montanalow

7 ms·

31.
by montanalow 4y ago
What may be missing from cloud, is alignment of incentives. If you waste more compute, you increase their profit margins. That would explain things the author questions like general latency increases.
32.
by montanalow 4y ago
I think what you're missing is that XGBoost is worthless without data to use for inference. That data can come from in process, or over the wire. One is fast, one is not.
33.
by montanalow 4y ago
As a contributor, I think it's interesting when comments focus on the language (Python vs Rust) vs the architecture (local vs remote). Inference is embarrassingly parallelizable, with Python Flask or Postgres replicas. I think the inte
34.
by montanalow 4y ago
Those well defined slices of pie you’re referring to are often exactly the political boundaries OP is referring to. This is the degenerate form of Conways Law.
35.
by montanalow 4y ago
We weren’t expecting 35x speedup. The difference was pretty surprising.
36.
by montanalow 4y ago
We're launching the PostgresML Gym today, a free/hosted version of Postgres with our machine learning extension built in. It enables you to train supervised ML algorithms (regression, classification) on real time data in your data
37.
by montanalow 4y ago
Go to huggingface.com and start with some of the tutorials. The operational/engineering skill sets alone are all you need to treat modern ML models like any other black box API/SDK.
38.
by montanalow 4y ago
Yep, they will use the specialized hardware by default if CUDA/cuBLAS has been installed on the system.
39.
by montanalow 4y ago
My point is that the ships telemetry was not an issue. Blaming electric vehicles with no basis, or lack of telemetry with clear evidence to the contrary is…something.
40.
by montanalow 4y ago
The entire crew was rescued. The article is nonsense.
41.
by montanalow 4y ago
This article is pure speculation by an insurance company that isn’t involved, and admit they have no idea what happened. FUD.
42.
by montanalow 4y ago
Landed a PR to enable GridSearchCV and RandomSearchCV. https://github.com/postgresml/postgresml/pull/83
43.
by montanalow 4y ago
Outside of academia, you need to be able to constantly adapt to changing training data. That is a constant pitfall for many projects that try to transition from offline to online.
44.
by montanalow 4y ago
Stateful services like Postgres cannot be rapidly scaled up/down to adjust to daily loads, even when those loads are very predictably busy at noon, but often sit idle overnight. Scheduling ML jobs while the db is typically wasting reso
45.
by montanalow 4y ago
We currently calculate those as applicable to classification and regression, but they are only displayed on a model detail page like here: https://demo.postgresml.org/models/1 https://demo.postgresml.org&#x2
46.
by montanalow 4y ago
Nice! Feel free to file any issues on Github if something doesn't work. I'd love to understand more about the internals of the hydra architecture.
47.
by montanalow 4y ago
This is an interesting benchmark I'll try to code up. Although, it seems a bit like an apples/oranges comparison, since a Dataframe in memory had to come from somewhere, either a CSV or database like Postgres, in which case I have
48.
by montanalow 4y ago
Custom proprietary models are on the TODO list. There are open questions about the mechanism to distribute the model params and code, which will lead to additional complexity, but I think it's something we'll eventually need to su
49.
by montanalow 4y ago
Yep! Part of the power of being inside the OLTP is that you can just create a VIEW of your training data, which could be anything from customer purchases, search results, whatever, and that VIEW can be re-scanned every time you do the train
50.
by montanalow 4y ago
This can be done "manually" by configuring Postgres replication and/or foreign data wrappers. We don't have a magic button for that, but if we have a few examples in the wild we can establish best practices and then put
51.
by montanalow 4y ago
Yep! traditional hypertuning techniques and automated broader surveys across multiple algos are both on the roadmap for "soon".
52.
by montanalow 4y ago
Awesome. I'll see what it'll take to get the extension running w/ Python 3.6. It's good to know what people's ecosystem dependencies look like.
53.
by montanalow 4y ago
OLAP vs OLTP will depend on your ML use case. Online predictions will likely be better served by an OLTP vs offline batch predictions being better served by OLAP. OLAP use cases often involve a lot of extra complexity out of the gate, and s
54.
by montanalow 4y ago
It’s probably a stretch to run GPT-3 inside a db, but most of the “deep learning” models I’ve run in more traditional environments are a few megabytes. That’s millions of params, but Morre’s law has been generous enough to us over the decad
55.
by montanalow 4y ago
The data split technique is one of the optional parameters for the call to ‘train’. Model degradation is a really interesting topic, that is hopefully made less difficult when retraining is trivialized, but we also want to add deeper analyt
56.
by montanalow 4y ago
We can’t control the extensions RDS allows to be installed, and they are historically conservative. Lev and I do have some fairly extensive experience with replication patterns to Postgres instances running in EC2. Foreign data wrappers are
57.
by montanalow 4y ago
Thanks! We’d still consider this an early stage project and would love your feedback for which features to prioritize. Our roadmap is only getting longer…
58.
by montanalow 4y ago
"Buy it again" is simply a name for the PostgresML project that the model is being trained for. There is deeper explanation in the README: https://github.com/postgresml/postgresml
59.
by montanalow 4y ago
That's exactly the target use case. Models make online predictions as part of Postgres queries, and can be periodically retrained in a cadence that makes sense for the particular data set. In my experience the real value of retraining
60.
by montanalow 5y ago
The strategy is similar to Madlib. I tried to deploy Madlib a year or so ago, and it didn't have the features I wanted and development seemed somewhat slow/stalled with vmware deprecating support for pymadlib. I got a wild hair la

More ›