Skip to content

Commit 80040bf

Browse files
mikepapadimmikepapadim
committed
Fix jbang
1 parent 5e5c0b1 commit 80040bf

3 files changed

Lines changed: 135 additions & 133 deletions

File tree

‎LlamaTornadoCli.java‎

Lines changed: 5 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -5,6 +5,8 @@
55
//DEPS io.github.beehive-lab:tornado-runtime:2.1.0
66

77
//SOURCES TornadoFlags.java
8+
// === Set to not get annoying warnings about annotation processing
9+
//JAVAC_OPTIONS -proc:full
810

911
// Compiler options
1012
//JAVAC_OPTIONS --enable-preview
@@ -87,7 +89,7 @@ public static void main(String[] args) throws IOException {
8789

8890
// Check if help requested
8991
if (args.length == 0 || hasHelpFlag(args)) {
90-
Options.printUsage(System.out);
92+
// Options.printUsage(System.out);
9193
System.exit(0);
9294
}
9395

@@ -137,8 +139,8 @@ private static boolean hasHelpFlag(String[] args) {
137139
private static void printBanner() {
138140
System.out.println("""
139141
╔══════════════════════════════════════════════════════════╗
140-
║ Llama-Tornado CLI - GPU-Accelerated LLM ║
141-
║ Powered by TornadoVM & Java 21 ║
142+
║ Llama-Tornado CLI - GPU-Accelerated LLM ║
143+
║ Powered by TornadoVM & Java 21 ║
142144
╚══════════════════════════════════════════════════════════╝
143145
""");
144146
}

‎src/main/java/org/beehive/gpullama3/Options.java‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -28,7 +28,7 @@ private static boolean getDefaultTornadoVM() {
2828
return Boolean.parseBoolean(System.getProperty("use.tornadovm", "false"));
2929
}
3030

31-
static void printUsage(PrintStream out) {
31+
public static void printUsage(PrintStream out) {
3232
out.println("Usage: jbang Llama3.java [options]");
3333
out.println();
3434
out.println("Options:");

‎src/main/java/org/beehive/gpullama3/cli/LlamaTornadoCli.java‎

Lines changed: 129 additions & 129 deletions
Original file line numberDiff line numberDiff line change
@@ -1,129 +1,129 @@
1-
package org.beehive.gpullama3.cli;
2-
3-
import org.beehive.gpullama3.Options;
4-
import org.beehive.gpullama3.auxiliary.LastRunMetrics;
5-
import org.beehive.gpullama3.inference.sampler.Sampler;
6-
import org.beehive.gpullama3.model.Model;
7-
8-
import java.io.IOException;
9-
10-
import static org.beehive.gpullama3.inference.sampler.Sampler.createSampler;
11-
import static org.beehive.gpullama3.model.loader.ModelLoader.loadModel;
12-
13-
/**
14-
* LlamaTornadoCli - Pure Java CLI for running llama-tornado models
15-
*
16-
* This class provides a standalone command-line interface for running LLaMA models
17-
* with TornadoVM acceleration. This version is compiled as part of the Maven build.
18-
*
19-
* For JBang usage, use the LlamaTornadoCli.java file in the root directory.
20-
*
21-
* Usage as compiled application:
22-
* java --enable-preview --add-modules jdk.incubator.vector \
23-
* -cp target/gpu-llama3-0.3.1.jar \
24-
* org.beehive.gpullama3.cli.LlamaTornadoCli \
25-
* --model path/to/model.gguf --prompt "Your prompt here"
26-
*
27-
* Examples:
28-
* # Interactive chat mode
29-
* java -cp target/gpu-llama3-0.3.1.jar \
30-
* org.beehive.gpullama3.cli.LlamaTornadoCli \
31-
* -m model.gguf --interactive
32-
*
33-
* # Single instruction mode
34-
* java -cp target/gpu-llama3-0.3.1.jar \
35-
* org.beehive.gpullama3.cli.LlamaTornadoCli \
36-
* -m model.gguf -p "Explain quantum computing"
37-
*
38-
* # With TornadoVM acceleration (requires TornadoVM runtime setup)
39-
* java -cp target/gpu-llama3-0.3.1.jar \
40-
* org.beehive.gpullama3.cli.LlamaTornadoCli \
41-
* -m model.gguf -p "Hello" --use-tornadovm true
42-
*/
43-
public class LlamaTornadoCli {
44-
45-
// Configuration flags
46-
public static final boolean USE_VECTOR_API = Boolean.parseBoolean(
47-
System.getProperty("llama.VectorAPI", "true"));
48-
public static final boolean SHOW_PERF_INTERACTIVE = Boolean.parseBoolean(
49-
System.getProperty("llama.ShowPerfInteractive", "true"));
50-
51-
/**
52-
* Run a single instruction and display the response
53-
*/
54-
private static void runSingleInstruction(Model model, Sampler sampler, Options options) {
55-
String response = model.runInstructOnce(sampler, options);
56-
System.out.println(response);
57-
if (SHOW_PERF_INTERACTIVE) {
58-
LastRunMetrics.printMetrics();
59-
}
60-
}
61-
62-
/**
63-
* Main entry point for the CLI application
64-
*
65-
* @param args command-line arguments (see Options.parseOptions for details)
66-
* @throws IOException if model loading fails
67-
*/
68-
public static void main(String[] args) throws IOException {
69-
// Print banner
70-
printBanner();
71-
72-
// Check if help requested
73-
if (args.length == 0 || hasHelpFlag(args)) {
74-
Options.printUsage(System.out);
75-
System.exit(0);
76-
}
77-
78-
try {
79-
// Parse options
80-
Options options = Options.parseOptions(args);
81-
82-
// Load model
83-
System.out.println("Loading model from: " + options.modelPath());
84-
Model model = loadModel(options);
85-
System.out.println("Model loaded successfully!");
86-
87-
// Create sampler
88-
Sampler sampler = createSampler(model, options);
89-
90-
// Run in interactive or single-instruction mode
91-
if (options.interactive()) {
92-
System.out.println("Starting interactive chat mode...");
93-
System.out.println("Type your messages below (Ctrl+C to exit):");
94-
System.out.println();
95-
model.runInteractive(sampler, options);
96-
} else {
97-
runSingleInstruction(model, sampler, options);
98-
}
99-
} catch (Exception e) {
100-
System.err.println("Error: " + e.getMessage());
101-
e.printStackTrace();
102-
System.exit(1);
103-
}
104-
}
105-
106-
/**
107-
* Check if help flag is present in arguments
108-
*/
109-
private static boolean hasHelpFlag(String[] args) {
110-
for (String arg : args) {
111-
if (arg.equals("--help") || arg.equals("-h")) {
112-
return true;
113-
}
114-
}
115-
return false;
116-
}
117-
118-
/**
119-
* Print ASCII banner
120-
*/
121-
private static void printBanner() {
122-
System.out.println("""
123-
╔══════════════════════════════════════════════════════════╗
124-
║ Llama-Tornado CLI - GPU-Accelerated LLM ║
125-
║ Powered by TornadoVM & Java 21 ║
126-
╚══════════════════════════════════════════════════════════╝
127-
""");
128-
}
129-
}
1+
//package org.beehive.gpullama3.cli;
2+
//
3+
//import org.beehive.gpullama3.Options;
4+
//import org.beehive.gpullama3.auxiliary.LastRunMetrics;
5+
//import org.beehive.gpullama3.inference.sampler.Sampler;
6+
//import org.beehive.gpullama3.model.Model;
7+
//
8+
//import java.io.IOException;
9+
//
10+
//import static org.beehive.gpullama3.inference.sampler.Sampler.createSampler;
11+
//import static org.beehive.gpullama3.model.loader.ModelLoader.loadModel;
12+
//
13+
///**
14+
// * LlamaTornadoCli - Pure Java CLI for running llama-tornado models
15+
// *
16+
// * This class provides a standalone command-line interface for running LLaMA models
17+
// * with TornadoVM acceleration. This version is compiled as part of the Maven build.
18+
// *
19+
// * For JBang usage, use the LlamaTornadoCli.java file in the root directory.
20+
// *
21+
// * Usage as compiled application:
22+
// * java --enable-preview --add-modules jdk.incubator.vector \
23+
// * -cp target/gpu-llama3-0.3.1.jar \
24+
// * org.beehive.gpullama3.cli.LlamaTornadoCli \
25+
// * --model path/to/model.gguf --prompt "Your prompt here"
26+
// *
27+
// * Examples:
28+
// * # Interactive chat mode
29+
// * java -cp target/gpu-llama3-0.3.1.jar \
30+
// * org.beehive.gpullama3.cli.LlamaTornadoCli \
31+
// * -m model.gguf --interactive
32+
// *
33+
// * # Single instruction mode
34+
// * java -cp target/gpu-llama3-0.3.1.jar \
35+
// * org.beehive.gpullama3.cli.LlamaTornadoCli \
36+
// * -m model.gguf -p "Explain quantum computing"
37+
// *
38+
// * # With TornadoVM acceleration (requires TornadoVM runtime setup)
39+
// * java -cp target/gpu-llama3-0.3.1.jar \
40+
// * org.beehive.gpullama3.cli.LlamaTornadoCli \
41+
// * -m model.gguf -p "Hello" --use-tornadovm true
42+
// */
43+
//public class LlamaTornadoCli {
44+
//
45+
// // Configuration flags
46+
// public static final boolean USE_VECTOR_API = Boolean.parseBoolean(
47+
// System.getProperty("llama.VectorAPI", "true"));
48+
// public static final boolean SHOW_PERF_INTERACTIVE = Boolean.parseBoolean(
49+
// System.getProperty("llama.ShowPerfInteractive", "true"));
50+
//
51+
// /**
52+
// * Run a single instruction and display the response
53+
// */
54+
// private static void runSingleInstruction(Model model, Sampler sampler, Options options) {
55+
// String response = model.runInstructOnce(sampler, options);
56+
// System.out.println(response);
57+
// if (SHOW_PERF_INTERACTIVE) {
58+
// LastRunMetrics.printMetrics();
59+
// }
60+
// }
61+
//
62+
// /**
63+
// * Main entry point for the CLI application
64+
// *
65+
// * @param args command-line arguments (see Options.parseOptions for details)
66+
// * @throws IOException if model loading fails
67+
// */
68+
// public static void main(String[] args) throws IOException {
69+
// // Print banner
70+
// printBanner();
71+
//
72+
// // Check if help requested
73+
// if (args.length == 0 || hasHelpFlag(args)) {
74+
//// Options.printUsage(System.out);
75+
// System.exit(0);
76+
// }
77+
//
78+
// try {
79+
// // Parse options
80+
// Options options = Options.parseOptions(args);
81+
//
82+
// // Load model
83+
// System.out.println("Loading model from: " + options.modelPath());
84+
// Model model = loadModel(options);
85+
// System.out.println("Model loaded successfully!");
86+
//
87+
// // Create sampler
88+
// Sampler sampler = createSampler(model, options);
89+
//
90+
// // Run in interactive or single-instruction mode
91+
// if (options.interactive()) {
92+
// System.out.println("Starting interactive chat mode...");
93+
// System.out.println("Type your messages below (Ctrl+C to exit):");
94+
// System.out.println();
95+
// model.runInteractive(sampler, options);
96+
// } else {
97+
// runSingleInstruction(model, sampler, options);
98+
// }
99+
// } catch (Exception e) {
100+
// System.err.println("Error: " + e.getMessage());
101+
// e.printStackTrace();
102+
// System.exit(1);
103+
// }
104+
// }
105+
//
106+
// /**
107+
// * Check if help flag is present in arguments
108+
// */
109+
// private static boolean hasHelpFlag(String[] args) {
110+
// for (String arg : args) {
111+
// if (arg.equals("--help") || arg.equals("-h")) {
112+
// return true;
113+
// }
114+
// }
115+
// return false;
116+
// }
117+
//
118+
// /**
119+
// * Print ASCII banner
120+
// */
121+
// private static void printBanner() {
122+
// System.out.println("""
123+
// ╔══════════════════════════════════════════════════════════╗
124+
// ║ Llama-Tornado CLI - GPU-Accelerated LLM ║
125+
// ║ Powered by TornadoVM & Java 21 ║
126+
// ╚══════════════════════════════════════════════════════════╝
127+
// """);
128+
// }
129+
//}

0 commit comments

Comments
 (0)