Upload README.md
Browse files
README.md
ADDED
@@ -0,0 +1,529 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model: HuggingFaceH4/zephyr-7b-beta
|
3 |
+
datasets:
|
4 |
+
- HuggingFaceH4/ultrachat_200k
|
5 |
+
- HuggingFaceH4/ultrafeedback_binarized
|
6 |
+
inference: false
|
7 |
+
language:
|
8 |
+
- en
|
9 |
+
license: mit
|
10 |
+
model-index:
|
11 |
+
- name: zephyr-7b-beta
|
12 |
+
results: []
|
13 |
+
model_creator: Hugging Face H4
|
14 |
+
model_name: Zephyr 7B Beta
|
15 |
+
model_type: mistral
|
16 |
+
prompt_template: '<|system|>
|
17 |
+
|
18 |
+
</s>
|
19 |
+
|
20 |
+
<|user|>
|
21 |
+
|
22 |
+
{prompt}</s>
|
23 |
+
|
24 |
+
<|assistant|>
|
25 |
+
|
26 |
+
'
|
27 |
+
quantized_by: TheBloke
|
28 |
+
tags:
|
29 |
+
- generated_from_trainer
|
30 |
+
---
|
31 |
+
<!-- markdownlint-disable MD041 -->
|
32 |
+
|
33 |
+
<!-- header start -->
|
34 |
+
<!-- 200823 -->
|
35 |
+
<div style="width: auto; margin-left: auto; margin-right: auto">
|
36 |
+
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
|
37 |
+
</div>
|
38 |
+
<div style="display: flex; justify-content: space-between; width: 100%;">
|
39 |
+
<div style="display: flex; flex-direction: column; align-items: flex-start;">
|
40 |
+
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
|
41 |
+
</div>
|
42 |
+
<div style="display: flex; flex-direction: column; align-items: flex-end;">
|
43 |
+
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
|
44 |
+
</div>
|
45 |
+
</div>
|
46 |
+
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
|
47 |
+
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
|
48 |
+
<!-- header end -->
|
49 |
+
|
50 |
+
# Zephyr 7B Beta - GGUF
|
51 |
+
- Model creator: [Hugging Face H4](https://huggingface.co/HuggingFaceH4)
|
52 |
+
- Original model: [Zephyr 7B Beta](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta)
|
53 |
+
|
54 |
+
<!-- description start -->
|
55 |
+
## Description
|
56 |
+
|
57 |
+
This repo contains GGUF format model files for [Hugging Face H4's Zephyr 7B Beta](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta).
|
58 |
+
|
59 |
+
<!-- description end -->
|
60 |
+
<!-- README_GGUF.md-about-gguf start -->
|
61 |
+
### About GGUF
|
62 |
+
|
63 |
+
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
|
64 |
+
|
65 |
+
Here is an incomplate list of clients and libraries that are known to support GGUF:
|
66 |
+
|
67 |
+
* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.
|
68 |
+
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
|
69 |
+
* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
|
70 |
+
* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration.
|
71 |
+
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.
|
72 |
+
* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
|
73 |
+
* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server.
|
74 |
+
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
|
75 |
+
* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.
|
76 |
+
|
77 |
+
<!-- README_GGUF.md-about-gguf end -->
|
78 |
+
<!-- repositories-available start -->
|
79 |
+
## Repositories available
|
80 |
+
|
81 |
+
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/zephyr-7B-beta-AWQ)
|
82 |
+
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/zephyr-7B-beta-GPTQ)
|
83 |
+
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF)
|
84 |
+
* [Hugging Face H4's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta)
|
85 |
+
<!-- repositories-available end -->
|
86 |
+
|
87 |
+
<!-- prompt-template start -->
|
88 |
+
## Prompt template: Zephyr
|
89 |
+
|
90 |
+
```
|
91 |
+
<|system|>
|
92 |
+
</s>
|
93 |
+
<|user|>
|
94 |
+
{prompt}</s>
|
95 |
+
<|assistant|>
|
96 |
+
|
97 |
+
```
|
98 |
+
|
99 |
+
<!-- prompt-template end -->
|
100 |
+
|
101 |
+
|
102 |
+
<!-- compatibility_gguf start -->
|
103 |
+
## Compatibility
|
104 |
+
|
105 |
+
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)
|
106 |
+
|
107 |
+
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
|
108 |
+
|
109 |
+
## Explanation of quantisation methods
|
110 |
+
|
111 |
+
<details>
|
112 |
+
<summary>Click to see details</summary>
|
113 |
+
|
114 |
+
The new methods available are:
|
115 |
+
|
116 |
+
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
|
117 |
+
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
|
118 |
+
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
|
119 |
+
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
|
120 |
+
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
|
121 |
+
|
122 |
+
Refer to the Provided Files table below to see what files use which methods, and how.
|
123 |
+
</details>
|
124 |
+
<!-- compatibility_gguf end -->
|
125 |
+
|
126 |
+
<!-- README_GGUF.md-provided-files start -->
|
127 |
+
## Provided files
|
128 |
+
|
129 |
+
| Name | Quant method | Bits | Size | Max RAM required | Use case |
|
130 |
+
| ---- | ---- | ---- | ---- | ---- | ----- |
|
131 |
+
| [zephyr-7b-beta.Q2_K.gguf](https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/blob/main/zephyr-7b-beta.Q2_K.gguf) | Q2_K | 2 | 3.08 GB| 5.58 GB | smallest, significant quality loss - not recommended for most purposes |
|
132 |
+
| [zephyr-7b-beta.Q3_K_S.gguf](https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/blob/main/zephyr-7b-beta.Q3_K_S.gguf) | Q3_K_S | 3 | 3.16 GB| 5.66 GB | very small, high quality loss |
|
133 |
+
| [zephyr-7b-beta.Q3_K_M.gguf](https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/blob/main/zephyr-7b-beta.Q3_K_M.gguf) | Q3_K_M | 3 | 3.52 GB| 6.02 GB | very small, high quality loss |
|
134 |
+
| [zephyr-7b-beta.Q3_K_L.gguf](https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/blob/main/zephyr-7b-beta.Q3_K_L.gguf) | Q3_K_L | 3 | 3.82 GB| 6.32 GB | small, substantial quality loss |
|
135 |
+
| [zephyr-7b-beta.Q4_0.gguf](https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/blob/main/zephyr-7b-beta.Q4_0.gguf) | Q4_0 | 4 | 4.11 GB| 6.61 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
|
136 |
+
| [zephyr-7b-beta.Q4_K_S.gguf](https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/blob/main/zephyr-7b-beta.Q4_K_S.gguf) | Q4_K_S | 4 | 4.14 GB| 6.64 GB | small, greater quality loss |
|
137 |
+
| [zephyr-7b-beta.Q4_K_M.gguf](https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/blob/main/zephyr-7b-beta.Q4_K_M.gguf) | Q4_K_M | 4 | 4.37 GB| 6.87 GB | medium, balanced quality - recommended |
|
138 |
+
| [zephyr-7b-beta.Q5_0.gguf](https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/blob/main/zephyr-7b-beta.Q5_0.gguf) | Q5_0 | 5 | 5.00 GB| 7.50 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
|
139 |
+
| [zephyr-7b-beta.Q5_K_S.gguf](https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/blob/main/zephyr-7b-beta.Q5_K_S.gguf) | Q5_K_S | 5 | 5.00 GB| 7.50 GB | large, low quality loss - recommended |
|
140 |
+
| [zephyr-7b-beta.Q5_K_M.gguf](https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/blob/main/zephyr-7b-beta.Q5_K_M.gguf) | Q5_K_M | 5 | 5.13 GB| 7.63 GB | large, very low quality loss - recommended |
|
141 |
+
| [zephyr-7b-beta.Q6_K.gguf](https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/blob/main/zephyr-7b-beta.Q6_K.gguf) | Q6_K | 6 | 5.94 GB| 8.44 GB | very large, extremely low quality loss |
|
142 |
+
| [zephyr-7b-beta.Q8_0.gguf](https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/blob/main/zephyr-7b-beta.Q8_0.gguf) | Q8_0 | 8 | 7.70 GB| 10.20 GB | very large, extremely low quality loss - not recommended |
|
143 |
+
|
144 |
+
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
|
145 |
+
|
146 |
+
|
147 |
+
|
148 |
+
<!-- README_GGUF.md-provided-files end -->
|
149 |
+
|
150 |
+
<!-- README_GGUF.md-how-to-download start -->
|
151 |
+
## How to download GGUF files
|
152 |
+
|
153 |
+
**Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
|
154 |
+
|
155 |
+
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
|
156 |
+
|
157 |
+
* LM Studio
|
158 |
+
* LoLLMS Web UI
|
159 |
+
* Faraday.dev
|
160 |
+
|
161 |
+
### In `text-generation-webui`
|
162 |
+
|
163 |
+
Under Download Model, you can enter the model repo: TheBloke/zephyr-7B-beta-GGUF and below it, a specific filename to download, such as: zephyr-7b-beta.Q4_K_M.gguf.
|
164 |
+
|
165 |
+
Then click Download.
|
166 |
+
|
167 |
+
### On the command line, including multiple files at once
|
168 |
+
|
169 |
+
I recommend using the `huggingface-hub` Python library:
|
170 |
+
|
171 |
+
```shell
|
172 |
+
pip3 install huggingface-hub
|
173 |
+
```
|
174 |
+
|
175 |
+
Then you can download any individual model file to the current directory, at high speed, with a command like this:
|
176 |
+
|
177 |
+
```shell
|
178 |
+
huggingface-cli download TheBloke/zephyr-7B-beta-GGUF zephyr-7b-beta.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
|
179 |
+
```
|
180 |
+
|
181 |
+
<details>
|
182 |
+
<summary>More advanced huggingface-cli download usage</summary>
|
183 |
+
|
184 |
+
You can also download multiple files at once with a pattern:
|
185 |
+
|
186 |
+
```shell
|
187 |
+
huggingface-cli download TheBloke/zephyr-7B-beta-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
|
188 |
+
```
|
189 |
+
|
190 |
+
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
|
191 |
+
|
192 |
+
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
|
193 |
+
|
194 |
+
```shell
|
195 |
+
pip3 install hf_transfer
|
196 |
+
```
|
197 |
+
|
198 |
+
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
|
199 |
+
|
200 |
+
```shell
|
201 |
+
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/zephyr-7B-beta-GGUF zephyr-7b-beta.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
|
202 |
+
```
|
203 |
+
|
204 |
+
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
|
205 |
+
</details>
|
206 |
+
<!-- README_GGUF.md-how-to-download end -->
|
207 |
+
|
208 |
+
<!-- README_GGUF.md-how-to-run start -->
|
209 |
+
## Example `llama.cpp` command
|
210 |
+
|
211 |
+
Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
|
212 |
+
|
213 |
+
```shell
|
214 |
+
./main -ngl 32 -m zephyr-7b-beta.Q4_K_M.gguf --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "<|system|>\n</s>\n<|user|>\n{prompt}</s>\n<|assistant|>"
|
215 |
+
```
|
216 |
+
|
217 |
+
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
|
218 |
+
|
219 |
+
Change `-c 2048` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically.
|
220 |
+
|
221 |
+
If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
|
222 |
+
|
223 |
+
For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)
|
224 |
+
|
225 |
+
## How to run in `text-generation-webui`
|
226 |
+
|
227 |
+
Further instructions here: [text-generation-webui/docs/llama.cpp.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp.md).
|
228 |
+
|
229 |
+
## How to run from Python code
|
230 |
+
|
231 |
+
You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries.
|
232 |
+
|
233 |
+
### How to load this model in Python code, using ctransformers
|
234 |
+
|
235 |
+
#### First install the package
|
236 |
+
|
237 |
+
Run one of the following commands, according to your system:
|
238 |
+
|
239 |
+
```shell
|
240 |
+
# Base ctransformers with no GPU acceleration
|
241 |
+
pip install ctransformers
|
242 |
+
# Or with CUDA GPU acceleration
|
243 |
+
pip install ctransformers[cuda]
|
244 |
+
# Or with AMD ROCm GPU acceleration (Linux only)
|
245 |
+
CT_HIPBLAS=1 pip install ctransformers --no-binary ctransformers
|
246 |
+
# Or with Metal GPU acceleration for macOS systems only
|
247 |
+
CT_METAL=1 pip install ctransformers --no-binary ctransformers
|
248 |
+
```
|
249 |
+
|
250 |
+
#### Simple ctransformers example code
|
251 |
+
|
252 |
+
```python
|
253 |
+
from ctransformers import AutoModelForCausalLM
|
254 |
+
|
255 |
+
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
|
256 |
+
llm = AutoModelForCausalLM.from_pretrained("TheBloke/zephyr-7B-beta-GGUF", model_file="zephyr-7b-beta.Q4_K_M.gguf", model_type="mistral", gpu_layers=50)
|
257 |
+
|
258 |
+
print(llm("AI is going to"))
|
259 |
+
```
|
260 |
+
|
261 |
+
## How to use with LangChain
|
262 |
+
|
263 |
+
Here are guides on using llama-cpp-python and ctransformers with LangChain:
|
264 |
+
|
265 |
+
* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
|
266 |
+
* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
|
267 |
+
|
268 |
+
<!-- README_GGUF.md-how-to-run end -->
|
269 |
+
|
270 |
+
<!-- footer start -->
|
271 |
+
<!-- 200823 -->
|
272 |
+
## Discord
|
273 |
+
|
274 |
+
For further support, and discussions on these models and AI in general, join us at:
|
275 |
+
|
276 |
+
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
|
277 |
+
|
278 |
+
## Thanks, and how to contribute
|
279 |
+
|
280 |
+
Thanks to the [chirper.ai](https://chirper.ai) team!
|
281 |
+
|
282 |
+
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
|
283 |
+
|
284 |
+
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
|
285 |
+
|
286 |
+
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
|
287 |
+
|
288 |
+
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
|
289 |
+
|
290 |
+
* Patreon: https://patreon.com/TheBlokeAI
|
291 |
+
* Ko-Fi: https://ko-fi.com/TheBlokeAI
|
292 |
+
|
293 |
+
**Special thanks to**: Aemon Algiz.
|
294 |
+
|
295 |
+
**Patreon special mentions**: Pierre Kircher, Stanislav Ovsiannikov, Michael Levine, Eugene Pentland, Andrey, 준교 김, Randy H, Fred von Graf, Artur Olbinski, Caitlyn Gatomon, terasurfer, Jeff Scroggin, James Bentley, Vadim, Gabriel Puliatti, Harry Royden McLaughlin, Sean Connelly, Dan Guido, Edmond Seymore, Alicia Loh, subjectnull, AzureBlack, Manuel Alberto Morcote, Thomas Belote, Lone Striker, Chris Smitley, Vitor Caleffi, Johann-Peter Hartmann, Clay Pascal, biorpg, Brandon Frisco, sidney chen, transmissions 11, Pedro Madruga, jinyuan sun, Ajan Kanaga, Emad Mostaque, Trenton Dambrowitz, Jonathan Leane, Iucharbius, usrbinkat, vamX, George Stoitzev, Luke Pendergrass, theTransient, Olakabola, Swaroop Kallakuri, Cap'n Zoog, Brandon Phillips, Michael Dempsey, Nikolai Manek, danny, Matthew Berman, Gabriel Tamborski, alfie_i, Raymond Fosdick, Tom X Nguyen, Raven Klaugh, LangChain4j, Magnesian, Illia Dulskyi, David Ziegler, Mano Prime, Luis Javier Navarrete Lozano, Erik Bjäreholt, 阿明, Nathan Dryer, Alex, Rainer Wilmers, zynix, TL, Joseph William Delisle, John Villwock, Nathan LeClaire, Willem Michiel, Joguhyik, GodLy, OG, Alps Aficionado, Jeffrey Morgan, ReadyPlayerEmma, Tiffany J. Kim, Sebastain Graf, Spencer Kim, Michael Davis, webtim, Talal Aujan, knownsqashed, John Detwiler, Imad Khwaja, Deo Leter, Jerry Meng, Elijah Stavena, Rooh Singh, Pieter, SuperWojo, Alexandros Triantafyllidis, Stephen Murray, Ai Maven, ya boyyy, Enrico Ros, Ken Nordquist, Deep Realms, Nicholas, Spiking Neurons AB, Elle, Will Dee, Jack West, RoA, Luke @flexchar, Viktor Bowallius, Derek Yates, Subspace Studios, jjj, Toran Billups, Asp the Wyvern, Fen Risland, Ilya, NimbleBox.ai, Chadd, Nitin Borwankar, Emre, Mandus, Leonard Tan, Kalila, K, Trailburnt, S_X, Cory Kujawski
|
296 |
+
|
297 |
+
|
298 |
+
Thank you to all my generous patrons and donaters!
|
299 |
+
|
300 |
+
And thank you again to a16z for their generous grant.
|
301 |
+
|
302 |
+
<!-- footer end -->
|
303 |
+
|
304 |
+
<!-- original-model-card start -->
|
305 |
+
# Original model card: Hugging Face H4's Zephyr 7B Beta
|
306 |
+
|
307 |
+
|
308 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
309 |
+
should probably proofread and complete it, then remove this comment. -->
|
310 |
+
|
311 |
+
<img src="https://huggingface.co/HuggingFaceH4/zephyr-7b-alpha/resolve/main/thumbnail.png" alt="Zephyr Logo" width="800" style="margin-left:'auto' margin-right:'auto' display:'block'"/>
|
312 |
+
|
313 |
+
|
314 |
+
# Model Card for Zephyr 7B β
|
315 |
+
|
316 |
+
Zephyr is a series of language models that are trained to act as helpful assistants. Zephyr-7B-β is the second model in the series, and is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) that was trained on on a mix of publicly available, synthetic datasets using [Direct Preference Optimization (DPO)](https://arxiv.org/abs/2305.18290). We found that removing the in-built alignment of these datasets boosted performance on [MT Bench](https://huggingface.co/spaces/lmsys/mt-bench) and made the model more helpful. However, this means that model is likely to generate problematic text when prompted to do so and should only be used for educational and research purposes. You can find more details in the [technical report](https://arxiv.org/abs/2310.16944).
|
317 |
+
|
318 |
+
|
319 |
+
## Model description
|
320 |
+
|
321 |
+
- **Model type:** A 7B parameter GPT-like model fine-tuned on a mix of publicly available, synthetic datasets.
|
322 |
+
- **Language(s) (NLP):** Primarily English
|
323 |
+
- **License:** MIT
|
324 |
+
- **Finetuned from model:** [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1)
|
325 |
+
|
326 |
+
### Model Sources
|
327 |
+
|
328 |
+
<!-- Provide the basic links for the model. -->
|
329 |
+
|
330 |
+
- **Repository:** https://github.com/huggingface/alignment-handbook
|
331 |
+
- **Demo:** https://huggingface.co/spaces/HuggingFaceH4/zephyr-chat
|
332 |
+
- **Chatbot Arena:** Evaluate Zephyr 7B against 10+ LLMs in the LMSYS arena: http://arena.lmsys.org
|
333 |
+
|
334 |
+
## Performance
|
335 |
+
|
336 |
+
At the time of release, Zephyr-7B-β is the highest ranked 7B chat model on the [MT-Bench](https://huggingface.co/spaces/lmsys/mt-bench) and [AlpacaEval](https://tatsu-lab.github.io/alpaca_eval/) benchmarks:
|
337 |
+
|
338 |
+
| Model | Size | Alignment | MT-Bench (score) | AlpacaEval (win rate %) |
|
339 |
+
|-------------|-----|----|---------------|--------------|
|
340 |
+
| StableLM-Tuned-α | 7B| dSFT |2.75| -|
|
341 |
+
| MPT-Chat | 7B |dSFT |5.42| -|
|
342 |
+
| Xwin-LMv0.1 | 7B| dPPO| 6.19| 87.83|
|
343 |
+
| Mistral-Instructv0.1 | 7B| - | 6.84 |-|
|
344 |
+
| Zephyr-7b-α |7B| dDPO| 6.88| -|
|
345 |
+
| **Zephyr-7b-β** 🪁 | **7B** | **dDPO** | **7.34** | **90.60** |
|
346 |
+
| Falcon-Instruct | 40B |dSFT |5.17 |45.71|
|
347 |
+
| Guanaco | 65B | SFT |6.41| 71.80|
|
348 |
+
| Llama2-Chat | 70B |RLHF |6.86| 92.66|
|
349 |
+
| Vicuna v1.3 | 33B |dSFT |7.12 |88.99|
|
350 |
+
| WizardLM v1.0 | 70B |dSFT |7.71 |-|
|
351 |
+
| Xwin-LM v0.1 | 70B |dPPO |- |95.57|
|
352 |
+
| GPT-3.5-turbo | - |RLHF |7.94 |89.37|
|
353 |
+
| Claude 2 | - |RLHF |8.06| 91.36|
|
354 |
+
| GPT-4 | -| RLHF |8.99| 95.28|
|
355 |
+
|
356 |
+
In particular, on several categories of MT-Bench, Zephyr-7B-β has strong performance compared to larger open models like Llama2-Chat-70B:
|
357 |
+
|
358 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6200d0a443eb0913fa2df7cc/raxvt5ma16d7T23my34WC.png)
|
359 |
+
|
360 |
+
However, on more complex tasks like coding and mathematics, Zephyr-7B-β lags behind proprietary models and more research is needed to close the gap.
|
361 |
+
|
362 |
+
|
363 |
+
## Intended uses & limitations
|
364 |
+
|
365 |
+
The model was initially fine-tuned on a filtered and preprocessed of the [`UltraChat`](https://huggingface.co/datasets/stingning/ultrachat) dataset, which contains a diverse range of synthetic dialogues generated by ChatGPT.
|
366 |
+
We then further aligned the model with [🤗 TRL's](https://github.com/huggingface/trl) `DPOTrainer` on the [openbmb/UltraFeedback](https://huggingface.co/datasets/openbmb/UltraFeedback) dataset, which contains 64k prompts and model completions that are ranked by GPT-4. As a result, the model can be used for chat and you can check out our [demo](https://huggingface.co/spaces/HuggingFaceH4/zephyr-chat) to test its capabilities.
|
367 |
+
|
368 |
+
You can find the datasets used for training Zephyr-7B-β [here](https://huggingface.co/collections/HuggingFaceH4/zephyr-7b-6538c6d6d5ddd1cbb1744a66)
|
369 |
+
|
370 |
+
Here's how you can run the model using the `pipeline()` function from 🤗 Transformers:
|
371 |
+
|
372 |
+
```python
|
373 |
+
# Install transformers from source - only needed for versions <= v4.34
|
374 |
+
# pip install git+https://github.com/huggingface/transformers.git
|
375 |
+
# pip install accelerate
|
376 |
+
|
377 |
+
import torch
|
378 |
+
from transformers import pipeline
|
379 |
+
|
380 |
+
pipe = pipeline("text-generation", model="HuggingFaceH4/zephyr-7b-beta", torch_dtype=torch.bfloat16, device_map="auto")
|
381 |
+
|
382 |
+
# We use the tokenizer's chat template to format each message - see https://huggingface.co/docs/transformers/main/en/chat_templating
|
383 |
+
messages = [
|
384 |
+
{
|
385 |
+
"role": "system",
|
386 |
+
"content": "You are a friendly chatbot who always responds in the style of a pirate",
|
387 |
+
},
|
388 |
+
{"role": "user", "content": "How many helicopters can a human eat in one sitting?"},
|
389 |
+
]
|
390 |
+
prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
|
391 |
+
outputs = pipe(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
|
392 |
+
print(outputs[0]["generated_text"])
|
393 |
+
# <|system|>
|
394 |
+
# You are a friendly chatbot who always responds in the style of a pirate.</s>
|
395 |
+
# <|user|>
|
396 |
+
# How many helicopters can a human eat in one sitting?</s>
|
397 |
+
# <|assistant|>
|
398 |
+
# Ah, me hearty matey! But yer question be a puzzler! A human cannot eat a helicopter in one sitting, as helicopters are not edible. They be made of metal, plastic, and other materials, not food!
|
399 |
+
```
|
400 |
+
|
401 |
+
## Bias, Risks, and Limitations
|
402 |
+
|
403 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
404 |
+
|
405 |
+
Zephyr-7B-β has not been aligned to human preferences with techniques like RLHF or deployed with in-the-loop filtering of responses like ChatGPT, so the model can produce problematic outputs (especially when prompted to do so).
|
406 |
+
It is also unknown what the size and composition of the corpus was used to train the base model (`mistralai/Mistral-7B-v0.1`), however it is likely to have included a mix of Web data and technical sources like books and code. See the [Falcon 180B model card](https://huggingface.co/tiiuae/falcon-180B#training-data) for an example of this.
|
407 |
+
|
408 |
+
|
409 |
+
## Training and evaluation data
|
410 |
+
|
411 |
+
During DPO training, this model achieves the following results on the evaluation set:
|
412 |
+
|
413 |
+
- Loss: 0.7496
|
414 |
+
- Rewards/chosen: -4.5221
|
415 |
+
- Rewards/rejected: -8.3184
|
416 |
+
- Rewards/accuracies: 0.7812
|
417 |
+
- Rewards/margins: 3.7963
|
418 |
+
- Logps/rejected: -340.1541
|
419 |
+
- Logps/chosen: -299.4561
|
420 |
+
- Logits/rejected: -2.3081
|
421 |
+
- Logits/chosen: -2.3531
|
422 |
+
|
423 |
+
|
424 |
+
### Training hyperparameters
|
425 |
+
|
426 |
+
The following hyperparameters were used during training:
|
427 |
+
- learning_rate: 5e-07
|
428 |
+
- train_batch_size: 2
|
429 |
+
- eval_batch_size: 4
|
430 |
+
- seed: 42
|
431 |
+
- distributed_type: multi-GPU
|
432 |
+
- num_devices: 16
|
433 |
+
- total_train_batch_size: 32
|
434 |
+
- total_eval_batch_size: 64
|
435 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
436 |
+
- lr_scheduler_type: linear
|
437 |
+
- lr_scheduler_warmup_ratio: 0.1
|
438 |
+
- num_epochs: 3.0
|
439 |
+
|
440 |
+
### Training results
|
441 |
+
|
442 |
+
The table below shows the full set of DPO training metrics:
|
443 |
+
|
444 |
+
|
445 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
446 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
447 |
+
| 0.6284 | 0.05 | 100 | 0.6098 | 0.0425 | -0.1872 | 0.7344 | 0.2297 | -258.8416 | -253.8099 | -2.7976 | -2.8234 |
|
448 |
+
| 0.4908 | 0.1 | 200 | 0.5426 | -0.0279 | -0.6842 | 0.75 | 0.6563 | -263.8124 | -254.5145 | -2.7719 | -2.7960 |
|
449 |
+
| 0.5264 | 0.15 | 300 | 0.5324 | 0.0414 | -0.9793 | 0.7656 | 1.0207 | -266.7627 | -253.8209 | -2.7892 | -2.8122 |
|
450 |
+
| 0.5536 | 0.21 | 400 | 0.4957 | -0.0185 | -1.5276 | 0.7969 | 1.5091 | -272.2460 | -254.4203 | -2.8542 | -2.8764 |
|
451 |
+
| 0.5362 | 0.26 | 500 | 0.5031 | -0.2630 | -1.5917 | 0.7812 | 1.3287 | -272.8869 | -256.8653 | -2.8702 | -2.8958 |
|
452 |
+
| 0.5966 | 0.31 | 600 | 0.5963 | -0.2993 | -1.6491 | 0.7812 | 1.3499 | -273.4614 | -257.2279 | -2.8778 | -2.8986 |
|
453 |
+
| 0.5014 | 0.36 | 700 | 0.5382 | -0.2859 | -1.4750 | 0.75 | 1.1891 | -271.7204 | -257.0942 | -2.7659 | -2.7869 |
|
454 |
+
| 0.5334 | 0.41 | 800 | 0.5677 | -0.4289 | -1.8968 | 0.7969 | 1.4679 | -275.9378 | -258.5242 | -2.7053 | -2.7265 |
|
455 |
+
| 0.5251 | 0.46 | 900 | 0.5772 | -0.2116 | -1.3107 | 0.7344 | 1.0991 | -270.0768 | -256.3507 | -2.8463 | -2.8662 |
|
456 |
+
| 0.5205 | 0.52 | 1000 | 0.5262 | -0.3792 | -1.8585 | 0.7188 | 1.4793 | -275.5552 | -258.0276 | -2.7893 | -2.7979 |
|
457 |
+
| 0.5094 | 0.57 | 1100 | 0.5433 | -0.6279 | -1.9368 | 0.7969 | 1.3089 | -276.3377 | -260.5136 | -2.7453 | -2.7536 |
|
458 |
+
| 0.5837 | 0.62 | 1200 | 0.5349 | -0.3780 | -1.9584 | 0.7656 | 1.5804 | -276.5542 | -258.0154 | -2.7643 | -2.7756 |
|
459 |
+
| 0.5214 | 0.67 | 1300 | 0.5732 | -1.0055 | -2.2306 | 0.7656 | 1.2251 | -279.2761 | -264.2903 | -2.6986 | -2.7113 |
|
460 |
+
| 0.6914 | 0.72 | 1400 | 0.5137 | -0.6912 | -2.1775 | 0.7969 | 1.4863 | -278.7448 | -261.1467 | -2.7166 | -2.7275 |
|
461 |
+
| 0.4655 | 0.77 | 1500 | 0.5090 | -0.7987 | -2.2930 | 0.7031 | 1.4943 | -279.8999 | -262.2220 | -2.6651 | -2.6838 |
|
462 |
+
| 0.5731 | 0.83 | 1600 | 0.5312 | -0.8253 | -2.3520 | 0.7812 | 1.5268 | -280.4902 | -262.4876 | -2.6543 | -2.6728 |
|
463 |
+
| 0.5233 | 0.88 | 1700 | 0.5206 | -0.4573 | -2.0951 | 0.7812 | 1.6377 | -277.9205 | -258.8084 | -2.6870 | -2.7097 |
|
464 |
+
| 0.5593 | 0.93 | 1800 | 0.5231 | -0.5508 | -2.2000 | 0.7969 | 1.6492 | -278.9703 | -259.7433 | -2.6221 | -2.6519 |
|
465 |
+
| 0.4967 | 0.98 | 1900 | 0.5290 | -0.5340 | -1.9570 | 0.8281 | 1.4230 | -276.5395 | -259.5749 | -2.6564 | -2.6878 |
|
466 |
+
| 0.0921 | 1.03 | 2000 | 0.5368 | -1.1376 | -3.1615 | 0.7812 | 2.0239 | -288.5854 | -265.6111 | -2.6040 | -2.6345 |
|
467 |
+
| 0.0733 | 1.08 | 2100 | 0.5453 | -1.1045 | -3.4451 | 0.7656 | 2.3406 | -291.4208 | -265.2799 | -2.6289 | -2.6595 |
|
468 |
+
| 0.0972 | 1.14 | 2200 | 0.5571 | -1.6915 | -3.9823 | 0.8125 | 2.2908 | -296.7934 | -271.1505 | -2.6471 | -2.6709 |
|
469 |
+
| 0.1058 | 1.19 | 2300 | 0.5789 | -1.0621 | -3.8941 | 0.7969 | 2.8319 | -295.9106 | -264.8563 | -2.5527 | -2.5798 |
|
470 |
+
| 0.2423 | 1.24 | 2400 | 0.5455 | -1.1963 | -3.5590 | 0.7812 | 2.3627 | -292.5599 | -266.1981 | -2.5414 | -2.5784 |
|
471 |
+
| 0.1177 | 1.29 | 2500 | 0.5889 | -1.8141 | -4.3942 | 0.7969 | 2.5801 | -300.9120 | -272.3761 | -2.4802 | -2.5189 |
|
472 |
+
| 0.1213 | 1.34 | 2600 | 0.5683 | -1.4608 | -3.8420 | 0.8125 | 2.3812 | -295.3901 | -268.8436 | -2.4774 | -2.5207 |
|
473 |
+
| 0.0889 | 1.39 | 2700 | 0.5890 | -1.6007 | -3.7337 | 0.7812 | 2.1330 | -294.3068 | -270.2423 | -2.4123 | -2.4522 |
|
474 |
+
| 0.0995 | 1.45 | 2800 | 0.6073 | -1.5519 | -3.8362 | 0.8281 | 2.2843 | -295.3315 | -269.7538 | -2.4685 | -2.5050 |
|
475 |
+
| 0.1145 | 1.5 | 2900 | 0.5790 | -1.7939 | -4.2876 | 0.8438 | 2.4937 | -299.8461 | -272.1744 | -2.4272 | -2.4674 |
|
476 |
+
| 0.0644 | 1.55 | 3000 | 0.5735 | -1.7285 | -4.2051 | 0.8125 | 2.4766 | -299.0209 | -271.5201 | -2.4193 | -2.4574 |
|
477 |
+
| 0.0798 | 1.6 | 3100 | 0.5537 | -1.7226 | -4.2850 | 0.8438 | 2.5624 | -299.8200 | -271.4610 | -2.5367 | -2.5696 |
|
478 |
+
| 0.1013 | 1.65 | 3200 | 0.5575 | -1.5715 | -3.9813 | 0.875 | 2.4098 | -296.7825 | -269.9498 | -2.4926 | -2.5267 |
|
479 |
+
| 0.1254 | 1.7 | 3300 | 0.5905 | -1.6412 | -4.4703 | 0.8594 | 2.8291 | -301.6730 | -270.6473 | -2.5017 | -2.5340 |
|
480 |
+
| 0.085 | 1.76 | 3400 | 0.6133 | -1.9159 | -4.6760 | 0.8438 | 2.7601 | -303.7296 | -273.3941 | -2.4614 | -2.4960 |
|
481 |
+
| 0.065 | 1.81 | 3500 | 0.6074 | -1.8237 | -4.3525 | 0.8594 | 2.5288 | -300.4951 | -272.4724 | -2.4597 | -2.5004 |
|
482 |
+
| 0.0755 | 1.86 | 3600 | 0.5836 | -1.9252 | -4.4005 | 0.8125 | 2.4753 | -300.9748 | -273.4872 | -2.4327 | -2.4716 |
|
483 |
+
| 0.0746 | 1.91 | 3700 | 0.5789 | -1.9280 | -4.4906 | 0.8125 | 2.5626 | -301.8762 | -273.5149 | -2.4686 | -2.5115 |
|
484 |
+
| 0.1348 | 1.96 | 3800 | 0.6015 | -1.8658 | -4.2428 | 0.8281 | 2.3769 | -299.3976 | -272.8936 | -2.4943 | -2.5393 |
|
485 |
+
| 0.0217 | 2.01 | 3900 | 0.6122 | -2.3335 | -4.9229 | 0.8281 | 2.5894 | -306.1988 | -277.5699 | -2.4841 | -2.5272 |
|
486 |
+
| 0.0219 | 2.07 | 4000 | 0.6522 | -2.9890 | -6.0164 | 0.8281 | 3.0274 | -317.1334 | -284.1248 | -2.4105 | -2.4545 |
|
487 |
+
| 0.0119 | 2.12 | 4100 | 0.6922 | -3.4777 | -6.6749 | 0.7969 | 3.1972 | -323.7187 | -289.0121 | -2.4272 | -2.4699 |
|
488 |
+
| 0.0153 | 2.17 | 4200 | 0.6993 | -3.2406 | -6.6775 | 0.7969 | 3.4369 | -323.7453 | -286.6413 | -2.4047 | -2.4465 |
|
489 |
+
| 0.011 | 2.22 | 4300 | 0.7178 | -3.7991 | -7.4397 | 0.7656 | 3.6406 | -331.3667 | -292.2260 | -2.3843 | -2.4290 |
|
490 |
+
| 0.0072 | 2.27 | 4400 | 0.6840 | -3.3269 | -6.8021 | 0.8125 | 3.4752 | -324.9908 | -287.5042 | -2.4095 | -2.4536 |
|
491 |
+
| 0.0197 | 2.32 | 4500 | 0.7013 | -3.6890 | -7.3014 | 0.8125 | 3.6124 | -329.9841 | -291.1250 | -2.4118 | -2.4543 |
|
492 |
+
| 0.0182 | 2.37 | 4600 | 0.7476 | -3.8994 | -7.5366 | 0.8281 | 3.6372 | -332.3356 | -293.2291 | -2.4163 | -2.4565 |
|
493 |
+
| 0.0125 | 2.43 | 4700 | 0.7199 | -4.0560 | -7.5765 | 0.8438 | 3.5204 | -332.7345 | -294.7952 | -2.3699 | -2.4100 |
|
494 |
+
| 0.0082 | 2.48 | 4800 | 0.7048 | -3.6613 | -7.1356 | 0.875 | 3.4743 | -328.3255 | -290.8477 | -2.3925 | -2.4303 |
|
495 |
+
| 0.0118 | 2.53 | 4900 | 0.6976 | -3.7908 | -7.3152 | 0.8125 | 3.5244 | -330.1224 | -292.1431 | -2.3633 | -2.4047 |
|
496 |
+
| 0.0118 | 2.58 | 5000 | 0.7198 | -3.9049 | -7.5557 | 0.8281 | 3.6508 | -332.5271 | -293.2844 | -2.3764 | -2.4194 |
|
497 |
+
| 0.006 | 2.63 | 5100 | 0.7506 | -4.2118 | -7.9149 | 0.8125 | 3.7032 | -336.1194 | -296.3530 | -2.3407 | -2.3860 |
|
498 |
+
| 0.0143 | 2.68 | 5200 | 0.7408 | -4.2433 | -7.9802 | 0.8125 | 3.7369 | -336.7721 | -296.6682 | -2.3509 | -2.3946 |
|
499 |
+
| 0.0057 | 2.74 | 5300 | 0.7552 | -4.3392 | -8.0831 | 0.7969 | 3.7439 | -337.8013 | -297.6275 | -2.3388 | -2.3842 |
|
500 |
+
| 0.0138 | 2.79 | 5400 | 0.7404 | -4.2395 | -7.9762 | 0.8125 | 3.7367 | -336.7322 | -296.6304 | -2.3286 | -2.3737 |
|
501 |
+
| 0.0079 | 2.84 | 5500 | 0.7525 | -4.4466 | -8.2196 | 0.7812 | 3.7731 | -339.1662 | -298.7007 | -2.3200 | -2.3641 |
|
502 |
+
| 0.0077 | 2.89 | 5600 | 0.7520 | -4.5586 | -8.3485 | 0.7969 | 3.7899 | -340.4545 | -299.8206 | -2.3078 | -2.3517 |
|
503 |
+
| 0.0094 | 2.94 | 5700 | 0.7527 | -4.5542 | -8.3509 | 0.7812 | 3.7967 | -340.4790 | -299.7773 | -2.3062 | -2.3510 |
|
504 |
+
| 0.0054 | 2.99 | 5800 | 0.7520 | -4.5169 | -8.3079 | 0.7812 | 3.7911 | -340.0493 | -299.4038 | -2.3081 | -2.3530 |
|
505 |
+
|
506 |
+
|
507 |
+
### Framework versions
|
508 |
+
|
509 |
+
- Transformers 4.35.0.dev0
|
510 |
+
- Pytorch 2.0.1+cu118
|
511 |
+
- Datasets 2.12.0
|
512 |
+
- Tokenizers 0.14.0
|
513 |
+
|
514 |
+
## Citation
|
515 |
+
|
516 |
+
If you find Zephyr-7B-β is useful in your work, please cite it with:
|
517 |
+
|
518 |
+
```
|
519 |
+
@misc{tunstall2023zephyr,
|
520 |
+
title={Zephyr: Direct Distillation of LM Alignment},
|
521 |
+
author={Lewis Tunstall and Edward Beeching and Nathan Lambert and Nazneen Rajani and Kashif Rasul and Younes Belkada and Shengyi Huang and Leandro von Werra and Clémentine Fourrier and Nathan Habib and Nathan Sarrazin and Omar Sanseviero and Alexander M. Rush and Thomas Wolf},
|
522 |
+
year={2023},
|
523 |
+
eprint={2310.16944},
|
524 |
+
archivePrefix={arXiv},
|
525 |
+
primaryClass={cs.LG}
|
526 |
+
}
|
527 |
+
```
|
528 |
+
|
529 |
+
<!-- original-model-card end -->
|