Update README.md
Browse files
README.md
CHANGED
@@ -85,4 +85,56 @@ def gen(old, new, max_new_tokens=200, temperature=0.45, top_p=0.90):
|
|
85 |
return [tokenizer.decode(out[len(toks[0]):], skip_special_tokens=True) for out in outs]
|
86 |
```
|
87 |
|
88 |
-
use the "gen" function with the old and new code
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
85 |
return [tokenizer.decode(out[len(toks[0]):], skip_special_tokens=True) for out in outs]
|
86 |
```
|
87 |
|
88 |
+
use the "gen" function with the old and new code
|
89 |
+
|
90 |
+
# Example:
|
91 |
+
```py
|
92 |
+
- import datasets
|
93 |
+
- from pathlib import Path
|
94 |
+
from code_editing.models import CodeLlamaEditModel, LlamaChatModel, EditModel, EditCommand, ChatAdaptorEditModel, OctoCoderChatModel, codellama_edit_prompt_diff, apply_rel_diff_trim, OpenAIChatModel, StarCoderCommitEditModel
|
95 |
+
from code_editing.humanevalpack import batch_prompts_from_example
|
96 |
+
from code_editing.utils import gunzip_json_write
|
97 |
+
from typing import List, Callable
|
98 |
+
from tqdm import tqdm
|
99 |
+
|
100 |
+
|
101 |
+
# NOTE: this is the factory for each model type. to add a new model type, add a new case here
|
102 |
+
# and implement it in models.py. Also, add a new case in the argument parser below.
|
103 |
+
- def model_factory(model_type: str, quantize=False, num_gpus=1) -> Callable[[str], EditModel]:
|
104 |
+
+ def model_factory(
|
105 |
+
+ model_type: str,
|
106 |
+
+ quantize=False,
|
107 |
+
+ num_gpus=1,
|
108 |
+
+ system_supported=True,
|
109 |
+
+ ) -> Callable[[str], EditModel]:
|
110 |
+
if model_type == "codellama" or model_type == "deepseek":
|
111 |
+
return CodeLlamaEditModel
|
112 |
+
elif model_type == "starcoder":
|
113 |
+
return StarCoderCommitEditModel
|
114 |
+
elif model_type == "codellama-diff":
|
115 |
+
return (lambda path: CodeLlamaEditModel(path, prompt_format=codellama_edit_prompt_diff, post_process=apply_rel_diff_trim))
|
116 |
+
elif model_type == "openai":
|
117 |
+
return (lambda path: ChatAdaptorEditModel(OpenAIChatModel(path)))
|
118 |
+
elif model_type == "codellama-chat":
|
119 |
+
- return (lambda path: ChatAdaptorEditModel(LlamaChatModel(path, quantization=quantize, num_gpus=num_gpus)))
|
120 |
+
+ return (lambda path: ChatAdaptorEditModel(LlamaChatModel(path, quantization=quantize, num_gpus=num_gpus, system_supported=system_supported)))
|
121 |
+
elif model_type == "octocoder":
|
122 |
+
return (lambda path: ChatAdaptorEditModel(OctoCoderChatModel(path, quantization=quantize, num_gpus=num_gpus)))
|
123 |
+
else:
|
124 |
+
raise ValueError(f"Unknown model type: {model_type}")
|
125 |
+
|
126 |
+
def complete_problem(example: EditCommand, model: EditModel, batch_size: int, completion_limit: int, **kwargs) -> List[str]:
|
127 |
+
batches = batch_prompts_from_example(example, batch_size, completion_limit)
|
128 |
+
|
129 |
+
completions = []
|
130 |
+
for batch in batches:
|
131 |
+
resps = model.generate(batch, **kwargs)
|
132 |
+
for resp in resps:
|
133 |
+
completions.append(resp["content"])
|
134 |
+
|
135 |
+
return completions
|
136 |
+
```
|
137 |
+
Produced:
|
138 |
+
```
|
139 |
+
Add system_supported argument to model_factory
|
140 |
+
```
|