|
277 | 277 | "## <span style=\"font-size:0.8em;\"> </span>\n", |
278 | 278 | "\n", |
279 | 279 | "<div style=\"background: linear-gradient(135deg, #667eea 0%, #764ba2 100%); color: white; padding: 20px; border-radius: 10px; margin: 20px 0;\">\n", |
280 | | - " <h3 style=\"margin: 0; text-align: center;\">📋Section 4. Creating Data Assets from Pipeline Output</h3>\n", |
281 | | - "</div>" |
282 | | - ] |
283 | | - }, |
284 | | - { |
285 | | - "cell_type": "markdown", |
286 | | - "metadata": {}, |
287 | | - "source": [ |
288 | | - "<p>Follow these steps to convert your pipeline outputs into re-usable data assets:</p>\n", |
289 | | - "\n", |
290 | | - "<ol style=\"color: #2c3e50; line-height: 1.8;\">\n", |
291 | | - "<li><strong style=\"color: #9b59b6;\">Access Pipeline Overview:</strong> Double-click on the completed pipeline component to view the pipeline overview.</li>\n", |
292 | | - "\n", |
293 | | - "<li><strong style=\"color: #9b59b6;\">Locate Output Asset:</strong> In the outputs section, find the data asset named <code style=\"background: #f8f9fa; padding: 2px 6px; border-radius: 3px;\">model_output</code></li>\n", |
294 | | - "\n", |
295 | | - "<li><strong style=\"color: #9b59b6;\">Browse Datastore:</strong> Click <span style=\"background: #2e53cc63; color: white; padding: 2px 8px; border-radius: 3px;\">View in datastores browse</span> in the bottom left</li>\n", |
296 | | - "\n", |
297 | | - "<li><strong style=\"color: #9b59b6;\">Navigate Structure:</strong> Explore the tree structure of all folders in the output directory</li>\n", |
298 | | - "\n", |
299 | | - "<li><strong style=\"color: #9b59b6;\">Create Asset:</strong> Navigate to the <code style=\"background: #f8f9fa; padding: 2px 6px; border-radius: 3px;\">model_output</code> folder, open the context menu, and select <span style=\"background: #2e53cc63; color: white; padding: 2px 8px; border-radius: 3px;\">Create a data asset</span></li>\n", |
300 | | - "\n", |
301 | | - "<li><strong style=\"color: #9b59b6;\">Configure Asset:</strong> Enter a name for the output asset in the modal window</li>\n", |
302 | | - "\n", |
303 | | - "<li><strong style=\"color: #9b59b6;\">Complete Creation:</strong> Click the <span style=\"background: #2e53cc63; color: white; padding: 2px 8px; border-radius: 3px;\">Create</span> button</li>\n", |
304 | | - "\n", |
305 | | - "<li><strong style=\"color: #9b59b6;\">Verify Access:</strong> The new data asset will now be visible in the data section of Azure ML Studio</li>\n", |
306 | | - "</ol>\n", |
307 | | - "\n", |
308 | | - "<p><em style=\"color: #7f8c8d;\">💡 These data assets can be reused across multiple workflows and experiments.</em></p>\n" |
309 | | - ] |
310 | | - }, |
311 | | - { |
312 | | - "cell_type": "markdown", |
313 | | - "metadata": {}, |
314 | | - "source": [ |
315 | | - "## <span style=\"font-size:0.8em;\"> </span>\n", |
316 | | - "\n", |
317 | | - "<div style=\"background: linear-gradient(135deg, #667eea 0%, #764ba2 100%); color: white; padding: 20px; border-radius: 10px; margin: 20px 0;\">\n", |
318 | | - " <h3 style=\"margin: 0; text-align: center;\">📊Section 5. Compare Model Performance across Base Model vs GRPO vs Reinforce++ </h3>\n", |
| 280 | + " <h3 style=\"margin: 0; text-align: center;\">📊Section 4. Compare Model Performance across Base Model vs GRPO vs Reinforce++ </h3>\n", |
319 | 281 | "</div>" |
320 | 282 | ] |
321 | 283 | }, |
|
342 | 304 | "<p><em>💡 The evaluation will help identify which RL method produces the most effective model for financial reasoning tasks.</em></p>" |
343 | 305 | ] |
344 | 306 | }, |
| 307 | + { |
| 308 | + "cell_type": "markdown", |
| 309 | + "metadata": {}, |
| 310 | + "source": [ |
| 311 | + "<p> We will now submit evaluation job, with grpo and rlpp model outputs </p>" |
| 312 | + ] |
| 313 | + }, |
345 | 314 | { |
346 | 315 | "cell_type": "code", |
347 | 316 | "execution_count": null, |
|
390 | 359 | "cell_type": "markdown", |
391 | 360 | "metadata": {}, |
392 | 361 | "source": [ |
393 | | - "<p> Fetch metrics from evalution run </p>" |
| 362 | + "<p> Now, lets fetch metrics from evalution run inorder to show comparison</p>" |
394 | 363 | ] |
395 | 364 | }, |
396 | 365 | { |
|
484 | 453 | "## <span style=\"font-size:0.8em;\"> </span>\n", |
485 | 454 | "\n", |
486 | 455 | "<div style=\"background: linear-gradient(135deg, #667eea 0%, #764ba2 100%); color: white; padding: 20px; border-radius: 10px; margin: 20px 0;\">\n", |
487 | | - " <h3 style=\"margin: 0; text-align: center;\">🧩Section 6. Create Draft Model for Speculative Decoding</h3>\n", |
| 456 | + " <h3 style=\"margin: 0; text-align: center;\">🧩Section 5. Create Draft Model for Speculative Decoding</h3>\n", |
488 | 457 | "</div>" |
489 | 458 | ] |
490 | 459 | }, |
|
533 | 502 | "## <span style=\"font-size:0.8em;\"> </span>\n", |
534 | 503 | "\n", |
535 | 504 | "<div style=\"background: linear-gradient(135deg, #667eea 0%, #764ba2 100%); color: white; padding: 20px; border-radius: 10px; margin: 20px 0;\">\n", |
536 | | - " <h3 style=\"margin: 0; text-align: center;\">🔄Section 7. Prepare Combined Model for Deployment</h3>\n", |
| 505 | + " <h3 style=\"margin: 0; text-align: center;\">🔄Section 6. Prepare Combined Model for Deployment</h3>\n", |
537 | 506 | "</div>\n", |
538 | 507 | "\n" |
539 | 508 | ] |
|
587 | 556 | "## <span style=\"font-size:0.8em;\"> </span>\n", |
588 | 557 | "\n", |
589 | 558 | "<div style=\"background: linear-gradient(135deg, #667eea 0%, #764ba2 100%); color: white; padding: 20px; border-radius: 10px; margin: 20px 0;\">\n", |
590 | | - " <h3 style=\"margin: 0; text-align: center;\">🚀Section 8. Deploy Speculative Decoding Endpoint</h3>\n", |
| 559 | + " <h3 style=\"margin: 0; text-align: center;\">🚀Section 7. Deploy Speculative Decoding Endpoint</h3>\n", |
591 | 560 | "</div>\n", |
592 | 561 | "\n" |
593 | 562 | ] |
|
627 | 596 | "## <span style=\"font-size:0.8em;\"> </span>\n", |
628 | 597 | "\n", |
629 | 598 | "<div style=\"background: linear-gradient(135deg, #667eea 0%, #764ba2 100%); color: white; padding: 20px; border-radius: 10px; margin: 20px 0;\">\n", |
630 | | - " <h3 style=\"margin: 0; text-align: center;\">🚀Section 9. Deploy Base Model Endpoint for Comparison</h3>\n", |
| 599 | + " <h3 style=\"margin: 0; text-align: center;\">🚀Section 8. Deploy Base Model Endpoint for Comparison</h3>\n", |
631 | 600 | "</div>\n", |
632 | 601 | "\n" |
633 | 602 | ] |
|
669 | 638 | "## <span style=\"font-size:0.8em;\"> </span>\n", |
670 | 639 | "\n", |
671 | 640 | "<div style=\"background: linear-gradient(135deg, #667eea 0%, #764ba2 100%); color: white; padding: 20px; border-radius: 10px; margin: 20px 0;\">\n", |
672 | | - " <h3 style=\"margin: 0; text-align: center;\">🧪Section 10. Test Deployment</h3>\n", |
| 641 | + " <h3 style=\"margin: 0; text-align: center;\">🧪Section 9. Test Deployment</h3>\n", |
673 | 642 | "</div>" |
674 | 643 | ] |
675 | 644 | }, |
|
704 | 673 | "## <span style=\"font-size:0.8em;\"> </span>\n", |
705 | 674 | "\n", |
706 | 675 | "<div style=\"background: linear-gradient(135deg, #667eea 0%, #764ba2 100%); color: white; padding: 20px; border-radius: 10px; margin: 20px 0;\">\n", |
707 | | - " <h3 style=\"margin: 0; text-align: center;\">📊Section 11. Performance Evaluation Pipeline</h3>\n", |
| 676 | + " <h3 style=\"margin: 0; text-align: center;\">📊Section 10. Performance Evaluation Pipeline</h3>\n", |
708 | 677 | "</div>" |
709 | 678 | ] |
710 | 679 | }, |
|
0 commit comments