Ай бұрын
Ай бұрын
Ай бұрын
2 ай бұрын
2 ай бұрын
Пікірлер
@GuilhermeNarcisoLee Ай бұрын
Thanks brother, really great content.
@MatijaGrcic Ай бұрын
Remember, technology transformation is a journey. Organize effectively before diving into the details of Generative AI.
@kevon217 Ай бұрын
Nice work!
@ruili7941 Ай бұрын
much more useful than the bertopic official document
@reanwithkimleng 2 ай бұрын
❤❤❤
@ShonDev 2 ай бұрын
Do you have any paper on that topic?
@ShonDev 2 ай бұрын
Very informative, thanks
@chandruchan2799 2 ай бұрын
1st like
@VasylHerman 2 ай бұрын
This is great stuff!
@jaysisodiya4360 2 ай бұрын
Impressive approach
@israrshohaaz6997 3 ай бұрын
Plese mention the link of data set
@lokeshdeshmukh8044 4 ай бұрын
from where we can get the dataset for the study purpose
@sindhujakannan6300 4 ай бұрын
Informative. Thank you 👍
@drabbassiamiracardologieme1746 4 ай бұрын
Thank you very much
@pkmisra769 4 ай бұрын
The best video ever !!
@mobitomy 6 ай бұрын
Nice one mate
@nikhilagrawal526 6 ай бұрын
In case, if we pass the general data which doesn't contains PHI it is still performing the de-identification of it. How can we solve this challenge ?
@nikhilagrawal526 6 ай бұрын
@Johnsnowlabs
@hudaalhusain6722 6 ай бұрын
ما شاء الله عليكم والله مجهود جبار وعمل عظيم جداً.. جزاكم الله كل خير والله يعطيكم الف عافية👏👏👏👏
@rickyS-D76 6 ай бұрын
Zain is the best in Multi-modal application and weaviate <3 💌
@wusetianawuzetian 6 ай бұрын
Does she know what is BPE and how to use it?
@rafikyahia7100 7 ай бұрын
I'm working on something similar, this is very helpful, thank you for the content!
@jukee67 7 ай бұрын
This is dystopian and insane. Don't worry I don't use anything outside of this platform and sharing anything like this will just get two words invented by the powers that be, said behind my back. In the end you people lose and in the after end up down near the 9th gate. Tik tok, tik tok...counting down. I mean no ill will in my comment. It is what it is at this point. You do it for the money. I am glad I am not working with or for any of these entities in intelligence and government. I have most of my soul intact midway thru life. Even I am willing to admit I was not perfect.
@kali786516 8 ай бұрын
both folks English is to hard to understand
@encycdatabelldiarycareertr5616 8 ай бұрын
Pls share the presentation deck for all sessions
@jimcastrigano4592 8 ай бұрын
Is this still supported?
@test-rg9ld 8 ай бұрын
Hey i need your slides please provide me as soon as possible, this is a request. Thank you.
@test-rg9ld 8 ай бұрын
hey can you pleasse give me the slides ?
@abderrahmane_mihoubi 8 ай бұрын
Good explanation Lawrence. Thanks
@mirzaimadbaig 9 ай бұрын
good work
@tameshhalder857 9 ай бұрын
Defragment missing :)
@johng5295 9 ай бұрын
Thanks in a million.
@axiomatic4059 10 ай бұрын
Great video. Where can we find the link to the notebooks?
@reslleygabriel 10 ай бұрын
Great content, thanks for sharing!
@johng5295 10 ай бұрын
Awesome.
@StanleyDenman 10 ай бұрын
I am quite confused. If you have to create a JSON table correlating a question with the "measure" and "diagnostic code" your solution just seems rules-based, and the creation of that table would be a herculean task.
@PenPixelLLP 10 ай бұрын
is the code open-source? if yes where can i find it? Thanks
@leauyn 10 ай бұрын
AttributeError: 'NLUPipeline' object has no attribute 'predict_embeds' when running FAISS.from_documents(pre_processed_docs, embeddings).as_retriever(), what may be the problem?
@ishas.7508 10 ай бұрын
Is this code available on GitHub
@StevenNess 11 ай бұрын
Very cool
@newyorkmathan 11 ай бұрын
Awesome presentation by Veysel Kokaman. The last leg was the most informative portion which confirmed my understanding that for 'Specific use cases you only need a SMALL model trained using specific data, rather than the 100+MB general purpose LLMs ''. Having your own data is the key for the success.
@harrykekgmail 11 ай бұрын
wow. thanks. Fantastic explanations. This is so practical.
@misterloafer5021 11 ай бұрын
Thank you for this. It was very informative. What are your views on using FactScore?
@teachingmachine 11 ай бұрын
It's all about drafting proper prompt Cocaman.. anyways thanks for the information, I will try to regenerate these summarisations on the models shown with improvised prompts.
@NAE333 11 ай бұрын
This had Nothing to do with insurance...
@JeisonJimenez-tb3nc Жыл бұрын
Hello Swagata, I really liked the video, thank you. I'm new to the LLM field, but I'm trying to train the GPT-2 model in a specific domain, in this case, an engineering domain. One of the big questions I have is how I should process my data (which are PDF documents) and in what format I should present this information to the model. This is so that after training, I can ask the model questions and it can respond accordingly.
@alimahmans9681 Жыл бұрын
I started researching RAG a week ago and was fortunate to find this video today Thank you from the bottom of my heart for your very rich content
@Prosper203 Жыл бұрын
good job👍... Please, can you share the presentation slide ?
@geoffreygordonashbrook1683 Жыл бұрын
Great level of detail in overview and coverage of scale and production-deployment issues (and costs). Many thanks! John Snow Labs does it again.
@dr.ahmetkorkmaz942 Жыл бұрын
It's nice to know what a great project we are involved in, congratulations Veysel!!!
@KemalCanKara Жыл бұрын
Thank you for great presentation. Are you going to share the notebook?