cfahlgren1 HF staff commited on
Commit
3ae54cc
1 Parent(s): 15917f6

update chatml

Browse files
src/snippets/alpaca_to_chatml.md CHANGED
@@ -11,13 +11,13 @@ code: |
11
  )
12
  SELECT
13
  [
14
- struct_pack(
15
- "from" := 'system',
16
- "value" := instruction
17
- ),
18
  struct_pack(
19
  "from" := 'user',
20
- "value" := input
 
 
 
 
21
  ),
22
  struct_pack(
23
  "from" := 'assistant',
@@ -26,7 +26,6 @@ code: |
26
  ] AS conversation
27
  FROM source_view
28
  WHERE instruction IS NOT NULL
29
- AND input IS NOT NULL
30
  AND output IS NOT NULL;
31
  ---
32
 
@@ -40,13 +39,13 @@ source_view AS (
40
  )
41
  SELECT
42
  [
43
- struct_pack(
44
- "from" := 'system',
45
- "value" := instruction
46
- ),
47
  struct_pack(
48
  "from" := 'user',
49
- "value" := input
 
 
 
 
50
  ),
51
  struct_pack(
52
  "from" := 'assistant',
@@ -55,7 +54,6 @@ SELECT
55
  ] AS conversation
56
  FROM source_view
57
  WHERE instruction IS NOT NULL
58
- AND input IS NOT NULL
59
  AND output IS NOT NULL;
60
  ```
61
 
@@ -83,6 +81,6 @@ Differences between Alpaca and ChatML Conversation Format:
83
  height="560px"
84
  ></iframe>
85
 
86
- You can run this query through via the `sql_console` in the Hugging Face Hub [here](https://huggingface.co/datasets/yahma/alpaca-cleaned?sql_console=true&sql=WITH+%0Asource_view+AS+%28%0A++SELECT+*+FROM+train++--+Change+%27train%27+to+your+desired+view+name+here%0A%29%0ASELECT+%0A++%5B%0A++++struct_pack%28%0A++++++%22from%22+%3A%3D+%27system%27%2C%0A++++++%22value%22+%3A%3D+instruction%0A++++%29%2C%0A++++struct_pack%28%0A++++++%22from%22+%3A%3D+%27user%27%2C%0A++++++%22value%22+%3A%3D+input%0A++++%29%2C%0A++++struct_pack%28%0A++++++%22from%22+%3A%3D+%27assistant%27%2C%0A++++++%22value%22+%3A%3D+output%0A++++%29%0A++%5D+AS+conversation%0AFROM+source_view%0AWHERE+instruction+IS+NOT+NULL+%0A++AND+input+IS+NOT+NULL+%0A++AND+output+IS+NOT+NULL%3B).
87
 
88
  ![Alpaca to ChatML](./alpaca-to-conversation.png)
 
11
  )
12
  SELECT
13
  [
 
 
 
 
14
  struct_pack(
15
  "from" := 'user',
16
+ "value" := CASE
17
+ WHEN input IS NOT NULL AND input != ''
18
+ THEN instruction || '\n' || input
19
+ ELSE instruction
20
+ END
21
  ),
22
  struct_pack(
23
  "from" := 'assistant',
 
26
  ] AS conversation
27
  FROM source_view
28
  WHERE instruction IS NOT NULL
 
29
  AND output IS NOT NULL;
30
  ---
31
 
 
39
  )
40
  SELECT
41
  [
 
 
 
 
42
  struct_pack(
43
  "from" := 'user',
44
+ "value" := CASE
45
+ WHEN input IS NOT NULL AND input != ''
46
+ THEN instruction || '\n' || input
47
+ ELSE instruction
48
+ END
49
  ),
50
  struct_pack(
51
  "from" := 'assistant',
 
54
  ] AS conversation
55
  FROM source_view
56
  WHERE instruction IS NOT NULL
 
57
  AND output IS NOT NULL;
58
  ```
59
 
 
81
  height="560px"
82
  ></iframe>
83
 
84
+ You can run this query through via the `sql_console` in the Hugging Face Hub [here](https://huggingface.co/datasets/yahma/alpaca-cleaned?row=16&sql_console=true&sql=--+Convert+Alpaca+format+to+Conversation+format%0AWITH+%0Asource_view+AS+%28%0A++SELECT+*+FROM+train++--+Change+%27train%27+to+your+desired+view+name+here%0A%29%0ASELECT+%0A++%5B%0A++++struct_pack%28%0A++++++%22from%22+%3A%3D+%27user%27%2C%0A++++++%22value%22+%3A%3D+CASE+%0A+++++++++++++++++++WHEN+input+IS+NOT+NULL+AND+input+%21%3D+%27%27+%0A+++++++++++++++++++THEN+instruction+%7C%7C+%27%5Cn%27+%7C%7C+input%0A+++++++++++++++++++ELSE+instruction%0A+++++++++++++++++END%0A++++%29%2C%0A++++struct_pack%28%0A++++++%22from%22+%3A%3D+%27assistant%27%2C%0A++++++%22value%22+%3A%3D+output%0A++++%29%0A++%5D+AS+conversation%0AFROM+source_view%0AWHERE+instruction+IS+NOT+NULL+%0A++AND+output+IS+NOT+NULL%3B).
85
 
86
  ![Alpaca to ChatML](./alpaca-to-conversation.png)
static/alpaca-to-conversation.png CHANGED