dangvansam commited on
Commit
9776ca8
·
verified ·
1 Parent(s): 3af22b8

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +64 -0
README.md ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: facebook/MobileLLM-R1-140M
3
+ library_name: transformers
4
+ model_name: >-
5
+ transformer-facebook-MobileLLM-R1-140M-ml-512-bs-32-ws-100-lr-1e-05-full_ft-merge_user_input_system_prompt
6
+ tags:
7
+ - generated_from_trainer
8
+ - sft
9
+ - trl
10
+ licence: license
11
+ pipeline_tag: text-classification
12
+ ---
13
+
14
+ # Model Card for transformer-facebook-MobileLLM-R1-140M-ml-512-bs-32-ws-100-lr-1e-05-full_ft-merge_user_input_system_prompt
15
+
16
+ This model is a fine-tuned version of [facebook/MobileLLM-R1-140M](https://huggingface.co/facebook/MobileLLM-R1-140M).
17
+ It has been trained using [TRL](https://github.com/huggingface/trl).
18
+
19
+ ## Quick start
20
+
21
+ ```python
22
+ from transformers import pipeline
23
+
24
+ question = """
25
+ You are a speaking turn-ending identifier. Your task is to identify whether the user's speaking turn is complete or not. Respond with `end` if the user's turn is complete, or `continue` if it is not.
26
+
27
+ User input: I want to
28
+ """
29
+ generator = pipeline("text-generation", model="None", device="cuda")
30
+ output = generator([{"role": "user", "content": question}], max_new_tokens=1, return_full_text=False)[0]
31
+ print(output["generated_text"])
32
+ ```
33
+
34
+ ## Training procedure
35
+
36
+ [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/dangvansam98/LLM-TURN-DETECTION/runs/ub528cqr)
37
+
38
+
39
+ This model was trained with SFT.
40
+
41
+ ### Framework versions
42
+
43
+ - TRL: 0.21.0
44
+ - Transformers: 4.55.2
45
+ - Pytorch: 2.6.0
46
+ - Datasets: 3.6.0
47
+ - Tokenizers: 0.21.4
48
+
49
+ ## Citations
50
+
51
+
52
+
53
+ Cite TRL as:
54
+
55
+ ```bibtex
56
+ @misc{vonwerra2022trl,
57
+ title = {{TRL: Transformer Reinforcement Learning}},
58
+ author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
59
+ year = 2020,
60
+ journal = {GitHub repository},
61
+ publisher = {GitHub},
62
+ howpublished = {\url{https://github.com/huggingface/trl}}
63
+ }
64
+ ```