Saksham

dabbling in ml

[Saksham @capablegupta]

thank you guys for a great show. @myheroacademia

[Saksham @capablegupta]

@burkov yeah , would like to know if there is any kind of comparison being made with LLMs ( upto 7B ) & openai for retrieval based answers.

[Saksham @capablegupta]

@jeremyphoward How to get the most out of open source LLMs for rag based tasks ( in terms of setting/identifying best parameters ) .

P.S : os LLMs ( falcon & llama-2 ) claim to have a close performance to chatGPT , but this isn't true when we tested it for rag workflow , your views ?

[Saksham @capablegupta]

@SamuelSzuchan Hi @SamuelSzuchan , I tested the https://t.co/2DsgeEpg6q , its great , work out of the box .

Have a few questions regarding implementation. Is there a way I can contact to founder or eng. team .

Thanks ,

[Saksham @capablegupta]

@carlcarrie GitHub link not working .

[Saksham @capablegupta]

@4real4reall Akame ga kill .

[Saksham @capablegupta]

@chickenbiryan This song in the movie is 🔥 https://t.co/xsKGVPcCVm

[Saksham @capablegupta]

@doesdatmaksense @sGx_tweets Can use https://t.co/UIPTcfm98O for data labelling/test set creation.
Has active learning , so hook it up with a pretrained model that is available & you're good to go.
Have used this in my recent NER for PHI( protected health information) labelling .

[Saksham @capablegupta]

@driscollis @KalobTaulien pyenv + virtualvenv . Maintains exact python version & manage virtual env.
Read below blog by @eprosenthal .
https://t.co/GGNYdgzoCj

[Saksham @capablegupta]

@JFPuget @LChoshen https://t.co/z6td123ZkB
I find this informative .

[Saksham @capablegupta]

@jmans89 @marktenenholtz https://t.co/WER0ebklBC
Search the topic and select the framework.

[Saksham @capablegupta]

@driscollis Not exactly the same . But I enjoyed the explanation .
https://t.co/f2mZz8Ugi8

[Saksham @capablegupta]

@giovannibassi I like the codesnap extension too. Great help for blogs and presentation.

[Saksham @capablegupta]

@marktenenholtz Hey @marktenenholtz , can you point out some resources ( repository / personal projects) which shows the best practices ( model monitoring /experiment tracking , automation CI/CD ) in ML projects. I haven't seen personal projects which combines these in their workflows.
Thanks .

[Saksham @capablegupta]

@kapoor_saumitra popOS .

[Saksham @capablegupta]

@abhi1thakur https://t.co/KwiI3gzTiP is a good choice , it has some integration available for GitHub. I think you'll like the tool .

[Saksham @capablegupta]

@benalidjamel6 @shreyacasmalert Assassination Classroom 💯

[Saksham @capablegupta]

@rothecoder The important thing according to me , it's okay to get intimidated by the stuff you're seeing & learning everyday , but never trick/ tell yourself that you'll never be good at it . It will take time , consistency & lot of "this sucks" kind of moment.
Happy Learning :)

[Saksham @capablegupta]

@aishwarya_2x21 https://t.co/PGMqYzkkWE has great set of questions and I kinda love the UI too :)

[Saksham @capablegupta]

@kormilitzin @huggingface @streamlit @Gradio I am also working on PHI detection model , looking forward to using the above in the model .