Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

feat: change request/response to json value #24

Merged
merged 3 commits into from
Apr 5, 2024
Merged

feat: change request/response to json value #24

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
2589744
feat: change request/response to json value
Cifko Apr 5, 2024
f3d1b29
fix clippy
Cifko Apr 5, 2024
6419af3
fixes
Cifko Apr 5, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
39 changes: 19 additions & 20 deletions atoma-inference/src/main.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,20 +3,16 @@ use std::time::Duration;
use ed25519_consensus::SigningKey as PrivateKey;
use hf_hub::api::sync::Api;
use inference::{
models::{
candle::mamba::MambaModel,
config::ModelsConfig,
types::{TextRequest, TextResponse},
},
models::{candle::mamba::MambaModel, config::ModelsConfig, types::TextRequest},
service::{ModelService, ModelServiceError},
};

#[tokio::main]
async fn main() -> Result<(), ModelServiceError> {
tracing_subscriber::fmt::init();

let (req_sender, req_receiver) = tokio::sync::mpsc::channel::<TextRequest>(32);
let (resp_sender, mut resp_receiver) = tokio::sync::mpsc::channel::<TextResponse>(32);
let (req_sender, req_receiver) = tokio::sync::mpsc::channel::<serde_json::Value>(32);
let (resp_sender, mut resp_receiver) = tokio::sync::mpsc::channel::<serde_json::Value>(32);

let model_config = ModelsConfig::from_file_path("../inference.toml".parse().unwrap());
let private_key_bytes =
Expand Down Expand Up @@ -44,19 +40,22 @@ async fn main() -> Result<(), ModelServiceError> {
tokio::time::sleep(Duration::from_millis(5000)).await;

req_sender
.send(TextRequest {
request_id: 0,
prompt: "Leon, the professional is a movie".to_string(),
model: "state-spaces/mamba-130m".to_string(),
max_tokens: 512,
temperature: Some(0.0),
random_seed: 42,
repeat_last_n: 64,
repeat_penalty: 1.1,
sampled_nodes: vec![pk],
top_p: Some(1.0),
top_k: 10,
})
.send(
serde_json::to_value(TextRequest {
request_id: 0,
prompt: "Leon, the professional is a movie".to_string(),
model: "state-spaces/mamba-130m".to_string(),
max_tokens: 512,
temperature: Some(0.0),
random_seed: 42,
repeat_last_n: 64,
repeat_penalty: 1.1,
sampled_nodes: vec![pk],
top_p: Some(1.0),
top_k: 10,
})
.unwrap(),
)
.await
.expect("Failed to send request");

Expand Down
112 changes: 49 additions & 63 deletions atoma-inference/src/model_thread.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,13 +11,13 @@ use tracing::{debug, error, info, warn};

use crate::{
apis::{ApiError, ApiTrait},
models::{config::ModelsConfig, ModelError, ModelId, ModelTrait, Request, Response},
models::{config::ModelsConfig, ModelError, ModelId, ModelTrait},
};

pub struct ModelThreadCommand<Req, Resp>(Req, oneshot::Sender<Resp>)
where
Req: Request,
Resp: Response;
pub struct ModelThreadCommand {
request: serde_json::Value,
response_sender: oneshot::Sender<serde_json::Value>,
}

#[derive(Debug, Error)]
pub enum ModelThreadError {
Expand All @@ -27,6 +27,8 @@ pub enum ModelThreadError {
ModelError(ModelError),
#[error("Core thread shutdown: `{0}`")]
Shutdown(RecvError),
#[error("Serde error: `{0}`")]
SerdeError(#[from] serde_json::Error),
}

impl From<ModelError> for ModelThreadError {
Expand All @@ -41,82 +43,68 @@ impl From<ApiError> for ModelThreadError {
}
}

pub struct ModelThreadHandle<Req, Resp>
where
Req: Request,
Resp: Response,
{
sender: mpsc::Sender<ModelThreadCommand<Req, Resp>>,
pub struct ModelThreadHandle {
sender: mpsc::Sender<ModelThreadCommand>,
join_handle: std::thread::JoinHandle<Result<(), ModelThreadError>>,
}

impl<Req, Resp> ModelThreadHandle<Req, Resp>
where
Req: Request,
Resp: Response,
{
impl ModelThreadHandle {
pub fn stop(self) {
drop(self.sender);
self.join_handle.join().ok();
}
}

pub struct ModelThread<M: ModelTrait, Req: Request, Resp: Response> {
pub struct ModelThread<M: ModelTrait> {
model: M,
receiver: mpsc::Receiver<ModelThreadCommand<Req, Resp>>,
receiver: mpsc::Receiver<ModelThreadCommand>,
}

impl<M, Req, Resp> ModelThread<M, Req, Resp>
impl<M> ModelThread<M>
where
M: ModelTrait<Input = Req::ModelInput, Output = Resp::ModelOutput>,
Req: Request,
Resp: Response,
M: ModelTrait,
{
pub fn run(mut self, public_key: PublicKey) -> Result<(), ModelThreadError> {
pub fn run(mut self, _public_key: PublicKey) -> Result<(), ModelThreadError> {
debug!("Start Model thread");

while let Ok(command) = self.receiver.recv() {
let ModelThreadCommand(request, sender) = command;

if !request.is_node_authorized(&public_key) {
error!("Current node, with verification key = {:?} is not authorized to run request with id = {}", public_key, request.request_id());
continue;
}

let model_input = request.into_model_input();
let ModelThreadCommand {
request,
response_sender,
} = command;

// TODO: Implement node authorization
// if !request.is_node_authorized(&public_key) {
Cifko marked this conversation as resolved.
Show resolved Hide resolved
// error!("Current node, with verification key = {:?} is not authorized to run request with id = {}", public_key, request.request_id());
// continue;
// }

let model_input = serde_json::from_value(request).unwrap();
let model_output = self
.model
.run(model_input)
.map_err(ModelThreadError::ModelError)?;
let response = Resp::from_model_output(model_output);
sender.send(response).ok();
let response = serde_json::to_value(model_output)?;
response_sender.send(response).ok();
}

Ok(())
}
}

pub struct ModelThreadDispatcher<Req, Resp>
where
Req: Request,
Resp: Response,
{
model_senders: HashMap<ModelId, mpsc::Sender<ModelThreadCommand<Req, Resp>>>,
pub(crate) responses: FuturesUnordered<oneshot::Receiver<Resp>>,
pub struct ModelThreadDispatcher {
model_senders: HashMap<ModelId, mpsc::Sender<ModelThreadCommand>>,
pub(crate) responses: FuturesUnordered<oneshot::Receiver<serde_json::Value>>,
}

impl<Req, Resp> ModelThreadDispatcher<Req, Resp>
where
Req: Clone + Request,
Resp: Response,
{
impl ModelThreadDispatcher {
pub(crate) fn start<M, F>(
config: ModelsConfig,
public_key: PublicKey,
) -> Result<(Self, Vec<ModelThreadHandle<Req, Resp>>), ModelThreadError>
) -> Result<(Self, Vec<ModelThreadHandle>), ModelThreadError>
where
F: ApiTrait + Send + Sync + 'static,
M: ModelTrait<Input = Req::ModelInput, Output = Resp::ModelOutput> + Send + 'static,
M: ModelTrait, //<Input = Req::ModelInput, Output = Resp::ModelOutput> + Send + 'static,
{
let api_key = config.api_key();
let storage_path = config.storage_path();
Expand All @@ -129,19 +117,16 @@ where
info!("Spawning new thread for model: {}", model_config.model_id());
let api = api.clone();

let (model_sender, model_receiver) = mpsc::channel::<ModelThreadCommand<_, _>>();
let (model_sender, model_receiver) = mpsc::channel::<ModelThreadCommand>();
let model_name = model_config.model_id().clone();
model_senders.insert(model_name.clone(), model_sender.clone());

let join_handle = std::thread::spawn(move || {
info!("Fetching files for model: {model_name}");
let filenames = api.fetch(model_name, model_config.revision())?;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This should be changed to use the fetch model trait method

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I had a problem, this was my debug line, removed

let x = serde_json::from_value(model_config.params().clone()).unwrap();

let model = M::load(
filenames,
model_config.precision(),
model_config.device_id(),
)?;
let model = M::load(filenames, x, model_config.device_id())?;
let model_thread = ModelThread {
model,
receiver: model_receiver,
Expand Down Expand Up @@ -170,10 +155,12 @@ where
Ok((model_dispatcher, handles))
}

fn send(&self, command: ModelThreadCommand<Req, Resp>) {
let request = command.0.clone();
let model_id = request.requested_model();
fn send(&self, command: ModelThreadCommand) {
let request = command.request.clone();
let model_id = request.get("model").unwrap().as_str().unwrap().to_string();
println!("model_id {model_id}");

println!("{:?}", self.model_senders);
let sender = self
.model_senders
.get(&model_id)
Expand All @@ -185,14 +172,13 @@ where
}
}

impl<T, U> ModelThreadDispatcher<T, U>
where
T: Clone + Request,
U: Response,
{
pub(crate) fn run_inference(&self, request: T) {
impl ModelThreadDispatcher {
pub(crate) fn run_inference(&self, request: serde_json::Value) {
let (sender, receiver) = oneshot::channel();
self.send(ModelThreadCommand(request, sender));
self.send(ModelThreadCommand {
request,
response_sender: sender,
});
self.responses.push(receiver);
}
}
3 changes: 2 additions & 1 deletion atoma-inference/src/models/candle/falcon.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -49,14 +49,15 @@ impl ModelTrait for FalconModel {
type Fetch = ();
type Input = TextModelInput;
type Output = String;
type Load = PrecisionBits;

fn fetch(_fetch: &Self::Fetch) -> Result<(), ModelError> {
Ok(())
}

fn load(
filenames: Vec<PathBuf>,
precision: PrecisionBits,
precision: Self::Load,
device_id: usize,
) -> Result<Self, ModelError>
where
Expand Down
Loading