koesn commited on
Commit
a5c0a56
1 Parent(s): a03e642

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +118 -0
README.md CHANGED
@@ -1,3 +1,121 @@
1
  ---
 
 
2
  license: mit
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ tags:
3
+ - legal
4
  license: mit
5
+ language:
6
+ - en
7
  ---
8
+ # Saul-Instruct-v1
9
+
10
+ ## Description
11
+ This repo contains GGUF format model files for Saul-Instruct-v1.
12
+
13
+ ## Files Provided
14
+ | Name | Quant | Bits | File Size | Remark |
15
+ | ---------------------------- | ------- | ---- | --------- | -------------------------------- |
16
+ | saul-instruct-v1.IQ3_S.gguf | IQ3_S | 3 | 3.18 GB | 3.44 bpw quantization |
17
+ | saul-instruct-v1.IQ3_M.gguf | IQ3_M | 3 | 3.28 GB | 3.66 bpw quantization mix |
18
+ | saul-instruct-v1.Q4_0.gguf | Q4_0 | 4 | 4.11 GB | 3.56G, +0.2166 ppl |
19
+ | saul-instruct-v1.IQ4_NL.gguf | IQ4_NL | 4 | 4.16 GB | 4.25 bpw non-linear quantization |
20
+ | saul-instruct-v1.Q4_K_M.gguf | Q4_K_M | 4 | 4.37 GB | 3.80G, +0.0532 ppl |
21
+ | saul-instruct-v1.Q5_K_M.gguf | Q5_K_M | 5 | 5.13 GB | 4.45G, +0.0122 ppl |
22
+ | saul-instruct-v1.Q6_K.gguf | Q6_K | 6 | 5.94 GB | 5.15G, +0.0008 ppl |
23
+ | saul-instruct-v1.Q8_0.gguf | Q8_0 | 8 | 7.70 GB | 6.70G, +0.0004 ppl |
24
+
25
+ ## Parameters
26
+ | path | type | architecture | rope_theta | sliding_win | max_pos_embed |
27
+ | ----------------------- | ------- | ------------------ | ---------- | ----------- | ------------- |
28
+ | Equall/Saul-Instruct-v1 | mistral | MistralForCausalLM | 10000 | 4096 | 32768 |
29
+
30
+ ## Benchmarks
31
+ See original model card.
32
+
33
+ # Original Model Card
34
+
35
+ ---
36
+ library_name: transformers
37
+ tags:
38
+ - legal
39
+ license: mit
40
+ language:
41
+ - en
42
+ ---
43
+
44
+ # Equall/Saul-Instruct-v1
45
+
46
+ This is the instruct model for Equall/Saul-Instruct-v1, a large instruct language model tailored for Legal domain. This model is obtained by continue pretraining of Mistral-7B.
47
+
48
+ Checkout our website and register https://equall.ai/
49
+
50
+
51
+ ![image/png](https://cdn-uploads.huggingface.co/production/uploads/644a900e3a619fe72b14af0f/OU4Y3s-WckYKMN4fQkNiS.png)
52
+
53
+ ## Model Details
54
+
55
+ ### Model Description
56
+
57
+ <!-- Provide a longer summary of what this model is. -->
58
+
59
+ This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
60
+
61
+ - **Developed by:** Equall.ai in collaboration with CentraleSupelec, Sorbonne Université, Instituto Superior Técnico and NOVA School of Law
62
+ - **Model type:** 7B
63
+ - **Language(s) (NLP):** English
64
+ - **License:** MIT
65
+
66
+ ### Model Sources
67
+
68
+ <!-- Provide the basic links for the model. -->
69
+
70
+ - **Paper:** https://arxiv.org/abs/2403.03883
71
+
72
+ ## Uses
73
+
74
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
75
+ You can use it for legal use cases that involves generation.
76
+
77
+ Here's how you can run the model using the pipeline() function from 🤗 Transformers:
78
+
79
+ ```python
80
+
81
+ # Install transformers from source - only needed for versions <= v4.34
82
+ # pip install git+https://github.com/huggingface/transformers.git
83
+ # pip install accelerate
84
+
85
+ import torch
86
+ from transformers import pipeline
87
+
88
+ pipe = pipeline("text-generation", model="Equall/Saul-Instruct-v1", torch_dtype=torch.bfloat16, device_map="auto")
89
+ # We use the tokenizer’s chat template to format each message - see https://huggingface.co/docs/transformers/main/en/chat_templating
90
+ messages = [
91
+ {"role": "user", "content": "[YOUR QUERY GOES HERE]"},
92
+ ]
93
+ prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
94
+ outputs = pipe(prompt, max_new_tokens=256, do_sample=False)
95
+ print(outputs[0]["generated_text"])
96
+ ```
97
+
98
+ ## Bias, Risks, and Limitations
99
+
100
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
101
+
102
+
103
+ This model is built upon the technology of LLM, which comes with inherent limitations. It may occasionally generate inaccurate or nonsensical outputs. Furthermore, being a 7B model, it's anticipated to exhibit less robust performance compared to larger models, such as the 70B variant.
104
+
105
+ ## Citation
106
+
107
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
108
+
109
+ **BibTeX:**
110
+
111
+
112
+ ```bibtex
113
+ @misc{colombo2024saullm7b,
114
+ title={SaulLM-7B: A pioneering Large Language Model for Law},
115
+ author={Pierre Colombo and Telmo Pessoa Pires and Malik Boudiaf and Dominic Culver and Rui Melo and Caio Corro and Andre F. T. Martins and Fabrizio Esposito and Vera Lúcia Raposo and Sofia Morgado and Michael Desa},
116
+ year={2024},
117
+ eprint={2403.03883},
118
+ archivePrefix={arXiv},
119
+ primaryClass={cs.CL}
120
+ }
121
+ ```