TheBloke commited on
Commit
deeb884
1 Parent(s): 4c38fe0

Upload README.md

Browse files
Files changed (1) hide show
  1. README.md +216 -46
README.md CHANGED
@@ -1,6 +1,23 @@
1
  ---
 
2
  inference: false
3
  license: other
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4
  ---
5
 
6
  <!-- header start -->
@@ -20,103 +37,197 @@ license: other
20
  <hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
21
  <!-- header end -->
22
 
23
- # Camel AI's CAMEL 13B Combined Data GPTQ
 
 
24
 
25
- These files are GPTQ 4bit model files for [Camel AI's CAMEL 13B Combined Data](https://huggingface.co/camel-ai/CAMEL-13B-Combined-Data).
 
26
 
27
- It is the result of quantising to 4bit using [AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ).
28
 
 
 
 
 
29
  ## Repositories available
30
 
31
- * [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/CAMEL-13B-Combined-Data-GPTQ)
32
- * [2, 3, 4, 5, 6 and 8-bit GGML models for CPU+GPU inference](https://huggingface.co/TheBloke/CAMEL-13B-Combined-Data-GGML)
33
- * [Unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/CAMEL-13B-Combined-Data-fp16)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
34
 
35
- ## How to easily download and use this model in text-generation-webui
 
 
 
 
 
 
36
 
37
- Please make sure you're using the latest version of text-generation-webui
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
38
 
39
  1. Click the **Model tab**.
40
  2. Under **Download custom model or LoRA**, enter `TheBloke/CAMEL-13B-Combined-Data-GPTQ`.
 
 
41
  3. Click **Download**.
42
- 4. The model will start downloading. Once it's finished it will say "Done"
43
  5. In the top left, click the refresh icon next to **Model**.
44
  6. In the **Model** dropdown, choose the model you just downloaded: `CAMEL-13B-Combined-Data-GPTQ`
45
  7. The model will automatically load, and is now ready for use!
46
  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.
47
- * Note that you do not need to set GPTQ parameters any more. These are set automatically from the file `quantize_config.json`.
48
  9. Once you're ready, click the **Text Generation tab** and enter a prompt to get started!
 
49
 
 
50
  ## How to use this GPTQ model from Python code
51
 
52
- First make sure you have [AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ) installed:
 
 
 
 
 
 
 
 
 
53
 
54
- `pip install auto-gptq`
 
 
 
 
 
 
 
 
 
 
 
 
 
55
 
56
- Then try the following example code:
57
 
58
  ```python
59
- from transformers import AutoTokenizer, pipeline, logging
60
- from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
61
- import argparse
62
 
63
  model_name_or_path = "TheBloke/CAMEL-13B-Combined-Data-GPTQ"
64
- model_basename = "camel-13b-combined-GPTQ-4bit-128g.no-act.order"
65
-
66
- use_triton = False
 
 
 
67
 
68
  tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
69
 
70
- model = AutoGPTQForCausalLM.from_quantized(model_name_or_path,
71
- model_basename=model_basename,
72
- use_safetensors=True,
73
- trust_remote_code=False,
74
- device="cuda:0",
75
- use_triton=use_triton,
76
- quantize_config=None)
77
-
78
  prompt = "Tell me about AI"
79
- prompt_template=f'''### Human: {prompt}
80
- ### Assistant:'''
 
 
 
 
 
 
81
 
82
  print("\n\n*** Generate:")
83
 
84
  input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
85
- output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
86
  print(tokenizer.decode(output[0]))
87
 
88
  # Inference can also be done using transformers' pipeline
89
 
90
- # Prevent printing spurious transformers error when using pipeline with AutoGPTQ
91
- logging.set_verbosity(logging.CRITICAL)
92
-
93
  print("*** Pipeline:")
94
  pipe = pipeline(
95
  "text-generation",
96
  model=model,
97
  tokenizer=tokenizer,
98
  max_new_tokens=512,
 
99
  temperature=0.7,
100
  top_p=0.95,
101
- repetition_penalty=1.15
 
102
  )
103
 
104
  print(pipe(prompt_template)[0]['generated_text'])
105
  ```
 
106
 
107
- ## Provided files
108
-
109
- **camel-13b-combined-GPTQ-4bit-128g.no-act.order.safetensors**
110
 
111
- This will work with AutoGPTQ and CUDA versions of GPTQ-for-LLaMa. There are reports of issues with Triton mode of recent GPTQ-for-LLaMa. If you have issues, please use AutoGPTQ instead.
112
 
113
- It was created with group_size 128 to increase inference accuracy, but without --act-order (desc_act) to increase compatibility and improve inference speed.
114
 
115
- * `camel-13b-combined-GPTQ-4bit-128g.no-act.order.safetensors`
116
- * Works with AutoGPTQ in CUDA or Triton modes.
117
- * Works with GPTQ-for-LLaMa in CUDA mode. May have issues with GPTQ-for-LLaMa Triton mode.
118
- * Works with text-generation-webui, including one-click-installers.
119
- * Parameters: Groupsize = 128. Act Order / desc_act = False.
120
 
121
  <!-- footer start -->
122
  <!-- 200823 -->
@@ -126,10 +237,12 @@ For further support, and discussions on these models and AI in general, join us
126
 
127
  [TheBloke AI's Discord server](https://discord.gg/theblokeai)
128
 
129
- ## Thanks, and how to contribute.
130
 
131
  Thanks to the [chirper.ai](https://chirper.ai) team!
132
 
 
 
133
  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.
134
 
135
  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.
@@ -141,7 +254,7 @@ Donaters will get priority support on any and all AI/LLM/model questions and req
141
 
142
  **Special thanks to**: Aemon Algiz.
143
 
144
- **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
145
 
146
 
147
  Thank you to all my generous patrons and donaters!
@@ -152,6 +265,63 @@ And thank you again to a16z for their generous grant.
152
 
153
  # Original model card: Camel AI's CAMEL 13B Combined Data
154
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
155
  CAMEL-13B-Combined-Data is a chat large language model obtained by finetuning LLaMA-13B model on a total of 229K conversations collected through our [CAMEL](https://arxiv.org/abs/2303.17760) framework, 100K English public conversations from ShareGPT that can be found [here](https://github.com/lm-sys/FastChat/issues/90#issuecomment-1493250773), and 52K instructions from Alpaca dataset that can be found [here](https://github.com/tatsu-lab/stanford_alpaca/blob/761dc5bfbdeeffa89b8bff5d038781a4055f796a/alpaca_data.json). We evaluate our model offline using EleutherAI's language model evaluation harness used by Huggingface's Open LLM Benchmark. CAMEL<sup>*</sup>-13B scores an average of **58.1**, outperfroming LLaMA-30B (58.3), and on par with LLaMA-65B(58.1)!
156
 
157
  | Model | size | ARC-C (25 shots, acc_norm) | HellaSwag (10 shots, acc_norm) | MMLU (5 shots, acc_norm) | TruthfulQA (0 shot, mc2) | Average | Delta |
 
1
  ---
2
+ base_model: https://huggingface.co/camel-ai/CAMEL-13B-Combined-Data
3
  inference: false
4
  license: other
5
+ model_creator: CAMEL
6
+ model_name: CAMEL 13B Combined Data
7
+ model_type: llama
8
+ prompt_template: 'Below is an instruction that describes a task. Write a response
9
+ that appropriately completes the request.
10
+
11
+
12
+ ### Instruction:
13
+
14
+ {prompt}
15
+
16
+
17
+ ### Response:
18
+
19
+ '
20
+ quantized_by: TheBloke
21
  ---
22
 
23
  <!-- header start -->
 
37
  <hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
38
  <!-- header end -->
39
 
40
+ # CAMEL 13B Combined Data - GPTQ
41
+ - Model creator: [CAMEL](https://huggingface.co/camel-ai)
42
+ - Original model: [CAMEL 13B Combined Data](https://huggingface.co/camel-ai/CAMEL-13B-Combined-Data)
43
 
44
+ <!-- description start -->
45
+ ## Description
46
 
47
+ This repo contains GPTQ model files for [Camel AI's CAMEL 13B Combined Data](https://huggingface.co/camel-ai/CAMEL-13B-Combined-Data).
48
 
49
+ 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.
50
+
51
+ <!-- description end -->
52
+ <!-- repositories-available start -->
53
  ## Repositories available
54
 
55
+ * [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/CAMEL-13B-Combined-Data-AWQ)
56
+ * [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/CAMEL-13B-Combined-Data-GPTQ)
57
+ * [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/CAMEL-13B-Combined-Data-GGUF)
58
+ * [CAMEL's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/CAMEL-13B-Combined-Data-fp16)
59
+ <!-- repositories-available end -->
60
+
61
+ <!-- prompt-template start -->
62
+ ## Prompt template: Alpaca
63
+
64
+ ```
65
+ Below is an instruction that describes a task. Write a response that appropriately completes the request.
66
+
67
+ ### Instruction:
68
+ {prompt}
69
+
70
+ ### Response:
71
+
72
+ ```
73
+
74
+ <!-- prompt-template end -->
75
+ <!-- licensing start -->
76
+ ## Licensing
77
+
78
+ The creator of the source model has listed its license as `other`, and this quantization has therefore used that same license.
79
+
80
+ 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.
81
+
82
+ In the meantime, any questions regarding licensing, and in particular how these two licenses might interact, should be directed to the original model repository: [Camel AI's CAMEL 13B Combined Data](https://huggingface.co/camel-ai/CAMEL-13B-Combined-Data).
83
+ <!-- licensing end -->
84
+ <!-- README_GPTQ.md-provided-files start -->
85
+ ## Provided files and GPTQ parameters
86
+
87
+ Multiple quantisation parameters are provided, to allow you to choose the best one for your hardware and requirements.
88
+
89
+ Each separate quant is in a different branch. See below for instructions on fetching from different branches.
90
+
91
+ 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.
92
+
93
+ <details>
94
+ <summary>Explanation of GPTQ parameters</summary>
95
 
96
+ - Bits: The bit size of the quantised model.
97
+ - GS: GPTQ group size. Higher numbers use less VRAM, but have lower quantisation accuracy. "None" is the lowest possible value.
98
+ - 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.
99
+ - Damp %: A GPTQ parameter that affects how samples are processed for quantisation. 0.01 is default, but 0.1 results in slightly better accuracy.
100
+ - 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).
101
+ - 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.
102
+ - ExLlama Compatibility: Whether this file can be loaded with ExLlama, which currently only supports Llama models in 4-bit.
103
 
104
+ </details>
105
+
106
+ | Branch | Bits | GS | Act Order | Damp % | GPTQ Dataset | Seq Len | Size | ExLlama | Desc |
107
+ | ------ | ---- | -- | --------- | ------ | ------------ | ------- | ---- | ------- | ---- |
108
+ | [main](https://huggingface.co/TheBloke/CAMEL-13B-Combined-Data-GPTQ/tree/main) | 4 | 128 | No | 0.01 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 2048 | 8.11 GB | Yes | 4-bit, without Act Order and group size 128g. |
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/CAMEL-13B-Combined-Data-GPTQ:main`
116
+ - With Git, you can clone a branch with:
117
+ ```
118
+ git clone --single-branch --branch main https://huggingface.co/TheBloke/CAMEL-13B-Combined-Data-GPTQ
119
+ ```
120
+ - In Python Transformers code, the branch is the `revision` parameter; see below.
121
+ <!-- README_GPTQ.md-download-from-branches end -->
122
+ <!-- README_GPTQ.md-text-generation-webui start -->
123
+ ## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
124
+
125
+ Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
126
+
127
+ 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.
128
 
129
  1. Click the **Model tab**.
130
  2. Under **Download custom model or LoRA**, enter `TheBloke/CAMEL-13B-Combined-Data-GPTQ`.
131
+ - To download from a specific branch, enter for example `TheBloke/CAMEL-13B-Combined-Data-GPTQ:main`
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".
135
  5. In the top left, click the refresh icon next to **Model**.
136
  6. In the **Model** dropdown, choose the model you just downloaded: `CAMEL-13B-Combined-Data-GPTQ`
137
  7. The model will automatically load, and is now ready for use!
138
  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.
139
+ * 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`.
140
  9. Once you're ready, click the **Text Generation tab** and enter a prompt to get started!
141
+ <!-- README_GPTQ.md-text-generation-webui end -->
142
 
143
+ <!-- README_GPTQ.md-use-from-python start -->
144
  ## How to use this GPTQ model from Python code
145
 
146
+ ### Install the necessary packages
147
+
148
+ Requires: Transformers 4.32.0 or later, Optimum 1.12.0 or later, and AutoGPTQ 0.4.2 or later.
149
+
150
+ ```shell
151
+ pip3 install transformers>=4.32.0 optimum>=1.12.0
152
+ pip3 install auto-gptq --extra-index-url https://huggingface.github.io/autogptq-index/whl/cu118/ # Use cu117 if on CUDA 11.7
153
+ ```
154
+
155
+ If you have problems installing AutoGPTQ using the pre-built wheels, install it from source instead:
156
 
157
+ ```shell
158
+ pip3 uninstall -y auto-gptq
159
+ git clone https://github.com/PanQiWei/AutoGPTQ
160
+ cd AutoGPTQ
161
+ pip3 install .
162
+ ```
163
+
164
+ ### For CodeLlama models only: you must use Transformers 4.33.0 or later.
165
+
166
+ If 4.33.0 is not yet released when you read this, you will need to install Transformers from source:
167
+ ```shell
168
+ pip3 uninstall -y transformers
169
+ pip3 install git+https://github.com/huggingface/transformers.git
170
+ ```
171
 
172
+ ### You can then use the following code
173
 
174
  ```python
175
+ from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
 
 
176
 
177
  model_name_or_path = "TheBloke/CAMEL-13B-Combined-Data-GPTQ"
178
+ # To use a different branch, change revision
179
+ # For example: revision="main"
180
+ model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
181
+ device_map="auto",
182
+ trust_remote_code=True,
183
+ revision="main")
184
 
185
  tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
186
 
 
 
 
 
 
 
 
 
187
  prompt = "Tell me about AI"
188
+ prompt_template=f'''Below is an instruction that describes a task. Write a response that appropriately completes the request.
189
+
190
+ ### Instruction:
191
+ {prompt}
192
+
193
+ ### Response:
194
+
195
+ '''
196
 
197
  print("\n\n*** Generate:")
198
 
199
  input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
200
+ output = model.generate(inputs=input_ids, temperature=0.7, do_sample=True, top_p=0.95, top_k=40, max_new_tokens=512)
201
  print(tokenizer.decode(output[0]))
202
 
203
  # Inference can also be done using transformers' pipeline
204
 
 
 
 
205
  print("*** Pipeline:")
206
  pipe = pipeline(
207
  "text-generation",
208
  model=model,
209
  tokenizer=tokenizer,
210
  max_new_tokens=512,
211
+ do_sample=True,
212
  temperature=0.7,
213
  top_p=0.95,
214
+ top_k=40,
215
+ repetition_penalty=1.1
216
  )
217
 
218
  print(pipe(prompt_template)[0]['generated_text'])
219
  ```
220
+ <!-- README_GPTQ.md-use-from-python end -->
221
 
222
+ <!-- README_GPTQ.md-compatibility start -->
223
+ ## Compatibility
 
224
 
225
+ 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).
226
 
227
+ [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.
228
 
229
+ [Huggingface Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) is compatible with all GPTQ models.
230
+ <!-- README_GPTQ.md-compatibility end -->
 
 
 
231
 
232
  <!-- footer start -->
233
  <!-- 200823 -->
 
237
 
238
  [TheBloke AI's Discord server](https://discord.gg/theblokeai)
239
 
240
+ ## Thanks, and how to contribute
241
 
242
  Thanks to the [chirper.ai](https://chirper.ai) team!
243
 
244
+ Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
245
+
246
  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.
247
 
248
  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.
 
254
 
255
  **Special thanks to**: Aemon Algiz.
256
 
257
+ **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
258
 
259
 
260
  Thank you to all my generous patrons and donaters!
 
265
 
266
  # Original model card: Camel AI's CAMEL 13B Combined Data
267
 
268
+
269
+ <!-- header start -->
270
+ <div style="width: 100%;">
271
+ <img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
272
+ </div>
273
+ <div style="display: flex; justify-content: space-between; width: 100%;">
274
+ <div style="display: flex; flex-direction: column; align-items: flex-start;">
275
+ <p><a href="https://discord.gg/Jq4vkcDakD">Chat & support: my new Discord server</a></p>
276
+ </div>
277
+ <div style="display: flex; flex-direction: column; align-items: flex-end;">
278
+ <p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
279
+ </div>
280
+ </div>
281
+ <!-- header end -->
282
+
283
+ # Camel AI's CAMEL 13B Combined Data fp16
284
+
285
+ These files are pytorch format fp16 model files for [Camel AI's CAMEL 13B Combined Data](https://huggingface.co/camel-ai/CAMEL-13B-Combined-Data).
286
+
287
+ It is the result of merging and/or converting the source repository to float16.
288
+
289
+ ## Repositories available
290
+
291
+ * [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/CAMEL-13B-Combined-Data-fp16)
292
+ * [2, 3, 4, 5, 6 and 8-bit GGML models for CPU+GPU inference](https://huggingface.co/TheBloke/CAMEL-13B-Combined-Data-GGML)
293
+ * [Unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/CAMEL-13B-Combined-Data-fp16)
294
+
295
+ <!-- footer start -->
296
+ ## Discord
297
+
298
+ For further support, and discussions on these models and AI in general, join us at:
299
+
300
+ [TheBloke AI's Discord server](https://discord.gg/Jq4vkcDakD)
301
+
302
+ ## Thanks, and how to contribute.
303
+
304
+ Thanks to the [chirper.ai](https://chirper.ai) team!
305
+
306
+ 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.
307
+
308
+ 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.
309
+
310
+ Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
311
+
312
+ * Patreon: https://patreon.com/TheBlokeAI
313
+ * Ko-Fi: https://ko-fi.com/TheBlokeAI
314
+
315
+ **Special thanks to**: Luke from CarbonQuill, Aemon Algiz, Dmitriy Samsonov.
316
+
317
+ **Patreon special mentions**: Oscar Rangel, Eugene Pentland, Talal Aujan, Cory Kujawski, Luke, Asp the Wyvern, Ai Maven, Pyrater, Alps Aficionado, senxiiz, Willem Michiel, Junyu Yang, trip7s trip, Sebastain Graf, Joseph William Delisle, Lone Striker, Jonathan Leane, Johann-Peter Hartmann, David Flickinger, Spiking Neurons AB, Kevin Schuppel, Mano Prime, Dmitriy Samsonov, Sean Connelly, Nathan LeClaire, Alain Rossmann, Fen Risland, Derek Yates, Luke Pendergrass, Nikolai Manek, Khalefa Al-Ahmad, Artur Olbinski, John Detwiler, Ajan Kanaga, Imad Khwaja, Trenton Dambrowitz, Kalila, vamX, webtim, Illia Dulskyi.
318
+
319
+ Thank you to all my generous patrons and donaters!
320
+
321
+ <!-- footer end -->
322
+
323
+ # Original model card: Camel AI's CAMEL 13B Combined Data
324
+
325
  CAMEL-13B-Combined-Data is a chat large language model obtained by finetuning LLaMA-13B model on a total of 229K conversations collected through our [CAMEL](https://arxiv.org/abs/2303.17760) framework, 100K English public conversations from ShareGPT that can be found [here](https://github.com/lm-sys/FastChat/issues/90#issuecomment-1493250773), and 52K instructions from Alpaca dataset that can be found [here](https://github.com/tatsu-lab/stanford_alpaca/blob/761dc5bfbdeeffa89b8bff5d038781a4055f796a/alpaca_data.json). We evaluate our model offline using EleutherAI's language model evaluation harness used by Huggingface's Open LLM Benchmark. CAMEL<sup>*</sup>-13B scores an average of **58.1**, outperfroming LLaMA-30B (58.3), and on par with LLaMA-65B(58.1)!
326
 
327
  | Model | size | ARC-C (25 shots, acc_norm) | HellaSwag (10 shots, acc_norm) | MMLU (5 shots, acc_norm) | TruthfulQA (0 shot, mc2) | Average | Delta |