Running LLMs locally on MacOS
phil at pgpotter.co.uk
Wed Aug 30 10:07:12 UTC 2023
Interesting, thanks for pointing it out.
Based in Chester in the UK.
On 30/08/2023 10:45, Sten-Erik Björling wrote:
> Here is an example from one of the more complete and fast models of codellama2:
> ➜ ~ ollama run phind-codellama
>>>> Give an example of how to define a stored procedure in PostgreSQL.
> Here is an example of how to create a simple stored procedure in PostgreSQL:
> CREATE OR REPLACE PROCEDURE update_salary(emp_id INT, percent_increase NUMERIC)
> LANGUAGE plpgsql
> AS $$
> UPDATE employees
> SET salary = salary + (salary * percent_increase/100)
> WHERE employee_id = emp_id;
> This stored procedure updates the salary of an employee by a given percentage. It takes two arguments - `emp_id` which is the ID of the employee to be updated and `percent_increase`, which is the increase percent value.
> The procedure is written in plpgsql, PostgreSQL's default procedural language. The procedure updates the salary by increasing it with the product of the current salary and the given percentage divided by 100 (to convert from a percentage to a decimal number).
> A bit more helpful than Google.
> B is billion of nodes/parameters, Running the larger models demand a lot of memory. The model above demand that the computer should have at least 32 GB memory if you want to have other systems active.
> Take care, all the best…
>> 30 aug. 2023 kl. 11:30 skrev Phil (OmnisList)<phil at pgpotter.co.uk>:
>> Have you played around with the Code Llama?
>> How in fact does it compare to just a simple Google search for getting answers?
>> I see "Continue" is within VScode, and can generate code for you...
>> Do you know what do the B figures mean? 7B, 13B, 34B and even 500B?
>> IT says that code Llama is a 7B that needs 16GB Memory...
>> Phil Potter
>> Based in Chester in the UK.
>> On 30/08/2023 06:55, Sten-Erik Björling wrote:
>>> Recommend that you look into ollama.ai - a solution to run Meta LLMs locally on your Mac. Several models supported - especially the coding support models and the uncensored ones are interesting.
>>> Demand a lot of memory for the larger ones, quite good performance - normal reservations apply…
>>> Take care all, all the best…
>>> Sten-Erik Björling
>>> Enviro Data
>>> Kyrkogatan 5A 2 tr
>>> SE-972 32 Luleå
>>> E-Mail:s-e.bjorling at enviro.se
>>> Mobile: +46-70-655 11 72
>>> Wire: @stenerikbjorling
>>> Skype: stenerikbjorling
>>> iChat:stene at icloud.com
>>> Signal: +46 70 655 11 72
>>> FaceTime:stene at icloud.com
>>> Telegram: @stenerikbjorling
>>> Hotmail /Messenger:stenerikbjorling at hotmail.com
>>> GMail:stenerikbjorling at gmail.com
>>> This email and any files transmitted with it are confidential, may be legally privileged and are intended solely for the use of the individual or entity to whom they are addressed. If you are not the intended recipient, please note that any use, distribution, or reproduction of the contents of this email is strictly prohibited and may be unlawful. If you are not the intended recipient, please notify the sender by return email and destroy all copies of the original message, including any attachments. Thank you.
>>> Please note that we take reasonable precautions to prevent the transmission of viruses; however, we cannot guarantee that this email or its attachments are free from viruses. We only send and receive emails because we are not liable for any loss or damage resulting from the opening of this message and/or attachments.
>>> Manage your list subscriptions athttps://lists.omnis-dev.com
>>> Start a new message ->mailto:omnisdev-en at lists.omnis-dev.com
>> Manage your list subscriptions athttps://lists.omnis-dev.com
>> Start a new message ->mailto:omnisdev-en at lists.omnis-dev.com
> Manage your list subscriptions athttps://lists.omnis-dev.com
> Start a new message ->mailto:omnisdev-en at lists.omnis-dev.com
More information about the omnisdev-en