Commit
•
4538042
1
Parent(s):
809e1a0
Update README.md
Browse files
README.md
CHANGED
|
@@ -23,13 +23,13 @@ This is a test model and may generate incorrect responses. Use at your own risk.
|
|
| 23 |
|
| 24 |
## Train Details
|
| 25 |
|
| 26 |
-
Base: Qwen1.5-1.8B
|
| 27 |
-
Training Data: ~20k [code examples](https://huggingface.co/datasets/reciprocate/dpo_ultra-capybara-code_filtered-best)
|
| 28 |
-
Epochs: 1
|
| 29 |
-
Method: ORPO
|
| 30 |
-
Hardware: 2 x A40
|
| 31 |
-
Quantization: 4-bit QLora
|
| 32 |
-
Lora Rank/Alpha: 16
|
| 33 |
|
| 34 |
# Limitations
|
| 35 |
|
|
|
|
| 23 |
|
| 24 |
## Train Details
|
| 25 |
|
| 26 |
+
- Base: Qwen1.5-1.8B
|
| 27 |
+
- Training Data: ~20k [code examples](https://huggingface.co/datasets/reciprocate/dpo_ultra-capybara-code_filtered-best)
|
| 28 |
+
- Epochs: 1
|
| 29 |
+
- Method: ORPO
|
| 30 |
+
- Hardware: 2 x A40
|
| 31 |
+
- Quantization: 4-bit QLora
|
| 32 |
+
- Lora Rank/Alpha: 16
|
| 33 |
|
| 34 |
# Limitations
|
| 35 |
|