Change lora repository name in readme
Browse files
README.md
CHANGED
|
@@ -43,7 +43,7 @@ This Multi-Task model achieves **85.4% overall accuracy**, significantly outperf
|
|
| 43 |
| **Twitter Financial News** | **83.3%** | 80.5% |
|
| 44 |
| **FiQA** (Forums) | **81.5%** | 72.6% |
|
| 45 |
|
| 46 |
-
> **Note:** For edge deployment or low-memory environments, check out the [LoRA version](https://huggingface.co/pmatorras/
|
| 47 |
|
| 48 |
## Architecture
|
| 49 |
|
|
@@ -81,9 +81,3 @@ print(probabilities)
|
|
| 81 |
- **Optimizer**: AdamW
|
| 82 |
- **Loss Function**: Weighted sum of Cross-Entropy (Classification) and MSE (Regression).
|
| 83 |
- **Compute**: Trained on NVIDIA RTX 4050.
|
| 84 |
-
|
| 85 |
-
|
| 86 |
-
### Key Features of this Card:
|
| 87 |
-
- **YAML Header:** Enables the "Hosted Inference API" widget on the right side of the page so people can test it instantly.
|
| 88 |
-
- **Cross-Linking:** It links to your GitHub, your Website, and the LoRA version.
|
| 89 |
-
- **Usage Snippet:** Essential for developers to copy-paste and run.
|
|
|
|
| 43 |
| **Twitter Financial News** | **83.3%** | 80.5% |
|
| 44 |
| **FiQA** (Forums) | **81.5%** | 72.6% |
|
| 45 |
|
| 46 |
+
> **Note:** For edge deployment or low-memory environments, check out the [LoRA version](https://huggingface.co/pmatorras/financial-sentiment-analysis-lora) which reduces storage by 99% (5MB vs 420MB).
|
| 47 |
|
| 48 |
## Architecture
|
| 49 |
|
|
|
|
| 81 |
- **Optimizer**: AdamW
|
| 82 |
- **Loss Function**: Weighted sum of Cross-Entropy (Classification) and MSE (Regression).
|
| 83 |
- **Compute**: Trained on NVIDIA RTX 4050.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|