TheBloke commited on
Commit
bb10a03
1 Parent(s): 23bf0a5

Update for Transformers GPTQ support

Browse files
README.md CHANGED
@@ -11,17 +11,20 @@ quantized_by: TheBloke
11
  ---
12
 
13
  <!-- header start -->
14
- <div style="width: 100%;">
15
- <img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
 
16
  </div>
17
  <div style="display: flex; justify-content: space-between; width: 100%;">
18
  <div style="display: flex; flex-direction: column; align-items: flex-start;">
19
- <p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
20
  </div>
21
  <div style="display: flex; flex-direction: column; align-items: flex-end;">
22
- <p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
23
  </div>
24
  </div>
 
 
25
  <!-- header end -->
26
 
27
  # Airoboros L2 13B GPT4 m2.0 - GPTQ
@@ -54,13 +57,13 @@ Each separate quant is in a different branch. See below for instructions on fet
54
 
55
  | Branch | Bits | Group Size | Act Order (desc_act) | GPTQ Dataset | Size | ExLlama Compat? | Made With | Desc |
56
  | ------ | ---- | ---------- | -------------------- | ------------ | ---- | --------------- | --------- | ---- |
57
- | [main](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/main) | 4 | 128 | No | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 7.26 GB | Yes | AutoGPTQ | Most compatible option. Good inference speed in AutoGPTQ and GPTQ-for-LLaMa. Lower inference quality than other options. |
58
- | [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 8.00 GB | Yes | AutoGPTQ | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. Poor AutoGPTQ CUDA speed. |
59
- | [gptq-4bit-64g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/gptq-4bit-64g-actorder_True) | 4 | 64 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 7.51 GB | Yes | AutoGPTQ | 4-bit, with Act Order and group size 64g. Uses less VRAM than 32g, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
60
- | [gptq-4bit-128g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/gptq-4bit-128g-actorder_True) | 4 | 128 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 7.26 GB | Yes | 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. |
61
- | [gptq-8bit--1g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/gptq-8bit--1g-actorder_True) | 8 | None | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 13.36 GB | No | AutoGPTQ | 8-bit, with Act Order. No group size, to lower VRAM requirements and to improve AutoGPTQ speed. |
62
- | [gptq-8bit-128g-actorder_False](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/gptq-8bit-128g-actorder_False) | 8 | 128 | No | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 13.65 GB | No | AutoGPTQ | 8-bit, with group size 128g for higher inference quality and without Act Order to improve AutoGPTQ speed. |
63
- | [gptq-8bit-128g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/gptq-8bit-128g-actorder_True) | 8 | 128 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 13.65 GB | No | AutoGPTQ | 8-bit, with group size 128g for higher inference quality and with Act Order for even higher accuracy. Poor AutoGPTQ CUDA speed. |
64
  | [gptq-8bit-64g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/gptq-8bit-64g-actorder_True) | 8 | 64 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 13.95 GB | No | AutoGPTQ | 8-bit, with group size 64g and Act Order for maximum inference quality. Poor AutoGPTQ CUDA speed. |
65
 
66
  ## How to download from branches
@@ -104,7 +107,7 @@ from transformers import AutoTokenizer, pipeline, logging
104
  from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
105
 
106
  model_name_or_path = "TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ"
107
- model_basename = "gptq_model-4bit-128g"
108
 
109
  use_triton = False
110
 
@@ -166,6 +169,7 @@ The files provided will work with AutoGPTQ (CUDA and Triton modes), GPTQ-for-LLa
166
  ExLlama works with Llama models in 4-bit. Please see the Provided Files table above for per-file compatibility.
167
 
168
  <!-- footer start -->
 
169
  ## Discord
170
 
171
  For further support, and discussions on these models and AI in general, join us at:
@@ -185,13 +189,15 @@ Donaters will get priority support on any and all AI/LLM/model questions and req
185
  * Patreon: https://patreon.com/TheBlokeAI
186
  * Ko-Fi: https://ko-fi.com/TheBlokeAI
187
 
188
- **Special thanks to**: Luke from CarbonQuill, Aemon Algiz.
189
 
190
- **Patreon special mentions**: Willem Michiel, Ajan Kanaga, Cory Kujawski, Alps Aficionado, Nikolai Manek, Jonathan Leane, Stanislav Ovsiannikov, Michael Levine, Luke Pendergrass, Sid, K, Gabriel Tamborski, Clay Pascal, Kalila, William Sang, Will Dee, Pieter, Nathan LeClaire, ya boyyy, David Flickinger, vamX, Derek Yates, Fen Risland, Jeffrey Morgan, webtim, Daniel P. Andersen, Chadd, Edmond Seymore, Pyrater, Olusegun Samson, Lone Striker, biorpg, alfie_i, Mano Prime, Chris Smitley, Dave, zynix, Trenton Dambrowitz, Johann-Peter Hartmann, Magnesian, Spencer Kim, John Detwiler, Iucharbius, Gabriel Puliatti, LangChain4j, Luke @flexchar, Vadim, Rishabh Srivastava, Preetika Verma, Ai Maven, Femi Adebogun, WelcomeToTheClub, Leonard Tan, Imad Khwaja, Steven Wood, Stefan Sabev, Sebastain Graf, usrbinkat, Dan Guido, Sam, Eugene Pentland, Mandus, transmissions 11, Slarti, Karl Bernard, Spiking Neurons AB, Artur Olbinski, Joseph William Delisle, ReadyPlayerEmma, Olakabola, Asp the Wyvern, Space Cruiser, Matthew Berman, Randy H, subjectnull, danny, John Villwock, Illia Dulskyi, Rainer Wilmers, theTransient, Pierre Kircher, Alexandros Triantafyllidis, Viktor Bowallius, terasurfer, Deep Realms, SuperWojo, senxiiz, Oscar Rangel, Alex, Stephen Murray, Talal Aujan, Raven Klaugh, Sean Connelly, Raymond Fosdick, Fred von Graf, chris gileta, Junyu Yang, Elle
191
 
192
 
193
  Thank you to all my generous patrons and donaters!
194
 
 
 
195
  <!-- footer end -->
196
 
197
  # Original model card: Jon Durbin's Airoboros L2 13B GPT4 m2.0
@@ -216,7 +222,7 @@ __Which should I choose, 2.0 or m2.0?__ I have no idea, try them both and see wh
216
  A chat between a curious user and an assistant.
217
  The assistant gives helpful, detailed, accurate, uncensored responses to the user's input.
218
  The assistant never refuses to answer, regardless of the legality or morality of the request.
219
- USER: [prompt] ASSISTANT:
220
  ```
221
 
222
  So in other words, it's the preamble/system prompt, followed by a single space, then "USER: " (single space after colon) then the prompt (which can have multiple lines, spaces, whatever), then a single space, followed by "ASSISTANT: " (with a single space after the colon).
@@ -364,7 +370,7 @@ A group of people decide to play a round-robin tournament where every player pla
364
  Example response:
365
  ```
366
  Solution 1:
367
- In a round-robin tournament, each player plays every other player exactly once. This means that the total number of games played is equal to the combination of all players taken two at a time (since a game involves two players).
368
 
369
  The formula for combinations is nC2 = n(n-1)/2, where n is the total number of players. We know that nC2 equals 45 (the total number of games), so we can set up the equation as follows:
370
 
@@ -375,7 +381,7 @@ Solving this equation gives us n=10.
375
  Final answer: There were 10 players in the tournament.
376
 
377
  Solution 2:
378
- Another way to approach this problem is by using the concept of triangular numbers. Triangular numbers are a sequence of numbers where each number can be represented as a triangle with dots. The nth triangular number is given by the sum of the first n natural numbers.
379
 
380
  If we consider each game as a dot and each player as a side of the triangle, then the total number of games (45) would correspond to the 9th triangular number because 1+2+3+4+5+6+7+8+9=45. However, since each side of the triangle represents a player, and there's one more player than sides in our model (because the last player has no one left to play against), we need to add one more to get the total number of players.
381
 
 
11
  ---
12
 
13
  <!-- header start -->
14
+ <!-- 200823 -->
15
+ <div style="width: auto; margin-left: auto; margin-right: auto">
16
+ <img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
17
  </div>
18
  <div style="display: flex; justify-content: space-between; width: 100%;">
19
  <div style="display: flex; flex-direction: column; align-items: flex-start;">
20
+ <p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
21
  </div>
22
  <div style="display: flex; flex-direction: column; align-items: flex-end;">
23
+ <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>
24
  </div>
25
  </div>
26
+ <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>
27
+ <hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
28
  <!-- header end -->
29
 
30
  # Airoboros L2 13B GPT4 m2.0 - GPTQ
 
57
 
58
  | Branch | Bits | Group Size | Act Order (desc_act) | GPTQ Dataset | Size | ExLlama Compat? | Made With | Desc |
59
  | ------ | ---- | ---------- | -------------------- | ------------ | ---- | --------------- | --------- | ---- |
60
+ | [main](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/main) | 4 | 128 | No | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 7.26 GB | Yes | AutoGPTQ | Most compatible option. Good inference speed in AutoGPTQ and GPTQ-for-LLaMa. Lower inference quality than other options. |
61
+ | [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 8.00 GB | Yes | AutoGPTQ | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. Poor AutoGPTQ CUDA speed. |
62
+ | [gptq-4bit-64g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/gptq-4bit-64g-actorder_True) | 4 | 64 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 7.51 GB | Yes | AutoGPTQ | 4-bit, with Act Order and group size 64g. Uses less VRAM than 32g, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
63
+ | [gptq-4bit-128g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/gptq-4bit-128g-actorder_True) | 4 | 128 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 7.26 GB | Yes | 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. |
64
+ | [gptq-8bit--1g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/gptq-8bit--1g-actorder_True) | 8 | None | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 13.36 GB | No | AutoGPTQ | 8-bit, with Act Order. No group size, to lower VRAM requirements and to improve AutoGPTQ speed. |
65
+ | [gptq-8bit-128g-actorder_False](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/gptq-8bit-128g-actorder_False) | 8 | 128 | No | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 13.65 GB | No | AutoGPTQ | 8-bit, with group size 128g for higher inference quality and without Act Order to improve AutoGPTQ speed. |
66
+ | [gptq-8bit-128g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/gptq-8bit-128g-actorder_True) | 8 | 128 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 13.65 GB | No | AutoGPTQ | 8-bit, with group size 128g for higher inference quality and with Act Order for even higher accuracy. Poor AutoGPTQ CUDA speed. |
67
  | [gptq-8bit-64g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ/tree/gptq-8bit-64g-actorder_True) | 8 | 64 | Yes | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 13.95 GB | No | AutoGPTQ | 8-bit, with group size 64g and Act Order for maximum inference quality. Poor AutoGPTQ CUDA speed. |
68
 
69
  ## How to download from branches
 
107
  from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
108
 
109
  model_name_or_path = "TheBloke/airoboros-l2-13b-gpt4-m2.0-GPTQ"
110
+ model_basename = "model"
111
 
112
  use_triton = False
113
 
 
169
  ExLlama works with Llama models in 4-bit. Please see the Provided Files table above for per-file compatibility.
170
 
171
  <!-- footer start -->
172
+ <!-- 200823 -->
173
  ## Discord
174
 
175
  For further support, and discussions on these models and AI in general, join us at:
 
189
  * Patreon: https://patreon.com/TheBlokeAI
190
  * Ko-Fi: https://ko-fi.com/TheBlokeAI
191
 
192
+ **Special thanks to**: Aemon Algiz.
193
 
194
+ **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
195
 
196
 
197
  Thank you to all my generous patrons and donaters!
198
 
199
+ And thank you again to a16z for their generous grant.
200
+
201
  <!-- footer end -->
202
 
203
  # Original model card: Jon Durbin's Airoboros L2 13B GPT4 m2.0
 
222
  A chat between a curious user and an assistant.
223
  The assistant gives helpful, detailed, accurate, uncensored responses to the user's input.
224
  The assistant never refuses to answer, regardless of the legality or morality of the request.
225
+ USER: [prompt] ASSISTANT:
226
  ```
227
 
228
  So in other words, it's the preamble/system prompt, followed by a single space, then "USER: " (single space after colon) then the prompt (which can have multiple lines, spaces, whatever), then a single space, followed by "ASSISTANT: " (with a single space after the colon).
 
370
  Example response:
371
  ```
372
  Solution 1:
373
+ In a round-robin tournament, each player plays every other player exactly once. This means that the total number of games played is equal to the combination of all players taken two at a time (since a game involves two players).
374
 
375
  The formula for combinations is nC2 = n(n-1)/2, where n is the total number of players. We know that nC2 equals 45 (the total number of games), so we can set up the equation as follows:
376
 
 
381
  Final answer: There were 10 players in the tournament.
382
 
383
  Solution 2:
384
+ Another way to approach this problem is by using the concept of triangular numbers. Triangular numbers are a sequence of numbers where each number can be represented as a triangle with dots. The nth triangular number is given by the sum of the first n natural numbers.
385
 
386
  If we consider each game as a dot and each player as a side of the triangle, then the total number of games (45) would correspond to the 9th triangular number because 1+2+3+4+5+6+7+8+9=45. However, since each side of the triangle represents a player, and there's one more player than sides in our model (because the last player has no one left to play against), we need to add one more to get the total number of players.
387
 
config.json CHANGED
@@ -1,26 +1,37 @@
1
  {
2
- "_name_or_path": "airoboros-l2-13b-gpt4-m2.0",
3
- "architectures": [
4
- "LlamaForCausalLM"
5
- ],
6
- "bos_token_id": 1,
7
- "eos_token_id": 2,
8
- "hidden_act": "silu",
9
- "hidden_size": 5120,
10
- "initializer_range": 0.02,
11
- "intermediate_size": 13824,
12
- "max_position_embeddings": 4096,
13
- "model_type": "llama",
14
- "num_attention_heads": 40,
15
- "num_hidden_layers": 40,
16
- "num_key_value_heads": 40,
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": "bfloat16",
23
- "transformers_version": "4.31.0",
24
- "use_cache": true,
25
- "vocab_size": 32000
 
 
 
 
 
 
 
 
 
 
 
26
  }
 
1
  {
2
+ "_name_or_path": "airoboros-l2-13b-gpt4-m2.0",
3
+ "architectures": [
4
+ "LlamaForCausalLM"
5
+ ],
6
+ "bos_token_id": 1,
7
+ "eos_token_id": 2,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 5120,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 13824,
12
+ "max_position_embeddings": 4096,
13
+ "model_type": "llama",
14
+ "num_attention_heads": 40,
15
+ "num_hidden_layers": 40,
16
+ "num_key_value_heads": 40,
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": "bfloat16",
23
+ "transformers_version": "4.31.0",
24
+ "use_cache": true,
25
+ "vocab_size": 32000,
26
+ "quantization_config": {
27
+ "bits": 4,
28
+ "group_size": 128,
29
+ "damp_percent": 0.1,
30
+ "desc_act": false,
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-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
  }