Update README.md
Browse files
README.md
CHANGED
|
@@ -63,6 +63,12 @@ Workflow:
|
|
| 63 |
| 2. Assignment 2 Model | Fine-tuned on Silver + Gold (Simple LLM) | 0.4975369710 |
|
| 64 |
| 3. Assignment 3 Model | Fine-tuned on Silver + Gold (QLoRA) | 0.5006382068 |
|
| 65 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 66 |
## Intended Use
|
| 67 |
- Educational/research use for green patent classification experiments.
|
| 68 |
- Binary label output: non-green (0) vs green (1).
|
|
|
|
| 63 |
| 2. Assignment 2 Model | Fine-tuned on Silver + Gold (Simple LLM) | 0.4975369710 |
|
| 64 |
| 3. Assignment 3 Model | Fine-tuned on Silver + Gold (QLoRA) | 0.5006382068 |
|
| 65 |
|
| 66 |
+
### Reflection (2–3 sentences)
|
| 67 |
+
|
| 68 |
+
Compared to Assignment 2, the Assignment 3 QLoRA workflow produced a small improvement in eval macro F1 (+0.0031).
|
| 69 |
+
This indicates that the advanced data-generation approach provided a measurable but modest downstream gain over the simpler Assignment 2 setup in this run.
|
| 70 |
+
However, both fine-tuned pipelines remained substantially below the frozen-embedding baseline, suggesting that data quality and labeling strategy still dominate final performance.
|
| 71 |
+
|
| 72 |
## Intended Use
|
| 73 |
- Educational/research use for green patent classification experiments.
|
| 74 |
- Binary label output: non-green (0) vs green (1).
|