mateiz

6 ms·

1.
by mateiz 2y ago
This is a cool paper showing there is value in using an LLM multiple times, but in recent research we showed that with majority voting, quality can decrease past some point as you make more calls. Check out https://arxiv.org&#x
2.
by mateiz 8y ago
Don't MLflow Projects exactly meet this use case? A project lives in a Git repo, which can include both code and data, and specifies its software environment (currently Conda but will eventually also support Docker): https:/&#x2F
3.
by mateiz 8y ago
While MLflow doesn't submit jobs to Kubernetes for you, it should be possible to integrate it with your favorite scheduler to do that. MLflow is designed to accept experiment results from wherever you are running your code, so you can
4.
An Analysis of the DAWNBench v1 Benchmark for Deep Learning(dawn.cs.stanford.edu)
1 points by mateiz 8y ago|0 comments
5.
DAWNBench v1 Deep Learning Benchmark Results(dawn.cs.stanford.edu)
4 points by mateiz 8y ago|0 comments
6.
by mateiz 9y ago
Yup, this is a great explanation. Cuckoo hashes actually do much better in terms of load if you use buckets with multiple items, as we did here. The classical one with one item per location can get "stuck" with unresolvable cycles
7.
by mateiz 9y ago
We'd love to hear about the size of the payloads used. With the code we posted, larger records take 10-15 more ns per query, which is still faster than the numbers in the paper, but of course YMMV based on the table implementation and
8.
by mateiz 9y ago
You can see the table here, it's not a lot of code: https://github.com/stanford-futuredata/index-baselines/blob/... . Each bucket just has 8 keys and you want to test whether one of them is equal to the k
9.
by mateiz 9y ago
Matei Zaharia (one of the PIs on DAWN) here. Snorkel, MacroBase and ASAP are already being used in production at several companies, and we intend to continue publishing everything as open source. We only started this lab a year ago, so a lo
10.
by mateiz 9y ago
Shoot, now we have to implement all of those before someone else does :).
11.
Making Apache Spark the Fastest Open Source Streaming Engine(databricks.com)
7 points by mateiz 9y ago|0 comments
12.
by mateiz 10y ago
https://en.wikipedia.org/wiki/The_Hercules_Text
13.
Structured Streaming in Apache Spark 2.0(databricks.com)
8 points by mateiz 10y ago|0 comments
14.
Continuous Applications: Evolving Streaming in Apache Spark 2.0(databricks.com)
6 points by mateiz 10y ago|0 comments
15.
by mateiz 10y ago
Databricks -- San Francisco -- https://databricks.com * Software Engineer (ONSITE) * Software Engineer Intern (ONSITE) * Product Manager (ONSITE) Databricks was founded in 2013 by the team that started Apache Spark, meaning you
16.
Deep Learning with Spark and TensorFlow(databricks.com)
228 points by mateiz 11y ago|30 comments
17.
Diving into Spark Streaming's Execution Model(databricks.com)
48 points by mateiz 11y ago|0 comments
18.
Announcing SparkR: R on Spark(databricks.com)
4 points by mateiz 11y ago|0 comments
19.
Project Tungsten: Bringing Spark Closer to Bare Metal(databricks.com)
59 points by mateiz 11y ago|11 comments