Update README.md
Browse files
README.md
CHANGED
|
@@ -54,7 +54,7 @@ The EchoLLaMA pipeline integrates four specialized models:
|
|
| 54 |
The LLaMA-3.2-1B-Instruct model was fine-tuned using:
|
| 55 |
|
| 56 |
- **Technique**: Direct Preference Optimization (DPO) with LoRA
|
| 57 |
-
- **Dataset**: 2000 samples from COCO 2017 processed with DETR,
|
| 58 |
- **Chosen Responses**: Generated by DeepSeek-V3-0324
|
| 59 |
- **Rejected Responses**: Generated by pre-fine-tuned LLaMA-3.2-1B-Instruct
|
| 60 |
- **Training Parameters**:
|
|
|
|
| 54 |
The LLaMA-3.2-1B-Instruct model was fine-tuned using:
|
| 55 |
|
| 56 |
- **Technique**: Direct Preference Optimization (DPO) with LoRA
|
| 57 |
+
- **Dataset**: 2000 samples from COCO 2017 processed with DETR, and Moondream
|
| 58 |
- **Chosen Responses**: Generated by DeepSeek-V3-0324
|
| 59 |
- **Rejected Responses**: Generated by pre-fine-tuned LLaMA-3.2-1B-Instruct
|
| 60 |
- **Training Parameters**:
|