Latest navi open source refresh

latest code change including the global thread pool

Closes twitter/the-algorithm#452
Closes twitter/the-algorithm#505

navi/README.md

@@ -31,6 +31,11 @@ In navi/navi, you can run the following commands:
 - `scripts/run_onnx.sh` for [Onnx](https://onnx.ai/)
 
 Do note that you need to create a models directory and create some versions, preferably using epoch time, e.g., `1679693908377`.
+so the models structure looks like:
+  models/
+       -web_click
+        - 1809000
+        - 1809010
 
 ## Build
 You can adapt the above scripts to build using Cargo.

navi/dr_transform/Cargo.toml

@@ -3,7 +3,6 @@ name = "dr_transform"
 version = "0.1.0"
 edition = "2021"
 
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 [dependencies]
 serde = { version = "1.0", features = ["derive"] }
 serde_json = "1.0"
@@ -12,7 +11,6 @@ bpr_thrift = { path = "../thrift_bpr_adapter/thrift/"}
 segdense = { path = "../segdense/"}
 thrift = "0.17.0"
 ndarray = "0.15"
-ort = {git ="https://github.com/pykeio/ort.git", tag="v1.14.2"}
 base64 = "0.20.0"
 npyz = "0.7.2"
 log = "0.4.17"
@@ -21,6 +19,11 @@ prometheus = "0.13.1"
 once_cell = "1.17.0"
 rand = "0.8.5"
 itertools = "0.10.5"
+anyhow = "1.0.70"
+[target.'cfg(not(target_os="linux"))'.dependencies]
+ort = {git ="https://github.com/pykeio/ort.git", features=["profiling"], tag="v1.14.6"}
+[target.'cfg(target_os="linux")'.dependencies]
+ort = {git ="https://github.com/pykeio/ort.git", features=["profiling", "tensorrt", "cuda", "copy-dylibs"], tag="v1.14.6"}
 [dev-dependencies]
 criterion = "0.3.0"
 

navi/dr_transform/src/lib.rs

@@ -3,3 +3,4 @@ pub mod converter;
 #[cfg(test)]
 mod test;
 pub mod util;
+pub extern crate ort;

navi/navi/Cargo.toml

@@ -1,8 +1,7 @@
 [package]
 name = "navi"
-version = "2.0.42"
+version = "2.0.45"
 edition = "2021"
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [[bin]]
 name = "navi"
@@ -16,12 +15,19 @@ required-features=["torch"]
 name = "navi_onnx"
 path = "src/bin/navi_onnx.rs"
 required-features=["onnx"]
+[[bin]]
+name = "navi_onnx_test"
+path = "src/bin/bin_tests/navi_onnx_test.rs"
+[[bin]]
+name = "navi_torch_test"
+path = "src/bin/bin_tests/navi_torch_test.rs"
+required-features=["torch"]
 
 [features]
 default=[]
 navi_console=[]
 torch=["tch"]
-onnx=["ort"]
+onnx=[]
 tf=["tensorflow"]
 [dependencies]
 itertools = "0.10.5"
@@ -47,6 +53,7 @@ parking_lot = "0.12.1"
 rand = "0.8.5"
 rand_pcg = "0.3.1"
 random = "0.12.2"
+x509-parser = "0.15.0"
 sha256 = "1.0.3"
 tonic = { version = "0.6.2", features=['compression', 'tls'] }
 tokio = { version = "1.17.0", features = ["macros", "rt-multi-thread", "fs", "process"] }
@@ -55,16 +62,12 @@ npyz = "0.7.3"
 base64 = "0.21.0"
 histogram = "0.6.9"
 tch = {version = "0.10.3", optional = true}
-tensorflow = { version = "0.20.0", optional = true }
+tensorflow = { version = "0.18.0", optional = true }
 once_cell = {version = "1.17.1"}
 ndarray = "0.15"
 serde = "1.0.154"
 serde_json = "1.0.94"
 dr_transform = { path = "../dr_transform"}
-[target.'cfg(not(target_os="linux"))'.dependencies]
-ort = {git ="https://github.com/pykeio/ort.git", features=["profiling"], optional = true, tag="v1.14.2"}
-[target.'cfg(target_os="linux")'.dependencies]
-ort = {git ="https://github.com/pykeio/ort.git", features=["profiling", "tensorrt", "cuda", "copy-dylibs"], optional = true, tag="v1.14.2"}
 [build-dependencies]
 tonic-build = {version = "0.6.2", features=['prost', "compression"] }
 [profile.release]
@@ -74,3 +77,5 @@ ndarray-rand = "0.14.0"
 tokio-test = "*"
 assert_cmd = "2.0"
 criterion = "0.4.0"
+
+

navi/navi/scripts/run_onnx.sh

@@ -1,10 +1,9 @@
 #!/bin/sh
 #RUST_LOG=debug LD_LIBRARY_PATH=so/onnx/lib target/release/navi_onnx --port 30 --num-worker-threads 8 --intra-op-parallelism 8 --inter-op-parallelism 8 \
 RUST_LOG=info LD_LIBRARY_PATH=so/onnx/lib cargo run --bin navi_onnx --features onnx -- \
-    --port 30 --num-worker-threads 8 --intra-op-parallelism 8 --inter-op-parallelism 8 \
+    --port 8030 --num-worker-threads 8 \
     --model-check-interval-secs 30 \
-    --model-dir models/int8 \
-    --output caligrated_probabilities \
-    --input "" \
     --modelsync-cli "echo" \
-    --onnx-ep-options use_arena=true
+    --onnx-ep-options use_arena=true \
+    --model-dir models/prod_home --output caligrated_probabilities  --input "" --intra-op-parallelism 8 --inter-op-parallelism 8 --max-batch-size 1 --batch-time-out-millis 1 \
+    --model-dir models/prod_home1 --output caligrated_probabilities  --input "" --intra-op-parallelism 8 --inter-op-parallelism 8 --max-batch-size 1 --batch-time-out-millis 1 \

navi/navi/src/bin/navi_onnx.rs

@@ -1,11 +1,24 @@
 use anyhow::Result;
+use log::info;
 use navi::cli_args::{ARGS, MODEL_SPECS};
 use navi::onnx_model::onnx::OnnxModel;
 use navi::{bootstrap, metrics};
 
 fn main() -> Result<()> {
     env_logger::init();
-    assert_eq!(MODEL_SPECS.len(), ARGS.inter_op_parallelism.len());
+    info!("global: {:?}", ARGS.onnx_global_thread_pool_options);
+    let assert_session_params = if ARGS.onnx_global_thread_pool_options.is_empty() {
+        // std::env::set_var("OMP_NUM_THREADS", "1");
+        info!("now we use per session thread pool");
+        MODEL_SPECS.len()
+    }
+    else {
+       info!("now we use global thread pool");
+        0
+    };
+    assert_eq!(assert_session_params, ARGS.inter_op_parallelism.len());
+    assert_eq!(assert_session_params, ARGS.inter_op_parallelism.len());
+
     metrics::register_custom_metrics();
     bootstrap::bootstrap(OnnxModel::new)
 }

navi/navi/src/bootstrap.rs

@@ -207,6 +207,9 @@ impl<T: Model> PredictionService for PredictService<T> {
                         PredictResult::DropDueToOverload => Err(Status::resource_exhausted("")),
                         PredictResult::ModelNotFound(idx) => {
                             Err(Status::not_found(format!("model index {}", idx)))
+                        },
+                        PredictResult::ModelNotReady(idx) => {
+                            Err(Status::unavailable(format!("model index {}", idx)))
                         }
                         PredictResult::ModelVersionNotFound(idx, version) => Err(
                             Status::not_found(format!("model index:{}, version {}", idx, version)),

navi/navi/src/cli_args.rs

@@ -87,13 +87,11 @@ pub struct Args {
     pub intra_op_parallelism: Vec<String>,
     #[clap(
         long,
-        default_value = "14",
         help = "number of threads to parallelize computations of the graph"
     )]
     pub inter_op_parallelism: Vec<String>,
     #[clap(
         long,
-        default_value = "serving_default",
         help = "signature of a serving. only TF"
     )]
     pub serving_sig: Vec<String>,
@@ -107,6 +105,8 @@ pub struct Args {
         help = "max warmup records to use. warmup only implemented for TF"
     )]
     pub max_warmup_records: usize,
+    #[clap(long, value_parser = Args::parse_key_val::<String, String>, value_delimiter=',')]
+    pub onnx_global_thread_pool_options: Vec<(String, String)>,
     #[clap(
     long,
     default_value = "true",

navi/navi/src/lib.rs

@@ -146,6 +146,7 @@ pub enum PredictResult {
     Ok(Vec<TensorScores>, i64),
     DropDueToOverload,
     ModelNotFound(usize),
+    ModelNotReady(usize),
     ModelVersionNotFound(usize, i64),
 }
 

navi/navi/src/onnx_model.rs

@@ -13,21 +13,22 @@ pub mod onnx {
     use dr_transform::converter::{BatchPredictionRequestToTorchTensorConverter, Converter};
     use itertools::Itertools;
     use log::{debug, info};
-    use ort::environment::Environment;
+    use dr_transform::ort::environment::Environment;
-    use ort::session::Session;
+    use dr_transform::ort::session::Session;
-    use ort::tensor::InputTensor;
+    use dr_transform::ort::tensor::InputTensor;
-    use ort::{ExecutionProvider, GraphOptimizationLevel, SessionBuilder};
+    use dr_transform::ort::{ExecutionProvider, GraphOptimizationLevel, SessionBuilder};
+    use dr_transform::ort::LoggingLevel;
     use serde_json::Value;
     use std::fmt::{Debug, Display};
     use std::sync::Arc;
     use std::{fmt, fs};
     use tokio::time::Instant;
-
     lazy_static! {
         pub static ref ENVIRONMENT: Arc<Environment> = Arc::new(
             Environment::builder()
                 .with_name("onnx home")
-                .with_log_level(ort::LoggingLevel::Error)
+                .with_log_level(LoggingLevel::Error)
+                .with_global_thread_pool(ARGS.onnx_global_thread_pool_options.clone())
                 .build()
                 .unwrap()
         );
@@ -101,7 +102,9 @@ pub mod onnx {
             let meta_info = format!("{}/{}/{}", ARGS.model_dir[idx], version, META_INFO);
             let mut builder = SessionBuilder::new(&ENVIRONMENT)?
                 .with_optimization_level(GraphOptimizationLevel::Level3)?
-                .with_parallel_execution(ARGS.onnx_use_parallel_mode == "true")?
+                .with_parallel_execution(ARGS.onnx_use_parallel_mode == "true")?;
+            if ARGS.onnx_global_thread_pool_options.is_empty() {
+                builder = builder
                     .with_inter_threads(
                         utils::get_config_or(
                             model_config,
@@ -117,7 +120,12 @@ pub mod onnx {
                             &ARGS.intra_op_parallelism[idx],
                         )
                             .parse()?,
-                )?
+                    )?;
+            }
+            else {
+                builder = builder.with_disable_per_session_threads()?;
+            }
+            builder = builder
                 .with_memory_pattern(ARGS.onnx_use_memory_pattern == "true")?
                 .with_execution_providers(&OnnxModel::ep_choices())?;
             match &ARGS.profiling {

navi/navi/src/predict_service.rs

@@ -1,7 +1,7 @@
 use anyhow::{anyhow, Result};
 use arrayvec::ArrayVec;
 use itertools::Itertools;
-use log::{error, info, warn};
+use log::{error, info};
 use std::fmt::{Debug, Display};
 use std::string::String;
 use std::sync::Arc;
@@ -179,17 +179,17 @@ impl<T: Model> PredictService<T> {
         //initialize the latest version array
         let mut cur_versions = vec!["".to_owned(); MODEL_SPECS.len()];
         loop {
-            let config = utils::read_config(&meta_file).unwrap_or_else(|e| {
-                warn!("config file {} not found due to: {}", meta_file, e);
-                Value::Null
-            });
             info!("***polling for models***"); //nice deliminter
-            info!("config:{}", config);
             if let Some(ref cli) = ARGS.modelsync_cli {
                 if let Err(e) = call_external_modelsync(cli, &cur_versions).await {
                     error!("model sync cli running error:{}", e)
                 }
             }
+            let config = utils::read_config(&meta_file).unwrap_or_else(|e| {
+                info!("config file {} not found due to: {}", meta_file, e);
+                Value::Null
+            });
+            info!("config:{}", config);
             for (idx, cur_version) in cur_versions.iter_mut().enumerate() {
                 let model_dir = &ARGS.model_dir[idx];
                 PredictService::scan_load_latest_model_from_model_dir(
@@ -229,6 +229,11 @@ impl<T: Model> PredictService<T> {
             let no_more_msg = match msg {
                 Ok(PredictMessage::Predict(model_spec_at, version, val, resp, ts)) => {
                     if let Some(model_predictors) = all_model_predictors.get_mut(model_spec_at) {
+                        if model_predictors.is_empty() {
+                            resp.send(PredictResult::ModelNotReady(model_spec_at))
+                                .unwrap_or_else(|e| error!("cannot send back model not ready error: {:?}", e));
+                        }
+                        else {
                             match version {
                                 None => model_predictors[0].push(val, resp, ts),
                                 Some(the_version) => match model_predictors
@@ -246,9 +251,10 @@ impl<T: Model> PredictService<T> {
                                     Some(predictor) => predictor.push(val, resp, ts),
                                 },
                             }
+                        }
                     } else {
                         resp.send(PredictResult::ModelNotFound(model_spec_at))
-                            .unwrap_or_else(|e| error!("cannot send back model error: {:?}", e))
+                            .unwrap_or_else(|e| error!("cannot send back model not found error: {:?}", e))
                     }
                     MPSC_CHANNEL_SIZE.dec();
                     false

navi/segdense/Cargo.toml

@@ -3,9 +3,9 @@ name = "segdense"
 version = "0.1.0"
 edition = "2021"
 
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
+env_logger = "0.10.0"
 serde = { version = "1.0.104", features = ["derive"] }
 serde_json = "1.0.48"
 log = "0.4.17"