Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,137 @@
|
|
1 |
---
|
2 |
-
license:
|
|
|
|
|
|
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
+
license: gpl
|
3 |
+
datasets:
|
4 |
+
- nomic-ai/gpt4all-j-prompt-generations
|
5 |
+
language:
|
6 |
+
- en
|
7 |
+
inference: false
|
8 |
---
|
9 |
+
# GPT4All-13B-snoozy-GGML
|
10 |
+
|
11 |
+
These files are GGML format model files of [Nomic.AI's GPT4all-13B-snoozy](https://huggingface.co/nomic-ai/gpt4all-13b-snoozy).
|
12 |
+
|
13 |
+
GGML files are for CPU inference using [llama.cpp](https://github.com/ggerganov/llama.cpp).
|
14 |
+
|
15 |
+
## Repositories available
|
16 |
+
|
17 |
+
* [4bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/GPT4ALL-13B-snoozy-GPTQ).
|
18 |
+
* [4bit and 5bit GGML models for GPU inference](https://huggingface.co/TheBloke/GPT4ALL-13B-snoozy-GGML).
|
19 |
+
* [Nomic.AI's original model in float32 HF for GPU inference](https://huggingface.co/nomic-ai/gpt4all-13b-snoozy).
|
20 |
+
|
21 |
+
|
22 |
+
## Provided files
|
23 |
+
| Name | Quant method | Bits | Size | RAM required | Use case |
|
24 |
+
| ---- | ---- | ---- | ---- | ---- | ----- |
|
25 |
+
`GPT4All-13B-snoozy.q4_0.bin` | q4_0 | 4bit | 8.14GB | 10GB | Maximum compatibility |
|
26 |
+
`GPT4All-13B-snoozy.q4_2.bin` | q4_2 | 4bit | 8.14GB | 10GB | Best compromise between resources, speed and quality |
|
27 |
+
`GPT4All-13B-snoozy.q5_0.bin` | q5_0 | 5bit | 8.95GB | 11GB | Brand new 5bit method. Potentially higher quality than 4bit, at cost of slightly higher resources. |
|
28 |
+
`GPT4All-13B-snoozy.q5_1.bin` | q5_1 | 5bit | 9.76GB | 12GB | Brand new 5bit method. Slightly higher resource usage than q5_0.|
|
29 |
+
|
30 |
+
* The q4_0 file provides lower quality, but maximal compatibility. It will work with past and future versions of llama.cpp
|
31 |
+
* The q4_2 file offers the best combination of performance and quality. This format is still subject to change and there may be compatibility issues, see below.
|
32 |
+
* The q5_0 file is using brand new 5bit method released 26th April. This is the 5bit equivalent of q4_0.
|
33 |
+
* The q5_1 file is using brand new 5bit method released 26th April. This is the 5bit equivalent of q4_1.
|
34 |
+
|
35 |
+
## q4_2 compatibility
|
36 |
+
|
37 |
+
q4_2 is a relatively new 4bit quantisation method offering improved quality. However they are still under development and their formats are subject to change.
|
38 |
+
|
39 |
+
In order to use these files you will need to use recent llama.cpp code. And it's possible that future updates to llama.cpp could require that these files are re-generated.
|
40 |
+
|
41 |
+
If and when the q4_2 file no longer works with recent versions of llama.cpp I will endeavour to update it.
|
42 |
+
|
43 |
+
If you want to ensure guaranteed compatibility with a wide range of llama.cpp versions, use the q4_0 file.
|
44 |
+
|
45 |
+
## q5_0 and q5_1 compatibility
|
46 |
+
|
47 |
+
These new methods were released to llama.cpp on 26th April. You will need to pull the latest llama.cpp code and rebuild to be able to use them.
|
48 |
+
|
49 |
+
Don't expect any third-party UIs/tools to support them yet.
|
50 |
+
|
51 |
+
## How to run in `llama.cpp`
|
52 |
+
|
53 |
+
I use the following command line; adjust for your tastes and needs:
|
54 |
+
|
55 |
+
```
|
56 |
+
./main -t 12 -m GPT4All-13B-snoozy.q4_2.bin --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "Below is an instruction that describes a task. Write a response that appropriately completes the request.
|
57 |
+
### Instruction:
|
58 |
+
Write a story about llamas
|
59 |
+
### Response:"
|
60 |
+
```
|
61 |
+
Change `-t 12` to the number of physical CPU cores you have. For example if your system has 8 cores/16 threads, use `-t 8`.
|
62 |
+
|
63 |
+
If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
|
64 |
+
|
65 |
+
## How to run in `text-generation-webui`
|
66 |
+
|
67 |
+
Further instructions here: [text-generation-webui/docs/llama.cpp-models.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp-models.md).
|
68 |
+
|
69 |
+
Note: at this time text-generation-webui will not support the new q5 quantisation methods.
|
70 |
+
|
71 |
+
**Thireus** has written a [great guide on how to update it to the latest llama.cpp code](https://huggingface.co/TheBloke/wizardLM-7B-GGML/discussions/5) so that these files can be used in the UI.
|
72 |
+
|
73 |
+
## Repositories available
|
74 |
+
|
75 |
+
* [4bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/GPT4ALL-13B-snoozy-GPTQ).
|
76 |
+
* [4bit and 5bit GGML models for GPU inference](https://huggingface.co/TheBloke/GPT4ALL-13B-snoozy-GGML).
|
77 |
+
* [Nomic.AI's original model in float32 HF for GPU inference](https://huggingface.co/nomic-ai/gpt4all-13b-snoozy).
|
78 |
+
|
79 |
+
|
80 |
+
# Original Model Card for GPT4All-13b-snoozy
|
81 |
+
|
82 |
+
An Apache-2 licensed chatbot trained over a massive curated corpus of assistant interactions including word problems, multi-turn dialogue, code, poems, songs, and stories.
|
83 |
+
|
84 |
+
## Model Details
|
85 |
+
|
86 |
+
### Model Description
|
87 |
+
|
88 |
+
<!-- Provide a longer summary of what this model is. -->
|
89 |
+
|
90 |
+
This model has been finetuned from LLama 13B
|
91 |
+
|
92 |
+
- **Developed by:** [Nomic AI](https://home.nomic.ai)
|
93 |
+
- **Model Type:** A finetuned LLama 13B model on assistant style interaction data
|
94 |
+
- **Language(s) (NLP):** English
|
95 |
+
- **License:** Apache-2
|
96 |
+
- **Finetuned from model [optional]:** LLama 13B
|
97 |
+
|
98 |
+
This model was trained on `nomic-ai/gpt4all-j-prompt-generations` using `revision=v1.3-groovy`
|
99 |
+
|
100 |
+
### Model Sources [optional]
|
101 |
+
|
102 |
+
<!-- Provide the basic links for the model. -->
|
103 |
+
|
104 |
+
- **Repository:** [https://github.com/nomic-ai/gpt4all](https://github.com/nomic-ai/gpt4all)
|
105 |
+
- **Base Model Repository:** [https://github.com/facebookresearch/llama](https://github.com/facebookresearch/llama)
|
106 |
+
- **Demo [optional]:** [https://gpt4all.io/](https://gpt4all.io/)
|
107 |
+
|
108 |
+
|
109 |
+
### Results
|
110 |
+
|
111 |
+
Results on common sense reasoning benchmarks
|
112 |
+
|
113 |
+
```
|
114 |
+
Model BoolQ PIQA HellaSwag WinoGrande ARC-e ARC-c OBQA
|
115 |
+
----------------------- ---------- ---------- ----------- ------------ ---------- ---------- ----------
|
116 |
+
GPT4All-J 6B v1.0 73.4 74.8 63.4 64.7 54.9 36.0 40.2
|
117 |
+
GPT4All-J v1.1-breezy 74.0 75.1 63.2 63.6 55.4 34.9 38.4
|
118 |
+
GPT4All-J v1.2-jazzy 74.8 74.9 63.6 63.8 56.6 35.3 41.0
|
119 |
+
GPT4All-J v1.3-groovy 73.6 74.3 63.8 63.5 57.7 35.0 38.8
|
120 |
+
GPT4All-J Lora 6B 68.6 75.8 66.2 63.5 56.4 35.7 40.2
|
121 |
+
GPT4All LLaMa Lora 7B 73.1 77.6 72.1 67.8 51.1 40.4 40.2
|
122 |
+
GPT4All 13B snoozy *83.3* 79.2 75.0 *71.3* 60.9 44.2 43.4
|
123 |
+
Dolly 6B 68.8 77.3 67.6 63.9 62.9 38.7 41.2
|
124 |
+
Dolly 12B 56.7 75.4 71.0 62.2 *64.6* 38.5 40.4
|
125 |
+
Alpaca 7B 73.9 77.2 73.9 66.1 59.8 43.3 43.4
|
126 |
+
Alpaca Lora 7B 74.3 *79.3* 74.0 68.8 56.6 43.9 42.6
|
127 |
+
GPT-J 6B 65.4 76.2 66.2 64.1 62.2 36.6 38.2
|
128 |
+
LLama 7B 73.1 77.4 73.0 66.9 52.5 41.4 42.4
|
129 |
+
LLama 13B 68.5 79.1 *76.2* 70.1 60.0 *44.6* 42.2
|
130 |
+
Pythia 6.9B 63.5 76.3 64.0 61.1 61.3 35.2 37.2
|
131 |
+
Pythia 12B 67.7 76.6 67.3 63.8 63.9 34.8 38.0
|
132 |
+
Vicuña T5 81.5 64.6 46.3 61.8 49.3 33.3 39.4
|
133 |
+
Vicuña 13B 81.5 76.8 73.3 66.7 57.4 42.7 43.6
|
134 |
+
Stable Vicuña RLHF 82.3 78.6 74.1 70.9 61.0 43.5 *44.4*
|
135 |
+
StableLM Tuned 62.5 71.2 53.6 54.8 52.4 31.1 33.4
|
136 |
+
StableLM Base 60.1 67.4 41.2 50.1 44.9 27.0 32.0
|
137 |
+
```
|