TheBloke commited on
Commit
7fa11b8
1 Parent(s): 0483b8f

Update for Transformers GPTQ support

Browse files
README.md CHANGED
@@ -17,17 +17,20 @@ quantized_by: TheBloke
17
  ---
18
 
19
  <!-- header start -->
20
- <div style="width: 100%;">
21
- <img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
 
22
  </div>
23
  <div style="display: flex; justify-content: space-between; width: 100%;">
24
  <div style="display: flex; flex-direction: column; align-items: flex-start;">
25
- <p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
26
  </div>
27
  <div style="display: flex; flex-direction: column; align-items: flex-end;">
28
- <p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
29
  </div>
30
  </div>
 
 
31
  <!-- header end -->
32
 
33
  # StableBeluga 7B - GPTQ
@@ -66,11 +69,11 @@ Each separate quant is in a different branch. See below for instructions on fet
66
 
67
  | Branch | Bits | Group Size | Act Order (desc_act) | File Size | ExLlama Compatible? | Made With | Description |
68
  | ------ | ---- | ---------- | -------------------- | --------- | ------------------- | --------- | ----------- |
69
- | [main](https://huggingface.co/TheBloke/StableBeluga-7B-GPTQ/tree/main) | 4 | 128 | False | 3.90 GB | True | AutoGPTQ | Most compatible option. Good inference speed in AutoGPTQ and GPTQ-for-LLaMa. Lower inference quality than other options. |
70
- | [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/StableBeluga-7B-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | True | 4.28 GB | True | AutoGPTQ | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. Poor AutoGPTQ CUDA speed. |
71
- | [gptq-4bit-64g-actorder_True](https://huggingface.co/TheBloke/StableBeluga-7B-GPTQ/tree/gptq-4bit-64g-actorder_True) | 4 | 64 | True | 4.02 GB | True | AutoGPTQ | 4-bit, with Act Order and group size 64g. Uses less VRAM than 32g, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
72
- | [gptq-4bit-128g-actorder_True](https://huggingface.co/TheBloke/StableBeluga-7B-GPTQ/tree/gptq-4bit-128g-actorder_True) | 4 | 128 | True | 3.90 GB | True | AutoGPTQ | 4-bit, with Act Order and group size 128g. Uses even less VRAM than 64g, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
73
- | [gptq-8bit--1g-actorder_True](https://huggingface.co/TheBloke/StableBeluga-7B-GPTQ/tree/gptq-8bit--1g-actorder_True) | 8 | None | True | 7.01 GB | False | AutoGPTQ | 8-bit, with Act Order. No group size, to lower VRAM requirements and to improve AutoGPTQ speed. |
74
  | [gptq-8bit-128g-actorder_True](https://huggingface.co/TheBloke/StableBeluga-7B-GPTQ/tree/gptq-8bit-128g-actorder_True) | 8 | 128 | True | 7.16 GB | False | AutoGPTQ | 8-bit, with group size 128g for higher inference quality and with Act Order for even higher accuracy. Poor AutoGPTQ CUDA speed. |
75
 
76
  ## How to download from branches
@@ -114,7 +117,7 @@ from transformers import AutoTokenizer, pipeline, logging
114
  from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
115
 
116
  model_name_or_path = "TheBloke/StableBeluga-7B-GPTQ"
117
- model_basename = "gptq_model-4bit-128g"
118
 
119
  use_triton = False
120
 
@@ -182,6 +185,7 @@ The files provided will work with AutoGPTQ (CUDA and Triton modes), GPTQ-for-LLa
182
  ExLlama works with Llama models in 4-bit. Please see the Provided Files table above for per-file compatibility.
183
 
184
  <!-- footer start -->
 
185
  ## Discord
186
 
187
  For further support, and discussions on these models and AI in general, join us at:
@@ -201,13 +205,15 @@ Donaters will get priority support on any and all AI/LLM/model questions and req
201
  * Patreon: https://patreon.com/TheBlokeAI
202
  * Ko-Fi: https://ko-fi.com/TheBlokeAI
203
 
204
- **Special thanks to**: Luke from CarbonQuill, Aemon Algiz.
205
 
206
- **Patreon special mentions**: Slarti, Chadd, John Detwiler, Pieter, zynix, K, Mano Prime, ReadyPlayerEmma, Ai Maven, Leonard Tan, Edmond Seymore, Joseph William Delisle, Luke @flexchar, Fred von Graf, Viktor Bowallius, Rishabh Srivastava, Nikolai Manek, Matthew Berman, Johann-Peter Hartmann, ya boyyy, Greatston Gnanesh, Femi Adebogun, Talal Aujan, Jonathan Leane, terasurfer, David Flickinger, William Sang, Ajan Kanaga, Vadim, Artur Olbinski, Raven Klaugh, Michael Levine, Oscar Rangel, Randy H, Cory Kujawski, RoA, Dave, Alex, Alexandros Triantafyllidis, Fen Risland, Eugene Pentland, vamX, Elle, Nathan LeClaire, Khalefa Al-Ahmad, Rainer Wilmers, subjectnull, Junyu Yang, Daniel P. Andersen, SuperWojo, LangChain4j, Mandus, Kalila, Illia Dulskyi, Trenton Dambrowitz, Asp the Wyvern, Derek Yates, Jeffrey Morgan, Deep Realms, Imad Khwaja, Pyrater, Preetika Verma, biorpg, Gabriel Tamborski, Stephen Murray, Spiking Neurons AB, Iucharbius, Chris Smitley, Willem Michiel, Luke Pendergrass, Sebastain Graf, senxiiz, Will Dee, Space Cruiser, Karl Bernard, Clay Pascal, Lone Striker, transmissions 11, webtim, WelcomeToTheClub, Sam, theTransient, Pierre Kircher, chris gileta, John Villwock, Sean Connelly, Willian Hasse
207
 
208
 
209
  Thank you to all my generous patrons and donaters!
210
 
 
 
211
  <!-- footer end -->
212
 
213
  # Original model card: Stability AI's StableBeluga 7B
@@ -280,7 +286,7 @@ Beluga is a new technology that carries risks with use. Testing conducted to dat
280
 
281
  ```bibtext
282
  @misc{touvron2023llama,
283
- title={Llama 2: Open Foundation and Fine-Tuned Chat Models},
284
  author={Hugo Touvron and Louis Martin and Kevin Stone and Peter Albert and Amjad Almahairi and Yasmine Babaei and Nikolay Bashlykov and Soumya Batra and Prajjwal Bhargava and Shruti Bhosale and Dan Bikel and Lukas Blecher and Cristian Canton Ferrer and Moya Chen and Guillem Cucurull and David Esiobu and Jude Fernandes and Jeremy Fu and Wenyin Fu and Brian Fuller and Cynthia Gao and Vedanuj Goswami and Naman Goyal and Anthony Hartshorn and Saghar Hosseini and Rui Hou and Hakan Inan and Marcin Kardas and Viktor Kerkez and Madian Khabsa and Isabel Kloumann and Artem Korenev and Punit Singh Koura and Marie-Anne Lachaux and Thibaut Lavril and Jenya Lee and Diana Liskovich and Yinghai Lu and Yuning Mao and Xavier Martinet and Todor Mihaylov and Pushkar Mishra and Igor Molybog and Yixin Nie and Andrew Poulton and Jeremy Reizenstein and Rashi Rungta and Kalyan Saladi and Alan Schelten and Ruan Silva and Eric Michael Smith and Ranjan Subramanian and Xiaoqing Ellen Tan and Binh Tang and Ross Taylor and Adina Williams and Jian Xiang Kuan and Puxin Xu and Zheng Yan and Iliyan Zarov and Yuchen Zhang and Angela Fan and Melanie Kambadur and Sharan Narang and Aurelien Rodriguez and Robert Stojnic and Sergey Edunov and Thomas Scialom},
285
  year={2023},
286
  eprint={2307.09288},
@@ -291,7 +297,7 @@ Beluga is a new technology that carries risks with use. Testing conducted to dat
291
 
292
  ```bibtext
293
  @misc{mukherjee2023orca,
294
- title={Orca: Progressive Learning from Complex Explanation Traces of GPT-4},
295
  author={Subhabrata Mukherjee and Arindam Mitra and Ganesh Jawahar and Sahaj Agarwal and Hamid Palangi and Ahmed Awadallah},
296
  year={2023},
297
  eprint={2306.02707},
 
17
  ---
18
 
19
  <!-- header start -->
20
+ <!-- 200823 -->
21
+ <div style="width: auto; margin-left: auto; margin-right: auto">
22
+ <img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
23
  </div>
24
  <div style="display: flex; justify-content: space-between; width: 100%;">
25
  <div style="display: flex; flex-direction: column; align-items: flex-start;">
26
+ <p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
27
  </div>
28
  <div style="display: flex; flex-direction: column; align-items: flex-end;">
29
+ <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>
30
  </div>
31
  </div>
32
+ <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>
33
+ <hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
34
  <!-- header end -->
35
 
36
  # StableBeluga 7B - GPTQ
 
69
 
70
  | Branch | Bits | Group Size | Act Order (desc_act) | File Size | ExLlama Compatible? | Made With | Description |
71
  | ------ | ---- | ---------- | -------------------- | --------- | ------------------- | --------- | ----------- |
72
+ | [main](https://huggingface.co/TheBloke/StableBeluga-7B-GPTQ/tree/main) | 4 | 128 | False | 3.90 GB | True | AutoGPTQ | Most compatible option. Good inference speed in AutoGPTQ and GPTQ-for-LLaMa. Lower inference quality than other options. |
73
+ | [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/StableBeluga-7B-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | True | 4.28 GB | True | AutoGPTQ | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. Poor AutoGPTQ CUDA speed. |
74
+ | [gptq-4bit-64g-actorder_True](https://huggingface.co/TheBloke/StableBeluga-7B-GPTQ/tree/gptq-4bit-64g-actorder_True) | 4 | 64 | True | 4.02 GB | True | AutoGPTQ | 4-bit, with Act Order and group size 64g. Uses less VRAM than 32g, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
75
+ | [gptq-4bit-128g-actorder_True](https://huggingface.co/TheBloke/StableBeluga-7B-GPTQ/tree/gptq-4bit-128g-actorder_True) | 4 | 128 | True | 3.90 GB | True | AutoGPTQ | 4-bit, with Act Order and group size 128g. Uses even less VRAM than 64g, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
76
+ | [gptq-8bit--1g-actorder_True](https://huggingface.co/TheBloke/StableBeluga-7B-GPTQ/tree/gptq-8bit--1g-actorder_True) | 8 | None | True | 7.01 GB | False | AutoGPTQ | 8-bit, with Act Order. No group size, to lower VRAM requirements and to improve AutoGPTQ speed. |
77
  | [gptq-8bit-128g-actorder_True](https://huggingface.co/TheBloke/StableBeluga-7B-GPTQ/tree/gptq-8bit-128g-actorder_True) | 8 | 128 | True | 7.16 GB | False | AutoGPTQ | 8-bit, with group size 128g for higher inference quality and with Act Order for even higher accuracy. Poor AutoGPTQ CUDA speed. |
78
 
79
  ## How to download from branches
 
117
  from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
118
 
119
  model_name_or_path = "TheBloke/StableBeluga-7B-GPTQ"
120
+ model_basename = "model"
121
 
122
  use_triton = False
123
 
 
185
  ExLlama works with Llama models in 4-bit. Please see the Provided Files table above for per-file compatibility.
186
 
187
  <!-- footer start -->
188
+ <!-- 200823 -->
189
  ## Discord
190
 
191
  For further support, and discussions on these models and AI in general, join us at:
 
205
  * Patreon: https://patreon.com/TheBlokeAI
206
  * Ko-Fi: https://ko-fi.com/TheBlokeAI
207
 
208
+ **Special thanks to**: Aemon Algiz.
209
 
210
+ **Patreon special mentions**: Sam, theTransient, Jonathan Leane, Steven Wood, webtim, Johann-Peter Hartmann, Geoffrey Montalvo, Gabriel Tamborski, Willem Michiel, John Villwock, Derek Yates, Mesiah Bishop, Eugene Pentland, Pieter, Chadd, Stephen Murray, Daniel P. Andersen, terasurfer, Brandon Frisco, Thomas Belote, Sid, Nathan LeClaire, Magnesian, Alps Aficionado, Stanislav Ovsiannikov, Alex, Joseph William Delisle, Nikolai Manek, Michael Davis, Junyu Yang, K, J, Spencer Kim, Stefan Sabev, Olusegun Samson, transmissions 11, Michael Levine, Cory Kujawski, Rainer Wilmers, zynix, Kalila, Luke @flexchar, Ajan Kanaga, Mandus, vamX, Ai Maven, Mano Prime, Matthew Berman, subjectnull, Vitor Caleffi, Clay Pascal, biorpg, alfie_i, 阿明, Jeffrey Morgan, ya boyyy, Raymond Fosdick, knownsqashed, Olakabola, Leonard Tan, ReadyPlayerEmma, Enrico Ros, Dave, Talal Aujan, Illia Dulskyi, Sean Connelly, senxiiz, Artur Olbinski, Elle, Raven Klaugh, Fen Risland, Deep Realms, Imad Khwaja, Fred von Graf, Will Dee, usrbinkat, SuperWojo, Alexandros Triantafyllidis, Swaroop Kallakuri, Dan Guido, John Detwiler, Pedro Madruga, Iucharbius, Viktor Bowallius, Asp the Wyvern, Edmond Seymore, Trenton Dambrowitz, Space Cruiser, Spiking Neurons AB, Pyrater, LangChain4j, Tony Hughes, Kacper Wikieł, Rishabh Srivastava, David Ziegler, Luke Pendergrass, Andrey, Gabriel Puliatti, Lone Striker, Sebastain Graf, Pierre Kircher, Randy H, NimbleBox.ai, Vadim, danny, Deo Leter
211
 
212
 
213
  Thank you to all my generous patrons and donaters!
214
 
215
+ And thank you again to a16z for their generous grant.
216
+
217
  <!-- footer end -->
218
 
219
  # Original model card: Stability AI's StableBeluga 7B
 
286
 
287
  ```bibtext
288
  @misc{touvron2023llama,
289
+ title={Llama 2: Open Foundation and Fine-Tuned Chat Models},
290
  author={Hugo Touvron and Louis Martin and Kevin Stone and Peter Albert and Amjad Almahairi and Yasmine Babaei and Nikolay Bashlykov and Soumya Batra and Prajjwal Bhargava and Shruti Bhosale and Dan Bikel and Lukas Blecher and Cristian Canton Ferrer and Moya Chen and Guillem Cucurull and David Esiobu and Jude Fernandes and Jeremy Fu and Wenyin Fu and Brian Fuller and Cynthia Gao and Vedanuj Goswami and Naman Goyal and Anthony Hartshorn and Saghar Hosseini and Rui Hou and Hakan Inan and Marcin Kardas and Viktor Kerkez and Madian Khabsa and Isabel Kloumann and Artem Korenev and Punit Singh Koura and Marie-Anne Lachaux and Thibaut Lavril and Jenya Lee and Diana Liskovich and Yinghai Lu and Yuning Mao and Xavier Martinet and Todor Mihaylov and Pushkar Mishra and Igor Molybog and Yixin Nie and Andrew Poulton and Jeremy Reizenstein and Rashi Rungta and Kalyan Saladi and Alan Schelten and Ruan Silva and Eric Michael Smith and Ranjan Subramanian and Xiaoqing Ellen Tan and Binh Tang and Ross Taylor and Adina Williams and Jian Xiang Kuan and Puxin Xu and Zheng Yan and Iliyan Zarov and Yuchen Zhang and Angela Fan and Melanie Kambadur and Sharan Narang and Aurelien Rodriguez and Robert Stojnic and Sergey Edunov and Thomas Scialom},
291
  year={2023},
292
  eprint={2307.09288},
 
297
 
298
  ```bibtext
299
  @misc{mukherjee2023orca,
300
+ title={Orca: Progressive Learning from Complex Explanation Traces of GPT-4},
301
  author={Subhabrata Mukherjee and Arindam Mitra and Ganesh Jawahar and Sahaj Agarwal and Hamid Palangi and Ahmed Awadallah},
302
  year={2023},
303
  eprint={2306.02707},
config.json CHANGED
@@ -1,25 +1,36 @@
1
  {
2
- "architectures": [
3
- "LlamaForCausalLM"
4
- ],
5
- "bos_token_id": 1,
6
- "eos_token_id": 2,
7
- "hidden_act": "silu",
8
- "hidden_size": 4096,
9
- "initializer_range": 0.02,
10
- "intermediate_size": 11008,
11
- "max_position_embeddings": 4096,
12
- "model_type": "llama",
13
- "num_attention_heads": 32,
14
- "num_hidden_layers": 32,
15
- "num_key_value_heads": 32,
16
- "pad_token_id": 0,
17
- "pretraining_tp": 1,
18
- "rms_norm_eps": 1e-05,
19
- "rope_scaling": null,
20
- "tie_word_embeddings": false,
21
- "torch_dtype": "float16",
22
- "transformers_version": "4.32.0.dev0",
23
- "use_cache": true,
24
- "vocab_size": 32000
 
 
 
 
 
 
 
 
 
 
 
25
  }
 
1
  {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
4
+ ],
5
+ "bos_token_id": 1,
6
+ "eos_token_id": 2,
7
+ "hidden_act": "silu",
8
+ "hidden_size": 4096,
9
+ "initializer_range": 0.02,
10
+ "intermediate_size": 11008,
11
+ "max_position_embeddings": 4096,
12
+ "model_type": "llama",
13
+ "num_attention_heads": 32,
14
+ "num_hidden_layers": 32,
15
+ "num_key_value_heads": 32,
16
+ "pad_token_id": 0,
17
+ "pretraining_tp": 1,
18
+ "rms_norm_eps": 1e-05,
19
+ "rope_scaling": null,
20
+ "tie_word_embeddings": false,
21
+ "torch_dtype": "float16",
22
+ "transformers_version": "4.32.0.dev0",
23
+ "use_cache": true,
24
+ "vocab_size": 32000,
25
+ "quantization_config": {
26
+ "bits": 4,
27
+ "group_size": 128,
28
+ "damp_percent": 0.1,
29
+ "desc_act": false,
30
+ "sym": true,
31
+ "true_sequential": true,
32
+ "model_name_or_path": null,
33
+ "model_file_base_name": "model",
34
+ "quant_method": "gptq"
35
+ }
36
  }
gptq_model-4bit-128g.safetensors → model.safetensors RENAMED
File without changes
quantize_config.json CHANGED
@@ -6,5 +6,5 @@
6
  "sym": true,
7
  "true_sequential": true,
8
  "model_name_or_path": null,
9
- "model_file_base_name": null
10
  }
 
6
  "sym": true,
7
  "true_sequential": true,
8
  "model_name_or_path": null,
9
+ "model_file_base_name": "model"
10
  }