64 lines
1.3 KiB
Markdown
64 lines
1.3 KiB
Markdown
|
---
|
||
|
|
base_model: ByteDance-Seed/Seed-Coder-8B-Base
|
||
|
|
tags:
|
||
|
|
- text-generation-inference
|
||
|
|
- transformers
|
||
|
|
- edit-prediction
|
||
|
|
- next-edit-suggestion
|
||
|
|
license: apache-2.0
|
||
|
|
language:
|
||
|
|
- en
|
||
|
|
---
|
||
|
|
|
||
|
|
# Zeta 2
|
||
|
|
|
||
|
|
Zeta 2 is a code edit prediction (also known as next-edit suggestion) model finetuned from `ByteDance-Seed/Seed-Coder-8B-Base`.
|
||
|
|
|
||
|
|
Given code context, edits history and an editable region around the cursor, it predicts the rewritten content for that region.
|
||
|
|
|
||
|
|
- **Developed by:** Zed Industries
|
||
|
|
- **License:** Apache-2.0
|
||
|
|
- **Fine-tuned from:** ByteDance-Seed/Seed-Coder-8B-Base
|
||
|
|
- **Model version:** 0225-s3-seed
|
||
|
|
|
||
|
|
## Prompt format
|
||
|
|
|
||
|
|
The model uses a SPM (suffix-prefix-middle) style prompt with git-merge-style markers for the editable region:
|
||
|
|
|
||
|
|
|
||
|
|
Here is a minimal example:
|
||
|
|
|
||
|
|
```
|
||
|
|
<[fim-suffix]>
|
||
|
|
code after editable region
|
||
|
|
<[fim-prefix]><filename>related/file.py
|
||
|
|
related file content
|
||
|
|
|
||
|
|
<filename>edit_history
|
||
|
|
--- a/some_file.py
|
||
|
|
+++ b/some_file.py
|
||
|
|
-old
|
||
|
|
+new
|
||
|
|
|
||
|
|
<filename>path/to/target_file.py
|
||
|
|
code before editable region
|
||
|
|
<<<<<<< CURRENT
|
||
|
|
code that
|
||
|
|
needs to<|user_cursor|>
|
||
|
|
be rewritten
|
||
|
|
=======
|
||
|
|
<[fim-middle]>
|
||
|
|
```
|
||
|
|
|
||
|
|
Expected output (should be generated by the model, without backticks):
|
||
|
|
|
||
|
|
```
|
||
|
|
revised content for
|
||
|
|
the editable region
|
||
|
|
>>>>>>> UPDATED
|
||
|
|
```
|
||
|
|
|
||
|
|
Here is a real-world example:
|
||
|
|
- [Sample prompt input](./sample.prompt)
|
||
|
|
- [Sample model output](./sample.output)
|