readme update
Browse files
README.md
CHANGED
|
@@ -10,7 +10,7 @@ In tackling the challenge of *agent out-of-sync*, **SyncBench** is released as a
|
|
| 10 |
**SyncBench** for *agent out-of-sync* recovery:
|
| 11 |
- Task: Agent *out-of-sync*
|
| 12 |
- SyncBench
|
| 13 |
-
- In our current version, we construct *SyncBench* based on 21 popular repositories, while our dataset construction approach allows adaptive upsampling and downsampling for custom needs
|
| 14 |
- Expand to incorporate other repositories as data sources
|
| 15 |
- Downsampled to smaller ealuation subsets
|
| 16 |
- Datasets: Caller and Callee
|
|
@@ -23,7 +23,7 @@ In tackling the challenge of *agent out-of-sync*, **SyncBench** is released as a
|
|
| 23 |
- Fields:
|
| 24 |
- **instance_id:** SyncBench instance ID
|
| 25 |
- **instance_type:** data type (i.e., `caller` or `callee`)
|
| 26 |
-
- **task_type:** task type (i.e., `fail-to-pass
|
| 27 |
- **repo_url:** URL link of the source GitHub repository
|
| 28 |
- **commit:** Git commit ID
|
| 29 |
- **fm_type:** Function type (i.e., `function` or `method`)
|
|
|
|
| 10 |
**SyncBench** for *agent out-of-sync* recovery:
|
| 11 |
- Task: Agent *out-of-sync*
|
| 12 |
- SyncBench
|
| 13 |
+
- In our current version, we construct *SyncBench* based on 21 popular GitHub repositories, while our dataset construction approach allows adaptive upsampling and downsampling for custom needs
|
| 14 |
- Expand to incorporate other repositories as data sources
|
| 15 |
- Downsampled to smaller ealuation subsets
|
| 16 |
- Datasets: Caller and Callee
|
|
|
|
| 23 |
- Fields:
|
| 24 |
- **instance_id:** SyncBench instance ID
|
| 25 |
- **instance_type:** data type (i.e., `caller` or `callee`)
|
| 26 |
+
- **task_type:** task type (i.e., `fail-to-pass` or `pass-to-pass`)
|
| 27 |
- **repo_url:** URL link of the source GitHub repository
|
| 28 |
- **commit:** Git commit ID
|
| 29 |
- **fm_type:** Function type (i.e., `function` or `method`)
|