Upload configuration file
Browse files- config/code_alpaca.yaml +36 -9
config/code_alpaca.yaml
CHANGED
|
@@ -8,14 +8,41 @@ operators:
|
|
| 8 |
- id: drop_columns
|
| 9 |
config:
|
| 10 |
type: function
|
| 11 |
-
function: data_strategies.commons.
|
| 12 |
function_config:
|
| 13 |
-
|
| 14 |
-
-
|
| 15 |
-
-
|
| 16 |
-
- seed
|
| 17 |
-
- concepts
|
| 18 |
-
- prompt
|
| 19 |
-
- id
|
| 20 |
input_ids:
|
| 21 |
-
- load_hf_python
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
- id: drop_columns
|
| 9 |
config:
|
| 10 |
type: function
|
| 11 |
+
function: data_strategies.commons.keep_only_columns
|
| 12 |
function_config:
|
| 13 |
+
columns_to_keep:
|
| 14 |
+
- instruction
|
| 15 |
+
- response
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
input_ids:
|
| 17 |
+
- load_hf_python
|
| 18 |
+
- id: rename_task_column
|
| 19 |
+
config:
|
| 20 |
+
type: function
|
| 21 |
+
function: data_strategies.commons.force_rename_columns
|
| 22 |
+
function_config:
|
| 23 |
+
column_maps:
|
| 24 |
+
response: response_seed
|
| 25 |
+
instruction: instruction_seed
|
| 26 |
+
input_ids:
|
| 27 |
+
- drop_columns
|
| 28 |
+
- id: sample_dataset
|
| 29 |
+
config:
|
| 30 |
+
type: function
|
| 31 |
+
function: data_strategies.commons.uniform_sample_limited
|
| 32 |
+
function_config:
|
| 33 |
+
num_samples: 50000 # 1.25% of 10000 target from stage 1
|
| 34 |
+
input_ids:
|
| 35 |
+
- rename_task_column
|
| 36 |
+
- id: annotate
|
| 37 |
+
config:
|
| 38 |
+
type: completions
|
| 39 |
+
map: chat
|
| 40 |
+
map_config:
|
| 41 |
+
system_message: "You are a helpful assistant."
|
| 42 |
+
user_message_column: instruction_seed
|
| 43 |
+
output_column: response
|
| 44 |
+
model: gpt-4o-mini
|
| 45 |
+
temperature: 1.0
|
| 46 |
+
batch: True
|
| 47 |
+
input_ids:
|
| 48 |
+
- sample_dataset
|