https://github.com/celvoxes/thinkr
gpt-o1 like chain of thoughts with local LLMs in R
https://github.com/celvoxes/thinkr
chainofthought cot llm ollama r
Last synced: about 2 months ago
JSON representation
gpt-o1 like chain of thoughts with local LLMs in R
- Host: GitHub
- URL: https://github.com/celvoxes/thinkr
- Owner: CelVoxes
- Created: 2024-09-22T15:44:32.000Z (9 months ago)
- Default Branch: main
- Last Pushed: 2024-10-15T16:07:41.000Z (8 months ago)
- Last Synced: 2025-03-29T08:22:33.953Z (2 months ago)
- Topics: chainofthought, cot, llm, ollama, r
- Language: R
- Homepage:
- Size: 467 KB
- Stars: 32
- Watchers: 1
- Forks: 3
- Open Issues: 0
-
Metadata Files:
- Readme: README.Rmd
Awesome Lists containing this project
README
---
title: "thinkR"
output: github_document
---
thinkR is an R package that enables o-1 like chain of thoughts using ollama.
## Installation
To install ceLLama, use the following command:
```{r eval=FALSE}
devtools::install_github("eonurk/thinkR")
```
## Usage
#### Step 1: Install Ollama
Download [`Ollama`](https://ollama.com/).
#### Step 2: Choose Your Model
Select your preferred model. For instance, to run the Llama3 model, use the following terminal command:
```{bash eval=FALSE}
ollama run llama3.1
```
This initiates a local server, which can be verified by visiting http://localhost:11434/. The page should display "Ollama is running".
#### Step 3: Think!
> Q: How many 'R's are in strawberry?
```{r cache=TRUE, eval=FALSE}
library(thinkR)
## Usage example
ollama <- OllamaHandler$new(model = "llama3.1")
result <- generate_response("How many 'R's are in strawberry?", ollama)
```
Thinking...
```{r echo=FALSE}
result <- readRDS("llama3_1.results.rds")
cat(paste0(
sapply(result$steps, function(m) {
if (!is.null(m$title) && !is.null(m$content) && !is.null(m$thinking_time)) {
sprintf("%s\n%s\nTime: %s s\n\n", m$title, m$content, m$thinking_time)
}
}, USE.NAMES = FALSE)
))
```
```{r echo=FALSE}
total_thinking_time <- result$total_thinking_time
cat(paste0("Total thinking time: ", sprintf("%.2f", total_thinking_time), " s"))
```
> A: Therefore, the final answer is 3!
> [!NOTE]\
> This output is cherry-picked from llama3.1... because 70b was a lot to download for my internet connection.
## Acknowledgments
Cursor is a great tool <3
### Credits
- [g-1](https://github.com/bklieger-groq/g1)
- [multi-1](https://github.com/tcsenpai/multi1)
## License
This project is licensed under the MIT License