urcyanide

11 ms·

1.
Why the code agents should run inside the containers(envd.tensorchord.ai)
2 points by urcyanide 10mo ago|0 comments
2.
Rewrite the pre-commit with features requested from the community(github.com)
2 points by urcyanide 1y ago|1 comments
3.
by urcyanide 1y ago
Rewriting it in Rust not only brings high performance, but also adopts more feature requests from the community.
4.
Earth Observability with the Help of Vectors(blog.vectorchord.ai)
1 points by urcyanide 1y ago|0 comments
5.
by urcyanide 2y ago
Thanks for pointing out. I have updated the post. I use the opposite shift since I store the binary in u64 with a different endian from the C++ version. Sorry for the confusion, my bad.
6.
by urcyanide 2y ago
Sorry for the confusing example. The bitwise one is correct since I store the binary in u64 with a different endian from the C++ version. (this happens because the C++ version is using a numpy script to do the preprocessing) My bad, I shoul
7.
by urcyanide 2y ago
Yes, i&32 is more readable.
8.
by urcyanide 2y ago
Not really, i \in (0..dim).step_by(32), shift \in {0, 32}, but for `i % 32` it's 0..32
9.
by urcyanide 2y ago
The end2end QPS is about 3x of the 1st Rust version on both SIFT & GIST datasets.
10.
by urcyanide 2y ago
I also see the profiling shows that the "iter::next" takes a large percentage in the flamegraph. Are they the same reason?
11.
by urcyanide 2y ago
Thanks for pointing out, this can be optimized by the compiler when enabling opt-level=3
12.
by urcyanide 2y ago
It has the same QPS as the C++ version for GIST dataset. While Rust has more SIMD, C++ has const generic. I guess there is still some space for future improvement.
13.
Lessons learned from profiling an algorithm in Rust(blog.mapotofu.org)
155 points by urcyanide 2y ago|38 comments
14.
by urcyanide 3y ago
Some potential issues about Python multiprocessing https://blog.mapotofu.org/blogs/python-multiprocessing/ . COW is quite tricky. BTW, most of the related official Python docs doesn’t mention the usage under ‘spaw
15.
by urcyanide 4y ago
If you are trying to compare the performance as a ML service, maybe you should try to compare it with other ML model serving frameworks like https://github.com/mosecorg/mosec or https://github.com/bento
16.
Python Machine Learning Service Can Run Way More Faster(github.com)
1 points by urcyanide 4y ago|0 comments
17.
Bridging the gap between infra and machine learning: start from envd(envd.tensorchord.ai)
1 points by urcyanide 4y ago|0 comments
18.
by urcyanide 4y ago
Recently, I found another tool built for machine learning environment: https://github.com/tensorchord/envd . It integrates with buildkit and can share the config like Python functions cross the team. This can reduce a l
19.
Show HN: Mosec makes the machine learning model serving flexible and efficient(github.com)
4 points by urcyanide 5y ago|1 comments