2-day Samarkand trip
•Registan Square: The 3 madrasas boast the most incredible mosaics - did you know they were built using a special technique that makes the colors vibrant for centuries?
•Gur-e Amir: His actual tomb is hidden beneath a giant slab of black jade.
•Bibi-Khanym Mosque: This grand mosque is a marvel of tilework.
•Shah-i-Zinda: This necropolis blew us away with its intricately decorated mausoleums, each a vibrant testament to Samarkand's rich history.
•Siyob Bazaar: Perfected our bargaining skills at this bustling market, the city's beating heart.
•Ulugbek Observatory: We climbed to the top of this ancient observatory, imagining the stargazers of old. Did you know it was one of the most advanced observatories in its time?
To end the day, we capped it off with some laughter and rides at a local Family Park - the perfect way to unwind after a history-packed day!
NB: Although I am from Samarkand, I learned new interesting things about it again🙂
@ladybytes
•Registan Square: The 3 madrasas boast the most incredible mosaics - did you know they were built using a special technique that makes the colors vibrant for centuries?
•Gur-e Amir: His actual tomb is hidden beneath a giant slab of black jade.
•Bibi-Khanym Mosque: This grand mosque is a marvel of tilework.
•Shah-i-Zinda: This necropolis blew us away with its intricately decorated mausoleums, each a vibrant testament to Samarkand's rich history.
•Siyob Bazaar: Perfected our bargaining skills at this bustling market, the city's beating heart.
•Ulugbek Observatory: We climbed to the top of this ancient observatory, imagining the stargazers of old. Did you know it was one of the most advanced observatories in its time?
To end the day, we capped it off with some laughter and rides at a local Family Park - the perfect way to unwind after a history-packed day!
NB: Although I am from Samarkand, I learned new interesting things about it again🙂
@ladybytes
👍5❤1🥰1
Durov Telegram haqida
Informativ suhbat.
Moskva, Berlin, SA, Singapur emas, Dubai Telegram uchun bosh ofis.
Shunday katta platforma uchun atigi 30 ta xodim, lekin eng zo'r mutaxassislar.
Durov Telegramda ham Product Owner, ham HR, ham Designer va boshqa rollar. Akasi matematika, algoritmlashda pro.
Organik o'sish, hech qanday reklamalarsiz.
https://youtu.be/1Ut6RouSs0w?si=CkN5IQBp-Cugq1pS
@ladybytes
Informativ suhbat.
Moskva, Berlin, SA, Singapur emas, Dubai Telegram uchun bosh ofis.
Shunday katta platforma uchun atigi 30 ta xodim, lekin eng zo'r mutaxassislar.
Durov Telegramda ham Product Owner, ham HR, ham Designer va boshqa rollar. Akasi matematika, algoritmlashda pro.
Organik o'sish, hech qanday reklamalarsiz.
https://youtu.be/1Ut6RouSs0w?si=CkN5IQBp-Cugq1pS
@ladybytes
YouTube
Telegram Creator on Elon Musk, Resisting FBI Attacks, and Getting Mugged in California
Subscribe to our new Telegram channel: https://news.1rj.ru/str/TuckerCarlsonNetwork
The social media app Telegram has over 900 million users around the world. Its founder Pavel Durov sat down with us at his offices in Dubai for a rare interview.
Watch more here: …
The social media app Telegram has over 900 million users around the world. Its founder Pavel Durov sat down with us at his offices in Dubai for a rare interview.
Watch more here: …
👍8❤1
Open source vs Commercial Models
I’m making a list of things to consider when using open source models and commercial models. Here’s what I have currently. What else should I add?
Considerations for commercial models
1. Data privacy: employees might accidentally include company’s private data in the prompt, e.g. when Samsung employees accidentally leaked the company’s secrets using ChatGPT.
2. Functionality: proprietary models might have important features like function calling and JSON mode. However, most model providers have no or limited logprobs (log probabilities) API. Logprobs are very useful for classification tasks, evaluation (confidence scoring), and interpretability.
3. API cost: API calls can get expensive at scale.
4. Finetuning: model providers might not let you finetune their models. Off-the-shelf, commercial models might be better for your use case, but might not be as good as open source + finetuning.
5. Edge use cases: can’t work for use cases on devices that have no Internet connection.
Considerations for open source models
1. Data lineage/copyright: people are less likely to sue open source model builders for training on copyrighted data. However, if you use these models to make money, you can get in trouble.
2. Functionality: hosting your models gives you access to logprobs and other intermediate outputs. There are external tools that provide function calling and constrained sampling for certain open source models, but these features might be limited.
3. Engineering cost: hosting and optimizing large models takes nontrivial time, talent, and engineering effort. APIs are expensive, but engineering can be even more so. This can be mitigated by using model hosting services if they support the models you want to use.
4. Finetuning: in theory, you can finetune open source models, but it might not be easy to do so.
From LinkedIn
@ladybytes
I’m making a list of things to consider when using open source models and commercial models. Here’s what I have currently. What else should I add?
Considerations for commercial models
1. Data privacy: employees might accidentally include company’s private data in the prompt, e.g. when Samsung employees accidentally leaked the company’s secrets using ChatGPT.
2. Functionality: proprietary models might have important features like function calling and JSON mode. However, most model providers have no or limited logprobs (log probabilities) API. Logprobs are very useful for classification tasks, evaluation (confidence scoring), and interpretability.
3. API cost: API calls can get expensive at scale.
4. Finetuning: model providers might not let you finetune their models. Off-the-shelf, commercial models might be better for your use case, but might not be as good as open source + finetuning.
5. Edge use cases: can’t work for use cases on devices that have no Internet connection.
Considerations for open source models
1. Data lineage/copyright: people are less likely to sue open source model builders for training on copyrighted data. However, if you use these models to make money, you can get in trouble.
2. Functionality: hosting your models gives you access to logprobs and other intermediate outputs. There are external tools that provide function calling and constrained sampling for certain open source models, but these features might be limited.
3. Engineering cost: hosting and optimizing large models takes nontrivial time, talent, and engineering effort. APIs are expensive, but engineering can be even more so. This can be mitigated by using model hosting services if they support the models you want to use.
4. Finetuning: in theory, you can finetune open source models, but it might not be easy to do so.
From LinkedIn
@ladybytes
ImgBB
IMG-20240503-114333-152 hosted at ImgBB
Image IMG-20240503-114333-152 hosted in ImgBB
👍3💯1
Menga yoqqan kinolar
Akeelah and the bee
The pursuit of happyness
Homeless to Harvard
@ladybytes
Akeelah and the bee
The pursuit of happyness
Homeless to Harvard
Bu kinolar sizga biroz bo'lsada, ilm olishda, o'z yo'lingizda davom etishga turtki beradi.
@ladybytes
👍12❤2💯2
Tangible Tuesday
Ok, most people do mental activity but what about physical? In this case I try my best as well. For example, everyday I do jogging👟, every week nordic walking🥢 and darts🎯, sometimes table tennis🏓. Today I participated in the selection of table tennis club at our university. Some sweat, intensity, focus, energy I got & give. 2 loss & 1 win:)
Software Construction class. Some new info I learned: CUDA programming, julia sets, designing specs, debugging.
In May, I started #5 book of 2024 for me - Good to Great. From the beginning it is literally impressive to see how the author is proving his thoughts with deep researches.
That is for today.
@ladybytes
I believe that the synergy of physical and mental activity are both equally important to live healthy life.
Ok, most people do mental activity but what about physical? In this case I try my best as well. For example, everyday I do jogging👟, every week nordic walking🥢 and darts🎯, sometimes table tennis🏓. Today I participated in the selection of table tennis club at our university. Some sweat, intensity, focus, energy I got & give. 2 loss & 1 win:)
Software Construction class. Some new info I learned: CUDA programming, julia sets, designing specs, debugging.
In May, I started #5 book of 2024 for me - Good to Great. From the beginning it is literally impressive to see how the author is proving his thoughts with deep researches.
That is for today.
@ladybytes
👍5❤2🥰1
Bugun Yakshanba.
Savol-javob qilamiz.
Today is Sunday.
Let's do AMA (Ask Me Anything)
http://t.me/anonaskbot?start=ladybytes
©️meme - Latofat opa kanalidan olindi
Savol-javob qilamiz.
Today is Sunday.
Let's do AMA (Ask Me Anything)
http://t.me/anonaskbot?start=ladybytes
©️meme - Latofat opa kanalidan olindi
👍4🔥1
У тебя новое анонимное сообщение!
Assalomu alaykum. Startup ingizni davom ettiryapsizmi?
↩️ Свайпни для ответа.
Va alaykum assalom. Hali bir qarorga kelmadim. O'rganishim kerak bo'lgan ko'p narsalar bor ekan.
Assalomu alaykum. Startup ingizni davom ettiryapsizmi?
↩️ Свайпни для ответа.
Va alaykum assalom. Hali bir qarorga kelmadim. O'rganishim kerak bo'lgan ko'p narsalar bor ekan.
👍6
У тебя новое анонимное сообщение!
where are you from(region)?
↩️ Свайпни для ответа.
Samarkand
where are you from(region)?
↩️ Свайпни для ответа.
Samarkand
❤4