👱 Arc2Face: A Foundation Model of Human Faces

TL; DR: a large dataset of high-resolution facial images, as well as a face generation model trained on its basis, which:

▪capable of creating photorealistic generations in a few seconds
▪provides complete similarity of generations to the target image compared to other existing models
▪built on the basis of Stable Diffusion and can be configured for any generation options, for example, different poses / facial expressions, etc.

▪ Github: https://github.com/foivospar/Arc2Face

▪ Project: https://arc2face.github.io

▪ Demo: https://huggingface.co/spaces/FoivosPar/Arc2Face

▪ Paper: https://arxiv.org/abs/2403.11641

✅ Telegram: https://news.1rj.ru/str/DataScienceT

📁 Recent Advances in Generative Adversarial Networks for Gene Expression Data: A Comprehensive Review

📕 Journal: Mathematics (I.F.=2.4)
🗓 Publish year: 2023

🧑‍💻Authors: Minhyeok Lee
🏢University: Chung-Ang University, Republic of Korea

📎 Study the paper

✅ Telegram: https://news.1rj.ru/str/DataScienceT

⚡️ LLocalSearch: completely locally running meta search engine using LLM Agents

It is a completely local metasearch engine using LLM agents.

The user can ask a question, and the system will use a chain of AI agents to find the answer. The user can see the progress of the work and the final answer. No OpenAI or Google API keys are required.

▪ Github

✅ Telegram: https://news.1rj.ru/str/DataScienceT

🌉 Financial Datasets is an open source Python library that allows developers to create synthetic financial data sets using large language models (LLMs).

With this library, you can generate realistic financial data sets in 5 lines of code, based on SEC reports such as 10-Ks, 10-Qs and other financial reports.

Such datasets are useful for:
• LLM assessments
• LLM fine tuning
• testing of financial instruments
• and much more

The project is completely open source.

pip financial-datasets.

▪ GitHub : https://github.com/virattt/financial-datasets

▪ Example with code: https://colab.research.google.com/gist/virattt/f9b5a0ae82cc0caab57df5dedc2927c9/intro-financial-datasets.ipynb#scrollTo=K-b_1BPtJsS1

✅ Telegram: https://news.1rj.ru/str/DataScienceT

🔥 ECoDepth: SOTA Diffusive Mono-Depth 🔥

🤨 New SIDE model using a diffusion backbone conditioned on ViT embeddings. It's the new SOTA in SIDE. Source Code released 💙

👉 Review: https://t.ly/s2pbB

👉 Paper: https://lnkd.in/eYt5yr_q

😏 Code: https://lnkd.in/eEcyPQcd

✅ Telegram: https://news.1rj.ru/str/DataScienceT

🧑‍🎓 Study of Tensor Network Applications in Complex Networks

📕 Integrated master's thesis in engineering physics

🗓 Publish year: 2022

📎 Study Thesis: https://repositorio.ul.pt/bitstream/10451/57310/1/TM_Francisco_Costa.pdf

✅ Telegram: https://news.1rj.ru/str/DataScienceT

🎥 Camera control for text-to-video.

CameraCtrl is a model that provides precise control of the camera position, which allows you to accurately control camera angles and movements when generating a view.

▪ Github: https://github.com/hehao13/CameraCtrl

▪ Paper: http://arxiv.org/abs/2404.02101

▪ Project: https://hehao13.github.io/projects-CameraCtrl/

▪ Weights: https://huggingface.co/hehao13/CameraCtrl/tree/main

✅ Telegram: https://news.1rj.ru/str/DataScienceT

🔥 RAG From Scratch 🔥

RAG ( Retrieval Augmented Generation ) is a method of working with LLM, in which the user writes his questions, and the developer programmatically supplements information from external sources and submits everything entirely to the input of the language model. In other words, information is added to the language model in the context of the request, based on which the language model can provide the user with a more complete and accurate answer.

This is a huge list of materials that will help you better understand RAG from the ground up, starting with the basics of indexing, searching and generation. The playlist contains short videos (5-10 minutes) and notebooks with code.

📌 Rag from scratch.
▪ Repository:
https://github.com/langchain-ai/rag-from-scratch
▪ Video playlist:
https://youtube.com/playlist?list=PLfaIDFEXuae2LXbO1_PKyVJiQ23ZztA0x&feature=shared

📌 How RAG can change with long context LLMS.
▪ Video: https://youtube.com/watch?v=SsHUNfhF32s

📌 Adaptive Rag
▪ Video:
https://youtu.be/04ighIjMcAI
▪ Code:
https://github.com/langchain-ai/langgraph/blob/main/examples/rag/langgraph_adaptive_rag_cohere.ipynb
▪ Article: https://arxiv.org/abs/2403.14403

📌 Checking the relevance of documents and returning to the search.
▪ Video:
https://youtube.com/watch?v=E2shqsYwxck
▪ Code:
https://github.com/langchain-ai/langgraph/blob/main/examples/rag/langgraph_crag.ipynb
▪ Article: https://arxiv.org/pdf/2401.15884.pdf

📌 Bug fixes in RAG:
▪ Code: https://github.com/langchain-ai/langgraph/blob/main/examples/rag/langgraph_self_rag.ipynb
Article: https://arxiv.org/abs/2310.11511.pdf

📌 Various approaches to direct questions to the right data source:
▪ Video: https://youtu.be/pfpIndq7Fi8
▪ Code: https://github.com/langchain-ai/rag-from-scratch/blob/main/rag_from_scratch_10_and_11.ipynb

📌 Structuring requests
▪ Video: https://youtu.be/kl6NwWYxvbM
▪ Code: https://github.com/langchain-ai/rag-from-scratch/blob/main/rag_from_scratch_10_and_11.ipynb
▪ Blog: https://blog.langchain.dev/query-construction/
2/ Deep dive into graphDBs: https://blog.langchain.dev/enhancing-rag-based-applications-accuracy-by-constructing-and-leveraging-knowledge-graphs/
3/ Query structuring: https://python.langchain.com/docs/use_cases/query_analysis/techniques/structuring
4/ Self-search queries: https://python.langchain.com/docs/modules/data_connection/retrievers/self_query

📌 Multi -Representation Indexing
▪ Video: https://youtu.be/gTCU9I6QqCE
▪ Code: https://github.com/langchain-ai/rag-from-scratch/blob/main/rag_from_scratch_12_to_14.ipynb
▪ Article: https://arxiv.org/pdf/2312.06648.pdf

📌 Grouping documents by similarity.
▪ Video: https://youtu.be/z_6EeA2LDSw
▪ Code: https://github.com/langchain-ai/langchain/blob/master/cookbook/RAPTOR.ipynb
▪ Article: https://arxiv.org/pdf/2401.18059.pdf

📌 ColBERT
▪ Video: https://youtu.be/cN6S0Ehm7_8
▪ Code: https://github.com/langchain-ai/rag-from-scratch/blob/main/rag_from_scratch_12_to_14.ipynb
▪ Article: https://arxiv.org/abs/2004.12832

📌 Query Translation -- Multi Query
▪ Video: https://youtube.com/watch?v=JChPi0CRnDY
▪ Code: https://github.com/langchain-ai/rag-from-scratch/blob/main/rag_from_scratch_5_to_9.ipynb
▪ Article: https://arxiv.org/pdf/2305.14283.pdf

📌 RAG Fusion
▪ Video: https://youtube.com/watch?v=77qELPbNgxA
▪ Code: https://github.com/langchain-ai/rag-from-scratch/blob/main/rag_from_scratch_5_to_9.ipynb
▪ Code: https://github.com/Raudaschl/rag-fusion

📌 Query Translation -- Decomposition
▪ Video: https://youtube.com/watch?v=h0OPWlEOank
▪ Code: https://github.com/langchain-ai/rag-from-scratch/blob/main/rag_from_scratch_5_to_9.ipynb
▪ Articles: https://arxiv.org/pdf/2205.10625.pdf https://arxiv.org/pdf/2212.10509.pdf

📌 Query Translation -- Step Back
▪ Video: https://youtube.com/watch?v=xn1jEjRyJ2U
▪ Code: https://github.com/langchain-ai/rag-from-scratch/blob/main/rag_from_scratch_5_to_9.ipynb
▪ Article: https://arxiv.org/pdf/2310.06117.pdf

✅ Telegram: https://news.1rj.ru/str/DataScienceT

📌 Query Translation -- HyDE
▪ Video:
https://youtube.com/watch?v=SaDzIVkYqyY
▪ Code: https://github.com/langchain-ai/rag-from-scratch/blob/main/rag_from_scratch_5_to_9.ipynb
▪ Article: https://arxiv.org/abs/2212.10496

✅ Telegram: https://news.1rj.ru/str/DataScienceT