info/developer-roadmap
1
0
Fork 0
mirror of https://github.com/kamranahmedse/developer-roadmap.git synced 2025-08-13 20:54:16 +02:00
Code Issues Releases Wiki Activity

Rearrange visual

Browse Source
This commit is contained in:
Kamran Ahmed
2023-05-16 16:29:55 +01:00
parent 626026eebc
commit 4b76d0b7aa
1 changed files with 3 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/data/guides/introduction-to-llms.md
View File

@@ -26,6 +26,9 @@ Everyone these days is talking about LLMs, ChatGPT and what not. GitHub's [trend
LLM stands for "Large Language Model." These are advanced AI systems designed to understand and generate human-like text based on the input they receive. These models have been trained on vast amounts of text data and can perform a wide range of language-related tasks, such as answering questions, carrying out conversations, summarizing text, translating languages, and much more. LLM stands for "Large Language Model." These are advanced AI systems designed to understand and generate human-like text based on the input they receive. These models have been trained on vast amounts of text data and can perform a wide range of language-related tasks, such as answering questions, carrying out conversations, summarizing text, translating languages, and much more.
[![LLMS Visualized](/guides/llms.png)](https://twitter.com/kamrify/status/1658271217189634049)
OpenAI has been a major contributor to this space in the past few years with their models and research. However, there are other players in the market as well e.g. Meta with their [OPT](https://huggingface.co/facebook/opt-66b), [OPT-IML](https://huggingface.co/facebook/opt-iml-30b) and [LLaMA](https://ai.facebook.com/blog/large-language-model-llama-meta-ai/) models, Google released [FLAN-T5](https://huggingface.co/google/flan-t5-xxl) and [BERT](https://huggingface.co/bert-base-uncased), [StableLM](https://github.com/stability-AI/stableLM/) by Stability AI, [Alpaca](https://crfm.stanford.edu/2023/03/13/alpaca.html) at Stanford and there are many [other opensource models as well](https://github.com/Hannibal046/Awesome-LLM). OpenAI has been a major contributor to this space in the past few years with their models and research. However, there are other players in the market as well e.g. Meta with their [OPT](https://huggingface.co/facebook/opt-66b), [OPT-IML](https://huggingface.co/facebook/opt-iml-30b) and [LLaMA](https://ai.facebook.com/blog/large-language-model-llama-meta-ai/) models, Google released [FLAN-T5](https://huggingface.co/google/flan-t5-xxl) and [BERT](https://huggingface.co/bert-base-uncased), [StableLM](https://github.com/stability-AI/stableLM/) by Stability AI, [Alpaca](https://crfm.stanford.edu/2023/03/13/alpaca.html) at Stanford and there are many [other opensource models as well](https://github.com/Hannibal046/Awesome-LLM).
## Training an LLM Model ## Training an LLM Model
@@ -60,7 +63,6 @@ Instruction Tuned LLMs = Base LLMs + Further Tuning + RLHF
To build an Instruction Tuned LLM, a Base LLM is taken and is further trained using a large dataset covering sample "Instructions" and how the model should perform as a result of those instructions. The model is then fine-tuned using a technique called "Reinforcement Learning with Human Feedback" (RLHF) which allows the model to learn from human feedback and improve its performance over time. To build an Instruction Tuned LLM, a Base LLM is taken and is further trained using a large dataset covering sample "Instructions" and how the model should perform as a result of those instructions. The model is then fine-tuned using a technique called "Reinforcement Learning with Human Feedback" (RLHF) which allows the model to learn from human feedback and improve its performance over time.
[![LLMS Visualized](/guides/llms.png)](https://twitter.com/kamrify/status/1658271217189634049)
## Conclusion ## Conclusion