zhuxunyu commited on
Commit
8947f47
·
verified ·
1 Parent(s): a62081d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -5
README.md CHANGED
@@ -11,7 +11,7 @@ metrics:
11
 
12
  <!-- Provide a quick summary of what the model is/does. -->
13
 
14
- We use Mix Thoughts Distillation to distill mathematical reasoning ability from gpt-3.5-turbo to CodeT5+-770m-py.
15
 
16
  ### Model Description
17
 
@@ -47,7 +47,7 @@ When given a question, the prompt "System of linear equations: (Do not simplify)
47
 
48
  ```python
49
  from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
50
- checkpoint = "zhuxunyu/mtd-codet5p-770m-py"
51
  device = "cuda" # for GPU usage or "cpu" for CPU usage
52
  tokenizer = AutoTokenizer.from_pretrained(checkpoint)
53
  model = AutoModelForSeq2SeqLM.from_pretrained(checkpoint).to(device)
@@ -62,7 +62,7 @@ generation = tokenizer.decode(output, skip_special_tokens=True)
62
 
63
  ```python
64
  from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
65
- checkpoint = "zhuxunyu/mtd-codet5p-770m-py"
66
  device = "cuda" # for GPU usage or "cpu" for CPU usage
67
  tokenizer = AutoTokenizer.from_pretrained(checkpoint)
68
  model = AutoModelForSeq2SeqLM.from_pretrained(checkpoint).to(device)
@@ -76,7 +76,7 @@ generation = tokenizer.decode(output, skip_special_tokens=True)
76
 
77
  ```python
78
  from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
79
- checkpoint = "zhuxunyu/mtd-codet5p-770m-py"
80
  device = "cuda" # for GPU usage or "cpu" for CPU usage
81
  tokenizer = AutoTokenizer.from_pretrained(checkpoint)
82
  model = AutoModelForSeq2SeqLM.from_pretrained(checkpoint).to(device)
@@ -111,7 +111,7 @@ their corresponding reasoning processes are built as a training dataset, and we
111
  | PoT | 50.34 | 55.2 | 51.6 | 88.33 |
112
  | EoT | 48.21 | 52.81 | 55.7 | 70.16 |
113
  | CoT | 25.47 | 29.67 | 23.3 | 46.5 |
114
- | Mix_all | 50.56 | 55.34 | 52.3 | 88.83 |
115
 
116
 
117
 
 
11
 
12
  <!-- Provide a quick summary of what the model is/does. -->
13
 
14
+ We use Ensemble Thoughts Distillation to distill mathematical reasoning ability from gpt-3.5-turbo to CodeT5+-770m-py.
15
 
16
  ### Model Description
17
 
 
47
 
48
  ```python
49
  from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
50
+ checkpoint = "zhuxunyu/etd-codet5p-770m-py"
51
  device = "cuda" # for GPU usage or "cpu" for CPU usage
52
  tokenizer = AutoTokenizer.from_pretrained(checkpoint)
53
  model = AutoModelForSeq2SeqLM.from_pretrained(checkpoint).to(device)
 
62
 
63
  ```python
64
  from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
65
+ checkpoint = "zhuxunyu/etd-codet5p-770m-py"
66
  device = "cuda" # for GPU usage or "cpu" for CPU usage
67
  tokenizer = AutoTokenizer.from_pretrained(checkpoint)
68
  model = AutoModelForSeq2SeqLM.from_pretrained(checkpoint).to(device)
 
76
 
77
  ```python
78
  from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
79
+ checkpoint = "zhuxunyu/etd-codet5p-770m-py"
80
  device = "cuda" # for GPU usage or "cpu" for CPU usage
81
  tokenizer = AutoTokenizer.from_pretrained(checkpoint)
82
  model = AutoModelForSeq2SeqLM.from_pretrained(checkpoint).to(device)
 
111
  | PoT | 50.34 | 55.2 | 51.6 | 88.33 |
112
  | EoT | 48.21 | 52.81 | 55.7 | 70.16 |
113
  | CoT | 25.47 | 29.67 | 23.3 | 46.5 |
114
+ | Ensemble_all | 50.56 | 55.34 | 52.3 | 88.83 |
115
 
116
 
117