TheBloke commited on
Commit
6ddb81f
1 Parent(s): e84f26f

Upload README.md

Browse files
Files changed (1) hide show
  1. README.md +40 -17
README.md CHANGED
@@ -1,14 +1,25 @@
1
  ---
 
2
  inference: false
3
  language:
4
  - en
5
  library_name: transformers
6
- license: llama2
7
  model_creator: OpenLemur
8
- model_link: https://huggingface.co/OpenLemur/lemur-70b-chat-v1
9
  model_name: Lemur 70B Chat v1
10
  model_type: llama
11
  pipeline_tag: text-generation
 
 
 
 
 
 
 
 
 
 
 
12
  quantized_by: TheBloke
13
  tags:
14
  - text-generation
@@ -56,9 +67,9 @@ Multiple GPTQ parameter permutations are provided; see Provided Files below for
56
  <!-- repositories-available start -->
57
  ## Repositories available
58
 
 
59
  * [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ)
60
  * [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GGUF)
61
- * [2, 3, 4, 5, 6 and 8-bit GGML models for CPU+GPU inference (deprecated)](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GGML)
62
  * [OpenLemur's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/OpenLemur/lemur-70b-chat-v1)
63
  <!-- repositories-available end -->
64
 
@@ -75,7 +86,15 @@ Multiple GPTQ parameter permutations are provided; see Provided Files below for
75
  ```
76
 
77
  <!-- prompt-template end -->
 
 
 
 
78
 
 
 
 
 
79
  <!-- README_GPTQ.md-provided-files start -->
80
  ## Provided files and GPTQ parameters
81
 
@@ -100,22 +119,22 @@ All recent GPTQ files are made with AutoGPTQ, and all files in non-main branches
100
 
101
  | Branch | Bits | GS | Act Order | Damp % | GPTQ Dataset | Seq Len | Size | ExLlama | Desc |
102
  | ------ | ---- | -- | --------- | ------ | ------------ | ------- | ---- | ------- | ---- |
103
- | [main](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ/tree/main) | 4 | None | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 35.33 GB | Yes | Most compatible option. Good inference speed in AutoGPTQ and GPTQ-for-LLaMa. Lower inference quality than other options. |
104
- | [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 40.66 GB | Yes | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. Poor AutoGPTQ CUDA speed. |
105
- | [gptq-4bit-64g-actorder_True](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ/tree/gptq-4bit-64g-actorder_True) | 4 | 64 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 37.99 GB | Yes | 4-bit, with Act Order and group size 64g. Uses less VRAM than 32g, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
106
- | [gptq-4bit-128g-actorder_True](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ/tree/gptq-4bit-128g-actorder_True) | 4 | 128 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 36.65 GB | Yes | 4-bit, with Act Order and group size 128g. Uses even less VRAM than 64g, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
107
  | [gptq-3bit--1g-actorder_True](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ/tree/gptq-3bit--1g-actorder_True) | 3 | None | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 26.78 GB | No | 3-bit, with Act Order and no group size. Lowest possible VRAM requirements. May be lower quality than 3-bit 128g. |
108
- | [gptq-3bit-128g-actorder_True](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ/tree/gptq-3bit-128g-actorder_True) | 3 | 128 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 28.03 GB | No | 3-bit, with group size 128g and act-order. Higher quality than 128g-False but poor AutoGPTQ CUDA speed. |
109
 
110
  <!-- README_GPTQ.md-provided-files end -->
111
 
112
  <!-- README_GPTQ.md-download-from-branches start -->
113
  ## How to download from branches
114
 
115
- - In text-generation-webui, you can add `:branch` to the end of the download name, eg `TheBloke/Lemur-70B-Chat-v1-GPTQ:gptq-4bit-32g-actorder_True`
116
  - With Git, you can clone a branch with:
117
  ```
118
- git clone --single-branch --branch gptq-4bit-32g-actorder_True https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ
119
  ```
120
  - In Python Transformers code, the branch is the `revision` parameter; see below.
121
  <!-- README_GPTQ.md-download-from-branches end -->
@@ -128,7 +147,7 @@ It is strongly recommended to use the text-generation-webui one-click-installers
128
 
129
  1. Click the **Model tab**.
130
  2. Under **Download custom model or LoRA**, enter `TheBloke/Lemur-70B-Chat-v1-GPTQ`.
131
- - To download from a specific branch, enter for example `TheBloke/Lemur-70B-Chat-v1-GPTQ:gptq-4bit-32g-actorder_True`
132
  - see Provided Files above for the list of branches for each option.
133
  3. Click **Download**.
134
  4. The model will start downloading. Once it's finished it will say "Done".
@@ -176,10 +195,10 @@ from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
176
 
177
  model_name_or_path = "TheBloke/Lemur-70B-Chat-v1-GPTQ"
178
  # To use a different branch, change revision
179
- # For example: revision="gptq-4bit-32g-actorder_True"
180
  model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
181
- torch_dtype=torch.float16,
182
  device_map="auto",
 
183
  revision="main")
184
 
185
  tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
@@ -196,7 +215,7 @@ prompt_template=f'''<|im_start|>system
196
  print("\n\n*** Generate:")
197
 
198
  input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
199
- output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
200
  print(tokenizer.decode(output[0]))
201
 
202
  # Inference can also be done using transformers' pipeline
@@ -207,9 +226,11 @@ pipe = pipeline(
207
  model=model,
208
  tokenizer=tokenizer,
209
  max_new_tokens=512,
 
210
  temperature=0.7,
211
  top_p=0.95,
212
- repetition_penalty=1.15
 
213
  )
214
 
215
  print(pipe(prompt_template)[0]['generated_text'])
@@ -234,10 +255,12 @@ For further support, and discussions on these models and AI in general, join us
234
 
235
  [TheBloke AI's Discord server](https://discord.gg/theblokeai)
236
 
237
- ## Thanks, and how to contribute.
238
 
239
  Thanks to the [chirper.ai](https://chirper.ai) team!
240
 
 
 
241
  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.
242
 
243
  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.
@@ -249,7 +272,7 @@ Donaters will get priority support on any and all AI/LLM/model questions and req
249
 
250
  **Special thanks to**: Aemon Algiz.
251
 
252
- **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 Wikieł, Luke Pendergrass, John Detwiler, theTransient, Nathan LeClaire, Tiffany J. Kim, biorpg, Eugene Pentland, Stanislav Ovsiannikov, Fred von Graf, terasurfer, Kalila, Dan Guido, Nitin Borwankar, 阿明, 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 Bjäreholt, 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
253
 
254
 
255
  Thank you to all my generous patrons and donaters!
 
1
  ---
2
+ base_model: https://huggingface.co/OpenLemur/lemur-70b-chat-v1
3
  inference: false
4
  language:
5
  - en
6
  library_name: transformers
7
+ license: cc-by-nc-4.0
8
  model_creator: OpenLemur
 
9
  model_name: Lemur 70B Chat v1
10
  model_type: llama
11
  pipeline_tag: text-generation
12
+ prompt_template: '<|im_start|>system
13
+
14
+ {system_message}<|im_end|>
15
+
16
+ <|im_start|>user
17
+
18
+ {prompt}<|im_end|>
19
+
20
+ <|im_start|>assistant
21
+
22
+ '
23
  quantized_by: TheBloke
24
  tags:
25
  - text-generation
 
67
  <!-- repositories-available start -->
68
  ## Repositories available
69
 
70
+ * [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-AWQ)
71
  * [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ)
72
  * [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GGUF)
 
73
  * [OpenLemur's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/OpenLemur/lemur-70b-chat-v1)
74
  <!-- repositories-available end -->
75
 
 
86
  ```
87
 
88
  <!-- prompt-template end -->
89
+ <!-- licensing start -->
90
+ ## Licensing
91
+
92
+ The creator of the source model has listed its license as `cc-by-nc-4.0`, and this quantization has therefore used that same license.
93
 
94
+ As this model is based on Llama 2, it is also subject to the Meta Llama 2 license terms, and the license files for that are additionally included. It should therefore be considered as being claimed to be licensed under both licenses. I contacted Hugging Face for clarification on dual licensing but they do not yet have an official position. Should this change, or should Meta provide any feedback on this situation, I will update this section accordingly.
95
+
96
+ In the meantime, any questions regarding licensing, and in particular how these two licenses might interact, should be directed to the original model repository: [OpenLemur's Lemur 70B Chat v1](https://huggingface.co/OpenLemur/lemur-70b-chat-v1).
97
+ <!-- licensing end -->
98
  <!-- README_GPTQ.md-provided-files start -->
99
  ## Provided files and GPTQ parameters
100
 
 
119
 
120
  | Branch | Bits | GS | Act Order | Damp % | GPTQ Dataset | Seq Len | Size | ExLlama | Desc |
121
  | ------ | ---- | -- | --------- | ------ | ------------ | ------- | ---- | ------- | ---- |
122
+ | [main](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ/tree/main) | 4 | None | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 35.33 GB | Yes | 4-bit, with Act Order. No group size, to lower VRAM requirements. |
123
+ | [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 40.66 GB | Yes | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. |
124
+ | [gptq-4bit-64g-actorder_True](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ/tree/gptq-4bit-64g-actorder_True) | 4 | 64 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 37.99 GB | Yes | 4-bit, with Act Order and group size 64g. Uses less VRAM than 32g, but with slightly lower accuracy. |
125
+ | [gptq-4bit-128g-actorder_True](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ/tree/gptq-4bit-128g-actorder_True) | 4 | 128 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 36.65 GB | Yes | 4-bit, with Act Order and group size 128g. Uses even less VRAM than 64g, but with slightly lower accuracy. |
126
  | [gptq-3bit--1g-actorder_True](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ/tree/gptq-3bit--1g-actorder_True) | 3 | None | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 26.78 GB | No | 3-bit, with Act Order and no group size. Lowest possible VRAM requirements. May be lower quality than 3-bit 128g. |
127
+ | [gptq-3bit-128g-actorder_True](https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ/tree/gptq-3bit-128g-actorder_True) | 3 | 128 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 28.03 GB | No | 3-bit, with group size 128g and act-order. Higher quality than 128g-False. |
128
 
129
  <!-- README_GPTQ.md-provided-files end -->
130
 
131
  <!-- README_GPTQ.md-download-from-branches start -->
132
  ## How to download from branches
133
 
134
+ - In text-generation-webui, you can add `:branch` to the end of the download name, eg `TheBloke/Lemur-70B-Chat-v1-GPTQ:main`
135
  - With Git, you can clone a branch with:
136
  ```
137
+ git clone --single-branch --branch main https://huggingface.co/TheBloke/Lemur-70B-Chat-v1-GPTQ
138
  ```
139
  - In Python Transformers code, the branch is the `revision` parameter; see below.
140
  <!-- README_GPTQ.md-download-from-branches end -->
 
147
 
148
  1. Click the **Model tab**.
149
  2. Under **Download custom model or LoRA**, enter `TheBloke/Lemur-70B-Chat-v1-GPTQ`.
150
+ - To download from a specific branch, enter for example `TheBloke/Lemur-70B-Chat-v1-GPTQ:main`
151
  - see Provided Files above for the list of branches for each option.
152
  3. Click **Download**.
153
  4. The model will start downloading. Once it's finished it will say "Done".
 
195
 
196
  model_name_or_path = "TheBloke/Lemur-70B-Chat-v1-GPTQ"
197
  # To use a different branch, change revision
198
+ # For example: revision="main"
199
  model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
 
200
  device_map="auto",
201
+ trust_remote_code=False,
202
  revision="main")
203
 
204
  tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
 
215
  print("\n\n*** Generate:")
216
 
217
  input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
218
+ output = model.generate(inputs=input_ids, temperature=0.7, do_sample=True, top_p=0.95, top_k=40, max_new_tokens=512)
219
  print(tokenizer.decode(output[0]))
220
 
221
  # Inference can also be done using transformers' pipeline
 
226
  model=model,
227
  tokenizer=tokenizer,
228
  max_new_tokens=512,
229
+ do_sample=True,
230
  temperature=0.7,
231
  top_p=0.95,
232
+ top_k=40,
233
+ repetition_penalty=1.1
234
  )
235
 
236
  print(pipe(prompt_template)[0]['generated_text'])
 
255
 
256
  [TheBloke AI's Discord server](https://discord.gg/theblokeai)
257
 
258
+ ## Thanks, and how to contribute
259
 
260
  Thanks to the [chirper.ai](https://chirper.ai) team!
261
 
262
+ Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
263
+
264
  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.
265
 
266
  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.
 
272
 
273
  **Special thanks to**: Aemon Algiz.
274
 
275
+ **Patreon special mentions**: Alicia Loh, Stephen Murray, K, Ajan Kanaga, RoA, Magnesian, Deo Leter, Olakabola, Eugene Pentland, zynix, Deep Realms, Raymond Fosdick, Elijah Stavena, Iucharbius, Erik Bjäreholt, Luis Javier Navarrete Lozano, Nicholas, theTransient, John Detwiler, alfie_i, knownsqashed, Mano Prime, Willem Michiel, Enrico Ros, LangChain4j, OG, Michael Dempsey, Pierre Kircher, Pedro Madruga, James Bentley, Thomas Belote, Luke @flexchar, Leonard Tan, Johann-Peter Hartmann, Illia Dulskyi, Fen Risland, Chadd, S_X, Jeff Scroggin, Ken Nordquist, Sean Connelly, Artur Olbinski, Swaroop Kallakuri, Jack West, Ai Maven, David Ziegler, Russ Johnson, transmissions 11, John Villwock, Alps Aficionado, Clay Pascal, Viktor Bowallius, Subspace Studios, Rainer Wilmers, Trenton Dambrowitz, vamX, Michael Levine, 준교 김, Brandon Frisco, Kalila, Trailburnt, Randy H, Talal Aujan, Nathan Dryer, Vadim, 阿明, ReadyPlayerEmma, Tiffany J. Kim, George Stoitzev, Spencer Kim, Jerry Meng, Gabriel Tamborski, Cory Kujawski, Jeffrey Morgan, Spiking Neurons AB, Edmond Seymore, Alexandros Triantafyllidis, Lone Striker, Cap'n Zoog, Nikolai Manek, danny, ya boyyy, Derek Yates, usrbinkat, Mandus, TL, Nathan LeClaire, subjectnull, Imad Khwaja, webtim, Raven Klaugh, Asp the Wyvern, Gabriel Puliatti, Caitlyn Gatomon, Joseph William Delisle, Jonathan Leane, Luke Pendergrass, SuperWojo, Sebastain Graf, Will Dee, Fred von Graf, Andrey, Dan Guido, Daniel P. Andersen, Nitin Borwankar, Elle, Vitor Caleffi, biorpg, jjj, NimbleBox.ai, Pieter, Matthew Berman, terasurfer, Michael Davis, Alex, Stanislav Ovsiannikov
276
 
277
 
278
  Thank you to all my generous patrons and donaters!