a6188466 commited on
Commit
f2550de
·
verified ·
1 Parent(s): 5e1e2c5

docs(README): rewrite adapter READMEs for clarity and alignment

Browse files
Files changed (1) hide show
  1. README.md +20 -14
README.md CHANGED
@@ -17,9 +17,11 @@ language:
17
 
18
  ## 🧠 Model Details
19
 
20
- This is a parameter-efficient fine-tuning (PEFT) LoRA adapter trained on a structured DSL for recurrence interpretation, such as natural language like "every Monday at 10am" or "on the last Friday of each month". It translates natural recurrence expressions into structured function calls using a custom DSL (e.g., `WEEKLY(...)`, `MONTHLY_BY_WEEKDAY(...)`).
21
 
22
- This adapter is designed as a **demonstration and testbed** for structured recurrence parsing using the custom DSL.
 
 
23
 
24
  - **Model type:** LoRA adapter (PEFT)
25
  - **Language(s):** English
@@ -28,30 +30,26 @@ This adapter is designed as a **demonstration and testbed** for structured recur
28
 
29
  ## 🚀 Uses
30
 
31
- This adapter fine-tunes Phi-4-mini-instruct for parsing natural recurrence expressions into DSL format.
 
32
 
33
  Example prompt:
34
 
35
  ```
36
- $ You are a precise parser of recurring schedule expressions. Your only job is to translate natural language recurrence expressions into structured DSL function calls.
37
  > every second Tuesday of the month at 1pm
38
  < MONTHLY_BY_WEEKDAY(1, TU, 2, TIME(13, 0))
39
  ```
40
 
41
- It is a useful starting point for prototyping DSL-based schedulers, assistants, or natural language interfaces requiring structured recurrence interpretation.
42
-
43
- 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).
44
-
45
-
46
  ## ⚠️ Out-of-Scope Use
47
 
48
- This model has been fine-tuned specifically for recurrence conversion DSL as a demonstration and testbed. It is not intended for general-purpose dialogue or unrelated tasks.
49
 
50
  ## 🏗️ Training Details
51
 
52
  Trained on [a6188466/mini-recurrence-converter-dsl-dataset](https://huggingface.co/datasets/a6188466/mini-recurrence-converter-dsl-dataset) using the `dsl` adapter from [`fifo-tool-datasets`](https://github.com/gh9869827/fifo-tool-datasets) and [`fine_tune.py`](https://github.com/gh9869827/fifo-tool-airlock-model-env/blob/main/fifo_tool_airlock_model_env/fine_tuning/phi_4/fine_tune.py).
53
 
54
- - **Dataset:** 279 examples mapping natural language to DSL commands, including hand-curated and [synthetic](https://github.com/gh9869827/fifo-dev-dsl/blob/main/fifo_dev_dsl/domain_specific/mini_recurrence_converter_dsl/generate_synthetic_data.py) samples
55
  - **Epochs:** 15
56
  - **Batch size:** 1
57
  - **Precision:** bf16
@@ -88,15 +86,23 @@ Trained on [a6188466/mini-recurrence-converter-dsl-dataset](https://huggingface.
88
  ## ✅ Evaluation
89
 
90
  - **Eval set:** Natural language queries similar in structure and intent to the training examples
91
- - **Metric:** Functional equivalence — two DSL expressions are considered correct when they evaluate to the same result
92
  - **Results:**
93
- - **97.79%** on a held-out [test set](https://huggingface.co/datasets/a6188466/mini-recurrence-converter-dsl-dataset) (**221/226** passed), containing queries similar in structure and intent to the training examples, using the functional equivalence metric defined above.
94
  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.
95
 
96
- These results demonstrate strong accuracy on tested queries, but performance should be carefully evaluated in each individual use case.
97
 
98
  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)
99
 
 
 
 
 
 
 
 
 
100
  ## 🪪 License
101
 
102
  MIT License. See [LICENSE](LICENSE) for details.
 
17
 
18
  ## 🧠 Model Details
19
 
20
+ This model is a parameter-efficient fine-tuning (PEFT) LoRA adapter, designed as a **demonstration and testbed** for the [Mini Recurrence Converter DSL module](https://github.com/gh9869827/fifo-dev-dsl/tree/main/fifo_dev_dsl/domain_specific/mini_recurrence_converter_dsl).
21
 
22
+ It is provided for demonstration and experimentation purposes only.
23
+
24
+ It translates English recurrence expressions, such as "every Monday at 10am" or "on the last Friday of each month", into structured function calls using that custom DSL.
25
 
26
  - **Model type:** LoRA adapter (PEFT)
27
  - **Language(s):** English
 
30
 
31
  ## 🚀 Uses
32
 
33
+ This adapter fine-tunes `Phi-4-mini-instruct` for parsing natural recurrence expressions into DSL format. It supports the DSL functions defined
34
+ in the [Mini Recurrence Converter DSL module](https://github.com/gh9869827/fifo-dev-dsl/tree/main/fifo_dev_dsl/domain_specific/mini_recurrence_converter_dsl).
35
 
36
  Example prompt:
37
 
38
  ```
39
+ $ You are a precise parser of recurring schedule expressions. Your only job is to translate natural language recurrence expressions into structured DSL function calls such as WEEKLY(...) or MONTHLY_BY_WEEKDAY(...). Do not explain or elaborate. Only return the code.
40
  > every second Tuesday of the month at 1pm
41
  < MONTHLY_BY_WEEKDAY(1, TU, 2, TIME(13, 0))
42
  ```
43
 
 
 
 
 
 
44
  ## ⚠️ Out-of-Scope Use
45
 
46
+ This adapter has been fine-tuned specifically as a demonstration and testbed for the [Mini Recurrence Converter DSL module](https://github.com/gh9869827/fifo-dev-dsl/tree/main/fifo_dev_dsl/domain_specific/mini_recurrence_converter_dsl). It is not intended for general-purpose dialogue or unrelated tasks.
47
 
48
  ## 🏗️ Training Details
49
 
50
  Trained on [a6188466/mini-recurrence-converter-dsl-dataset](https://huggingface.co/datasets/a6188466/mini-recurrence-converter-dsl-dataset) using the `dsl` adapter from [`fifo-tool-datasets`](https://github.com/gh9869827/fifo-tool-datasets) and [`fine_tune.py`](https://github.com/gh9869827/fifo-tool-airlock-model-env/blob/main/fifo_tool_airlock_model_env/fine_tuning/phi_4/fine_tune.py).
51
 
52
+ - **Dataset:** 279 examples mapping English recurrence expressions to DSL commands, including hand-curated and [synthetic](https://github.com/gh9869827/fifo-dev-dsl/blob/main/fifo_dev_dsl/domain_specific/mini_recurrence_converter_dsl/generate_synthetic_data.py) samples
53
  - **Epochs:** 15
54
  - **Batch size:** 1
55
  - **Precision:** bf16
 
86
  ## ✅ Evaluation
87
 
88
  - **Eval set:** Natural language queries similar in structure and intent to the training examples
89
+ - **Metric:** Functional equivalence — two DSL expressions are considered correct if they evaluate to the same result
90
  - **Results:**
91
+ - **97.79%** on held-out [test set](https://huggingface.co/datasets/a6188466/mini-recurrence-converter-dsl-dataset) (**221/226** passed), containing queries similar in structure and intent to the training examples, using the functional equivalence metric defined above.
92
  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.
93
 
94
+ These results demonstrate acceptable performance on evaluated queries. However, as this is a demonstration adapter trained on a narrow dataset, performance should be carefully and independently evaluated in each individual use case.
95
 
96
  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)
97
 
98
+ ## ⚠️ Limitations
99
+
100
+ This adapter is intended solely as a **demonstration and testbed** for the [Mini Recurrence Converter DSL module](https://github.com/gh9869827/fifo-dev-dsl/tree/main/fifo_dev_dsl/domain_specific/mini_recurrence_converter_dsl). It should not be used beyond that scope.
101
+
102
+ It does not cover all phrasings or edge cases of English **recurrence expressions**. Instead, it focuses on illustrative examples that are supported by the module's [DSL functions](https://github.com/gh9869827/fifo-dev-dsl/tree/main/fifo_dev_dsl/domain_specific/mini_recurrence_converter_dsl#-dsl-syntax-and-examples).
103
+
104
+ This adapter is provided *as is*, without warranties or guarantees of any kind. It is intended for demonstration and experimentation only.
105
+
106
  ## 🪪 License
107
 
108
  MIT License. See [LICENSE](LICENSE) for details.