Upload README.md
Browse files
README.md
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
---
|
2 |
inference: false
|
3 |
-
license:
|
4 |
model_creator: WizardLM
|
5 |
model_link: https://huggingface.co/WizardLM/WizardLM-70B-V1.0
|
6 |
model_name: WizardLM 70B V1.0
|
@@ -9,17 +9,20 @@ quantized_by: TheBloke
|
|
9 |
---
|
10 |
|
11 |
<!-- header start -->
|
12 |
-
|
13 |
-
|
|
|
14 |
</div>
|
15 |
<div style="display: flex; justify-content: space-between; width: 100%;">
|
16 |
<div style="display: flex; flex-direction: column; align-items: flex-start;">
|
17 |
-
<p><a href="https://discord.gg/theblokeai">Chat & support:
|
18 |
</div>
|
19 |
<div style="display: flex; flex-direction: column; align-items: flex-end;">
|
20 |
-
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
|
21 |
</div>
|
22 |
</div>
|
|
|
|
|
23 |
<!-- header end -->
|
24 |
|
25 |
# WizardLM 70B V1.0 - GGML
|
@@ -30,6 +33,14 @@ quantized_by: TheBloke
|
|
30 |
|
31 |
This repo contains GGML format model files for [WizardLM's WizardLM 70B V1.0](https://huggingface.co/WizardLM/WizardLM-70B-V1.0).
|
32 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
33 |
GPU acceleration is now available for Llama 2 70B GGML files, with both CUDA (NVidia) and Metal (macOS). The following clients/libraries are known to work with these files, including with GPU acceleration:
|
34 |
* [llama.cpp](https://github.com/ggerganov/llama.cpp), commit `e76d630` and later.
|
35 |
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI.
|
@@ -41,19 +52,25 @@ GPU acceleration is now available for Llama 2 70B GGML files, with both CUDA (NV
|
|
41 |
## Repositories available
|
42 |
|
43 |
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GPTQ)
|
44 |
-
* [2, 3, 4, 5, 6 and 8-bit
|
|
|
45 |
* [WizardLM's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/WizardLM/WizardLM-70B-V1.0)
|
46 |
|
47 |
## Prompt template: Vicuna
|
48 |
|
49 |
```
|
50 |
A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions. USER: {prompt} ASSISTANT:
|
|
|
51 |
```
|
52 |
|
53 |
<!-- compatibility_ggml start -->
|
54 |
## Compatibility
|
55 |
|
56 |
-
###
|
|
|
|
|
|
|
|
|
57 |
|
58 |
Or one of the other tools and libraries listed above.
|
59 |
|
@@ -82,52 +99,24 @@ Refer to the Provided Files table below to see what files use which methods, and
|
|
82 |
| Name | Quant method | Bits | Size | Max RAM required | Use case |
|
83 |
| ---- | ---- | ---- | ---- | ---- | ----- |
|
84 |
| [wizardlm-70b-v1.0.ggmlv3.q2_K.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q2_K.bin) | q2_K | 2 | 28.96 GB| 31.46 GB | New k-quant method. Uses GGML_TYPE_Q4_K for the attention.vw and feed_forward.w2 tensors, GGML_TYPE_Q2_K for the other tensors. |
|
85 |
-
| [wizardlm-70b-v1.0.ggmlv3.q3_K_L.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q3_K_L.bin) | q3_K_L | 3 | 36.49 GB| 38.99 GB | New k-quant method. Uses GGML_TYPE_Q5_K for the attention.wv, attention.wo, and feed_forward.w2 tensors, else GGML_TYPE_Q3_K |
|
86 |
-
| [wizardlm-70b-v1.0.ggmlv3.q3_K_M.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q3_K_M.bin) | q3_K_M | 3 | 33.39 GB| 35.89 GB | New k-quant method. Uses GGML_TYPE_Q4_K for the attention.wv, attention.wo, and feed_forward.w2 tensors, else GGML_TYPE_Q3_K |
|
87 |
| [wizardlm-70b-v1.0.ggmlv3.q3_K_S.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q3_K_S.bin) | q3_K_S | 3 | 30.09 GB| 32.59 GB | New k-quant method. Uses GGML_TYPE_Q3_K for all tensors |
|
|
|
|
|
88 |
| [wizardlm-70b-v1.0.ggmlv3.q4_0.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q4_0.bin) | q4_0 | 4 | 38.80 GB| 41.30 GB | Original quant method, 4-bit. |
|
89 |
-
| [wizardlm-70b-v1.0.ggmlv3.q4_1.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q4_1.bin) | q4_1 | 4 | 43.12 GB| 45.62 GB | Original quant method, 4-bit. Higher accuracy than q4_0 but not as high as q5_0. However has quicker inference than q5 models. |
|
90 |
-
| [wizardlm-70b-v1.0.ggmlv3.q4_K_M.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q4_K_M.bin) | q4_K_M | 4 | 41.69 GB| 44.19 GB | New k-quant method. Uses GGML_TYPE_Q6_K for half of the attention.wv and feed_forward.w2 tensors, else GGML_TYPE_Q4_K |
|
91 |
| [wizardlm-70b-v1.0.ggmlv3.q4_K_S.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q4_K_S.bin) | q4_K_S | 4 | 39.18 GB| 41.68 GB | New k-quant method. Uses GGML_TYPE_Q4_K for all tensors |
|
|
|
|
|
92 |
| [wizardlm-70b-v1.0.ggmlv3.q5_0.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q5_0.bin) | q5_0 | 5 | 47.43 GB| 49.93 GB | Original quant method, 5-bit. Higher accuracy, higher resource usage and slower inference. |
|
93 |
-
| [wizardlm-70b-v1.0.ggmlv3.q5_K_M.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q5_K_M.bin) | q5_K_M | 5 | 49.03 GB| 51.53 GB | New k-quant method. Uses GGML_TYPE_Q6_K for half of the attention.wv and feed_forward.w2 tensors, else GGML_TYPE_Q5_K |
|
94 |
| [wizardlm-70b-v1.0.ggmlv3.q5_K_S.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q5_K_S.bin) | q5_K_S | 5 | 47.74 GB| 50.24 GB | New k-quant method. Uses GGML_TYPE_Q5_K for all tensors |
|
95 |
-
| wizardlm-70b-v1.0.ggmlv3.
|
96 |
-
| wizardlm-70b-v1.0.ggmlv3.q6_K.bin | q6_K | 6 | 56.59 GB | 59.09 GB | New k-quant method. Uses GGML_TYPE_Q8_K - 6-bit quantization - for all tensors |
|
97 |
-
| wizardlm-70b-v1.0.ggmlv3.q8_0.bin | q8_0 | 8 | 73.23 GB | 75.73 GB | Original llama.cpp quant method, 8-bit. Almost indistinguishable from float16. High resource use and slow. Not recommended for most users. |
|
98 |
|
99 |
**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.
|
100 |
|
101 |
-
|
102 |
-
|
103 |
-
**Note:** HF does not support uploading files larger than 50GB. Therefore I have uploaded the q6_K and q8_0 files as multi-part ZIP files. They are not compressed, they are just for storing a .bin file in two parts.
|
104 |
|
105 |
-
|
106 |
-
<summary>Click for instructions regarding q5_1, q6_K and q8_0 files</summary>
|
107 |
-
|
108 |
-
### q5_1
|
109 |
-
Please download:
|
110 |
-
* `wizardlm-70b-v1.0.ggmlv3.q5_1.zip`
|
111 |
-
* `wizardlm-70b-v1.0.ggmlv3.q5_1.z01`
|
112 |
-
|
113 |
-
### q6_K
|
114 |
-
Please download:
|
115 |
-
* `wizardlm-70b-v1.0.ggmlv3.q6_K.zip`
|
116 |
-
* `wizardlm-70b-v1.0.ggmlv3.q6_K.z01`
|
117 |
-
|
118 |
-
### q8_0
|
119 |
-
Please download:
|
120 |
-
* `wizardlm-70b-v1.0.ggmlv3.q8_0.zip`
|
121 |
-
* `wizardlm-70b-v1.0.ggmlv3.q8_0.z01`
|
122 |
-
|
123 |
-
Then extract the .zip archive. This will will expand both parts automatically. On Linux I found I had to use `7zip` - the basic `unzip` tool did not work. Example:
|
124 |
-
```
|
125 |
-
sudo apt update -y && sudo apt install 7zip
|
126 |
-
7zz x wizardlm-70b-v1.0.ggmlv3.q6_K.zip
|
127 |
-
```
|
128 |
-
</details>
|
129 |
|
130 |
-
|
131 |
|
132 |
I use the following command line; adjust for your tastes and needs:
|
133 |
|
@@ -151,6 +140,7 @@ For other parameters and how to use them, please refer to [the llama.cpp documen
|
|
151 |
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).
|
152 |
|
153 |
<!-- footer start -->
|
|
|
154 |
## Discord
|
155 |
|
156 |
For further support, and discussions on these models and AI in general, join us at:
|
@@ -170,13 +160,15 @@ Donaters will get priority support on any and all AI/LLM/model questions and req
|
|
170 |
* Patreon: https://patreon.com/TheBlokeAI
|
171 |
* Ko-Fi: https://ko-fi.com/TheBlokeAI
|
172 |
|
173 |
-
**Special thanks to**:
|
174 |
|
175 |
-
**Patreon special mentions**:
|
176 |
|
177 |
|
178 |
Thank you to all my generous patrons and donaters!
|
179 |
|
|
|
|
|
180 |
<!-- footer end -->
|
181 |
|
182 |
# Original model card: WizardLM's WizardLM 70B V1.0
|
@@ -188,12 +180,48 @@ Thank you to all my generous patrons and donaters!
|
|
188 |
|
189 |
|
190 |
<p align="center">
|
191 |
-
π€ <a href="https://huggingface.co/WizardLM" target="_blank">HF Repo</a> β’ π¦ <a href="https://twitter.com/WizardLM_AI" target="_blank">Twitter</a> β’ π <a href="https://arxiv.org/abs/2304.12244" target="_blank">[WizardLM]</a> β’ π <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> <br>
|
192 |
</p>
|
193 |
<p align="center">
|
194 |
-
π Join our <a href="https://discord.gg/
|
195 |
</p>
|
196 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
197 |
|
198 |
<font size=4>
|
199 |
|
@@ -205,7 +233,6 @@ Thank you to all my generous patrons and donaters!
|
|
205 |
| <sup>WizardLM-30B-V1.0</sup> | <sup>π€ <a href="https://huggingface.co/WizardLM/WizardLM-30B-V1.0" target="_blank">HF Link</a></sup> | | <sup>7.01</sup> | | | <sup>37.8 pass@1</sup>| <sup>Non-commercial</sup> |
|
206 |
| <sup>WizardLM-13B-V1.0</sup> | <sup>π€ <a href="https://huggingface.co/WizardLM/WizardLM-13B-V1.0" target="_blank">HF Link</a> </sup> | | <sup>6.35</sup> | <sup>75.31%</sup> | | <sup> 24.0 pass@1 </sup> | <sup>Non-commercial</sup>|
|
207 |
| <sup>WizardLM-7B-V1.0 </sup>| <sup>π€ <a href="https://huggingface.co/WizardLM/WizardLM-7B-V1.0" target="_blank">HF Link</a> </sup> |<sup> π <a href="https://arxiv.org/abs/2304.12244" target="_blank">[WizardLM]</a> </sup>| | | |<sup>19.1 pass@1 </sup>|<sup> Non-commercial</sup>|
|
208 |
-
| <sup>WizardCoder-15B-V1.0</sup> | <sup> π€ <a href="https://huggingface.co/WizardLM/WizardCoder-15B-V1.0" target="_blank">HF Link</a></sup> | <sup>π <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a></sup> | | ||<sup> 57.3 pass@1 </sup> | <sup> <a href="https://huggingface.co/spaces/bigcode/bigcode-model-license-agreement" target="_blank">OpenRAIL-M</a></sup> |
|
209 |
</font>
|
210 |
|
211 |
- π₯π₯π₯ [08/09/2023] We released **WizardLM-70B-V1.0** model.
|
@@ -224,5 +251,21 @@ Thank you to all my generous patrons and donaters!
|
|
224 |
<b>WizardLM</b> adopts the prompt format from <b>Vicuna</b> and supports **multi-turn** conversation. The prompt should be as following:
|
225 |
|
226 |
```
|
227 |
-
A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions. USER:
|
228 |
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
inference: false
|
3 |
+
license: llama2
|
4 |
model_creator: WizardLM
|
5 |
model_link: https://huggingface.co/WizardLM/WizardLM-70B-V1.0
|
6 |
model_name: WizardLM 70B V1.0
|
|
|
9 |
---
|
10 |
|
11 |
<!-- header start -->
|
12 |
+
<!-- 200823 -->
|
13 |
+
<div style="width: auto; margin-left: auto; margin-right: auto">
|
14 |
+
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
|
15 |
</div>
|
16 |
<div style="display: flex; justify-content: space-between; width: 100%;">
|
17 |
<div style="display: flex; flex-direction: column; align-items: flex-start;">
|
18 |
+
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
|
19 |
</div>
|
20 |
<div style="display: flex; flex-direction: column; align-items: flex-end;">
|
21 |
+
<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>
|
22 |
</div>
|
23 |
</div>
|
24 |
+
<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>
|
25 |
+
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
|
26 |
<!-- header end -->
|
27 |
|
28 |
# WizardLM 70B V1.0 - GGML
|
|
|
33 |
|
34 |
This repo contains GGML format model files for [WizardLM's WizardLM 70B V1.0](https://huggingface.co/WizardLM/WizardLM-70B-V1.0).
|
35 |
|
36 |
+
### Important note regarding GGML files.
|
37 |
+
|
38 |
+
The GGML format has now been superseded by GGUF. As of August 21st 2023, [llama.cpp](https://github.com/ggerganov/llama.cpp) no longer supports GGML models. Third party clients and libraries are expected to still support it for a time, but many may also drop support.
|
39 |
+
|
40 |
+
Please use the GGUF models instead.
|
41 |
+
|
42 |
+
### About GGML
|
43 |
+
|
44 |
GPU acceleration is now available for Llama 2 70B GGML files, with both CUDA (NVidia) and Metal (macOS). The following clients/libraries are known to work with these files, including with GPU acceleration:
|
45 |
* [llama.cpp](https://github.com/ggerganov/llama.cpp), commit `e76d630` and later.
|
46 |
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI.
|
|
|
52 |
## Repositories available
|
53 |
|
54 |
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GPTQ)
|
55 |
+
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGUF)
|
56 |
+
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU+GPU inference (deprecated)](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML)
|
57 |
* [WizardLM's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/WizardLM/WizardLM-70B-V1.0)
|
58 |
|
59 |
## Prompt template: Vicuna
|
60 |
|
61 |
```
|
62 |
A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions. USER: {prompt} ASSISTANT:
|
63 |
+
|
64 |
```
|
65 |
|
66 |
<!-- compatibility_ggml start -->
|
67 |
## Compatibility
|
68 |
|
69 |
+
### Works with llama.cpp [commit `e76d630`](https://github.com/ggerganov/llama.cpp/commit/e76d630df17e235e6b9ef416c45996765d2e36fb) until August 21st, 2023
|
70 |
+
|
71 |
+
Will not work with `llama.cpp` after commit [dadbed99e65252d79f81101a392d0d6497b86caa](https://github.com/ggerganov/llama.cpp/commit/dadbed99e65252d79f81101a392d0d6497b86caa).
|
72 |
+
|
73 |
+
For compatibility with latest llama.cpp, please use GGUF files instead.
|
74 |
|
75 |
Or one of the other tools and libraries listed above.
|
76 |
|
|
|
99 |
| Name | Quant method | Bits | Size | Max RAM required | Use case |
|
100 |
| ---- | ---- | ---- | ---- | ---- | ----- |
|
101 |
| [wizardlm-70b-v1.0.ggmlv3.q2_K.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q2_K.bin) | q2_K | 2 | 28.96 GB| 31.46 GB | New k-quant method. Uses GGML_TYPE_Q4_K for the attention.vw and feed_forward.w2 tensors, GGML_TYPE_Q2_K for the other tensors. |
|
|
|
|
|
102 |
| [wizardlm-70b-v1.0.ggmlv3.q3_K_S.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q3_K_S.bin) | q3_K_S | 3 | 30.09 GB| 32.59 GB | New k-quant method. Uses GGML_TYPE_Q3_K for all tensors |
|
103 |
+
| [wizardlm-70b-v1.0.ggmlv3.q3_K_M.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q3_K_M.bin) | q3_K_M | 3 | 33.39 GB| 35.89 GB | New k-quant method. Uses GGML_TYPE_Q4_K for the attention.wv, attention.wo, and feed_forward.w2 tensors, else GGML_TYPE_Q3_K |
|
104 |
+
| [wizardlm-70b-v1.0.ggmlv3.q3_K_L.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q3_K_L.bin) | q3_K_L | 3 | 36.49 GB| 38.99 GB | New k-quant method. Uses GGML_TYPE_Q5_K for the attention.wv, attention.wo, and feed_forward.w2 tensors, else GGML_TYPE_Q3_K |
|
105 |
| [wizardlm-70b-v1.0.ggmlv3.q4_0.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q4_0.bin) | q4_0 | 4 | 38.80 GB| 41.30 GB | Original quant method, 4-bit. |
|
|
|
|
|
106 |
| [wizardlm-70b-v1.0.ggmlv3.q4_K_S.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q4_K_S.bin) | q4_K_S | 4 | 39.18 GB| 41.68 GB | New k-quant method. Uses GGML_TYPE_Q4_K for all tensors |
|
107 |
+
| [wizardlm-70b-v1.0.ggmlv3.q4_K_M.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q4_K_M.bin) | q4_K_M | 4 | 41.69 GB| 44.19 GB | New k-quant method. Uses GGML_TYPE_Q6_K for half of the attention.wv and feed_forward.w2 tensors, else GGML_TYPE_Q4_K |
|
108 |
+
| [wizardlm-70b-v1.0.ggmlv3.q4_1.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q4_1.bin) | q4_1 | 4 | 43.12 GB| 45.62 GB | Original quant method, 4-bit. Higher accuracy than q4_0 but not as high as q5_0. However has quicker inference than q5 models. |
|
109 |
| [wizardlm-70b-v1.0.ggmlv3.q5_0.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q5_0.bin) | q5_0 | 5 | 47.43 GB| 49.93 GB | Original quant method, 5-bit. Higher accuracy, higher resource usage and slower inference. |
|
|
|
110 |
| [wizardlm-70b-v1.0.ggmlv3.q5_K_S.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q5_K_S.bin) | q5_K_S | 5 | 47.74 GB| 50.24 GB | New k-quant method. Uses GGML_TYPE_Q5_K for all tensors |
|
111 |
+
| [wizardlm-70b-v1.0.ggmlv3.q5_K_M.bin](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GGML/blob/main/wizardlm-70b-v1.0.ggmlv3.q5_K_M.bin) | q5_K_M | 5 | 49.03 GB| 51.53 GB | New k-quant method. Uses GGML_TYPE_Q6_K for half of the attention.wv and feed_forward.w2 tensors, else GGML_TYPE_Q5_K |
|
|
|
|
|
112 |
|
113 |
**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.
|
114 |
|
115 |
+
## How to run in `llama.cpp`
|
|
|
|
|
116 |
|
117 |
+
Make sure you are using `llama.cpp` from commit [dadbed99e65252d79f81101a392d0d6497b86caa](https://github.com/ggerganov/llama.cpp/commit/dadbed99e65252d79f81101a392d0d6497b86caa) or earlier.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
118 |
|
119 |
+
For compatibility with latest llama.cpp, please use GGUF files instead.
|
120 |
|
121 |
I use the following command line; adjust for your tastes and needs:
|
122 |
|
|
|
140 |
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).
|
141 |
|
142 |
<!-- footer start -->
|
143 |
+
<!-- 200823 -->
|
144 |
## Discord
|
145 |
|
146 |
For further support, and discussions on these models and AI in general, join us at:
|
|
|
160 |
* Patreon: https://patreon.com/TheBlokeAI
|
161 |
* Ko-Fi: https://ko-fi.com/TheBlokeAI
|
162 |
|
163 |
+
**Special thanks to**: Aemon Algiz.
|
164 |
|
165 |
+
**Patreon special mentions**: Russ Johnson, J, alfie_i, Alex, NimbleBox.ai, Chadd, Mandus, Nikolai Manek, Ken Nordquist, ya boyyy, Illia Dulskyi, Viktor Bowallius, vamX, Iucharbius, zynix, Magnesian, Clay Pascal, Pierre Kircher, Enrico Ros, Tony Hughes, Elle, Andrey, knownsqashed, Deep Realms, Jerry Meng, Lone Striker, Derek Yates, Pyrater, Mesiah Bishop, James Bentley, Femi Adebogun, Brandon Frisco, SuperWojo, Alps Aficionado, Michael Dempsey, Vitor Caleffi, Will Dee, Edmond Seymore, usrbinkat, LangChain4j, Kacper WikieΕ, Luke Pendergrass, John Detwiler, theTransient, Nathan LeClaire, Tiffany J. Kim, biorpg, Eugene Pentland, Stanislav Ovsiannikov, Fred von Graf, terasurfer, Kalila, Dan Guido, Nitin Borwankar, ιΏζ, Ai Maven, John Villwock, Gabriel Puliatti, Stephen Murray, Asp the Wyvern, danny, Chris Smitley, ReadyPlayerEmma, S_X, Daniel P. Andersen, Olakabola, Jeffrey Morgan, Imad Khwaja, Caitlyn Gatomon, webtim, Alicia Loh, Trenton Dambrowitz, Swaroop Kallakuri, Erik BjΓ€reholt, Leonard Tan, Spiking Neurons AB, Luke @flexchar, Ajan Kanaga, Thomas Belote, Deo Leter, RoA, Willem Michiel, transmissions 11, subjectnull, Matthew Berman, Joseph William Delisle, David Ziegler, Michael Davis, Johann-Peter Hartmann, Talal Aujan, senxiiz, Artur Olbinski, Rainer Wilmers, Spencer Kim, Fen Risland, Cap'n Zoog, Rishabh Srivastava, Michael Levine, Geoffrey Montalvo, Sean Connelly, Alexandros Triantafyllidis, Pieter, Gabriel Tamborski, Sam, Subspace Studios, Junyu Yang, Pedro Madruga, Vadim, Cory Kujawski, K, Raven Klaugh, Randy H, Mano Prime, Sebastain Graf, Space Cruiser
|
166 |
|
167 |
|
168 |
Thank you to all my generous patrons and donaters!
|
169 |
|
170 |
+
And thank you again to a16z for their generous grant.
|
171 |
+
|
172 |
<!-- footer end -->
|
173 |
|
174 |
# Original model card: WizardLM's WizardLM 70B V1.0
|
|
|
180 |
|
181 |
|
182 |
<p align="center">
|
183 |
+
π€ <a href="https://huggingface.co/WizardLM" target="_blank">HF Repo</a> β’π± <a href="https://github.com/nlpxucan/WizardLM" target="_blank">Github Repo</a> β’ π¦ <a href="https://twitter.com/WizardLM_AI" target="_blank">Twitter</a> β’ π <a href="https://arxiv.org/abs/2304.12244" target="_blank">[WizardLM]</a> β’ π <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> β’ π <a href="https://arxiv.org/abs/2308.09583" target="_blank">[WizardMath]</a> <br>
|
184 |
</p>
|
185 |
<p align="center">
|
186 |
+
π Join our <a href="https://discord.gg/VZjjHtWrKs" target="_blank">Discord</a>
|
187 |
</p>
|
188 |
|
189 |
+
## Unofficial Video Introductions
|
190 |
+
Thanks to the enthusiastic friends, their video introductions are more lively and interesting.
|
191 |
+
1. [NEW WizardLM 70b π₯ Giant Model...Insane Performance](https://www.youtube.com/watch?v=WdpiIXrO4_o)
|
192 |
+
2. [GET WizardLM NOW! 7B LLM KING That Can Beat ChatGPT! I'm IMPRESSED!](https://www.youtube.com/watch?v=SaJ8wyKMBds)
|
193 |
+
3. [WizardLM: Enhancing Large Language Models to Follow Complex Instructions](https://www.youtube.com/watch?v=I6sER-qivYk)
|
194 |
+
4. [WizardCoder AI Is The NEW ChatGPT's Coding TWIN!](https://www.youtube.com/watch?v=XjsyHrmd3Xo)
|
195 |
+
|
196 |
+
|
197 |
+
|
198 |
+
|
199 |
+
|
200 |
+
## News
|
201 |
+
|
202 |
+
- π₯π₯π₯[2023/08/26] We released **WizardCoder-Python-34B-V1.0** , which achieves the **73.2 pass@1** and surpasses **GPT4 (2023/03/15)**, **ChatGPT-3.5**, and **Claude2** on the [HumanEval Benchmarks](https://github.com/openai/human-eval). For more details, please refer to [WizardCoder](https://github.com/nlpxucan/WizardLM/tree/main/WizardCoder).
|
203 |
+
- [2023/06/16] We released **WizardCoder-15B-V1.0** , which surpasses **Claude-Plus (+6.8)**, **Bard (+15.3)** and **InstructCodeT5+ (+22.3)** on the [HumanEval Benchmarks](https://github.com/openai/human-eval). For more details, please refer to [WizardCoder](https://github.com/nlpxucan/WizardLM/tree/main/WizardCoder).
|
204 |
+
|
205 |
+
| Model | Checkpoint | Paper | HumanEval | MBPP | Demo | License |
|
206 |
+
| ----- |------| ---- |------|-------| ----- | ----- |
|
207 |
+
| WizardCoder-Python-34B-V1.0 | π€ <a href="https://huggingface.co/WizardLM/WizardCoder-Python-34B-V1.0" target="_blank">HF Link</a> | π <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> | 73.2 | 61.2 | [Demo](http://47.103.63.15:50085/) | <a href="https://ai.meta.com/resources/models-and-libraries/llama-downloads/" target="_blank">Llama2</a> |
|
208 |
+
| WizardCoder-15B-V1.0 | π€ <a href="https://huggingface.co/WizardLM/WizardCoder-15B-V1.0" target="_blank">HF Link</a> | π <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> | 59.8 |50.6 | -- | <a href="https://huggingface.co/spaces/bigcode/bigcode-model-license-agreement" target="_blank">OpenRAIL-M</a> |
|
209 |
+
| WizardCoder-Python-13B-V1.0 | π€ <a href="https://huggingface.co/WizardLM/WizardCoder-Python-13B-V1.0" target="_blank">HF Link</a> | π <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> | 64.0 | 55.6 | -- | <a href="https://ai.meta.com/resources/models-and-libraries/llama-downloads/" target="_blank">Llama2</a> |
|
210 |
+
| WizardCoder-Python-7B-V1.0 | π€ <a href="https://huggingface.co/WizardLM/WizardCoder-Python-7B-V1.0" target="_blank">HF Link</a> | π <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> | 55.5 | 51.6 | [Demo](http://47.103.63.15:50088/) | <a href="https://ai.meta.com/resources/models-and-libraries/llama-downloads/" target="_blank">Llama2</a> |
|
211 |
+
| WizardCoder-3B-V1.0 | π€ <a href="https://huggingface.co/WizardLM/WizardCoder-3B-V1.0" target="_blank">HF Link</a> | π <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> | 34.8 |37.4 | -- | <a href="https://huggingface.co/spaces/bigcode/bigcode-model-license-agreement" target="_blank">OpenRAIL-M</a> |
|
212 |
+
| WizardCoder-1B-V1.0 | π€ <a href="https://huggingface.co/WizardLM/WizardCoder-1B-V1.0" target="_blank">HF Link</a> | π <a href="https://arxiv.org/abs/2306.08568" target="_blank">[WizardCoder]</a> | 23.8 |28.6 | -- | <a href="https://huggingface.co/spaces/bigcode/bigcode-model-license-agreement" target="_blank">OpenRAIL-M</a> |
|
213 |
+
|
214 |
+
- π₯ [08/11/2023] We release **WizardMath** Models.
|
215 |
+
- π₯ Our **WizardMath-70B-V1.0** model slightly outperforms some closed-source LLMs on the GSM8K, including **ChatGPT 3.5**, **Claude Instant 1** and **PaLM 2 540B**.
|
216 |
+
- π₯ Our **WizardMath-70B-V1.0** model achieves **81.6 pass@1** on the [GSM8k Benchmarks](https://github.com/openai/grade-school-math), which is **24.8** points higher than the SOTA open-source LLM.
|
217 |
+
- π₯ Our **WizardMath-70B-V1.0** model achieves **22.7 pass@1** on the [MATH Benchmarks](https://github.com/hendrycks/math), which is **9.2** points higher than the SOTA open-source LLM.
|
218 |
+
|
219 |
+
| Model | Checkpoint | Paper | GSM8k | MATH |Online Demo| License|
|
220 |
+
| ----- |------| ---- |------|-------| ----- | ----- |
|
221 |
+
| WizardMath-70B-V1.0 | π€ <a href="https://huggingface.co/WizardLM/WizardMath-70B-V1.0" target="_blank">HF Link</a> | π <a href="https://arxiv.org/abs/2308.09583" target="_blank">[WizardMath]</a>| **81.6** | **22.7** |[Demo](http://47.103.63.15:50083/)| <a href="https://ai.meta.com/resources/models-and-libraries/llama-downloads/" target="_blank">Llama 2 </a> |
|
222 |
+
| WizardMath-13B-V1.0 | π€ <a href="https://huggingface.co/WizardLM/WizardMath-13B-V1.0" target="_blank">HF Link</a> | π <a href="https://arxiv.org/abs/2308.09583" target="_blank">[WizardMath]</a>| **63.9** | **14.0** |[Demo](http://47.103.63.15:50082/)| <a href="https://ai.meta.com/resources/models-and-libraries/llama-downloads/" target="_blank">Llama 2 </a> |
|
223 |
+
| WizardMath-7B-V1.0 | π€ <a href="https://huggingface.co/WizardLM/WizardMath-7B-V1.0" target="_blank">HF Link</a> | π <a href="https://arxiv.org/abs/2308.09583" target="_blank">[WizardMath]</a>| **54.9** | **10.7** | [Demo](http://47.103.63.15:50080/)| <a href="https://ai.meta.com/resources/models-and-libraries/llama-downloads/" target="_blank">Llama 2 </a>|
|
224 |
+
|
225 |
|
226 |
<font size=4>
|
227 |
|
|
|
233 |
| <sup>WizardLM-30B-V1.0</sup> | <sup>π€ <a href="https://huggingface.co/WizardLM/WizardLM-30B-V1.0" target="_blank">HF Link</a></sup> | | <sup>7.01</sup> | | | <sup>37.8 pass@1</sup>| <sup>Non-commercial</sup> |
|
234 |
| <sup>WizardLM-13B-V1.0</sup> | <sup>π€ <a href="https://huggingface.co/WizardLM/WizardLM-13B-V1.0" target="_blank">HF Link</a> </sup> | | <sup>6.35</sup> | <sup>75.31%</sup> | | <sup> 24.0 pass@1 </sup> | <sup>Non-commercial</sup>|
|
235 |
| <sup>WizardLM-7B-V1.0 </sup>| <sup>π€ <a href="https://huggingface.co/WizardLM/WizardLM-7B-V1.0" target="_blank">HF Link</a> </sup> |<sup> π <a href="https://arxiv.org/abs/2304.12244" target="_blank">[WizardLM]</a> </sup>| | | |<sup>19.1 pass@1 </sup>|<sup> Non-commercial</sup>|
|
|
|
236 |
</font>
|
237 |
|
238 |
- π₯π₯π₯ [08/09/2023] We released **WizardLM-70B-V1.0** model.
|
|
|
251 |
<b>WizardLM</b> adopts the prompt format from <b>Vicuna</b> and supports **multi-turn** conversation. The prompt should be as following:
|
252 |
|
253 |
```
|
254 |
+
A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions. USER: Hi ASSISTANT: Hello.</s>USER: Who are you? ASSISTANT: I am WizardLM.</s>......
|
255 |
```
|
256 |
+
|
257 |
+
## Inference WizardLM Demo Script
|
258 |
+
|
259 |
+
We provide the inference WizardLM demo code [here](https://github.com/nlpxucan/WizardLM/tree/main/demo).
|
260 |
+
|
261 |
+
|
262 |
+
β<b>To commen concern about dataset:</b>
|
263 |
+
|
264 |
+
Recently, there have been clear changes in the open-source policy and regulations of our overall organization's code, data, and models.
|
265 |
+
|
266 |
+
|
267 |
+
Despite this, we have still worked hard to obtain opening the weights of the model first, but the data involves stricter auditing and is in review with our legal team .
|
268 |
+
|
269 |
+
Our researchers have no authority to publicly release them without authorization.
|
270 |
+
|
271 |
+
Thank you for your understanding.
|