TheBloke commited on
Commit
1c20af9
1 Parent(s): b251c20

Upload README.md

Browse files
Files changed (1) hide show
  1. README.md +103 -57
README.md CHANGED
@@ -1,7 +1,11 @@
1
  ---
2
  inference: false
3
- license: other
 
 
 
4
  model_type: llama
 
5
  ---
6
 
7
  <!-- header start -->
@@ -21,111 +25,150 @@ model_type: llama
21
  <hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
22
  <!-- header end -->
23
 
24
- # ConceptofMind's LLongMA 2 7B GPTQ
 
 
25
 
26
- These files are GPTQ model files for [ConceptofMind's LLongMA 2 7B](https://huggingface.co/conceptofmind/LLongMA-2-7b).
 
27
 
28
- Multiple GPTQ parameter permutations are provided; see Provided Files below for details of the options provided, their parameters, and the software used to create them.
29
 
 
30
 
 
 
31
  ## Repositories available
32
 
33
  * [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/LLongMA-2-7B-GPTQ)
34
- * [2, 3, 4, 5, 6 and 8-bit GGML models for CPU+GPU inference](https://huggingface.co/TheBloke/LLongMA-2-7B-GGML)
35
- * [Original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/conceptofmind/LLongMA-2-7b)
 
 
36
 
37
- ## Prompt template: Unknown
 
38
 
39
  ```
40
  {prompt}
 
41
  ```
42
 
43
- ## Provided files
 
 
 
44
 
45
  Multiple quantisation parameters are provided, to allow you to choose the best one for your hardware and requirements.
46
 
47
  Each separate quant is in a different branch. See below for instructions on fetching from different branches.
48
 
49
- | Branch | Bits | Group Size | Act Order (desc_act) | File Size | ExLlama Compatible? | Made With | Description |
50
- | ------ | ---- | ---------- | -------------------- | --------- | ------------------- | --------- | ----------- |
51
- | 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. |
52
- | 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. |
53
- | 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. |
54
- | 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, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
55
- | 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. |
56
- | gptq-8bit-128g-actorder_False | 8 | 128 | False | 7.16 GB | False | AutoGPTQ | 8-bit, with group size 128g for higher inference quality and without Act Order to improve AutoGPTQ speed. |
57
- | 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. |
58
- | gptq-8bit-64g-actorder_True | 8 | 64 | True | 7.31 GB | False | AutoGPTQ | 8-bit, with group size 64g and Act Order for maximum inference quality. Poor AutoGPTQ CUDA speed. |
 
 
59
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
60
  ## How to download from branches
61
 
62
  - In text-generation-webui, you can add `:branch` to the end of the download name, eg `TheBloke/LLongMA-2-7B-GPTQ:gptq-4bit-32g-actorder_True`
63
  - With Git, you can clone a branch with:
64
  ```
65
- git clone --branch gptq-4bit-32g-actorder_True https://huggingface.co/TheBloke/LLongMA-2-7B-GPTQ`
66
  ```
67
  - In Python Transformers code, the branch is the `revision` parameter; see below.
68
-
 
69
  ## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
70
 
71
  Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
72
 
73
- It is strongly recommended to use the text-generation-webui one-click-installers unless you know how to make a manual install.
74
 
75
  1. Click the **Model tab**.
76
  2. Under **Download custom model or LoRA**, enter `TheBloke/LLongMA-2-7B-GPTQ`.
77
  - To download from a specific branch, enter for example `TheBloke/LLongMA-2-7B-GPTQ:gptq-4bit-32g-actorder_True`
78
  - see Provided Files above for the list of branches for each option.
79
  3. Click **Download**.
80
- 4. The model will start downloading. Once it's finished it will say "Done"
81
  5. In the top left, click the refresh icon next to **Model**.
82
  6. In the **Model** dropdown, choose the model you just downloaded: `LLongMA-2-7B-GPTQ`
83
  7. The model will automatically load, and is now ready for use!
84
  8. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right.
85
- * Note that you do not need to set GPTQ parameters any more. These are set automatically from the file `quantize_config.json`.
86
  9. Once you're ready, click the **Text Generation tab** and enter a prompt to get started!
 
87
 
 
88
  ## How to use this GPTQ model from Python code
89
 
90
- First make sure you have [AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ) installed:
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
91
 
92
- `GITHUB_ACTIONS=true pip install auto-gptq`
 
 
 
 
93
 
94
- Then try the following example code:
95
 
96
  ```python
97
- from transformers import AutoTokenizer, pipeline, logging
98
- from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
99
 
100
  model_name_or_path = "TheBloke/LLongMA-2-7B-GPTQ"
101
- model_basename = "model"
102
-
103
- use_triton = False
 
 
 
104
 
105
  tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
106
 
107
- model = AutoGPTQForCausalLM.from_quantized(model_name_or_path,
108
- model_basename=model_basename,
109
- use_safetensors=True,
110
- trust_remote_code=False,
111
- device="cuda:0",
112
- use_triton=use_triton,
113
- quantize_config=None)
114
-
115
- """
116
- To download from a specific branch, use the revision parameter, as in this example:
117
-
118
- model = AutoGPTQForCausalLM.from_quantized(model_name_or_path,
119
- revision="gptq-4bit-32g-actorder_True",
120
- model_basename=model_basename,
121
- use_safetensors=True,
122
- trust_remote_code=False,
123
- device="cuda:0",
124
- quantize_config=None)
125
- """
126
-
127
  prompt = "Tell me about AI"
128
  prompt_template=f'''{prompt}
 
129
  '''
130
 
131
  print("\n\n*** Generate:")
@@ -136,9 +179,6 @@ print(tokenizer.decode(output[0]))
136
 
137
  # Inference can also be done using transformers' pipeline
138
 
139
- # Prevent printing spurious transformers error when using pipeline with AutoGPTQ
140
- logging.set_verbosity(logging.CRITICAL)
141
-
142
  print("*** Pipeline:")
143
  pipe = pipeline(
144
  "text-generation",
@@ -152,12 +192,17 @@ pipe = pipeline(
152
 
153
  print(pipe(prompt_template)[0]['generated_text'])
154
  ```
 
155
 
 
156
  ## Compatibility
157
 
158
- The files provided will work with AutoGPTQ (CUDA and Triton modes), GPTQ-for-LLaMa (only CUDA has been tested), and Occ4m's GPTQ-for-LLaMa fork.
 
 
159
 
160
- ExLlama works with Llama models in 4-bit. Please see the Provided Files table above for per-file compatibility.
 
161
 
162
  <!-- footer start -->
163
  <!-- 200823 -->
@@ -182,7 +227,7 @@ Donaters will get priority support on any and all AI/LLM/model questions and req
182
 
183
  **Special thanks to**: Aemon Algiz.
184
 
185
- **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
186
 
187
 
188
  Thank you to all my generous patrons and donaters!
@@ -193,6 +238,7 @@ And thank you again to a16z for their generous grant.
193
 
194
  # Original model card: ConceptofMind's LLongMA 2 7B
195
 
 
196
  LLongMA-2, a suite of Llama-2 models, trained at 8k context length using linear positional interpolation scaling. The model was trained in collaboration with Emozilla of NousResearch and Kaiokendev.
197
 
198
  We worked directly with Kaiokendev, to extend the context length of the Llama-2 7b model through fine-tuning. The models pass all our evaluations and maintain the same perplexity at 8k extrapolation surpassing the performance of other recent methodologies.
 
1
  ---
2
  inference: false
3
+ license: llama2
4
+ model_creator: Enrico Shippole
5
+ model_link: https://huggingface.co/conceptofmind/LLongMA-2-7b
6
+ model_name: LLongMA 2 7B
7
  model_type: llama
8
+ quantized_by: TheBloke
9
  ---
10
 
11
  <!-- header start -->
 
25
  <hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
26
  <!-- header end -->
27
 
28
+ # LLongMA 2 7B - GPTQ
29
+ - Model creator: [Enrico Shippole](https://huggingface.co/conceptofmind)
30
+ - Original model: [LLongMA 2 7B](https://huggingface.co/conceptofmind/LLongMA-2-7b)
31
 
32
+ <!-- description start -->
33
+ ## Description
34
 
35
+ This repo contains GPTQ model files for [ConceptofMind's LLongMA 2 7B](https://huggingface.co/conceptofmind/LLongMA-2-7b).
36
 
37
+ Multiple GPTQ parameter permutations are provided; see Provided Files below for details of the options provided, their parameters, and the software used to create them.
38
 
39
+ <!-- description end -->
40
+ <!-- repositories-available start -->
41
  ## Repositories available
42
 
43
  * [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/LLongMA-2-7B-GPTQ)
44
+ * [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/LLongMA-2-7B-GGUF)
45
+ * [2, 3, 4, 5, 6 and 8-bit GGML models for CPU+GPU inference (deprecated)](https://huggingface.co/TheBloke/LLongMA-2-7B-GGML)
46
+ * [Enrico Shippole's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/conceptofmind/LLongMA-2-7b)
47
+ <!-- repositories-available end -->
48
 
49
+ <!-- prompt-template start -->
50
+ ## Prompt template: None
51
 
52
  ```
53
  {prompt}
54
+
55
  ```
56
 
57
+ <!-- prompt-template end -->
58
+
59
+ <!-- README_GPTQ.md-provided-files start -->
60
+ ## Provided files and GPTQ parameters
61
 
62
  Multiple quantisation parameters are provided, to allow you to choose the best one for your hardware and requirements.
63
 
64
  Each separate quant is in a different branch. See below for instructions on fetching from different branches.
65
 
66
+ All recent GPTQ files are made with AutoGPTQ, and all files in non-main branches are made with AutoGPTQ. Files in the `main` branch which were uploaded before August 2023 were made with GPTQ-for-LLaMa.
67
+
68
+ <details>
69
+ <summary>Explanation of GPTQ parameters</summary>
70
+
71
+ - Bits: The bit size of the quantised model.
72
+ - GS: GPTQ group size. Higher numbers use less VRAM, but have lower quantisation accuracy. "None" is the lowest possible value.
73
+ - Act Order: True or False. Also known as `desc_act`. True results in better quantisation accuracy. Some GPTQ clients have had issues with models that use Act Order plus Group Size, but this is generally resolved now.
74
+ - Damp %: A GPTQ parameter that affects how samples are processed for quantisation. 0.01 is default, but 0.1 results in slightly better accuracy.
75
+ - GPTQ dataset: The dataset used for quantisation. Using a dataset more appropriate to the model's training can improve quantisation accuracy. Note that the GPTQ dataset is not the same as the dataset used to train the model - please refer to the original model repo for details of the training dataset(s).
76
+ - Sequence Length: The length of the dataset sequences used for quantisation. Ideally this is the same as the model sequence length. For some very long sequence models (16+K), a lower sequence length may have to be used. Note that a lower sequence length does not limit the sequence length of the quantised model. It only impacts the quantisation accuracy on longer inference sequences.
77
+ - ExLlama Compatibility: Whether this file can be loaded with ExLlama, which currently only supports Llama models in 4-bit.
78
 
79
+ </details>
80
+
81
+ | Branch | Bits | GS | Act Order | Damp % | GPTQ Dataset | Seq Len | Size | ExLlama | Desc |
82
+ | ------ | ---- | -- | --------- | ------ | ------------ | ------- | ---- | ------- | ---- |
83
+ | [main](https://huggingface.co/TheBloke/LLongMA-2-7B-GPTQ/tree/main) | 4 | 128 | No | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 3.90 GB | Yes | Most compatible option. Good inference speed in AutoGPTQ and GPTQ-for-LLaMa. Lower inference quality than other options. |
84
+ | [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/LLongMA-2-7B-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 4.28 GB | Yes | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. Poor AutoGPTQ CUDA speed. |
85
+ | [gptq-4bit-64g-actorder_True](https://huggingface.co/TheBloke/LLongMA-2-7B-GPTQ/tree/gptq-4bit-64g-actorder_True) | 4 | 64 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 4.02 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. |
86
+ | [gptq-4bit-128g-actorder_True](https://huggingface.co/TheBloke/LLongMA-2-7B-GPTQ/tree/gptq-4bit-128g-actorder_True) | 4 | 128 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 3.90 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. |
87
+ | [gptq-8bit--1g-actorder_True](https://huggingface.co/TheBloke/LLongMA-2-7B-GPTQ/tree/gptq-8bit--1g-actorder_True) | 8 | None | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 7.01 GB | No | 8-bit, with Act Order. No group size, to lower VRAM requirements and to improve AutoGPTQ speed. |
88
+ | [gptq-8bit-128g-actorder_False](https://huggingface.co/TheBloke/LLongMA-2-7B-GPTQ/tree/gptq-8bit-128g-actorder_False) | 8 | 128 | No | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 7.16 GB | No | 8-bit, with group size 128g for higher inference quality and without Act Order to improve AutoGPTQ speed. |
89
+ | [gptq-8bit-128g-actorder_True](https://huggingface.co/TheBloke/LLongMA-2-7B-GPTQ/tree/gptq-8bit-128g-actorder_True) | 8 | 128 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 7.16 GB | No | 8-bit, with group size 128g for higher inference quality and with Act Order for even higher accuracy. Poor AutoGPTQ CUDA speed. |
90
+ | [gptq-8bit-64g-actorder_True](https://huggingface.co/TheBloke/LLongMA-2-7B-GPTQ/tree/gptq-8bit-64g-actorder_True) | 8 | 64 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 7.31 GB | No | 8-bit, with group size 64g and Act Order for even higher inference quality. Poor AutoGPTQ CUDA speed. |
91
+
92
+ <!-- README_GPTQ.md-provided-files end -->
93
+
94
+ <!-- README_GPTQ.md-download-from-branches start -->
95
  ## How to download from branches
96
 
97
  - In text-generation-webui, you can add `:branch` to the end of the download name, eg `TheBloke/LLongMA-2-7B-GPTQ:gptq-4bit-32g-actorder_True`
98
  - With Git, you can clone a branch with:
99
  ```
100
+ git clone --single-branch --branch gptq-4bit-32g-actorder_True https://huggingface.co/TheBloke/LLongMA-2-7B-GPTQ
101
  ```
102
  - In Python Transformers code, the branch is the `revision` parameter; see below.
103
+ <!-- README_GPTQ.md-download-from-branches end -->
104
+ <!-- README_GPTQ.md-text-generation-webui start -->
105
  ## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
106
 
107
  Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
108
 
109
+ It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install.
110
 
111
  1. Click the **Model tab**.
112
  2. Under **Download custom model or LoRA**, enter `TheBloke/LLongMA-2-7B-GPTQ`.
113
  - To download from a specific branch, enter for example `TheBloke/LLongMA-2-7B-GPTQ:gptq-4bit-32g-actorder_True`
114
  - see Provided Files above for the list of branches for each option.
115
  3. Click **Download**.
116
+ 4. The model will start downloading. Once it's finished it will say "Done".
117
  5. In the top left, click the refresh icon next to **Model**.
118
  6. In the **Model** dropdown, choose the model you just downloaded: `LLongMA-2-7B-GPTQ`
119
  7. The model will automatically load, and is now ready for use!
120
  8. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right.
121
+ * Note that you do not need to and should not set manual GPTQ parameters any more. These are set automatically from the file `quantize_config.json`.
122
  9. Once you're ready, click the **Text Generation tab** and enter a prompt to get started!
123
+ <!-- README_GPTQ.md-text-generation-webui end -->
124
 
125
+ <!-- README_GPTQ.md-use-from-python start -->
126
  ## How to use this GPTQ model from Python code
127
 
128
+ ### Install the necessary packages
129
+
130
+ Requires: Transformers 4.32.0 or later, Optimum 1.12.0 or later, and AutoGPTQ 0.4.2 or later.
131
+
132
+ ```shell
133
+ pip3 install transformers>=4.32.0 optimum>=1.12.0
134
+ pip3 install auto-gptq --extra-index-url https://huggingface.github.io/autogptq-index/whl/cu118/ # Use cu117 if on CUDA 11.7
135
+ ```
136
+
137
+ If you have problems installing AutoGPTQ using the pre-built wheels, install it from source instead:
138
+
139
+ ```shell
140
+ pip3 uninstall -y auto-gptq
141
+ git clone https://github.com/PanQiWei/AutoGPTQ
142
+ cd AutoGPTQ
143
+ pip3 install .
144
+ ```
145
+
146
+ ### For CodeLlama models only: you must use Transformers 4.33.0 or later.
147
 
148
+ If 4.33.0 is not yet released when you read this, you will need to install Transformers from source:
149
+ ```shell
150
+ pip3 uninstall -y transformers
151
+ pip3 install git+https://github.com/huggingface/transformers.git
152
+ ```
153
 
154
+ ### You can then use the following code
155
 
156
  ```python
157
+ from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
 
158
 
159
  model_name_or_path = "TheBloke/LLongMA-2-7B-GPTQ"
160
+ # To use a different branch, change revision
161
+ # For example: revision="gptq-4bit-32g-actorder_True"
162
+ model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
163
+ torch_dtype=torch.float16,
164
+ device_map="auto",
165
+ revision="main")
166
 
167
  tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
168
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
169
  prompt = "Tell me about AI"
170
  prompt_template=f'''{prompt}
171
+
172
  '''
173
 
174
  print("\n\n*** Generate:")
 
179
 
180
  # Inference can also be done using transformers' pipeline
181
 
 
 
 
182
  print("*** Pipeline:")
183
  pipe = pipeline(
184
  "text-generation",
 
192
 
193
  print(pipe(prompt_template)[0]['generated_text'])
194
  ```
195
+ <!-- README_GPTQ.md-use-from-python end -->
196
 
197
+ <!-- README_GPTQ.md-compatibility start -->
198
  ## Compatibility
199
 
200
+ The files provided are tested to work with AutoGPTQ, both via Transformers and using AutoGPTQ directly. They should also work with [Occ4m's GPTQ-for-LLaMa fork](https://github.com/0cc4m/KoboldAI).
201
+
202
+ [ExLlama](https://github.com/turboderp/exllama) is compatible with Llama models in 4-bit. Please see the Provided Files table above for per-file compatibility.
203
 
204
+ [Huggingface Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) is compatible with all GPTQ models.
205
+ <!-- README_GPTQ.md-compatibility end -->
206
 
207
  <!-- footer start -->
208
  <!-- 200823 -->
 
227
 
228
  **Special thanks to**: Aemon Algiz.
229
 
230
+ **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
231
 
232
 
233
  Thank you to all my generous patrons and donaters!
 
238
 
239
  # Original model card: ConceptofMind's LLongMA 2 7B
240
 
241
+
242
  LLongMA-2, a suite of Llama-2 models, trained at 8k context length using linear positional interpolation scaling. The model was trained in collaboration with Emozilla of NousResearch and Kaiokendev.
243
 
244
  We worked directly with Kaiokendev, to extend the context length of the Llama-2 7b model through fine-tuning. The models pass all our evaluations and maintain the same perplexity at 8k extrapolation surpassing the performance of other recent methodologies.