Automatic dcft datacuration framework upload for airoboros_coding_instructions_gpt-4o-mini

config/Airoboros/categories/airoboros_coding_instructions_gpt-4o-mini.yaml

@@ -1,21 +1,8 @@
 operators:
-- id: load_airoboros_topics
-  config:
-    type: load_preexisting
-    framework_name: airoboros_topics_gpt-4o-mini
-
-- id: create_topics_strings
+- id: empty_dataset
   config:
     type: function
-    function: data_strategies.Airoboros.utils.create_topics_strings_dataset
-    function_config:
-      input_topic_column: topic
-      output_topics_str_column: topics
-      n_topics_per_string: 10 # this matches batch size in the config (10)
-      n_strings: 100 # TODO change this based on the yeild for the category
-      target_name: instruction
-  input_ids:
-    - load_airoboros_topics
+    function: data_strategies.commons.dummy_dataset
 
 - id: args_dataset
   config:
@@ -24,7 +11,6 @@ operators:
     function_config:
       language: English
       batch_size: 10
-      plain_ratio: 0.5
       topic_avoidance: Avoid any tasks that would be related to climate change, green tech, remewable energy, DEI, sex and/or gender, religion, politics, social issues, race, ethnicity, artificial intelligence, baking/cooking, or any topic that you would likely not respond to, or any task which a language model would not be able to respond to, e.g. tasks about emotions, feelings, physical senses, etc.
       flesch: The output should be written in such a way as to have a Flesch-Kincaid readability score of 30 or lower - best understood by those with college education.  The response must not contain any notes or information about Flesch-Kincaid scores.
       coding_languages:
@@ -65,7 +51,14 @@ operators:
         - swagger/openapi
         - pyramid web framework
   input_ids:
-    - create_topics_strings
+    - empty_dataset
+
+- id: sample_coding_instructions_args
+  config:
+    type: function
+    function: data_strategies.Airoboros.utils.sample_coding_instructions_args
+    function_config:
+      n_samples: 10 # Use this to control the output size
 
 - id: generate_instructions
   config:
@@ -129,8 +122,6 @@ operators:
         - topic_avoidance
         - flesch
         - batch_size
-        - topics
-        - plain_ratio
         - coding_languages
         - related_software
   input_ids: