doc(README): unify usage section and refine evaluation to highlight prototyping intent and need for case-by-case validation
Browse files
README.md
CHANGED
|
@@ -26,8 +26,6 @@ This is a parameter-efficient fine-tuning (PEFT) LoRA adapter trained on a struc
|
|
| 26 |
|
| 27 |
## 🚀 Uses
|
| 28 |
|
| 29 |
-
### Direct Use
|
| 30 |
-
|
| 31 |
This adapter enhances Phi-4-mini-instruct for parsing natural recurrence expressions into DSL format.
|
| 32 |
|
| 33 |
Example prompt:
|
|
@@ -38,9 +36,7 @@ $ You are a precise parser of recurring schedule expressions. Your only job is t
|
|
| 38 |
< MONTHLY_BY_WEEKDAY(1, TU, 2, TIME(13, 0))
|
| 39 |
```
|
| 40 |
|
| 41 |
-
|
| 42 |
-
|
| 43 |
-
Useful for prototyping DSL-based schedulers, assistants, or natural language interfaces requiring structured recurrence interpretation.
|
| 44 |
|
| 45 |
The DSL produced by this model is defined and executed by the [Mini Recurrence Converter DSL module](https://github.com/gh9869827/fifo-dev-dsl/tree/main/fifo_dev_dsl/domain_specific/mini_recurrence_converter_dsl).
|
| 46 |
|
|
@@ -61,7 +57,7 @@ print(parsed)
|
|
| 61 |
# Output: structured recurrence rule object
|
| 62 |
```
|
| 63 |
|
| 64 |
-
|
| 65 |
|
| 66 |
This model is specialized for recurrence conversion DSL. It is not intended for general-purpose dialogue or unrelated tasks.
|
| 67 |
|
|
@@ -111,7 +107,7 @@ Trained on [a6188466/mini-recurrence-converter-dsl-dataset](https://huggingface.
|
|
| 111 |
- **97.79%** on held-out [test set](https://huggingface.co/datasets/a6188466/mini-recurrence-converter-dsl-dataset) (**221/226** passed)
|
| 112 |
The test set includes 26 hand-curated and 200 [synthetic](https://github.com/gh9869827/fifo-dev-dsl/blob/main/fifo_dev_dsl/domain_specific/mini_recurrence_converter_dsl/generate_synthetic_data.py) examples.
|
| 113 |
|
| 114 |
-
These results demonstrate strong accuracy on tested queries,
|
| 115 |
|
| 116 |
Evaluation script: [`evaluate_mini_recurrence_converter_dsl_model.py`](https://github.com/gh9869827/fifo-dev-dsl/blob/main/fifo_dev_dsl/domain_specific/mini_recurrence_converter_dsl/evaluate_mini_recurrence_converter_dsl_model.py)
|
| 117 |
|
|
|
|
| 26 |
|
| 27 |
## 🚀 Uses
|
| 28 |
|
|
|
|
|
|
|
| 29 |
This adapter enhances Phi-4-mini-instruct for parsing natural recurrence expressions into DSL format.
|
| 30 |
|
| 31 |
Example prompt:
|
|
|
|
| 36 |
< MONTHLY_BY_WEEKDAY(1, TU, 2, TIME(13, 0))
|
| 37 |
```
|
| 38 |
|
| 39 |
+
It is a useful starting point for prototyping DSL-based schedulers, assistants, or natural language interfaces requiring structured recurrence interpretation.
|
|
|
|
|
|
|
| 40 |
|
| 41 |
The DSL produced by this model is defined and executed by the [Mini Recurrence Converter DSL module](https://github.com/gh9869827/fifo-dev-dsl/tree/main/fifo_dev_dsl/domain_specific/mini_recurrence_converter_dsl).
|
| 42 |
|
|
|
|
| 57 |
# Output: structured recurrence rule object
|
| 58 |
```
|
| 59 |
|
| 60 |
+
## Out-of-Scope Use
|
| 61 |
|
| 62 |
This model is specialized for recurrence conversion DSL. It is not intended for general-purpose dialogue or unrelated tasks.
|
| 63 |
|
|
|
|
| 107 |
- **97.79%** on held-out [test set](https://huggingface.co/datasets/a6188466/mini-recurrence-converter-dsl-dataset) (**221/226** passed)
|
| 108 |
The test set includes 26 hand-curated and 200 [synthetic](https://github.com/gh9869827/fifo-dev-dsl/blob/main/fifo_dev_dsl/domain_specific/mini_recurrence_converter_dsl/generate_synthetic_data.py) examples.
|
| 109 |
|
| 110 |
+
These results demonstrate strong accuracy on tested queries, but performance should be carefully evaluated in each individual use case.
|
| 111 |
|
| 112 |
Evaluation script: [`evaluate_mini_recurrence_converter_dsl_model.py`](https://github.com/gh9869827/fifo-dev-dsl/blob/main/fifo_dev_dsl/domain_specific/mini_recurrence_converter_dsl/evaluate_mini_recurrence_converter_dsl_model.py)
|
| 113 |
|