วีดีโอ

lovely

Bros DSE awareness is 0

Distillkit - AIM Arcee-SuperNova Event: colab.research.google.com/drive/1Fm47h787h6ROX-ylQTml2yN5wmWs0fje?usp=sharing Event Slides: www.canva.com/design/DAGSb1zA42E/YZL0ZegE01nNX3DDKNYqCg/view?DAGSb1zA42E&

Trong thang đánh giá kỹ thuật Chunking thì Chunking theo ngữ nghĩa và chunking theo agent được đánh giá ở cấp 4 và 5. Thực nghiệm cho thấy chunking agentic sử dụng LLMs cho kết quả cao nhất. Cấp 1: Tách ký tự - Các đoạn dữ liệu ký tự tĩnh đơn giản Cấp 2: Tách văn bản ký tự đệ quy - Chia nhỏ đệ quy dựa trên danh sách các dấu phân cách Cấp 3: Tách theo từng loại tài liệu - Các phương pháp chia nhỏ khác nhau cho các loại tài liệu khác nhau (PDF, Python, Markdown) Cấp 4: Tách ngữ nghĩa - Chia nhỏ dựa trên embedding. Kỹ thuật này chia đoạn văn bản thành các đoạn nhỏ dựa trên ngữ nghĩa, thay vì chỉ dựa vào độ dài cố định. Cấp 5: Tách dùng agent - Agentic Chunker: Agentic Chunker tự động nhóm các propositions (mệnh đề) có liên quan vào các chunks (nhóm). Khi thêm một proposition mới, hệ thống sẽ xác định xem có nên thêm nó vào một chunk hiện có hay tạo một chunk mới.

Trong thang đánh giá kỹ thuật Chunking thì Chunking theo ngữ nghĩa và chunking theo agent được đánh giá ở cấp 4 và 5. Thực nghiệm cho thấy chunking agentic sử dụng LLMs cho kết quả cao nhất. Cấp 1: Tách ký tự - Các đoạn dữ liệu ký tự tĩnh đơn giản Cấp 2: Tách văn bản ký tự đệ quy - Chia nhỏ đệ quy dựa trên danh sách các dấu phân cách Cấp 3: Tách theo từng loại tài liệu - Các phương pháp chia nhỏ khác nhau cho các loại tài liệu khác nhau (PDF, Python, Markdown) Cấp 4: Tách ngữ nghĩa - Chia nhỏ dựa trên embedding. Kỹ thuật này chia đoạn văn bản thành các đoạn nhỏ dựa trên ngữ nghĩa, thay vì chỉ dựa vào độ dài cố định. Cấp 5: Tách dùng agent - Agentic Chunker: Agentic Chunker tự động nhóm các propositions (mệnh đề) có liên quan vào các chunks (nhóm). Khi thêm một proposition mới, hệ thống sẽ xác định xem có nên thêm nó vào một chunk hiện có hay tạo một chunk mới.

It's looks the like Llama Index's version of the Microsoft Prompt Flow😅

bro. it's 1,999 USD for pple like me from asian countries can you please adjust it to at least PPP. Course is good but at 1999 I can have my colg degree here in India. Ideal price is 299-499 USD

Love this. Thanks for explaining this step by step (pun not intended).

Can you please share the Jupyter file (Colab link)?

AI Makerspace: Activation Aware Weight Quantization (AWQ): colab.research.google.com/drive/1eCwenXmSd7u8ZM3TSqIDm4V7BZ_3K6dA?usp=sharing Event Slides: www.canva.com/design/DAGRxxAiqtw/MIy6aqafzIfThBRBTPc86Q/view?DAGRxxAiqtw&

Limited mindset, solved it on a consumer product in like 4 days. non deterministic vs deterministic system validation 🙏

Great topics! Sounds like lots of people must be in the phase of moving from demo to prod. Exciting times! I’m in the midst of this now after pitching my RAG POC to corporate execs. I found that streaming the LLM outputs with pipeline that involves functions with multiple query tools is not easy! Need to get asynchronous streaming working next. Another topic for prod is going to be switching to FastAPI and a React widget for the web team, moving to AWS or Azure and integrating with their tracing and guardrail tools. Also on the agenda is red teaming on dev website before launch. Also need admin panel and a scraping & indexing pipeline if context is from your company website content (mine is starting with blogs, etc.)

👍

That's very helpful to me. Thank you

OMG such a helpful sesh. I had heard these terms before but never understood how they work together and what they mean. Love the care you put into explaining important concepts and illustrating them with real examples.

I like the wiz's new look

Top as always

Repo: github.com/AI-Maker-Space/Chainlit-Event-AIM/tree/main Event Slides: www.canva.com/design/DAGRHbLsvx8/EVQN5H1yJtzutJBqHbTBrw/view?DAGRHbLsvx8&

Hello :) I have an Idea to improve similarity search with human feedback: You could enhance search results by incorporating feedback on what was expected during a search query. Specifically, if a search result is unsatisfactory, you can adjust the position of the embeddings in the Milvus database based on this feedback. For example, if a relevant embedding was not found during a similarity search, you could move the desired but unfound embedding vector closer to the search embedding. This adjustment would make it more likely to be retrieved correctly in future searches with similar prompts. Steps to Implement This: 1.Identify Missing Embeddings: When a search result is not satisfactory, identify the embeddings that should have been found. 2.Adjust the Embeddings: Move the unfound embeddings closer to the search embedding by slightly adjusting their vectors. This aims to improve their chances of being recognized in future 3.searches. Reinforcement Learning: Utilize Reinforcement Learning to automate this process. The agent can learn which adjustments to the embeddings lead to better search results. I think you have a pretty good understanding for my idea. Does this approach sound feasible? Whats your oppinion?

No comments yet? I’m claiming first! ❤

Okay this short I like tbh

very cool tutorial, thanks you much

LlamaIndex Workflows - AI Makerspace: colab.research.google.com/drive/1p5L8pWzJG4KuopFNoc70OZiNpFkdYVCf?usp=sharing Event Slides: www.canva.com/design/DAGQd1a2yLY/5dKKRMR_f67OvUr-ADhnlg/view?DAGQd1a2yLY&

Good

Hold on MFker did you just use a example Question about Person(s) Living where abouts, During a live stream...which in terms probably for sure wasnt scripted......wtf is all that about. You know what im talking about

Top

This is awesome Daniel!! Very excited for the future of unsloth as more people get into it! You are making these tool accessible for everyone ❤

This is awesome Daniel!! Very excited for the future of unsloth as more people get into it! You are making these tool accessible for everyone ❤

very nice lecture. it is totally super clear!

Very valuable webinar!

AI Makerspace - Unsloth: colab.research.google.com/drive/1oqd8PrUx_6NPlNCP2TLxMYn_z44rNrM1?usp=sharing Continued pretraining - Llama 3.1 8b: colab.research.google.com/drive/1UHCo6cHQmCpmbdgZIx5qI0BeFEw8lgpX?usp=sharing Event Slides: www.canva.com/design/DAGPz2eciHc/Gl8A-t8i4mGK5ZDDml9e4g/view?DAGPz2eciHc&

Thanks for the video. The moment i saw your compromising with 3.5 for something I said, well..

Excited to see this...

The field of AI research on multi-agent systems is exciting. SmythOS is a platform that focuses on creating and overseeing AI agents.

DSPy Agent Example - AvaTaR: colab.research.google.com/drive/1G00JiyXnFHrxdgYM7V7Vm3ywCqNCewZE?usp=sharing Event Slides: www.canva.com/design/DAGPJ9Pe9fM/4lZHKckrPx3IWD7ZCmZEkw/view?DAGPJ9Pe9fM&

Link to the colab?

Are the slides available?

Good explain

In which live stream you have explained this fully ?

Event Slides: www.canva.com/design/DAGOfx2zV48/eUJk97tu-CwHZDtIPtSdbA/view?DAGOfx2zV48&

Insightful video on multi-agent systems. For advanced orchestration, explore SymthOS. #MultiAgentSystems #AI #SymthOS #Innovation

Interesting perspectives on combining several AI agents to solve complicated problems. eager to investigate these frameworks. #Innovation #Technology #MultiAgentSystems #AI

Great video. But one thing I am curious about is why is the input to fine tuning in reverse? I mean the asking the peft_model to generate instruction given response. How does one know a priori that is how the input ought to be preprocessed this way? I am trying to build a peft_model using the same base Mitral-7b, but in my case the data set is "fingpt-sentiment-train". This is a tweet with 5 different classes of sentiments. I am just passing the data set as is (with some pre-processing), i.e., give the tweet and get sentiment. Cheers Ram

Flowise has a langraph no code integration. It’s dope :)

When llm don’t know the answer there is almost zero chance “she” told you about it. just gives you a wrong answer, and this is here interesting things start.

Great working its very important. Congratulations my friends

Thanks again guys!

Is there a student self paced style?

Question - this approach seems different and more advanced than the embedding training session on the llama papers with llama index. Is that one still a valuable approach, or has the world moved on?

Fantastic as usual!