Hugging Face's logo

RichardErkhov
/
FredZhang7_-_distilgpt2-stable-diffusion-gguf

GGUF
Inference Endpoints
Model card Files Community
RichardErkhov commited on
Commit
baf89ab
•
1 Parent(s): ac0c272

uploaded readme

Browse files
Files changed (1) hide show
  1. README.md +108 -0
README.md ADDED
@@ -0,0 +1,108 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ distilgpt2-stable-diffusion - GGUF
11
+ - Model creator: https://huggingface.co/FredZhang7/
12
+ - Original model: https://huggingface.co/FredZhang7/distilgpt2-stable-diffusion/
13
+
14
+
15
+ | Name | Quant method | Size |
16
+ | ---- | ---- | ---- |
17
+ | [distilgpt2-stable-diffusion.Q2_K.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q2_K.gguf) | Q2_K | 0.06GB |
18
+ | [distilgpt2-stable-diffusion.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.IQ3_XS.gguf) | IQ3_XS | 0.07GB |
19
+ | [distilgpt2-stable-diffusion.IQ3_S.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.IQ3_S.gguf) | IQ3_S | 0.07GB |
20
+ | [distilgpt2-stable-diffusion.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q3_K_S.gguf) | Q3_K_S | 0.07GB |
21
+ | [distilgpt2-stable-diffusion.IQ3_M.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.IQ3_M.gguf) | IQ3_M | 0.07GB |
22
+ | [distilgpt2-stable-diffusion.Q3_K.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q3_K.gguf) | Q3_K | 0.07GB |
23
+ | [distilgpt2-stable-diffusion.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q3_K_M.gguf) | Q3_K_M | 0.07GB |
24
+ | [distilgpt2-stable-diffusion.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q3_K_L.gguf) | Q3_K_L | 0.07GB |
25
+ | [distilgpt2-stable-diffusion.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.IQ4_XS.gguf) | IQ4_XS | 0.07GB |
26
+ | [distilgpt2-stable-diffusion.Q4_0.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q4_0.gguf) | Q4_0 | 0.08GB |
27
+ | [distilgpt2-stable-diffusion.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.IQ4_NL.gguf) | IQ4_NL | 0.08GB |
28
+ | [distilgpt2-stable-diffusion.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q4_K_S.gguf) | Q4_K_S | 0.08GB |
29
+ | [distilgpt2-stable-diffusion.Q4_K.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q4_K.gguf) | Q4_K | 0.08GB |
30
+ | [distilgpt2-stable-diffusion.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q4_K_M.gguf) | Q4_K_M | 0.08GB |
31
+ | [distilgpt2-stable-diffusion.Q4_1.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q4_1.gguf) | Q4_1 | 0.08GB |
32
+ | [distilgpt2-stable-diffusion.Q5_0.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q5_0.gguf) | Q5_0 | 0.09GB |
33
+ | [distilgpt2-stable-diffusion.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q5_K_S.gguf) | Q5_K_S | 0.09GB |
34
+ | [distilgpt2-stable-diffusion.Q5_K.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q5_K.gguf) | Q5_K | 0.09GB |
35
+ | [distilgpt2-stable-diffusion.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q5_K_M.gguf) | Q5_K_M | 0.09GB |
36
+ | [distilgpt2-stable-diffusion.Q5_1.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q5_1.gguf) | Q5_1 | 0.09GB |
37
+ | [distilgpt2-stable-diffusion.Q6_K.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q6_K.gguf) | Q6_K | 0.1GB |
38
+ | [distilgpt2-stable-diffusion.Q8_0.gguf](https://huggingface.co/RichardErkhov/FredZhang7_-_distilgpt2-stable-diffusion-gguf/blob/main/distilgpt2-stable-diffusion.Q8_0.gguf) | Q8_0 | 0.12GB |
39
+
40
+
41
+
42
+
43
+ Original model description:
44
+ ---
45
+ license: creativeml-openrail-m
46
+ tags:
47
+ - stable-diffusion
48
+ - prompt-generator
49
+ - distilgpt2
50
+ datasets:
51
+ - FredZhang7/krea-ai-prompts
52
+ - Gustavosta/Stable-Diffusion-Prompts
53
+ - bartman081523/stable-diffusion-discord-prompts
54
+ widget:
55
+ - text: "amazing"
56
+ - text: "a photo of"
57
+ - text: "a sci-fi"
58
+ - text: "a portrait of"
59
+ - text: "a person standing"
60
+ - text: "a boy watching"
61
+ ---
62
+ # DistilGPT2 Stable Diffusion Model Card
63
+
64
+
65
+ <a href="https://huggingface.co/FredZhang7/distilgpt2-stable-diffusion-v2"> <font size="4"> <bold> Version 2 is here! </bold> </font> </a>
66
+
67
+
68
+ DistilGPT2 Stable Diffusion is a text generation model used to generate creative and coherent prompts for text-to-image models, given any text.
69
+ This model was finetuned on 2.03 million descriptive stable diffusion prompts from [Stable Diffusion discord](https://huggingface.co/datasets/bartman081523/stable-diffusion-discord-prompts), [Lexica.art](https://huggingface.co/datasets/Gustavosta/Stable-Diffusion-Prompts), and (my hand-picked) [Krea.ai](https://huggingface.co/datasets/FredZhang7/krea-ai-prompts). I filtered the hand-picked prompts based on the output results from Stable Diffusion v1.4.
70
+
71
+ Compared to other prompt generation models using GPT2, this one runs with 50% faster forwardpropagation and 40% less disk space & RAM.
72
+
73
+
74
+ ### PyTorch
75
+
76
+ ```bash
77
+ pip install --upgrade transformers
78
+ ```
79
+
80
+ ```python
81
+ from transformers import GPT2Tokenizer, GPT2LMHeadModel
82
+
83
+ # load the pretrained tokenizer
84
+ tokenizer = GPT2Tokenizer.from_pretrained('distilgpt2')
85
+ tokenizer.add_special_tokens({'pad_token': '[PAD]'})
86
+ tokenizer.max_len = 512
87
+
88
+ # load the fine-tuned model
89
+ model = GPT2LMHeadModel.from_pretrained('FredZhang7/distilgpt2-stable-diffusion')
90
+
91
+ # generate text using fine-tuned model
92
+ from transformers import pipeline
93
+ nlp = pipeline('text-generation', model=model, tokenizer=tokenizer)
94
+ ins = "a beautiful city"
95
+
96
+ # generate 10 samples
97
+ outs = nlp(ins, max_length=80, num_return_sequences=10)
98
+
99
+ # print the 10 samples
100
+ for i in range(len(outs)):
101
+ outs[i] = str(outs[i]['generated_text']).replace(' ', '')
102
+ print('\033[96m' + ins + '\033[0m')
103
+ print('\033[93m' + '\n\n'.join(outs) + '\033[0m')
104
+ ```
105
+
106
+ Example Output:
107
+ ![Example Output](./prompt-examples.png)
108
+