JacopoAbate commited on
Commit
92fdfb2
1 Parent(s): be49822

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +4 -4
README.md CHANGED
@@ -17,7 +17,7 @@ metrics:
17
  # Model Information
18
 
19
 
20
- Phi3-instruct-128k-ORPO is an updated version of [Phi-3-mini-128k-instruct](https://huggingface.co/microsoft/Phi-3-mini-128k-instruct), aligned with ORPO and QLora.
21
 
22
  - It's trained on [ultrafeedback-binarized-preferences-cleaned](https://huggingface.co/datasets/argilla/ultrafeedback-binarized-preferences-cleaned).
23
 
@@ -43,8 +43,8 @@ from transformers import AutoModelForCausalLM, AutoTokenizer
43
 
44
  device = "cpu" # if you want to use the gpu make sure to have cuda toolkit installed and change this to "cuda"
45
 
46
- model = AutoModelForCausalLM.from_pretrained("MoxoffSpA/Phi3-instruct-128k-ORPO")
47
- tokenizer = AutoTokenizer.from_pretrained("MoxoffSpA/Phi3-instruct-128k-ORPO")
48
 
49
  question = """Quanto è alta la torre di Pisa?"""
50
  context = """
@@ -78,7 +78,7 @@ print(trimmed_output)
78
 
79
  ## Bias, Risks and Limitations
80
 
81
- Phi3-instruct-128k-ORPO has not been aligned to human preferences for safety within the RLHF phase or deployed with in-the-loop filtering of
82
  responses like ChatGPT, so the model can produce problematic outputs (especially when prompted to do so). It is also unknown what the size and composition
83
  of the corpus was used to train the base model, however it is likely to have included a mix of Web data and technical sources
84
  like books and code.
 
17
  # Model Information
18
 
19
 
20
+ Phi-3-mini-128k-instruct-ORPO is an updated version of [Phi-3-mini-128k-instruct](https://huggingface.co/microsoft/Phi-3-mini-128k-instruct), aligned with ORPO and QLora.
21
 
22
  - It's trained on [ultrafeedback-binarized-preferences-cleaned](https://huggingface.co/datasets/argilla/ultrafeedback-binarized-preferences-cleaned).
23
 
 
43
 
44
  device = "cpu" # if you want to use the gpu make sure to have cuda toolkit installed and change this to "cuda"
45
 
46
+ model = AutoModelForCausalLM.from_pretrained("MoxoffSpA/Phi-3-mini-128k-instruct-ORPO")
47
+ tokenizer = AutoTokenizer.from_pretrained("MoxoffSpA/Phi-3-mini-128k-instruct-ORPO")
48
 
49
  question = """Quanto è alta la torre di Pisa?"""
50
  context = """
 
78
 
79
  ## Bias, Risks and Limitations
80
 
81
+ Phi-3-mini-128k-instruct-ORPO has not been aligned to human preferences for safety within the RLHF phase or deployed with in-the-loop filtering of
82
  responses like ChatGPT, so the model can produce problematic outputs (especially when prompted to do so). It is also unknown what the size and composition
83
  of the corpus was used to train the base model, however it is likely to have included a mix of Web data and technical sources
84
  like books and code.