dmayhem93 commited on
Commit
0c00a64
1 Parent(s): e4944ca

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +105 -0
README.md ADDED
@@ -0,0 +1,105 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-nc-4.0
3
+ datasets:
4
+ - conceptofmind/cot_submix_original
5
+ - conceptofmind/flan2021_submix_original
6
+ - conceptofmind/t0_submix_original
7
+ - conceptofmind/niv2_submix_original
8
+ language:
9
+ - en
10
+ pipeline_tag: text-generation
11
+ ---
12
+ # FreeWilly
13
+
14
+ ## Model Description
15
+
16
+ `FreeWilly2` is a Llama2 70B model finetuned on an Orca style Dataset
17
+
18
+ ## Usage
19
+
20
+ Start chatting with `FreeWilly2` using the following code snippet:
21
+
22
+ ```python
23
+ import torch
24
+ from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
25
+
26
+ tokenizer = AutoTokenizer.from_pretrained("stabilityai/FreeWilly2, use_fast=False)
27
+ model = AutoModelForCausalLM.from_pretrained("stabilityai/FreeWilly2", torch_dtype=torch.float16, low_cpu_mem_usage=True, use_accelerate=True)
28
+ generator = pipeline(model=model, tokenizer=tokenizer)
29
+ system_prompt = "### System:\nYou are Free Willy, an AI that follows instructions extremely well. Help as much as you can. Remember, be safe, and don't do anything illegal.\n\n"
30
+
31
+ message = "Write me a poem please"
32
+ prompt = f"{system_prompt}### User: {message}\n\n### Assistant:\n"
33
+
34
+ output = generator(prompt, num_return_sequences=1, do_sample=True, top_p=0.95, top_k=0)
35
+ print(output)
36
+ ```
37
+
38
+ FreeWilly should be used with this prompt format:
39
+ ```
40
+ ### System:
41
+ This is a system prompt, please behave and help the user.
42
+
43
+ ### User:
44
+ Your prompt here
45
+
46
+ ### Assistant
47
+ The output of FreeWilly2
48
+ ```
49
+
50
+ ## Model Details
51
+
52
+ * **Developed by**: [Stability AI](https://stability.ai/)
53
+ * **Model type**: FreeWilly is an auto-regressive language model fine-tuned on Llama2 70B.
54
+ * **Language(s)**: English
55
+ * **Library**: [HuggingFace Transformers](https://github.com/huggingface/transformers)
56
+ * **License**: Fine-tuned checkpoints (`FreeWilly2`) is licensed under the Non-Commercial Creative Commons license ([CC BY-NC-4.0](https://creativecommons.org/licenses/by-nc/4.0/))
57
+ * **Contact**: For questions and comments about the model, please email `[email protected]`
58
+
59
+ ### Training Dataset
60
+
61
+ `FreeWilly2` is trained on our internal Orca-style dataset
62
+
63
+ ### Training Procedure
64
+
65
+ Models are learned via supervised fine-tuning on the aforementioned datasets, trained in mixed-precision (BF16), and optimized with AdamW. We outline the following hyperparameters:
66
+
67
+ | Dataset | Batch Size | Learning Rate |Learning Rate Decay| Warm-up | Weight Decay | Betas |
68
+ |-------------------|------------|---------------|-------------------|---------|--------------|-------------|
69
+ | Orca pt1 packed | 256 | 3e-5 | Cosine to 3e-6 | 100 | 1e-6 | (0.9, 0.95) |
70
+ | Orca pt2 unpacked | 512 | 3e-5 | Cosine to 3e-6 | 100 | 1e-6 | (0.9, 0.95) |
71
+
72
+ ## Use and Limitations
73
+
74
+ ### Intended Use
75
+
76
+ These models are intended for research only, in adherence with the [CC BY-NC-4.0](https://creativecommons.org/licenses/by-nc/4.0/) license.
77
+
78
+ ### Limitations and bias
79
+
80
+ Although the aforementioned dataset helps to steer the base language models into "safer" distributions of text, not all biases and toxicity can be mitigated through fine-tuning. We ask that users be mindful of such potential issues that can arise in generated responses. Do not treat model outputs as substitutes for human judgment or as sources of truth. Please use it responsibly.
81
+
82
+ ## Citations
83
+
84
+ ```bibtext
85
+ @misc{touvron2023llama,
86
+ title={Llama 2: Open Foundation and Fine-Tuned Chat Models},
87
+ author={Hugo Touvron and Louis Martin and Kevin Stone and Peter Albert and Amjad Almahairi and Yasmine Babaei and Nikolay Bashlykov and Soumya Batra and Prajjwal Bhargava and Shruti Bhosale and Dan Bikel and Lukas Blecher and Cristian Canton Ferrer and Moya Chen and Guillem Cucurull and David Esiobu and Jude Fernandes and Jeremy Fu and Wenyin Fu and Brian Fuller and Cynthia Gao and Vedanuj Goswami and Naman Goyal and Anthony Hartshorn and Saghar Hosseini and Rui Hou and Hakan Inan and Marcin Kardas and Viktor Kerkez and Madian Khabsa and Isabel Kloumann and Artem Korenev and Punit Singh Koura and Marie-Anne Lachaux and Thibaut Lavril and Jenya Lee and Diana Liskovich and Yinghai Lu and Yuning Mao and Xavier Martinet and Todor Mihaylov and Pushkar Mishra and Igor Molybog and Yixin Nie and Andrew Poulton and Jeremy Reizenstein and Rashi Rungta and Kalyan Saladi and Alan Schelten and Ruan Silva and Eric Michael Smith and Ranjan Subramanian and Xiaoqing Ellen Tan and Binh Tang and Ross Taylor and Adina Williams and Jian Xiang Kuan and Puxin Xu and Zheng Yan and Iliyan Zarov and Yuchen Zhang and Angela Fan and Melanie Kambadur and Sharan Narang and Aurelien Rodriguez and Robert Stojnic and Sergey Edunov and Thomas Scialom},
88
+ year={2023},
89
+ eprint={2307.09288},
90
+ archivePrefix={arXiv},
91
+ primaryClass={cs.CL}
92
+ }
93
+ ```
94
+
95
+ ```bibtext
96
+ @misc{mukherjee2023orca,
97
+ title={Orca: Progressive Learning from Complex Explanation Traces of GPT-4},
98
+ author={Subhabrata Mukherjee and Arindam Mitra and Ganesh Jawahar and Sahaj Agarwal and Hamid Palangi and Ahmed Awadallah},
99
+ year={2023},
100
+ eprint={2306.02707},
101
+ archivePrefix={arXiv},
102
+ primaryClass={cs.CL}
103
+ }
104
+ ```
105
+