feat(dynamo-run): Download models from HF, smart model defaults (#126)

- Any engine can take the name of a Hugging Face repository. It will be downloaded before calling the engine.

- The default engine (previously always mistralrs) depends on what is compiled in.

- Text can be piped in and will result in a single run of the model.

All of those together mean if you build with `--features vllm` you can do this and it will download the model and run it with vllm, answer your question, and exit:
```
echo "What is the capital of Costa Rica?"  | dynamo-run Qwen/Qwen2.5-3B-Instruct
```
Co-authored-by: Ryan McCormick <rmccormick@nvidia.com>

launch/Cargo.lock

@@ -307,6 +307,31 @@ version = "1.4.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "ace50bade8e6234aa140d9a2f552bbee1db4d353f69b8217bc503490fc1a9f26"
 
+[[package]]
+name = "aws-lc-rs"
+version = "1.12.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5e4e8200b9a4a5801a769d50eeabc05670fec7e959a8cb7a63a93e4e519942ae"
+dependencies = [
+ "aws-lc-sys",
+ "paste",
+ "zeroize",
+]
+
+[[package]]
+name = "aws-lc-sys"
+version = "0.26.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0f9dd2e03ee80ca2822dd6ea431163d2ef259f2066a4d6ccaca6d9dcb386aa43"
+dependencies = [
+ "bindgen 0.69.5",
+ "cc",
+ "cmake",
+ "dunce",
+ "fs_extra",
+ "paste",
+]
+
 [[package]]
 name = "axum"
 version = "0.7.9"
@@ -674,6 +699,7 @@ dependencies = [
  "num_cpus",
  "rand",
  "reqwest",
+ "rustls",
  "serde",
  "serde_json",
  "thiserror 1.0.69",
@@ -1390,6 +1416,12 @@ version = "1.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "aac81fa3e28d21450aa4d2ac065992ba96a1d7303efbce51a95f4fd175b67562"
 
+[[package]]
+name = "dunce"
+version = "1.0.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "92773504d58c093f6de2459af4af33faa518c13451eb8f2b5698ed3d36e7c813"
+
 [[package]]
 name = "dyn-clone"
 version = "1.0.19"
@@ -1468,6 +1500,7 @@ dependencies = [
  "async-openai",
  "async-stream",
  "async-trait",
+ "candle-hf-hub",
  "clap",
  "dialoguer",
  "dynamo-llm",
@@ -1887,6 +1920,12 @@ dependencies = [
  "percent-encoding",
 ]
 
+[[package]]
+name = "fs_extra"
+version = "1.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "42703706b716c37f96a77aea830392ad231f44c9e9a67872fa5548707e11b11c"
+
 [[package]]
 name = "fuchsia-zircon"
 version = "0.3.3"
@@ -2371,6 +2410,7 @@ dependencies = [
  "tokio",
  "tokio-rustls",
  "tower-service",
+ "webpki-roots",
 ]
 
 [[package]]
@@ -4537,6 +4577,7 @@ dependencies = [
  "wasm-bindgen-futures",
  "wasm-streams",
  "web-sys",
+ "webpki-roots",
  "windows-registry",
 ]
 
@@ -4647,6 +4688,7 @@ version = "0.23.23"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "47796c98c480fce5406ef69d1c76378375492c3b0a0de587be0c1d9feb12f395"
 dependencies = [
+ "aws-lc-rs",
  "log",
  "once_cell",
  "ring",
@@ -4705,6 +4747,7 @@ version = "0.102.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "64ca1bc8749bd4cf37b5ce386cc146580777b4e8572c7b97baf22c83f444bee9"
 dependencies = [
+ "aws-lc-rs",
  "ring",
  "rustls-pki-types",
  "untrusted",

launch/dynamo-run/Cargo.toml

@@ -36,6 +36,7 @@ anyhow = "1"
 async-openai = "0.27.2"
 async-stream = { version = "0.3" }
 async-trait = { version = "0.1" }
+candle-hf-hub = { version = "0.3.3", default-features = false, features = ["tokio", "rustls-tls"] }
 clap = { version = "4.5", features = ["derive", "env"] }
 dialoguer = { version = "0.11", default-features = false, features = ["editor", "history"] }
 futures = { version = "0.3" }

launch/dynamo-run/README.md

@@ -38,9 +38,11 @@ If you have an `HF_TOKEN` environment variable set, this will download Qwen2.5 3
 dynamo-run Qwen/Qwen2.5-3B-Instruct
 ```
 
+The parameter can be the ID of a HuggingFace repository (it will be downloaded), a GGUF file, or a folder containing safetensors, config.json, etc (a locally checked out HuggingFace repository).
+
 ## Download a model from Hugging Face
 
-For example one of these should be fast and good quality on almost any machine: https://huggingface.co/bartowski/Llama-3.2-3B-Instruct-GGUF
+One of these should be fast and good quality on almost any machine: https://huggingface.co/bartowski/Llama-3.2-3B-Instruct-GGUF
 
 ## Run
 
@@ -349,3 +351,10 @@ DYN_TOKEN_ECHO_DELAY_MS=1 dynamo-run in=http out=echo_full
 ```
 
 The default delay is 10ms, which produces approximately 100 tokens per second.
+
+## Defaults
+
+The input defaults to `in=text`.
+
+The output will default to whatever engine you have compiled in (so depending on `--features`). If all features
+are enabled at build time, then the default is currently `out=vllm`.

launch/dynamo-run/src/hub.rs

+// SPDX-FileCopyrightText: Copyright (c) 2024-2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+// SPDX-License-Identifier: Apache-2.0
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use candle_hf_hub::api::tokio::ApiBuilder;
+use std::path::{Path, PathBuf};
+
+const IGNORED: [&str; 3] = [".gitattributes", "LICENSE", "README.md"];
+
+/// Attempt to download a model from Hugging Face
+/// Returns the directory it is in
+pub async fn from_hf(name: &Path) -> anyhow::Result<PathBuf> {
+    let api = ApiBuilder::new().with_progress(true).build()?;
+
+    let repo = api.model(name.display().to_string());
+    let info = repo.info().await?;
+    let mut p = PathBuf::new();
+    for sib in info.siblings {
+        if IGNORED.contains(&sib.rfilename.as_str()) || is_image(&sib.rfilename) {
+            continue;
+        }
+        p = repo.get(&sib.rfilename).await?;
+    }
+    match p.parent() {
+        Some(p) => Ok(p.to_path_buf()),
+        None => Err(anyhow::anyhow!("Invalid HF cache path: {}", p.display())),
+    }
+}
+
+fn is_image(s: &str) -> bool {
+    s.ends_with(".png") || s.ends_with("PNG") || s.ends_with(".jpg") || s.ends_with("JPG")
+}

launch/dynamo-run/src/input/text.rs

@@ -31,7 +31,7 @@ use dynamo_runtime::{
 };
 use futures::StreamExt;
 use std::{
-    io::{ErrorKind, Read, Write},
+    io::{ErrorKind, Write},
     sync::Arc,
 };
 
@@ -40,12 +40,10 @@ use crate::EngineConfig;
 /// Max response tokens for each single query. Must be less than model context size.
 const MAX_TOKENS: u32 = 8192;
 
-/// Output of `isatty` if the fd is indeed a TTY
-const IS_A_TTY: i32 = 1;
-
 pub async fn run(
     runtime: Runtime,
     cancel_token: CancellationToken,
+    single_prompt: Option<String>,
     engine_config: EngineConfig,
 ) -> anyhow::Result<()> {
     let (service_name, engine, inspect_template): (
@@ -75,7 +73,7 @@ pub async fn run(
             service_name,
             engine,
         } => {
-            tracing::info!("Model: {service_name}");
+            tracing::debug!("Model: {service_name}");
             (service_name, engine, false)
         }
         EngineConfig::StaticCore {
@@ -101,12 +99,19 @@ pub async fn run(
                 .link(preprocessor.backward_edge())?
                 .link(frontend)?;
 
-            tracing::info!("Model: {service_name} with pre-processing");
+            tracing::debug!("Model: {service_name} with pre-processing");
             (service_name, pipeline, true)
         }
         EngineConfig::None => unreachable!(),
     };
-    main_loop(cancel_token, &service_name, engine, inspect_template).await
+    main_loop(
+        cancel_token,
+        &service_name,
+        engine,
+        single_prompt,
+        inspect_template,
+    )
+    .await
 }
 
 #[allow(deprecated)]
@@ -114,20 +119,17 @@ async fn main_loop(
     cancel_token: CancellationToken,
     service_name: &str,
     engine: OpenAIChatCompletionsStreamingEngine,
+    mut initial_prompt: Option<String>,
     _inspect_template: bool,
 ) -> anyhow::Result<()> {
+    if initial_prompt.is_none() {
         tracing::info!("Ctrl-c to exit");
+    }
     let theme = dialoguer::theme::ColorfulTheme::default();
 
-    let mut initial_prompt = if unsafe { libc::isatty(libc::STDIN_FILENO) == IS_A_TTY } {
-        None
-    } else {
-        // Something piped in, use that as initial prompt
-        let mut input = String::new();
-        std::io::stdin().read_to_string(&mut input).unwrap();
-        Some(input)
-    };
-
+    // Initial prompt is the pipe case: `echo "Hello" | dynamo-run ..`
+    // We run that single prompt and exit
+    let single = initial_prompt.is_some();
     let mut history = dialoguer::BasicHistory::default();
     let mut messages = vec![];
     while !cancel_token.is_cancelled() {
@@ -236,6 +238,10 @@ async fn main_loop(
             },
         );
         messages.push(assistant_message);
+
+        if single {
+            break;
+        }
     }
     println!();
     Ok(())

launch/dynamo-run/src/lib.rs

@@ -13,6 +13,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::io::Read;
 #[cfg(any(feature = "vllm", feature = "sglang"))]
 use std::{future::Future, pin::Pin};
 
@@ -24,6 +25,7 @@ use dynamo_runtime::protocols::Endpoint;
 
 mod flags;
 pub use flags::Flags;
+mod hub;
 mod input;
 #[cfg(any(feature = "vllm", feature = "sglang"))]
 mod net;
@@ -80,7 +82,7 @@ pub async fn run(
     let cancel_token = runtime.primary_token();
 
     // Turn relative paths into absolute paths
-    let model_path = flags
+    let mut model_path = flags
         .model_path_pos
         .clone()
         .or(flags.model_path_flag.clone())
@@ -91,8 +93,9 @@ pub async fn run(
                 Some(p)
             }
         });
+
     // Serve the model under the name provided, or the name of the GGUF file or HF repo.
-    let model_name = flags
+    let mut model_name = flags
         .model_name
         .clone()
         .or_else(|| {
@@ -108,6 +111,18 @@ pub async fn run(
                 None
             }
         });
+
+    // If it's an HF repo download it
+    if let Some(inner_model_path) = model_path.as_ref() {
+        if !inner_model_path.exists() {
+            model_name = inner_model_path
+                .iter()
+                .last()
+                .map(|s| s.to_string_lossy().to_string());
+            model_path = Some(hub::from_hf(inner_model_path).await?);
+        }
+    }
+
     // Load the model deployment card, if any
     // Only used by some engines, so without those feature flags it's unused.
     #[allow(unused_variables)]
@@ -373,7 +388,19 @@ pub async fn run(
             crate::input::http::run(runtime.clone(), flags.http_port, engine_config).await?;
         }
         Input::Text => {
-            crate::input::text::run(runtime.clone(), cancel_token.clone(), engine_config).await?;
+            crate::input::text::run(runtime.clone(), cancel_token.clone(), None, engine_config)
+                .await?;
+        }
+        Input::Stdin => {
+            let mut prompt = String::new();
+            std::io::stdin().read_to_string(&mut prompt).unwrap();
+            crate::input::text::run(
+                runtime.clone(),
+                cancel_token.clone(),
+                Some(prompt),
+                engine_config,
+            )
+            .await?;
         }
         Input::Endpoint(path) => {
             crate::input::endpoint::run(runtime.clone(), path, engine_config).await?;

launch/dynamo-run/src/main.rs

@@ -31,14 +31,6 @@ Example:
 
 "#;
 
-const DEFAULT_IN: Input = Input::Text;
-
-#[cfg(feature = "mistralrs")]
-const DEFAULT_OUT: Output = Output::MistralRs;
-
-#[cfg(not(feature = "mistralrs"))]
-const DEFAULT_OUT: Output = Output::EchoFull;
-
 const ZMQ_SOCKET_PREFIX: &str = "dyn";
 
 const USAGE: &str = "USAGE: dynamo-run in=[http|text|dyn://<path>|none] out=[mistralrs|sglang|llamacpp|vllm|trtllm|echo_full|echo_core|pystr:<engine.py>|pytok:<engine.py>] [--http-port 8080] [--model-path <path>] [--model-name <served-model-name>] [--model-config <hf-repo>] [--tensor-parallel-size=1] [--num-nodes=1] [--node-rank=0] [--leader-addr=127.0.0.1:9876] [--base-gpu-id=0]";
@@ -159,14 +151,18 @@ async fn wrapper(runtime: dynamo_runtime::Runtime) -> anyhow::Result<()> {
             non_flag_params += 1;
             x
         }
-        None => DEFAULT_IN,
+        None => Input::default(),
     };
     let out_opt = match out_opt {
         Some(x) => {
             non_flag_params += 1;
             x
         }
-        None => DEFAULT_OUT,
+        None => {
+            let default_engine = Output::default(); // smart default based on feature flags
+            tracing::debug!("Using engine: {default_engine}");
+            default_engine
+        }
     };
 
     // Clap skips the first argument expecting it to be the binary name, so add it back

launch/dynamo-run/src/opt.rs

@@ -13,7 +13,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-use std::fmt;
+use std::{fmt, io::IsTerminal as _};
 
 use crate::ENDPOINT_SCHEME;
 
@@ -22,7 +22,10 @@ pub enum Input {
     /// Run an OpenAI compatible HTTP server
     Http,
 
-    /// Read prompt from stdin
+    /// Single prompt on stdin
+    Stdin,
+
+    /// Interactive chat
     Text,
 
     /// Pull requests from a namespace/component/endpoint path.
@@ -41,6 +44,7 @@ impl TryFrom<&str> for Input {
         match s {
             "http" => Ok(Input::Http),
             "text" => Ok(Input::Text),
+            "stdin" => Ok(Input::Stdin),
             "none" => Ok(Input::None),
             endpoint_path if endpoint_path.starts_with(ENDPOINT_SCHEME) => {
                 let path = endpoint_path.strip_prefix(ENDPOINT_SCHEME).unwrap();
@@ -56,6 +60,7 @@ impl fmt::Display for Input {
         let s = match self {
             Input::Http => "http",
             Input::Text => "text",
+            Input::Stdin => "stdin",
             Input::Endpoint(path) => path,
             Input::None => "none",
         };
@@ -63,6 +68,16 @@ impl fmt::Display for Input {
     }
 }
 
+impl Default for Input {
+    fn default() -> Self {
+        if std::io::stdin().is_terminal() {
+            Input::Text
+        } else {
+            Input::Stdin
+        }
+    }
+}
+
 pub enum Output {
     /// Accept un-preprocessed requests, echo the prompt back as the response
     EchoFull,
@@ -185,3 +200,36 @@ impl fmt::Display for Output {
         write!(f, "{s}")
     }
 }
+
+/// Returns the engine to use if user did not say on cmd line
+/// Uses whatever was compiled in, with a priority ordering.
+#[allow(unused_assignments, unused_mut)]
+impl Default for Output {
+    fn default() -> Self {
+        // Default if no engines
+        let mut out = Output::EchoFull;
+
+        // Runs everywhere but needs local CUDA to build
+        #[cfg(feature = "mistralrs")]
+        {
+            out = Output::MistralRs;
+        }
+
+        #[cfg(feature = "llamacpp")]
+        {
+            out = Output::LlamaCpp;
+        }
+
+        #[cfg(feature = "sglang")]
+        {
+            out = Output::SgLang;
+        }
+
+        #[cfg(feature = "vllm")]
+        {
+            out = Output::Vllm;
+        }
+
+        out
+    }
+}

lib/llm/src/engines/vllm/worker.rs

@@ -336,7 +336,7 @@ async fn start_vllm(
         let mut lines = stdout.lines();
         while let Ok(Some(line)) = lines.next_line().await {
             let mut line_parts = line.splitn(4, ' ');
-            let log_level = line_parts.next().unwrap_or_default();
+            let mut log_level = line_parts.next().unwrap_or_default();
             // Skip date (0) and time (1). Print last (2) which is everything else.
             let line = line_parts.nth(2).unwrap_or_default();
             if line.starts_with("custom_op.py:68") {
@@ -344,10 +344,14 @@ async fn start_vllm(
                 // custom_op.py:68] custom op <the op> enabled
                 continue;
             }
+            if line.contains("ERROR") {
+                log_level = "ERROR";
+            }
             match log_level {
                 "DEBUG" => tracing::debug!("VLLM: {line}"),
-                "INFO" => tracing::info!("VLLM: {line}"),
+                "INFO" => tracing::debug!("VLLM: {line}"), // VLLM is noisy
                 "WARNING" => tracing::warn!("VLLM: {line}"),
+                "ERROR" => tracing::error!("VLLM: {line}"),
                 level => tracing::info!("VLLM: {level} {line}"),
             }
         }

lib/runtime/src/worker.rs

@@ -153,7 +153,7 @@ impl Worker {
 
             match &result {
                 Ok(_) => {
-                    tracing::info!("Application shutdown successfully");
+                    tracing::debug!("Application shutdown successfully");
                 }
                 Err(e) => {
                     tracing::error!("Application shutdown with error: {:?}", e);
@@ -200,7 +200,7 @@ async fn signal_handler(cancel_token: CancellationToken) -> Result<()> {
             tracing::info!("SIGTERM received, starting graceful shutdown");
         },
         _ = cancel_token.cancelled() => {
-            tracing::info!("CancellationToken triggered; shutting down");
+            tracing::debug!("CancellationToken triggered; shutting down");
         },
     }