crowwork

15 ms·

31.
Web LLM(github.com)
6 points by crowwork 3y ago|0 comments
32.
by crowwork 4y ago
Webgpu will ship this year, so it will be more widely available pretty soon
33.
by crowwork 4y ago
checkout https://mlc.ai/web-stable-diffusion , which is builds on top of Apache TVM and brings in models from PyTorch2.0, ONNX and other means into the ML compilation flow
34.
Web Stable Diffusion(github.com)
254 points by crowwork 4y ago|41 comments
35.
by crowwork 4y ago
FP16 is already in the spec and hopefully will ship this year we believe
36.
by crowwork 4y ago
On the current setup(M2 and metal), it should run as fast as the local native environment.
37.
by crowwork 4y ago
The current demo uses CLIP model from openai, which is likely what you are looking for
38.
by crowwork 4y ago
on apple M2max, it takes 20sec, and with the fp16 support, there will be opportunities for improvement likely soon this eyar
39.
Running Stable Diffusion fully in browser with WebGPU(mlc.ai)
42 points by crowwork 4y ago|19 comments
40.
by crowwork 4y ago
This project brings stable diffusion models to web browsers. Everything runs inside the browser with no server support. Please check out our GitHub repo to see how we did it. There is also a demo which you can try out.
41.
TinyML – How TVM Is Taming Tiny(tvm.apache.org)
6 points by crowwork 6y ago|0 comments
42.
Compiling Machine Learning to WASM and WebGPU with Apache TVM(tvm.apache.org)
4 points by crowwork 6y ago|0 comments
43.
FFI Navigator: Language Server for Cross Language FFI Calls(github.com)
2 points by crowwork 7y ago|0 comments
44.
TVM and Deep Learning Compilation Conference 2019 Videos and Slides
2 points by crowwork 7y ago|0 comments
45.
by crowwork 7y ago
As TVM continuously demonstrates improvements to the efficiency of deep learning execution, it has become clear that PyTorch stands to benefit from directly leveraging the compiler stack. A major tenet of PyTorch is providing seamless and r
46.
by crowwork 7y ago
With learning-based program optimizer, we can competitive performance on benchmark models and significant boost on emerging models against TensorRT(int8).
47.
Automating Optimization of Quantized Deep Learning Models on CUDA(tvm.ai)
13 points by crowwork 7y ago|2 comments
48.
by crowwork 8y ago
“TVM is right for the Apache Software Foundation, and the Apache Software Foundation is right for TVM: One thing the ASF excels at is enabling collaboration across organizations, and encouraging collaboration even among competitors. With co
49.
by crowwork 8y ago
TVM is an open-source deep learning compiler stack for CPUs, GPUs, and specialized accelerators. It aims to close the gap between the productivity-focused deep learning frameworks, and the performance- or efficiency-oriented hardware backen
50.
Golang Runtime for Deep Learning Deployment in TVM(tvm.ai)
2 points by crowwork 8y ago|1 comments
51.
by crowwork 8y ago
Engineering and design contain tradeoffs, and one is better than other is quite subjective and must put into the context. We would certainly welcome healthy discussions, where I would be more than happy to talk about the different design ch
52.
by crowwork 8y ago
please stop doing this, it is sad to see fires being set up on threads which belong to Halide. TVM benefit a lot from its Halide ancestry, and we are deeply grateful for that. I personally made mistake initially, which I deeply regretted, f
53.
Automating Generation of Low Precision Deep Learning Operators(tvm.ai)
1 points by crowwork 8y ago|0 comments
54.
by crowwork 8y ago
indeed, they refer to GPU (kernel) programs
55.
by crowwork 8y ago
TVM comitter here, we have benefited a lot from the Halide community in particular its IR, and we are very grateful of that. I personally think it is wrong to plug-in the ads here as this post is about Halide. There is no good or bad choic
56.
by crowwork 8y ago
The Versatile Tensor Accelerator (VTA) is an extension of the TVM framework designed to advance deep learning and hardware innovation. - docs https://docs.tvm.ai/vta/ - techreport https://arxiv.org/abs&
57.
VTA: An Open, Customizable Deep Learning Acceleration Stack(tvm.ai)
23 points by crowwork 8y ago|2 comments
58.
by crowwork 8y ago
the benchmarks are not about GEMM, but real-world deep learning workloads which could have very different characteristics from GEMM
59.
by crowwork 9y ago
great discussion about the difference between mobile and normal gpu
60.
by crowwork 9y ago
part of TVM https://github.com/dmlc/tvm is built with primitives in Halide. Halide is indeed super cool and TVM benefit a lot from its experience. While Halide optimizes CPU and image processing workload well. TVM also

More ›