fblgit commited on
Commit
9cd3099
1 Parent(s): 05aa86a

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +22 -8
README.md CHANGED
@@ -31,7 +31,7 @@ dataset_info:
31
  num_bytes: 1140631.4
32
  num_examples: 25000
33
  download_size: 27928049
34
- dataset_size: 35294604.0
35
  configs:
36
  - config_name: default
37
  data_files:
@@ -41,13 +41,23 @@ configs:
41
  path: data/test-*
42
  tags:
43
  - math
 
 
 
 
 
 
 
 
44
  ---
45
 
46
  # Simple Math
47
 
48
  Just like my teacher gave me homework, i thought maybe we can also add some of these basics on the trainings of our models.
49
 
50
- It was created with very simple code that is in the repo, if you add more complex operations and so.. please share the code :D thank you
 
 
51
 
52
  ## Does it Works?
53
  ### 34BEAGLES Evaluation:
@@ -96,19 +106,23 @@ hf (pretrained=/data/models/34BeaglesMath-v1,dtype=bfloat16,trust_remote_code=Tr
96
  | - stem |N/A |none | 0|acc |0.6708|± |0.1236|
97
  ```
98
 
99
- I think it works. Not too difficult, not too easy, as a curriculum works well :)
 
 
 
 
 
 
100
 
101
- Note to contributors:
102
  **thank you to those contributing on the experiment with beautiful commits and good spirit**
103
- * The model needs some splits
104
- * The complexity has to be gradual as show in experiments
105
  * Feel free to contribute on the readme Evaluation tests.
106
  * Lets aim to build an ablation & paper together. All contributors will be cited.
107
- * Add your log entry on the version so we can keep a track, thanks.
108
-
109
 
110
  ## Versions
111
  ```
 
112
  24.01.24 Added gradual complexity on a separate script
113
  20-23.01.24 Multiple contributions with operations and increased complexity on the main generator script.
114
  ```
 
31
  num_bytes: 1140631.4
32
  num_examples: 25000
33
  download_size: 27928049
34
+ dataset_size: 35294604
35
  configs:
36
  - config_name: default
37
  data_files:
 
41
  path: data/test-*
42
  tags:
43
  - math
44
+ - finance
45
+ license: cc-by-nc-nd-4.0
46
+ task_categories:
47
+ - text-generation
48
+ - question-answering
49
+ pretty_name: Simple Math
50
+ size_categories:
51
+ - 100K<n<1M
52
  ---
53
 
54
  # Simple Math
55
 
56
  Just like my teacher gave me homework, i thought maybe we can also add some of these basics on the trainings of our models.
57
 
58
+ It was created with very simple code that is in the repo, if you add more complex operations and so.. **please share the code** :D thank you
59
+
60
+ Current Code Version: 20240127.fblgit (A modification over @win10 for progressive and DPO operation)
61
 
62
  ## Does it Works?
63
  ### 34BEAGLES Evaluation:
 
106
  | - stem |N/A |none | 0|acc |0.6708|± |0.1236|
107
  ```
108
 
109
+ But it gets better, because when increasing length and complexity, the marks are even superior:
110
+ ```
111
+ |Tasks|Version| Filter |n-shot| Metric |Value | |Stderr|
112
+ |-----|-------|----------|-----:|-----------|-----:|---|-----:|
113
+ |gsm8k|Yaml |get-answer| 5|exact_match|0.6611|± | 0.013|
114
+ ```
115
+ On a 3.20% GSM Improvement compared to its base model.
116
 
117
+ ## Note to contributors:
118
  **thank you to those contributing on the experiment with beautiful commits and good spirit**
119
+
 
120
  * Feel free to contribute on the readme Evaluation tests.
121
  * Lets aim to build an ablation & paper together. All contributors will be cited.
 
 
122
 
123
  ## Versions
124
  ```
125
+ 27.01.24 Added new code to generate the dataset, seed 42 and now also generates DPO.
126
  24.01.24 Added gradual complexity on a separate script
127
  20-23.01.24 Multiple contributions with operations and increased complexity on the main generator script.
128
  ```