RichardErkhov commited on
Commit
e0eba7d
1 Parent(s): ff5b1b4

uploaded readme

Browse files
Files changed (1) hide show
  1. README.md +106 -0
README.md ADDED
@@ -0,0 +1,106 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Quantization made by Richard Erkhov.
2
+
3
+ [Github](https://github.com/RichardErkhov)
4
+
5
+ [Discord](https://discord.gg/pvy7H8DZMG)
6
+
7
+ [Request more models](https://github.com/RichardErkhov/quant_request)
8
+
9
+
10
+ Mistral-7B-Instruct-v0.2 - bnb 4bits
11
+ - Model creator: https://huggingface.co/mistralai/
12
+ - Original model: https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2/
13
+
14
+
15
+
16
+
17
+ Original model description:
18
+ ---
19
+ license: apache-2.0
20
+ pipeline_tag: text-generation
21
+ tags:
22
+ - finetuned
23
+ inference: true
24
+ widget:
25
+ - messages:
26
+ - role: user
27
+ content: What is your favorite condiment?
28
+ ---
29
+
30
+ # Model Card for Mistral-7B-Instruct-v0.2
31
+
32
+ The Mistral-7B-Instruct-v0.2 Large Language Model (LLM) is an instruct fine-tuned version of the Mistral-7B-v0.2.
33
+
34
+ Mistral-7B-v0.2 has the following changes compared to Mistral-7B-v0.1
35
+ - 32k context window (vs 8k context in v0.1)
36
+ - Rope-theta = 1e6
37
+ - No Sliding-Window Attention
38
+
39
+ For full details of this model please read our [paper](https://arxiv.org/abs/2310.06825) and [release blog post](https://mistral.ai/news/la-plateforme/).
40
+
41
+ ## Instruction format
42
+
43
+ In order to leverage instruction fine-tuning, your prompt should be surrounded by `[INST]` and `[/INST]` tokens. The very first instruction should begin with a begin of sentence id. The next instructions should not. The assistant generation will be ended by the end-of-sentence token id.
44
+
45
+ E.g.
46
+ ```
47
+ text = "<s>[INST] What is your favourite condiment? [/INST]"
48
+ "Well, I'm quite partial to a good squeeze of fresh lemon juice. It adds just the right amount of zesty flavour to whatever I'm cooking up in the kitchen!</s> "
49
+ "[INST] Do you have mayonnaise recipes? [/INST]"
50
+ ```
51
+
52
+ This format is available as a [chat template](https://huggingface.co/docs/transformers/main/chat_templating) via the `apply_chat_template()` method:
53
+
54
+ ```python
55
+ from transformers import AutoModelForCausalLM, AutoTokenizer
56
+
57
+ device = "cuda" # the device to load the model onto
58
+
59
+ model = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-7B-Instruct-v0.2")
60
+ tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-Instruct-v0.2")
61
+
62
+ messages = [
63
+ {"role": "user", "content": "What is your favourite condiment?"},
64
+ {"role": "assistant", "content": "Well, I'm quite partial to a good squeeze of fresh lemon juice. It adds just the right amount of zesty flavour to whatever I'm cooking up in the kitchen!"},
65
+ {"role": "user", "content": "Do you have mayonnaise recipes?"}
66
+ ]
67
+
68
+ encodeds = tokenizer.apply_chat_template(messages, return_tensors="pt")
69
+
70
+ model_inputs = encodeds.to(device)
71
+ model.to(device)
72
+
73
+ generated_ids = model.generate(model_inputs, max_new_tokens=1000, do_sample=True)
74
+ decoded = tokenizer.batch_decode(generated_ids)
75
+ print(decoded[0])
76
+ ```
77
+
78
+ ## Troubleshooting
79
+ - If you see the following error:
80
+ ```
81
+ Traceback (most recent call last):
82
+ File "", line 1, in
83
+ File "/transformers/models/auto/auto_factory.py", line 482, in from_pretrained
84
+ config, kwargs = AutoConfig.from_pretrained(
85
+ File "/transformers/models/auto/configuration_auto.py", line 1022, in from_pretrained
86
+ config_class = CONFIG_MAPPING[config_dict["model_type"]]
87
+ File "/transformers/models/auto/configuration_auto.py", line 723, in getitem
88
+ raise KeyError(key)
89
+ KeyError: 'mistral'
90
+ ```
91
+
92
+ Installing transformers from source should solve the issue
93
+ pip install git+https://github.com/huggingface/transformers
94
+
95
+ This should not be required after transformers-v4.33.4.
96
+
97
+ ## Limitations
98
+
99
+ The Mistral 7B Instruct model is a quick demonstration that the base model can be easily fine-tuned to achieve compelling performance.
100
+ It does not have any moderation mechanisms. We're looking forward to engaging with the community on ways to
101
+ make the model finely respect guardrails, allowing for deployment in environments requiring moderated outputs.
102
+
103
+ ## The Mistral AI Team
104
+
105
+ Albert Jiang, Alexandre Sablayrolles, Arthur Mensch, Blanche Savary, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Emma Bou Hanna, Florian Bressand, Gianna Lengyel, Guillaume Bour, Guillaume Lample, Lélio Renard Lavaud, Louis Ternon, Lucile Saulnier, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Théophile Gervet, Thibaut Lavril, Thomas Wang, Timothée Lacroix, William El Sayed.
106
+