Yes. I forgot to mention.
On Tue, Jun 12, 2018 at 2:06 AM Donald Szeto wrote:
> On Mon, Jun 11, 2018 at 9:38 AM Digambar Bhat
> wrote:
>
>> You don't need to deploy same engine again and again. You just deploy
>> once and train whenever you want. Deployed instance will automatically
>> point
As an example, reduce this number to 10 or even 5:
https://github.com/apache/predictionio-template-recommender/blob/develop/engine.json#L15
On Mon, Jun 11, 2018 at 1:35 PM KRISH MEHTA wrote:
> Hi I am using MyRecommendation Template. What do you suggest me doing?
>
> On Jun 11, 2018, at 1:33
1. What Spark version are you using? After taking a brief look, there seems
to be changes in how classpaths are being passed into the latest version of
Spark.
2. When pio-shell starts, it should echo the classpath being. Does that
include a directory that contains your hbase-site.xml?
3. In your
On Mon, Jun 11, 2018 at 9:38 AM Digambar Bhat
wrote:
> You don't need to deploy same engine again and again. You just deploy once
> and train whenever you want. Deployed instance will automatically point to
> newly trained model as hot swap happens.
>
To clarify, this happens with certain
Hi I am using MyRecommendation Template. What do you suggest me doing?
> On Jun 11, 2018, at 1:33 PM, Donald Szeto wrote:
>
> Which template are you using? This is usually caused by too many iterations
> on a single machine. The parameter is usually called `numIterations` in
> `engine.json`.
We are using PIO 0.12.1. Are there additional parameters that I should be
sending in?
Thanks for the response.
--Cliff.
On Mon, Jun 11, 2018 at 8:29 PM, Donald Szeto wrote:
> Hi Cliff,
>
> What PIO version are you using? We fixed a similar issue in
>
Which template are you using? This is usually caused by too many iterations
on a single machine. The parameter is usually called `numIterations` in
`engine.json`.
Regards,
Donald
On Mon, Jun 11, 2018 at 10:22 AM KRISH MEHTA
wrote:
> When I am running Pio train I am encountering the error given
Thanks a lot for your response.
> On Jun 11, 2018, at 1:31 PM, Donald Szeto wrote:
>
> Hi Krish,
>
> This warning does not impede functionality, but suggesting you could achieve
> further performance boost if you install numerical libraries native to your
> machine's architecture. For more
Hi Krish,
This warning does not impede functionality, but suggesting you could
achieve further performance boost if you install numerical libraries native
to your machine's architecture. For more info, this is a good read:
Hi Cliff,
What PIO version are you using? We fixed a similar issue in
https://issues.apache.org/jira/browse/PIO-72. If it is happening still in
0.12.0+ we will need to investigate.
Regards,
Donald
On Tue, Jun 5, 2018 at 1:35 PM Miller, Clifford <
clifford.mil...@phoenix-opsgroup.com> wrote:
>
It seems to be a problem related to your JVM setup:
https://stackoverflow.com/questions/6784463/error-trustanchors-parameter-must-be-non-empty
In your case, `pio build` invoked the underlying build tool (sbt) and was
trying to download other libraries from the Internet, but failed at that
point.
Actually if you are using the Universal Recommender you only need to deploy
once as long as the engine.json does not change. The hot swap happens as
@Digambar says and there is literally no downtime. If you are using any of the
other recommenders you do have to re-deploy after every train but
You don't need to deploy same engine again and again. You just deploy once
and train whenever you want. Deployed instance will automatically point to
newly trained model as hot swap happens.
Regards,
Digambar
On Mon 11 Jun, 2018, 10:02 PM KRISH MEHTA, wrote:
> Hi,
> I have just started using
13 matches
Mail list logo