TPUs are still a bit of a pain (to say the least!) to get working with PyTorch, and PyTorch is the easiest way still to get stuff done, on the whole.
COMPUTING
-
Perplexity Launches pplx-api LLM Platform with Mistral Llama2
By
–
Introducing pplx-api, our LLM API which serves Mistral and Llama2 models with blazing speed and throughput. pplx-api is in public beta for our Pro subscribers! We partnered with @nvidia and @awscloud to build our proprietary inference. Learn more: https://
pplx.ai/introducing-pp
lx-api
… -
vRAN L1 Acceleration Delivers Performance and Energy Efficiency
By
–
As the RAN evolves from traditional closed architectures to an open & software-defined architecture, a full inline acceleration of vRAN L1 delivers best-in-class performance and energy efficiency. https://
nvda.ws/46yNv4g -
JAX Library Enables Computer Vision on Spherical Surfaces
By
–
Applying computer vision models designed for planar images to data projected on spherical surfaces is challenging. Here we present an open-source library in JAX to solve the challenges of rotation and regular sampling for state-of-the-art performance → https://t.co/wXdIpkmtDy pic.twitter.com/0mC7PdeLW4
— Google AI (@GoogleAI) 4 octobre 2023Applying computer vision models designed for planar images to data projected on spherical surfaces is challenging. Here we present an open-source library in JAX to solve the challenges of rotation and regular sampling for state-of-the-art performance → https://
goo.gle/46z3vD7 -
Fractional Fourier Neural Operator Advances Global Local Features
By
–
Nice work on Fractional Fourier neural operator. Fractional Fourier transform incorporates both time and frequency content that allows for both global and local features to be captured. Fun fact: my undergrad thesis used Fractional transform.
-
IBM API Connect Wins Red Dot and Stratus Design Awards
By
–
In recognition of excellence in user-centric design, the @IBM API Connect team earns two prestigious industry awards: – The @reddot Design Award for #UX and #UX – The Stratus Award by @BigAwards Sichao Wu, IBM outlines contributing factors: https://
ibm.co/3ZG2XJw -
BNB Quantized Layers Initialization Performance Issues
By
–
The issue is the creation of the bnb quantised layers, not the adapters. (They are fast to init anyway since they're small.)
-
Model Sharding Challenges for Large GPU Deployment
By
–
Not sure if there's a way to do that with model sharding — which is necessary if your model is too big to fit on the GPU.
-
Python contextlib.contextmanager decorator explained
By
–
BTW this is a great example of the use of the handy `contextlib.contextmanager` decorator. Check out the docs if you haven't used it before: https://
docs.python.org/3/library/cont
extlib.html#contextlib.contextmanager
… -
COO Discusses RISC-V Strategy and Tenstorrent Board Role
By
–
Our COO Keith Witek sits down with @IanCutress of @TechTechPotato to talk bringing IP to market, building with #riscv and joining the @tenstorrent board. Watch the full video here –>