diff --git a/blog.mdx b/blog.mdx index 800deb1..8a95a45 100644 --- a/blog.mdx +++ b/blog.mdx @@ -5,7 +5,7 @@ description: "Posts about the great work we are doing at Anarchy" ## Posts -- **[Guided Distillation](https://anarchy.ai/blog/guided_distillation)** Read about how we finetuned a Pythia model to constrain the output space of a 70 million parameter Pythia model! +- **[Guided Distillation](https://anarchy.ai/blog/guided_distillation)** Read about how we fine-tuned a Pythia model to constrain the output space of a 70 million-parameter Pythia model! - **[The REBEL Agent](https://anarchy.ai/blog/rebel)** Read about the REBEL agent, a tool that empowers LLMs to use external tools to solve compositional questions! diff --git a/blog/guided_distillation.mdx b/blog/guided_distillation.mdx index f1af950..c7ca1af 100644 --- a/blog/guided_distillation.mdx +++ b/blog/guided_distillation.mdx @@ -3,7 +3,7 @@ title: 🎻 Guided Distillation description: 'Finetune a Pythia 70M model to only output within a ' --- -The paper Efficient Guided Generation for Large Language Models provides a state machine based approach to constrain the output of a Large Language Model. To use the code from this paper, we used the Outlines python library. Using this library we were able to add three types of guiding to LLM generation on our LLM-VM repositiory +The paper Efficient Guided Generation for Large Language Models provides a state machine-based approach to constrain the output of a Large Language Model. To use the code from this paper, we used the Outlines Python library. Using this library we were able to add three types of guiding to LLM generation on our LLM-VM repository. - **Regex** - The output of the underlying LLM will follow the inputted regex. - **Type** - The output of the underlying LLM will be of the inputted type. @@ -14,7 +14,7 @@ The paper Efficient Guided Generation for Large Language Models provides a state from llm_vm.client import Client import os -# Instantiate the client specifying which LLM you want to use. Gudied generation always uses an open source version of GPT2. +# Instantiate the client specifying which LLM you want to use. Guided generation always uses an open source version of GPT2. client = Client() # Regex-based guided generation @@ -33,12 +33,12 @@ response = client.complete(prompt = 'How many presidents has the USA had?', print(response) ``` -In this blog we discuss how we distilled the GPT2 model down to a 70M parameter Pythia model and were able to limit the constrain the output space of the Pythia model. +In this blog, we discuss how we distilled the GPT2 model down to a 70M parameter Pythia model and were able to limit constrain of the output space of the Pythia model. -Saving finetuned LLMs can result in a large amount of storage space. Make sure to keep an eye on how much you're saving +Saving fine-tuned LLMs can result in a large amount of storage space. Make sure to keep an eye on how much you're saving ## References https://arxiv.org/pdf/2307.09702v4.pdf diff --git a/contributing/community_development.mdx b/contributing/community_development.mdx index 774625f..985dc46 100644 --- a/contributing/community_development.mdx +++ b/contributing/community_development.mdx @@ -11,7 +11,7 @@ The first and most importantly step is to [join our discord](https://discord.com Our aim is for this community to be not just for organization, but a teaching community. Don't be afraid to speak up with questions, and don't shame people for asking them! -There are a mix of experts and n00bs there. +There is a mix of experts and n00bs there. ### Becoming an official contributor! @@ -28,6 +28,7 @@ We reserve the right to revoke contributor status at any time for any reason (pl ## Bounty Program Some tickets we'll pay you for closing! Look at the [ticket](https://github.com/anarchy-ai/LLM-VM/issues) labels to see how much the bounty is. + Note we can't guarantee that if you close a ticket we can pay you (only if you are from a non-US sanctioned country), and it is up to you to reach out and be annoying about this. To get started, it is essential to read the [full guide in discord](https://discord.com/channels/1075227138766147656/1147542772824408074) diff --git a/contributing/getting_involved.mdx b/contributing/getting_involved.mdx index b75edad..7840848 100644 --- a/contributing/getting_involved.mdx +++ b/contributing/getting_involved.mdx @@ -1,11 +1,11 @@ --- title: "🔥 Become an Anarchist" -description: "Excited to develop new democratized AGI? Us too!" +description: "Excited to develop a new democratized AGI? Us too!" --- -This project is at it's very earliest stages, but we still welcome support! -We have a number of ways to get involved, from [working for us](jobs), to working on [tickets directly](community_development). +This project is at its very earliest stages, but we still welcome support! +We have a number of ways to get involved, from [working for us](jobs) to working on [tickets directly](community_development). -By far the easiest way to get a job is to work on tickets directly and become a top contributor and get noticed! +By far the easiest way to get a job is to work on tickets directly. Become a top contributor and get noticed! -As we want to encourage open-source AGI development, we have a bounty program to pay for open development \ No newline at end of file +As we want to encourage open-source AGI development, we have a bounty program to pay for open development. diff --git a/contributing/jobs.mdx b/contributing/jobs.mdx index b98733b..032ea85 100644 --- a/contributing/jobs.mdx +++ b/contributing/jobs.mdx @@ -1,6 +1,6 @@ --- title: "👷 Work for us" -description: "Here we describe the process for getting a job and where to find open-jobs" +description: "Here we describe the process for getting a job and where to find open jobs" --- Right now we're very early stage so we have limited jobs available. @@ -19,5 +19,5 @@ Nevertheless, we're always on the lookout for exceptional talent. * Compilers and GPU/driver engineers * DevRel Profiles -We welcome applicants with non traditional backgrounds who strongly believe in the mission and want -to assist with marketing and sales. \ No newline at end of file +We welcome applicants with non-traditional backgrounds who strongly believe in the mission and want +to assist with marketing and sales. diff --git a/get_started/install.mdx b/get_started/install.mdx index 805f5a5..8a6d577 100644 --- a/get_started/install.mdx +++ b/get_started/install.mdx @@ -1,19 +1,19 @@ --- title: 📀 Install -description: "You can install everything you need easily from the command line. Make sure check the [system requirements](system) and that you have [Python](https://www.python.org/downloads/release/python-3114/) installed." +description: "You can install everything you need easily from the command line. Make sure to check the [system requirements](system) and that you have [Python](https://www.python.org/downloads/release/python-3114/) installed." --- ## Installation **Use git and pip to quickly install the LLM-VM** ### Get the code -You can download the code from our [Github](https://github.com/anarchy-ai/LLM-VM/archive/refs/heads/main.zip) repo manually, or navigate a CLI to your perferred directory and run: +You can download the code from our [Github](https://github.com/anarchy-ai/LLM-VM/archive/refs/heads/main.zip) repo manually, or navigate a CLI to your preferred directory and run: ```bash CLI Install git clone https://github.com/anarchy-ai/LLM-VM.git ``` ### Install the module -Once you've downloaded the code, getting it to work with python is easy, open a CLI and +Once you've downloaded the code, getting it to work with Python is easy, open a CLI and ```bash Standard Install cd LLM-VM @@ -27,7 +27,7 @@ pip3 install --editable . This will install both the LLM-VM and the HTTP server. -Installing with the ```--editable``` flag allows your changes to the codebase take effect without rebuilding with ```pip3 install .``` +Installing with the ```--editable``` flag allows your changes to the codebase to take effect without rebuilding with ```pip3 install .``` - \ No newline at end of file + diff --git a/get_started/quickstart/agents.mdx b/get_started/quickstart/agents.mdx index b3477c4..7923f4e 100644 --- a/get_started/quickstart/agents.mdx +++ b/get_started/quickstart/agents.mdx @@ -6,7 +6,7 @@ description: 'Our AI agents expand what you can do with LLMs!' ## The REBEL Agent **REcursion Based Extensible LLM** -Our REBEL agent takes a novel approach to answering complex questions. Using recursive reasoning, REBEL expands what LLMs can do with problem decomposition and tool use. In this way, we are able to answer questions requiring data LLMs were not directly trained on. +Our REBEL agent takes a novel approach to answering complex questions. Using recursive reasoning, REBEL expands on what LLMs can do with problem decomposition and tool use. In this way, we are able to answer questions requiring data that LLMs were not directly trained on. ### Running REBEL **Getting started with REBEL is easy** @@ -23,7 +23,7 @@ response = client.complete( prompt = 'Is it warmer in Paris or Timbuktu and what are the temperatures in either city?', context='', openai_key=os.getenv("OPENAI_API_KEY"), #for REBEL we need an OpenAI key - tools=[{'description': 'Find the weather at a location and returns it in celcius.', #this tool list contains only one dictionary, therefore only one tool + tools=[{'description': 'Find the weather at a location and returns it in celsius.', #this tool list contains only one dictionary, therefore only one tool 'dynamic_params': { "latitude": 'latitude of as a float', "longitude": 'the longitude as a float' @@ -39,7 +39,7 @@ Tools are defined by dictionaries that are added to the list ```tools```. These |Field| Type | Description| |-|-|-| |```'description'```| string | A description of what the tool does| -|```'dynamic_params'```| dictionary | A dictionary containing key value pairs (paramter name : description) of the API endpoint's mutable parameters that need to be set by REBEL in order to answer a query| +|```'dynamic_params'```| dictionary | A dictionary containing key-value pairs (parameter name : description) of the API endpoint's mutable parameters that need to be set by REBEL in order to answer a query| |```'method'```| string | ```GET``` or ```POST```, whichever is the type of the API endpoint| |```'url'```| string | URL of the API endpoint that the given tool specifies| |```'static_params'```| dictionary | Any parameters that are constant between all API calls. An API key/token is an example of this| diff --git a/get_started/quickstart/completions.mdx b/get_started/quickstart/completions.mdx index 48001de..d283b6f 100644 --- a/get_started/quickstart/completions.mdx +++ b/get_started/quickstart/completions.mdx @@ -42,7 +42,7 @@ print(response) -Using OpenAI's models require an OpenAI API Key and may result in costs not associated with Anarchy's LLM-VM +Using OpenAI's models requires an OpenAI API Key and may result in costs not associated with Anarchy's LLM-VM ## Supported Models We support several open LLM model families. You can see which ones and the default models used below. @@ -53,4 +53,4 @@ Just replace ```'chat_gpt'``` in the example with your desired model and you're For more information on selecting models visit our [Local LLMs](local_llms) section. - \ No newline at end of file + diff --git a/get_started/quickstart/finetuning.mdx b/get_started/quickstart/finetuning.mdx index bce9782..d0799b0 100644 --- a/get_started/quickstart/finetuning.mdx +++ b/get_started/quickstart/finetuning.mdx @@ -1,15 +1,15 @@ --- -title: 🎻 Finetuning -description: 'Here we describe how to finetune LLMs and load previously trained models' +title: 🎻 Fine-tuning +description: 'Here we describe how to fine-tune LLMs and load previously trained models' --- -## Finetuning Models -The LLM-VM uses a data synthesis feature to engage a teacher/student model for finetuning super specific small models from larger more general ones through a simple process. +## Fine-tuning Models +The LLM-VM uses a data synthesis feature to engage a teacher/student model for fine-tuning super-specific small models from larger more general ones through a simple process. 1. A scalable call is made to OpenAI (or other large parameter LLM) - 'give me fifty examples like the following query....' -2. The response is parsed into a training set for finetuning -3. The ```small_model``` is trained on the synthezied data set and saved +2. The response is parsed into a training set for fine-tuning +3. The ```small_model``` is trained on the synthesized data set and saved ```python Finetuning a Model # import our client @@ -28,11 +28,11 @@ response=client.complete(prompt="Answer question Q. ", finetune=True,) ``` -### Loading a finetuned LLM from disk -Loading a finetuned model into the LLM-VM for completions is very straightforward. +### Loading a fine-tuned LLM from disk +Loading a fine-tuned model into the LLM-VM for completions is very straightforward. -1. Specify the family of finetuned models in "big_model" -2. Specify the specific filename of the finetuned model you wish to load +1. Specify the family of fine-tuned models in "big_model" +2. Specify the specific filename of the fine-tuned model you wish to load 3. Run the ```load_finetune(model_filename)``` class function to load the model 4. Generate completions in the usual way ```python Loading a Finetuned Model @@ -43,7 +43,7 @@ import os # Instantiate the client specifying which LLM you want to use client=Client(big_model='pythia') -# specify the file name of the finetuned model to load +# Specify the file name of the finetuned model to load model_name='.pt' client.load_finetune(model_name) @@ -56,4 +56,4 @@ print(response) Saving finetuned LLMs can result in a large amount of storage space. Make sure to keep an eye on how much you're saving - \ No newline at end of file + diff --git a/get_started/quickstart/local_llms.mdx b/get_started/quickstart/local_llms.mdx index a6219ff..d34f44f 100644 --- a/get_started/quickstart/local_llms.mdx +++ b/get_started/quickstart/local_llms.mdx @@ -3,9 +3,9 @@ title: 🧰 Local LLMs description: 'Here we explain how to choose non-default LLMs for use in the LLM-VM' --- -The LLM-VM suports several default models intended to make experimentation with LLMs accessible to everyone, but if you have the memory required, larger parameter models will perform far better! +The LLM-VM supports several default models intended to make experimentation with LLMs accessible to everyone, but if you have the memory required, larger parameter models will perform far better! -Here is an example if you want to use a large and small neo model for your teacher and student, and you have enough ram: +Here is an example if you want to use a large and small neo model for your teacher and student, and you have enough RAM: ```python Loading Non-default models # import our client from llm_vm.client import Client @@ -22,6 +22,6 @@ response = client.complete(prompt = 'What is Anarchy?', context = '') print(response) # Anarchy is a political philosophy that advocates no government... ``` -Now, keep in mind the [gpt-neox-20b](https://huggingface.co/EleutherAI/gpt-neox-20b) is almost 42 GB in size, and would require at least as much RAM to use, in addition to the [gpt-neo-125m](https://huggingface.co/EleutherAI/gpt-neo-125m) which is another ~ 0.5 GB. +Now, keep in mind the [gpt-neox-20b](https://huggingface.co/EleutherAI/gpt-neox-20b) is almost 42 GB in size and would require at least as much RAM to use, in addition to the [gpt-neo-125m](https://huggingface.co/EleutherAI/gpt-neo-125m) which is another ~ 0.5 GB. - \ No newline at end of file + diff --git a/get_started/quickstart/quickstart.mdx b/get_started/quickstart/quickstart.mdx index 9148121..d3803c1 100644 --- a/get_started/quickstart/quickstart.mdx +++ b/get_started/quickstart/quickstart.mdx @@ -10,7 +10,7 @@ description: 'Our LLM-VM lets get going at your speed with the completion endpoi # import our client from llm_vm.client import Client -# Select which LLM you want to use, here we have openAI's +# Select which LLM you want to use, here we have OpenAI's client=Client(big_model='chat_gpt') # Put in your prompt and go! @@ -20,7 +20,7 @@ response=client.complete(prompt='What is Anarchy?', print(response) # Anarchy is a political ideology that advocates for the absence of government... ``` -Our Client allows you to interact with OpenAI's completion endpoint seemlessly. +Our Client allows you to interact with OpenAI's completion endpoint seamlessly. ## Locally Run an LLM @@ -72,9 +72,9 @@ print(response) # Anarchy is a political philosophy that advocates no government... ``` -## Finetuning -### Finetuning an LLM -**Finetune intelligently with LLM-VM** +## Fine-tuning +### Fine-tuning an LLM +**Fine-tune intelligently with LLM-VM** Our strategy pairs a small LLM with a larger LLM in a student/teacher relationship. @@ -97,9 +97,9 @@ response=client.complete(prompt="Answer question Q. ", ``` ### Loading a finetuned LLM from disk -**Quickly test and deploy your finetuned models with LLM-VM** +**Quickly test and deploy your fine-tuned models with LLM-VM** -Just specify the parent model, your finetuned filename, and you're almost there! +Just specify the parent model and your finetuned filename, and you're almost there! ```python # import our client from llm_vm.client import Client @@ -118,4 +118,4 @@ response=client.complete(prompt='What is anarchy?', print(response) ``` - \ No newline at end of file + diff --git a/get_started/quickstart/server.mdx b/get_started/quickstart/server.mdx index a832760..49ec253 100644 --- a/get_started/quickstart/server.mdx +++ b/get_started/quickstart/server.mdx @@ -6,12 +6,12 @@ description: 'Here you can find instructions on setting up our HTTP endpoint for ## LLM-VM Server ### Get up and running -All the functionality described in previous sections can be accessed almost entirely through an API call. First start ther server with the following command. +All the functionality described in previous sections can be accessed almost entirely through an API call. First start the server with the following command. ```bash llm_vm_server ``` -This will spin up the flask server with settings specifed in your ```settings.toml``` file! +This will spin up the flask server with settings specified in your ```settings.toml``` file! ### Specifying your models **Save your development environment in a ```setting.toml``` file** @@ -23,4 +23,4 @@ SMALL_MODEL = "bloom" HOST = "127.0.0.1" ``` The server pulls from this file to determine which LLMs will be used, allowing for reliable behavior every server launch. - \ No newline at end of file + diff --git a/get_started/system.mdx b/get_started/system.mdx index cb7420b..2349809 100644 --- a/get_started/system.mdx +++ b/get_started/system.mdx @@ -35,9 +35,9 @@ pip3 install . pip3 install --editable . ``` -Installing with the ```--editable``` flag allows your changes to the codebase take effect without rebuilding with ```pip3 install .``` +Installing with the ```--editable``` flag allows your changes to the codebase to take effect without rebuilding with ```pip3 install .``` ### Supported Models -Which model you choose to use will be the biggest impact on your system's requirements. +Which model you choose to use will have the biggest impact on your system's requirements. LLM-VM comes with support for some default models intended to get people experimenting with LLMs. You can see their specifications here. @@ -49,4 +49,4 @@ We recommend making sure you have enough RAM to load the models. Our section on [Local LLMs](quickstart/local_llms) has more information on specifying LLM models. - \ No newline at end of file + diff --git a/reference/endpoint/create.mdx b/reference/endpoint/create.mdx index 7e7f86b..64be10a 100644 --- a/reference/endpoint/create.mdx +++ b/reference/endpoint/create.mdx @@ -25,7 +25,7 @@ The contents of the user group This is the internal ID for this user group. You don't need to record this - information, since you will not need to use it. + information since you will not need to use it. @@ -44,10 +44,10 @@ The contents of the user group This is the environment tag of the user group. Possible values are 'Customer' - and 'Testing'. User group id's must be unique to each environment, so you can - not create multiple user groups with with same id. If you have a production - customer and a test user group with the same id, you will be required to label - one as 'Customer' and another as 'Testing' + and 'Testing'. User group IDs must be unique to each environment, so you can + not create multiple user groups with the same ID. If you have a production + customer and a test user group with the same ID you will be required to label + one as 'Customer' and another as 'Testing'. diff --git a/reference/endpoint/get.mdx b/reference/endpoint/get.mdx index ce95f65..07c326c 100644 --- a/reference/endpoint/get.mdx +++ b/reference/endpoint/get.mdx @@ -15,8 +15,8 @@ description: "This endpoint gets or creates a new user." - This is a JSON mapping of schema id to either the data source that this user group should be - associated with or id of the datasource you provided when creating it. + This is a JSON mapping of schema ID to either the data source that this user group should be + associated with or the ID of the data source you provided when creating it. @@ -41,7 +41,7 @@ The contents of the user group - This is the internal ID for this user group. You don't need to record this information, since + This is the internal ID for this user group. You don't need to record this information since you will not need to use it. diff --git a/reference/endpoint/update.mdx b/reference/endpoint/update.mdx index 4304987..73406ac 100644 --- a/reference/endpoint/update.mdx +++ b/reference/endpoint/update.mdx @@ -15,8 +15,8 @@ description: "This endpoint updates an existing user." - This is a JSON mapping of schema id to either the data source that this user - group should be associated with or id of the datasource you provided when + This is a JSON mapping of schema ID to either the data source that this user + group should be associated with or the ID of the data source you provided when creating it. diff --git a/reference/getting_started.mdx b/reference/getting_started.mdx index 2b1f62e..a9b2725 100644 --- a/reference/getting_started.mdx +++ b/reference/getting_started.mdx @@ -16,14 +16,14 @@ To install the LLM-VM you simply need to download this repository and install it This will install both the library and test-server. ### Generating Completions -Our LLM-VM gets you working directly with popular LLMs locally in just 3 lines. Once you've installed (as above), just load your model and start generating! +Our LLM-VM gets you working directly with popular LLMs locally in just 3 lines. Once you've installed it (as above), just load your model and start generating! ```python # import our client from llm_vm.client import Client -# Select which LLM you want to use, here we have openAI's +# Select which LLM you want to use, here we have OpenAI's client = Client(big_model = 'chat_gpt') # Put in your prompt and go! diff --git a/welcome/contact.mdx b/welcome/contact.mdx index 4c3b68f..03cc3de 100644 --- a/welcome/contact.mdx +++ b/welcome/contact.mdx @@ -9,5 +9,5 @@ Hop on our [discord](https://discord.gg/qaFf7S373c)! ## Customer -If you are currently using, or would like to use our product in production and are willing to pay for either service or hosting, +If you are currently using or would like to use our product in production and are willing to pay for either service or hosting, feel free to [book a meeting with us directly](https://calendly.com/anarchy-ai/customer-call) diff --git a/welcome/llm_vm.mdx b/welcome/llm_vm.mdx index 1cf5d94..6adb3a3 100644 --- a/welcome/llm_vm.mdx +++ b/welcome/llm_vm.mdx @@ -1,16 +1,16 @@ --- title: '💁‍♂️ What is an LLM-VM?' -description: 'The Large Language Model Virtual Machine(LLM-VM) is a optimized backend for managing LLMs. It serves to facilitate communication and coordination between data, language models(running on the CPU), prompts, and various tools to streamline AGI development.' +description: 'The Large Language Model Virtual Machine(LLM-VM) is an optimized backend for managing LLMs. It serves to facilitate communication and coordination between data, language models(running on the CPU), prompts, and various tools to streamline AGI development.' --- ## Features of the LLM-VM - **Implicit Agents** - The Anarchy LLM-VM can be set up to use external tools through our agents such as [REBEL](../get_started/quickstart/agents) just by supplying tool descriptions! -- **Inference Optimization** - The Anarchy LLM-VM is optimized from agent level all the way to assembly on known LLM architectures to get the most bang for your buck. With state of the art batching, sparse inference and quantization, distillation, and multi-level colocation, we aim to provide the fastest framework available. +- **Inference Optimization** - The Anarchy LLM-VM is optimized from agent level all the way to assembly on known LLM architectures to get the most bang for your buck. With state-of-the-art batching, sparse inference and quantization, distillation, and multi-level colocation, we aim to provide the fastest framework available. -- **Task Auto-Optimization** - The Anarchy LLM-VM will analyze your use cases for repetitive tasks where it can activate student-teacher distillation to train a super-efficient small model from a larger more general model without loosing accuracy. It can furthermore take advantage of data-synthesis techniques to improve results. +- **Task Auto-Optimization** - The Anarchy LLM-VM will analyze your use cases for repetitive tasks where it can activate student-teacher distillation to train a super-efficient small model from a larger more general model without losing accuracy. It can furthermore take advantage of data-synthesis techniques to improve results. -- **Library Callable** - Our library can be used from any python codebase directly. +- **Library Callable** - Our library can be used from any Python codebase directly. - **HTTP Endpoint** - We've provided an HTTP standalone server to handle completion requests via a convenient API. @@ -21,12 +21,12 @@ description: 'The Large Language Model Virtual Machine(LLM-VM) is a optimized ba - **Lower costs** - Running models locally can reduce the pay-as-you-go costs of development and testing. -- **Flexibility** - Anarchy allows you to rapidly switch between popular models so you can pinpoint the exact right tool for your project. +- **Flexibility** - Anarchy allows you to rapidly switch between popular models so you can pinpoint the right tool for your project. -- **Community** - Join our active community of highly motivated developers and engineers working passionately to democratize AGI +- **Community** - Join our active community of highly motivated developers and engineers working passionately to democratize AGI. ## Goals -The LLM-VM wants to centralize and optimize the functionalities of modern completion endpoints in an opinionated way, allowing for the efficient batching of calls that might otherwise be extremely costly across multiple endpoints +The LLM-VM wants to centralize and optimize the functionalities of modern completion endpoints in an opinionated way, allowing for the efficient batching of calls that might otherwise be extremely costly across multiple endpoints. -We want to make the LLM-VM model and architecture agnostic. We want to create a backend that gives you an optimized solution regardless of which model you choose, and which architecture and hardware solution you choose to run it on +We want to make the LLM-VM model and architecture agnostic. We want to create a backend that gives you an optimized solution regardless of which model you choose, and which architecture and hardware solution you choose to run it on. diff --git a/welcome/roadmap.mdx b/welcome/roadmap.mdx index 9754213..9c1c7a3 100644 --- a/welcome/roadmap.mdx +++ b/welcome/roadmap.mdx @@ -7,12 +7,12 @@ description: "This project is in BETA and developing rapidly. We're eager to get **We're actively working on delivering these features** -- **Live Data Augmentation** - You will be able to provide a live updating data-set and the Anarchy LLM-VM will fine-tune your models or work with a vector DB to provide up-to-date information with citations +- **Live Data Augmentation** - You will be able to provide a live updating data-set and the Anarchy LLM-VM will fine-tune your models or work with a vector database to provide up-to-date information with citations. -- **Web Playground** - You will be able to run the Anarchy LLM-VM and test it's outputs from the browser. +- **Web Playground** - You will be able to run the Anarchy LLM-VM and test its outputs from the browser. -- **Load-Balancing and Orchestration** -- If you have multiple LLMs or providers you'd like to utilize, you will be able to hand them to the Anarchy LLM-VM to automatically figure out which to work with and when to optimize your uptime or your costs +- **Load-Balancing and Orchestration** -- If you have multiple LLMs or providers you'd like to utilize, you will be able to hand them to the Anarchy LLM-VM to automatically figure out which to work with and when to optimize your uptime or your costs. -- **Output Templating** - You can ensure that the LLM only outputs data in specific formats and fills in variables from a template with either regular expressions, LMQL, or OpenAI's template language +- **Output Templating** - You can ensure that the LLM only outputs data in specific formats and fills in variables from a template with either regular expressions, LMQL, or OpenAI's template language. -- **Persistent Stateful Memory** - The Anarchy LLM-VM can remember a user's conversation history and react accordingly +- **Persistent Stateful Memory** - The Anarchy LLM-VM can remember a user's conversation history and react accordingly. diff --git a/welcome/who_are_we.mdx b/welcome/who_are_we.mdx index e2f8baf..a129310 100644 --- a/welcome/who_are_we.mdx +++ b/welcome/who_are_we.mdx @@ -1,6 +1,6 @@ --- title: "🏴‍☠️ Organization" -description: "As self proclaimed anarchists, we eschew organization, but alas, we do live in a society." +description: "As self-proclaimed anarchists, we eschew organization, but alas we do live in a society." --- # Core Team @@ -23,7 +23,7 @@ A shout-out to our stellar community members! # 🍼 Funding -We're a [YC (W23)]("https://www.ycombinator.com/companies/anarchy") company. In addition, we've raised 2m from various investors on the premise of building an OSS LLM Infra company. In particular, these investors have been magnificent to work with: +We're a [YC (W23)]("https://www.ycombinator.com/companies/anarchy") company. In addition, we've raised 2m from various investors to build an OSS LLM Infra company. In particular, these investors have been magnificent to work with: - **[Fellows Fund](https://www.fellowsfundvc.com/), [Amino Capital](https://www.aminocapital.com/), [Venrex](https://www.venrex.partners/), [Futureland Ventures](https://www.futurelandventures.com/)** - **[Yash Tulsani](https://www.linkedin.com/in/yash-tulsani/), [Timothy Chen](https://www.linkedin.com/in/timchen/), [Arup Chakrabarti](https://www.linkedin.com/in/arupchak/)** - **[Keith Sherry](https://www.linkedin.com/in/keithtsherry/), [Ahmed Medhat](https://www.linkedin.com/in/ahmedmedhatm/), and [Jonathan Paulson](https://www.linkedin.com/in/jonathan-paulson-aa57203a/).** diff --git a/welcome/why_anarchy.mdx b/welcome/why_anarchy.mdx index 9bc4b3b..46d9adc 100644 --- a/welcome/why_anarchy.mdx +++ b/welcome/why_anarchy.mdx @@ -5,16 +5,16 @@ description: "We're on a mission to democratize AI at whatever the cost. We're h ## Current Projects -- **[LLM-VM](https://anarchy.ai/welcome/llm_vm)** The LLM-VM is an LLM agnostic JIT for natural language. Specifically, it uses LLMs to convert conversational natural language into a dynamic series of LLM and IO commands. You provide the underlying provider(s), actions (APIs, code-hooks) and their descriptions, data-sources (PDFs, websites...), and the LLM-VM will take care of load-balancing, fine-tuning, natural language compilation and tool-selection. +- **[LLM-VM](https://anarchy.ai/welcome/llm_vm)** The LLM-VM is an LLM agnostic JIT for natural language. Specifically, it uses LLMs to convert conversational natural language into a dynamic series of LLM and IO commands. You provide the underlying provider(s), actions (APIs, code-hooks) and their descriptions, data sources (PDFs, websites...), and the LLM-VM will take care of load-balancing, fine-tuning, natural language compilation, and tool selection. - **[chat.dev](https://chat.dev)** This is our managed hosting service for the LLM-VM. Here you can create an optimized LLM endpoint with access to your APIs and data. ## Story -Anarchy was born out of exaspiration. We'd been AI safety researchers, hackers and long-time GNU beneficiaries, +Anarchy was born out of exasperation. We'd been AI safety researchers, hackers, and long-time GNU beneficiaries, who saw the gestalt of our domains being annexed by the fascist rhetoric of so-called "doomers" afraid -of "long-termist" sci-fi hypotheticals. To us, the encroachment of regulations on our self-expression -was a far more dire threat. We decided then to eschew such viewpoints and do what such extremists would consider +of "long-termist" sci-fi hypotheticals. To us, the encroachment of regulations on our self-expression +was a far more dire threat. We decided then to eschew such viewpoints and do what such extremists would consider most anarchic: give AI access to the world. ## Our promises @@ -23,7 +23,7 @@ most anarchic: give AI access to the world. **2. We will work to accelerate the commoditization of AI.** -**3. We will not judge your use-cases.** (provided it is legal) +**3. We will not judge your use cases.** (provided it is legal) ## 💸 How will we make money? 💸