2025 at 8:09 PM
There are errors everywhere in the paper, incomplete formulas, unnecessary repetitions. No legitimate ML researcher would train a deep learning model in PyTorch
- on CUDA...
- on 120 samples...
- with 115 features...
- then visualize results in Excel.
As a reviewer, 5-min quick scan -> nope.
- on CUDA...
- on 120 samples...
- with 115 features...
- then visualize results in Excel.
As a reviewer, 5-min quick scan -> nope.
Meet GAN Man Vinit — adversarial ML specialist, malware detection builder, LLM safety engineer, and overall model bouncer. If you want the real name of PyTorch Pope, reach out to me & I’m happy to introduce you.
November 27, 2025 at 12:57 PM
Meet GAN Man Vinit — adversarial ML specialist, malware detection builder, LLM safety engineer, and overall model bouncer. If you want the real name of PyTorch Pope, reach out to me & I’m happy to introduce you.
Hands-On Machine Learning with Scikit-Learn and PyTorch added an online appendix on State-Space Models (SSMs)
ageron.github.io/homlp/HOMLP_...
ageron.github.io/homlp/HOMLP_...
November 26, 2025 at 2:47 PM
Hands-On Machine Learning with Scikit-Learn and PyTorch added an online appendix on State-Space Models (SSMs)
ageron.github.io/homlp/HOMLP_...
ageron.github.io/homlp/HOMLP_...
Hi. Just to clarify a point that might or might not be worth mentioning: LLMs are not coded. They're trained. The actual code for an LLM is a few lines of PyTorch. The program used to create the model has little influence on how the model turns out. That all comes down to training.
November 26, 2025 at 7:05 PM
Hi. Just to clarify a point that might or might not be worth mentioning: LLMs are not coded. They're trained. The actual code for an LLM is a few lines of PyTorch. The program used to create the model has little influence on how the model turns out. That all comes down to training.
@semianalysis.substack.com.web.brid.gy Dylan Patel on the AMD ROCm PyTorch support.
November 25, 2025 at 11:21 AM
@semianalysis.substack.com.web.brid.gy Dylan Patel on the AMD ROCm PyTorch support.
EGGROLL (Evolution Guided General Optimization via Low-rank Learning)
This project demonstrates integer-only training of a language model directly on the CPU, completely bypassing the need for GPUs, floating-point arithmetic, or heavy ML frameworks like PyTorch or JAX.
This project demonstrates integer-only training of a language model directly on the CPU, completely bypassing the need for GPUs, floating-point arithmetic, or heavy ML frameworks like PyTorch or JAX.
November 24, 2025 at 10:40 PM
EGGROLL (Evolution Guided General Optimization via Low-rank Learning)
This project demonstrates integer-only training of a language model directly on the CPU, completely bypassing the need for GPUs, floating-point arithmetic, or heavy ML frameworks like PyTorch or JAX.
This project demonstrates integer-only training of a language model directly on the CPU, completely bypassing the need for GPUs, floating-point arithmetic, or heavy ML frameworks like PyTorch or JAX.