TheBloke commited on
Commit
4e8d76e
1 Parent(s): 123edfd

Update for Transformers GPTQ support

Browse files
README.md CHANGED
@@ -13,17 +13,20 @@ pipeline_tag: text-generation
13
  ---
14
 
15
  <!-- header start -->
16
- <div style="width: 100%;">
17
- <img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
 
18
  </div>
19
  <div style="display: flex; justify-content: space-between; width: 100%;">
20
  <div style="display: flex; flex-direction: column; align-items: flex-start;">
21
- <p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
22
  </div>
23
  <div style="display: flex; flex-direction: column; align-items: flex-end;">
24
- <p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
25
  </div>
26
  </div>
 
 
27
  <!-- header end -->
28
 
29
  # StableBeluga 2 - GGML
@@ -64,14 +67,14 @@ Each separate quant is in a different branch. See below for instructions on fet
64
 
65
  | Branch | Bits | Group Size | Act Order (desc_act) | File Size | ExLlama Compatible? | Made With | Description |
66
  | ------ | ---- | ---------- | -------------------- | --------- | ------------------- | --------- | ----------- |
67
- | main | 4 | None | True | 35.33 GB | True | AutoGPTQ | Most compatible option. Good inference speed in AutoGPTQ and GPTQ-for-LLaMa. Lower inference quality than other options. |
68
- | gptq-4bit-32g-actorder_True | 4 | 32 | True | 40.66 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. |
69
- | gptq-4bit-128g-actorder_True | 4 | 128 | True | 36.65 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. |
70
- | gptq-4bit-64g-actorder_True | 4 | 64 | True | 37.99 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. |
71
- | gptq-3bit--1g-actorder_True | 3 | None | True | 26.78 GB | False | AutoGPTQ | 3-bit, with Act Order and no group size. Lowest possible VRAM requirements. May be lower quality than 3-bit 128g. |
72
- | gptq-3bit-128g-actorder_False | 3 | 128 | False | 28.03 GB | False | AutoGPTQ | 3-bit, with group size 128g but no act-order. Slightly higher VRAM requirements than 3-bit None. |
73
- | gptq-3bit-128g-actorder_True | 3 | 128 | True | 28.03 GB | False | AutoGPTQ | 3-bit, with group size 128g and act-order. Higher quality than 128g-False but poor AutoGPTQ CUDA speed. |
74
- | gptq-3bit-64g-actorder_True | 3 | 64 | True | 29.30 GB | False | AutoGPTQ | 3-bit, with group size 64g and act-order. Highest quality 3-bit option. Poor AutoGPTQ CUDA speed. |
75
  | gptq-4bit-128g-actorder_False | 4 | 128 | False | 36.65 GB | True | AutoGPTQ | 4-bit, without Act Order and group size 128g. |
76
 
77
  ## How to download from branches
@@ -115,7 +118,7 @@ from transformers import AutoTokenizer, pipeline, logging
115
  from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
116
 
117
  model_name_or_path = "TheBloke/StableBeluga2-70B-GPTQ"
118
- model_basename = "gptq_model-4bit--1g"
119
 
120
  use_triton = False
121
 
@@ -184,6 +187,7 @@ The files provided will work with AutoGPTQ (CUDA and Triton modes), GPTQ-for-LLa
184
  ExLlama works with Llama models in 4-bit. Please see the Provided Files table above for per-file compatibility.
185
 
186
  <!-- footer start -->
 
187
  ## Discord
188
 
189
  For further support, and discussions on these models and AI in general, join us at:
@@ -203,13 +207,15 @@ Donaters will get priority support on any and all AI/LLM/model questions and req
203
  * Patreon: https://patreon.com/TheBlokeAI
204
  * Ko-Fi: https://ko-fi.com/TheBlokeAI
205
 
206
- **Special thanks to**: Luke from CarbonQuill, Aemon Algiz.
207
 
208
- **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
209
 
210
 
211
  Thank you to all my generous patrons and donaters!
212
 
 
 
213
  <!-- footer end -->
214
 
215
  # Original model card: Stability AI's StableBeluga 2
@@ -282,7 +288,7 @@ Beluga is a new technology that carries risks with use. Testing conducted to dat
282
 
283
  ```bibtext
284
  @misc{touvron2023llama,
285
- title={Llama 2: Open Foundation and Fine-Tuned Chat Models},
286
  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},
287
  year={2023},
288
  eprint={2307.09288},
@@ -293,11 +299,11 @@ Beluga is a new technology that carries risks with use. Testing conducted to dat
293
 
294
  ```bibtext
295
  @misc{mukherjee2023orca,
296
- title={Orca: Progressive Learning from Complex Explanation Traces of GPT-4},
297
  author={Subhabrata Mukherjee and Arindam Mitra and Ganesh Jawahar and Sahaj Agarwal and Hamid Palangi and Ahmed Awadallah},
298
  year={2023},
299
  eprint={2306.02707},
300
  archivePrefix={arXiv},
301
  primaryClass={cs.CL}
302
  }
303
- ```
 
13
  ---
14
 
15
  <!-- header start -->
16
+ <!-- 200823 -->
17
+ <div style="width: auto; margin-left: auto; margin-right: auto">
18
+ <img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
19
  </div>
20
  <div style="display: flex; justify-content: space-between; width: 100%;">
21
  <div style="display: flex; flex-direction: column; align-items: flex-start;">
22
+ <p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
23
  </div>
24
  <div style="display: flex; flex-direction: column; align-items: flex-end;">
25
+ <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>
26
  </div>
27
  </div>
28
+ <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>
29
+ <hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
30
  <!-- header end -->
31
 
32
  # StableBeluga 2 - GGML
 
67
 
68
  | Branch | Bits | Group Size | Act Order (desc_act) | File Size | ExLlama Compatible? | Made With | Description |
69
  | ------ | ---- | ---------- | -------------------- | --------- | ------------------- | --------- | ----------- |
70
+ | main | 4 | None | True | 35.33 GB | True | AutoGPTQ | Most compatible option. Good inference speed in AutoGPTQ and GPTQ-for-LLaMa. Lower inference quality than other options. |
71
+ | gptq-4bit-32g-actorder_True | 4 | 32 | True | 40.66 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. |
72
+ | gptq-4bit-128g-actorder_True | 4 | 128 | True | 36.65 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-4bit-64g-actorder_True | 4 | 64 | True | 37.99 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. |
74
+ | gptq-3bit--1g-actorder_True | 3 | None | True | 26.78 GB | False | AutoGPTQ | 3-bit, with Act Order and no group size. Lowest possible VRAM requirements. May be lower quality than 3-bit 128g. |
75
+ | gptq-3bit-128g-actorder_False | 3 | 128 | False | 28.03 GB | False | AutoGPTQ | 3-bit, with group size 128g but no act-order. Slightly higher VRAM requirements than 3-bit None. |
76
+ | gptq-3bit-128g-actorder_True | 3 | 128 | True | 28.03 GB | False | AutoGPTQ | 3-bit, with group size 128g and act-order. Higher quality than 128g-False but poor AutoGPTQ CUDA speed. |
77
+ | gptq-3bit-64g-actorder_True | 3 | 64 | True | 29.30 GB | False | AutoGPTQ | 3-bit, with group size 64g and act-order. Highest quality 3-bit option. Poor AutoGPTQ CUDA speed. |
78
  | gptq-4bit-128g-actorder_False | 4 | 128 | False | 36.65 GB | True | AutoGPTQ | 4-bit, without Act Order and group size 128g. |
79
 
80
  ## How to download from branches
 
118
  from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
119
 
120
  model_name_or_path = "TheBloke/StableBeluga2-70B-GPTQ"
121
+ model_basename = "model"
122
 
123
  use_triton = False
124
 
 
187
  ExLlama works with Llama models in 4-bit. Please see the Provided Files table above for per-file compatibility.
188
 
189
  <!-- footer start -->
190
+ <!-- 200823 -->
191
  ## Discord
192
 
193
  For further support, and discussions on these models and AI in general, join us at:
 
207
  * Patreon: https://patreon.com/TheBlokeAI
208
  * Ko-Fi: https://ko-fi.com/TheBlokeAI
209
 
210
+ **Special thanks to**: Aemon Algiz.
211
 
212
+ **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
213
 
214
 
215
  Thank you to all my generous patrons and donaters!
216
 
217
+ And thank you again to a16z for their generous grant.
218
+
219
  <!-- footer end -->
220
 
221
  # Original model card: Stability AI's StableBeluga 2
 
288
 
289
  ```bibtext
290
  @misc{touvron2023llama,
291
+ title={Llama 2: Open Foundation and Fine-Tuned Chat Models},
292
  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},
293
  year={2023},
294
  eprint={2307.09288},
 
299
 
300
  ```bibtext
301
  @misc{mukherjee2023orca,
302
+ title={Orca: Progressive Learning from Complex Explanation Traces of GPT-4},
303
  author={Subhabrata Mukherjee and Arindam Mitra and Ganesh Jawahar and Sahaj Agarwal and Hamid Palangi and Ahmed Awadallah},
304
  year={2023},
305
  eprint={2306.02707},
306
  archivePrefix={arXiv},
307
  primaryClass={cs.CL}
308
  }
309
+ ```
config.json CHANGED
@@ -1,26 +1,37 @@
1
  {
2
- "_name_or_path": "/fsx/dakota/orca/tmp_orca",
3
- "architectures": [
4
- "LlamaForCausalLM"
5
- ],
6
- "bos_token_id": 1,
7
- "eos_token_id": 2,
8
- "hidden_act": "silu",
9
- "hidden_size": 8192,
10
- "initializer_range": 0.02,
11
- "intermediate_size": 28672,
12
- "max_position_embeddings": 4096,
13
- "model_type": "llama",
14
- "num_attention_heads": 64,
15
- "num_hidden_layers": 80,
16
- "num_key_value_heads": 8,
17
- "pad_token_id": 0,
18
- "pretraining_tp": 1,
19
- "rms_norm_eps": 1e-05,
20
- "rope_scaling": null,
21
- "tie_word_embeddings": false,
22
- "torch_dtype": "float32",
23
- "transformers_version": "4.32.0.dev0",
24
- "use_cache": true,
25
- "vocab_size": 32000
 
 
 
 
 
 
 
 
 
 
 
26
  }
 
1
  {
2
+ "_name_or_path": "/fsx/dakota/orca/tmp_orca",
3
+ "architectures": [
4
+ "LlamaForCausalLM"
5
+ ],
6
+ "bos_token_id": 1,
7
+ "eos_token_id": 2,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 8192,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 28672,
12
+ "max_position_embeddings": 4096,
13
+ "model_type": "llama",
14
+ "num_attention_heads": 64,
15
+ "num_hidden_layers": 80,
16
+ "num_key_value_heads": 8,
17
+ "pad_token_id": 0,
18
+ "pretraining_tp": 1,
19
+ "rms_norm_eps": 1e-05,
20
+ "rope_scaling": null,
21
+ "tie_word_embeddings": false,
22
+ "torch_dtype": "float32",
23
+ "transformers_version": "4.32.0.dev0",
24
+ "use_cache": true,
25
+ "vocab_size": 32000,
26
+ "quantization_config": {
27
+ "bits": 4,
28
+ "group_size": -1,
29
+ "damp_percent": 0.01,
30
+ "desc_act": true,
31
+ "sym": true,
32
+ "true_sequential": true,
33
+ "model_name_or_path": null,
34
+ "model_file_base_name": "model",
35
+ "quant_method": "gptq"
36
+ }
37
  }
gptq_model-4bit--1g.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
  }