Merge pull request #115 from premAI-io/fix-links

_config.yml

@@ -139,6 +139,7 @@ sphinx:
     - https://www.crcv.ucf.edu/data/UCF101.php
     - https://www.pinecone.io/learn/series/faiss/locality-sensitive-hashing
     - https://chat.lmsys.org
+    - https://platform.openai.com/docs/.*
     linkcheck_allowed_redirects:
       https://doi.org/.*/.*: https://.*
       https://codespaces.new/.*: https://github.com/login.*

index.md

@@ -225,7 +225,7 @@ LLM
   A [Large Language Model](https://en.wikipedia.org/wiki/Large_language_model) is neural network (often a {term}`transformer` containing billions of parameters) designed to perform tasks in natural language via [fine tuning](<https://en.wikipedia.org/wiki/Fine-tuning_(machine_learning)>) or [prompt engineering](https://en.wikipedia.org/wiki/Prompt_engineering).
 
 MLOps
-  [Machine Learning Operations](https://blogs.nvidia.com/blog/2020/09/03/what-is-mlops): best practices to run AI using software products & cloud services
+  [Machine Learning Operations](https://blogs.nvidia.com/blog/what-is-mlops): best practices to run AI using software products & cloud services
 
 MoE
   [Mixture-of-Experts](https://en.wikipedia.org/wiki/Mixture_of_experts) is a technique which uses one or more specialist model(s) from a collection of models ("experts") to solve general problems. Not that this is different from [ensemble](https://en.wikipedia.org/wiki/Ensemble_learning) models (which combine results from all models).

mlops-engines.md

@@ -84,7 +84,7 @@ Cons:
 
 ## vLLM
 
-This is an open-source project created by researchers at Berkeley to improve the performance of LLM inferencing. [vLLM](https://vllm.ai) primarily optimises LLM throughput via methods like PagedAttention and Continuous Batching. The project is fairly new and there is ongoing development.
+This is an open-source project created by researchers at Berkeley to improve the performance of LLM inferencing. https://github.com/vllm-project/vllm primarily optimises LLM throughput via methods like PagedAttention and Continuous Batching. The project is fairly new and there is ongoing development.
 
 Pros:
 

models.md

@@ -7,7 +7,7 @@
 Some ideas:
 
 - [The History of Open-Source LLMs: Better Base Models (part 2)](https://cameronrwolfe.substack.com/p/the-history-of-open-source-llms-better) (LLaMA, MPT, Falcon, LLaMA-2)
-- [Papers I've read this week, Mixture of Experts edition](https://finbarrtimbers.substack.com/p/papers-ive-read-this-week-mixture) (conditional routing models)
+- [Papers I've read this week, Mixture of Experts edition](https://www.artfintel.com/p/papers-ive-read-this-week-mixture) (conditional routing models)
 - [AI and Memory Wall](https://medium.com/riselab/ai-and-memory-wall-2cb4265cb0b8)
 - https://github.com/imaurer/awesome-decentralized-llm
 - https://github.com/huggingface/transformers/blob/main/awesome-transformers.md
@@ -45,7 +45,7 @@ ChatGPT is sensitive to tweaks to the input phrasing or attempting the same prom
 
 #### GPT-4
 
-[GPT-4 is a language model developed by OpenAI](https://openai.com/research/gpt-4). It is the successor to GPT-3 and has been made publicly available via the paid chatbot product ChatGPT Plus and via OpenAI's API. It is a large multimodal model that can accept image and text inputs and emit text outputs, [though multimodal capabilities aren't released to the public yet](https://analyticsindiamag.com/what-happened-to-multimodal-gpt-4). It exhibits human-level performance on various professional and academic benchmarks and can follow complex instructions in natural language and solve difficult problems with accuracy. It can handle input prompts of up to 32k tokens, which is a significant increase from GPT-3.5's 4k tokens. It can solve complex mathematical and scientific problems beyond the capabilities of GPT-3.5, such as advanced calculus problems or simulating chemical reactions [more effectively than its predecessor](https://www.searchenginejournal.com/gpt-4-vs-gpt-3-5/482463). It is more reliable, creative, and able to handle much more nuanced instructions than GPT-3.5.
+[GPT-4 is a language model developed by OpenAI](https://openai.com/research/gpt-4). It is the successor to GPT-3 and has been made publicly available via the paid chatbot product ChatGPT Plus and via OpenAI's API. It is a large multimodal model that can accept image and text inputs and emit text outputs, [though multimodal capabilities aren't released to the public yet](http://analyticsindiamag.com/what-happened-to-multimodal-gpt-4/). It exhibits human-level performance on various professional and academic benchmarks and can follow complex instructions in natural language and solve difficult problems with accuracy. It can handle input prompts of up to 32k tokens, which is a significant increase from GPT-3.5's 4k tokens. It can solve complex mathematical and scientific problems beyond the capabilities of GPT-3.5, such as advanced calculus problems or simulating chemical reactions [more effectively than its predecessor](https://www.searchenginejournal.com/gpt-4-vs-gpt-3-5/482463). It is more reliable, creative, and able to handle much more nuanced instructions than GPT-3.5.
 
 Despite its capabilities, [GPT-4 still sometimes "hallucinates"](https://www.reddit.com/r/ChatGPT/comments/12fmrcd/examples_of_gpt4_hallucination) facts and makes reasoning errors.
 
@@ -111,7 +111,7 @@ Before [](#chatgpt)'s (GPT-3.5) public release we had [GPT-3](https://en.wikiped
 There has been a few visible marks across modalities of AI models, highly catalysing growth of open source:
 
 - [Meta AI launches LLaMA](https://ai.meta.com/blog/large-language-model-llama-meta-ai), open sourcing the code but not the weights.
-- [StabilityAI released Stable Diffusion](https://stability.ai/blog/stable-diffusion-announcement).
+- [StabilityAI released Stable Diffusion](https://stability.ai/news/stable-diffusion-announcement).
 
 #### [Stable Diffusion](https://registry.premai.io/detail.html?service=stable-diffusion-1-5)
 
@@ -371,15 +371,15 @@ Till now we've mostly been looking at LLMs in general and not other models, let'
 
 #### [Stable Diffusion XL](https://registry.premai.io/detail.html?service=stable-diffusion-xl-with-refiner)
 
-[StabilityAI released Stable Diffusion XL 1.0 (SDXL)](https://stability.ai/blog/stable-diffusion-sdxl-1-announcement) models on 26th July, being current State of the Art for text-to-image and image-to-image generation open sourced models. They released a [base model](https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0) and a [refinement model](https://huggingface.co/stabilityai/stable-diffusion-xl-refiner-1.0) which is used to improve the visual fidelity of samples generated by SDXL.
+[StabilityAI released Stable Diffusion XL 1.0 (SDXL)](https://stability.ai/news/stable-diffusion-sdxl-1-announcement) models on 26th July, being current State of the Art for text-to-image and image-to-image generation open sourced models. They released a [base model](https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0) and a [refinement model](https://huggingface.co/stabilityai/stable-diffusion-xl-refiner-1.0) which is used to improve the visual fidelity of samples generated by SDXL.
 
 Few months back they released Stable-diffusion-xl {cite}`podell2023sdxl` [base](https://huggingface.co/stabilityai/stable-diffusion-xl-base-0.9) and [refinement](https://huggingface.co/stabilityai/stable-diffusion-xl-refiner-0.9) models versioned as 0.9, where license permitting only research purpose usages.
 
 SDXL consistently surpasses all previous versions of Stable Diffusion models by a significant margin:
 
 ```{figure} https://static.premai.io/book/models_sdxl-winrate.png
 :width: 60%
-[SDXL Winrate](https://stability.ai/blog/stable-diffusion-sdxl-1-announcement)
+[SDXL Winrate](https://stability.ai/news/stable-diffusion-sdxl-1-announcement)
 ```
 
 ##### Uniqueness

references.md

@@ -7,7 +7,6 @@
 
 - "Catching up on the weird world of LLMs" (summary of the last few years) https://simonwillison.net/2023/Aug/3/weird-world-of-llms
 - "Open challenges in LLM research" (exciting post title but mediocre content) https://huyenchip.com/2023/08/16/llm-research-open-challenges.html
-- "AI forest" https://www.michaeldempsey.me/blog/2023/07/18/the-dark-forest-of-rd-and-capital-deployment-in-ai
 - https://github.com/zeno-ml/zeno-build/tree/main/examples/analysis_gpt_mt/report
 - "Patterns for Building LLM-based Systems & Products" (Evals, RAG, fine-tuning, caching, guardrails, defensive UX, and collecting user feedback) https://eugeneyan.com/writing/llm-patterns
 

sdk.md

@@ -46,7 +46,7 @@ The list of vector stores that LangChain supports can be found [here](https://ap
 
 ### Models
 
-This is the heart of most LLMs, where the core functionality resides. There are broadly [2 different types of models](https://python.langchain.com/docs/modules/model_io/models) which LangChain integrates with:
+This is the heart of most LLMs, where the core functionality resides. There are broadly [2 different types of models](https://python.langchain.com/docs/modules/model_io) which LangChain integrates with:
 
 - **Language**: Inputs & outputs are `string`s
 - **Chat**: Run on top of a Language model. Inputs are a list of chat messages, and output is a chat message