Triangle104 commited on
Commit
666ab34
·
verified ·
1 Parent(s): 0745f57

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +88 -0
README.md CHANGED
@@ -17,6 +17,94 @@ tags:
17
  This model was converted to GGUF format from [`prithivMLmods/GWQ-9B-Preview2`](https://huggingface.co/prithivMLmods/GWQ-9B-Preview2) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space.
18
  Refer to the [original model card](https://huggingface.co/prithivMLmods/GWQ-9B-Preview2) for more details on the model.
19
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
20
  ## Use with llama.cpp
21
  Install llama.cpp through brew (works on Mac and Linux)
22
 
 
17
  This model was converted to GGUF format from [`prithivMLmods/GWQ-9B-Preview2`](https://huggingface.co/prithivMLmods/GWQ-9B-Preview2) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space.
18
  Refer to the [original model card](https://huggingface.co/prithivMLmods/GWQ-9B-Preview2) for more details on the model.
19
 
20
+ ---
21
+
22
+ Chain of Continuous Thought Synthetic Dataset, which enhances its
23
+ ability to perform reasoning, multi-step problem solving, and logical
24
+ inferences.
25
+
26
+
27
+ Text Generation:
28
+ The model is ideal for
29
+ creative writing tasks such as generating poems, stories, and essays. It
30
+ can also be used for generating code comments, documentation, and
31
+ markdown files.
32
+
33
+
34
+ Instruction Following:
35
+ GWQ’s
36
+ instruction-tuned variant is suitable for generating responses based on
37
+ user instructions, making it useful for virtual assistants, tutoring
38
+ systems, and automated customer support.
39
+
40
+
41
+ Domain-Specific Applications:
42
+ Thanks to its
43
+ modular design and open-source nature, the model can be fine-tuned for
44
+ specific tasks like legal document summarization, medical record
45
+ analysis, or financial report generation.
46
+
47
+
48
+
49
+
50
+
51
+
52
+
53
+
54
+
55
+ Limitations of GWQ2
56
+
57
+
58
+
59
+
60
+ Resource Requirements:
61
+ Although lightweight
62
+ compared to larger models, the 9B parameter size still requires
63
+ significant computational resources, including GPUs with large memory
64
+ for inference.
65
+
66
+
67
+ Knowledge Cutoff:
68
+ The model’s pre-training
69
+ data may not include recent information, making it less effective for
70
+ answering queries on current events or newly developed topics.
71
+
72
+
73
+ Bias in Outputs:
74
+ Since the model is trained
75
+ on publicly available datasets, it may inherit biases present in those
76
+ datasets, leading to potentially biased or harmful outputs in sensitive
77
+ contexts.
78
+
79
+
80
+ Hallucinations:
81
+ Like other large language
82
+ models, GWQ can occasionally generate incorrect or nonsensical
83
+ information, especially when asked for facts or reasoning outside its
84
+ training scope.
85
+
86
+
87
+ Lack of Common-Sense Reasoning:
88
+ While GWQ is
89
+ fine-tuned for reasoning, it may still struggle with tasks requiring
90
+ deep common-sense knowledge or nuanced understanding of human behavior
91
+ and emotions.
92
+
93
+
94
+ Dependency on Fine-Tuning:
95
+ For optimal
96
+ performance on domain-specific tasks, fine-tuning on relevant datasets
97
+ is required, which demands additional computational resources and
98
+ expertise.
99
+
100
+
101
+ Context Length Limitation:
102
+ The model’s
103
+ ability to process long documents is limited by its maximum context
104
+ window size. If the input exceeds this limit, truncation may lead to
105
+ loss of important information.
106
+
107
+ ---
108
  ## Use with llama.cpp
109
  Install llama.cpp through brew (works on Mac and Linux)
110