TheBloke commited on
Commit
1109a33
1 Parent(s): 49a7b15

Upload README.md

Browse files
Files changed (1) hide show
  1. README.md +358 -0
README.md ADDED
@@ -0,0 +1,358 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: deepseek-ai/deepseek-coder-33b-instruct
3
+ inference: false
4
+ license: other
5
+ license_link: LICENSE
6
+ license_name: deepseek
7
+ model_creator: DeepSeek
8
+ model_name: Deepseek Coder 33B Instruct
9
+ model_type: deepseek
10
+ prompt_template: 'You are an AI programming assistant, utilizing the Deepseek Coder
11
+ model, developed by Deepseek Company, and you only answer questions related to computer
12
+ science. For politically sensitive questions, security and privacy issues, and other
13
+ non-computer science questions, you will refuse to answer.
14
+
15
+ ### Instruction:
16
+
17
+ {prompt}
18
+
19
+ ### Response:
20
+
21
+ '
22
+ quantized_by: TheBloke
23
+ ---
24
+ <!-- markdownlint-disable MD041 -->
25
+
26
+ <!-- header start -->
27
+ <!-- 200823 -->
28
+ <div style="width: auto; margin-left: auto; margin-right: auto">
29
+ <img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
30
+ </div>
31
+ <div style="display: flex; justify-content: space-between; width: 100%;">
32
+ <div style="display: flex; flex-direction: column; align-items: flex-start;">
33
+ <p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
34
+ </div>
35
+ <div style="display: flex; flex-direction: column; align-items: flex-end;">
36
+ <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>
37
+ </div>
38
+ </div>
39
+ <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>
40
+ <hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
41
+ <!-- header end -->
42
+
43
+ # Deepseek Coder 33B Instruct - GGUF
44
+ - Model creator: [DeepSeek](https://huggingface.co/deepseek-ai)
45
+ - Original model: [Deepseek Coder 33B Instruct](https://huggingface.co/deepseek-ai/deepseek-coder-33b-instruct)
46
+
47
+ <!-- description start -->
48
+ ## Description
49
+
50
+ This repo contains GGUF format model files for [DeepSeek's Deepseek Coder 33B Instruct](https://huggingface.co/deepseek-ai/deepseek-coder-33b-instruct).
51
+
52
+ These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/).
53
+
54
+ <!-- description end -->
55
+ <!-- README_GGUF.md-about-gguf start -->
56
+ ### About GGUF
57
+
58
+ GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
59
+
60
+ Here is an incomplete list of clients and libraries that are known to support GGUF:
61
+
62
+ * [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.
63
+ * [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
64
+ * [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
65
+ * [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration.
66
+ * [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.
67
+ * [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
68
+ * [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server.
69
+ * [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
70
+ * [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.
71
+
72
+ <!-- README_GGUF.md-about-gguf end -->
73
+ <!-- repositories-available start -->
74
+ ## Repositories available
75
+
76
+ * [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-AWQ)
77
+ * [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GPTQ)
78
+ * [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GGUF)
79
+ * [DeepSeek's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/deepseek-ai/deepseek-coder-33b-instruct)
80
+ <!-- repositories-available end -->
81
+
82
+ <!-- prompt-template start -->
83
+ ## Prompt template: DeepSeek
84
+
85
+ ```
86
+ You are an AI programming assistant, utilizing the Deepseek Coder model, developed by Deepseek Company, and you only answer questions related to computer science. For politically sensitive questions, security and privacy issues, and other non-computer science questions, you will refuse to answer.
87
+ ### Instruction:
88
+ {prompt}
89
+ ### Response:
90
+
91
+ ```
92
+
93
+ <!-- prompt-template end -->
94
+
95
+
96
+ <!-- compatibility_gguf start -->
97
+ ## Compatibility
98
+
99
+ These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)
100
+
101
+ They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
102
+
103
+ ## Explanation of quantisation methods
104
+
105
+ <details>
106
+ <summary>Click to see details</summary>
107
+
108
+ The new methods available are:
109
+
110
+ * GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
111
+ * GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
112
+ * GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
113
+ * GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
114
+ * GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
115
+
116
+ Refer to the Provided Files table below to see what files use which methods, and how.
117
+ </details>
118
+ <!-- compatibility_gguf end -->
119
+
120
+ <!-- README_GGUF.md-provided-files start -->
121
+ ## Provided files
122
+
123
+ | Name | Quant method | Bits | Size | Max RAM required | Use case |
124
+ | ---- | ---- | ---- | ---- | ---- | ----- |
125
+ | [deepseek-coder-33b-instruct.Q2_K.gguf](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GGUF/blob/main/deepseek-coder-33b-instruct.Q2_K.gguf) | Q2_K | 2 | 14.03 GB| 16.53 GB | smallest, significant quality loss - not recommended for most purposes |
126
+ | [deepseek-coder-33b-instruct.Q3_K_S.gguf](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GGUF/blob/main/deepseek-coder-33b-instruct.Q3_K_S.gguf) | Q3_K_S | 3 | 14.42 GB| 16.92 GB | very small, high quality loss |
127
+ | [deepseek-coder-33b-instruct.Q3_K_M.gguf](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GGUF/blob/main/deepseek-coder-33b-instruct.Q3_K_M.gguf) | Q3_K_M | 3 | 16.07 GB| 18.57 GB | very small, high quality loss |
128
+ | [deepseek-coder-33b-instruct.Q3_K_L.gguf](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GGUF/blob/main/deepseek-coder-33b-instruct.Q3_K_L.gguf) | Q3_K_L | 3 | 17.56 GB| 20.06 GB | small, substantial quality loss |
129
+ | [deepseek-coder-33b-instruct.Q4_0.gguf](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GGUF/blob/main/deepseek-coder-33b-instruct.Q4_0.gguf) | Q4_0 | 4 | 18.82 GB| 21.32 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
130
+ | [deepseek-coder-33b-instruct.Q4_K_S.gguf](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GGUF/blob/main/deepseek-coder-33b-instruct.Q4_K_S.gguf) | Q4_K_S | 4 | 18.89 GB| 21.39 GB | small, greater quality loss |
131
+ | [deepseek-coder-33b-instruct.Q4_K_M.gguf](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GGUF/blob/main/deepseek-coder-33b-instruct.Q4_K_M.gguf) | Q4_K_M | 4 | 19.94 GB| 22.44 GB | medium, balanced quality - recommended |
132
+ | [deepseek-coder-33b-instruct.Q5_0.gguf](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GGUF/blob/main/deepseek-coder-33b-instruct.Q5_0.gguf) | Q5_0 | 5 | 22.96 GB| 25.46 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
133
+ | [deepseek-coder-33b-instruct.Q5_K_S.gguf](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GGUF/blob/main/deepseek-coder-33b-instruct.Q5_K_S.gguf) | Q5_K_S | 5 | 22.96 GB| 25.46 GB | large, low quality loss - recommended |
134
+ | [deepseek-coder-33b-instruct.Q5_K_M.gguf](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GGUF/blob/main/deepseek-coder-33b-instruct.Q5_K_M.gguf) | Q5_K_M | 5 | 23.54 GB| 26.04 GB | large, very low quality loss - recommended |
135
+ | [deepseek-coder-33b-instruct.Q6_K.gguf](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GGUF/blob/main/deepseek-coder-33b-instruct.Q6_K.gguf) | Q6_K | 6 | 27.36 GB| 29.86 GB | very large, extremely low quality loss |
136
+ | [deepseek-coder-33b-instruct.Q8_0.gguf](https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GGUF/blob/main/deepseek-coder-33b-instruct.Q8_0.gguf) | Q8_0 | 8 | 35.43 GB| 37.93 GB | very large, extremely low quality loss - not recommended |
137
+
138
+ **Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
139
+
140
+
141
+
142
+ <!-- README_GGUF.md-provided-files end -->
143
+
144
+ <!-- README_GGUF.md-how-to-download start -->
145
+ ## How to download GGUF files
146
+
147
+ **Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
148
+
149
+ The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
150
+
151
+ * LM Studio
152
+ * LoLLMS Web UI
153
+ * Faraday.dev
154
+
155
+ ### In `text-generation-webui`
156
+
157
+ Under Download Model, you can enter the model repo: TheBloke/deepseek-coder-33B-instruct-GGUF and below it, a specific filename to download, such as: deepseek-coder-33b-instruct.Q4_K_M.gguf.
158
+
159
+ Then click Download.
160
+
161
+ ### On the command line, including multiple files at once
162
+
163
+ I recommend using the `huggingface-hub` Python library:
164
+
165
+ ```shell
166
+ pip3 install huggingface-hub
167
+ ```
168
+
169
+ Then you can download any individual model file to the current directory, at high speed, with a command like this:
170
+
171
+ ```shell
172
+ huggingface-cli download TheBloke/deepseek-coder-33B-instruct-GGUF deepseek-coder-33b-instruct.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
173
+ ```
174
+
175
+ <details>
176
+ <summary>More advanced huggingface-cli download usage</summary>
177
+
178
+ You can also download multiple files at once with a pattern:
179
+
180
+ ```shell
181
+ huggingface-cli download TheBloke/deepseek-coder-33B-instruct-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
182
+ ```
183
+
184
+ For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
185
+
186
+ To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
187
+
188
+ ```shell
189
+ pip3 install hf_transfer
190
+ ```
191
+
192
+ And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
193
+
194
+ ```shell
195
+ HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/deepseek-coder-33B-instruct-GGUF deepseek-coder-33b-instruct.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
196
+ ```
197
+
198
+ Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
199
+ </details>
200
+ <!-- README_GGUF.md-how-to-download end -->
201
+
202
+ <!-- README_GGUF.md-how-to-run start -->
203
+ ## Example `llama.cpp` command
204
+
205
+ Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
206
+
207
+ ```shell
208
+ ./main -ngl 32 -m deepseek-coder-33b-instruct.Q4_K_M.gguf --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "You are an AI programming assistant, utilizing the Deepseek Coder model, developed by Deepseek Company, and you only answer questions related to computer science. For politically sensitive questions, security and privacy issues, and other non-computer science questions, you will refuse to answer.\n### Instruction:\n{prompt}\n### Response:"
209
+ ```
210
+
211
+ Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
212
+
213
+ Change `-c 2048` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically.
214
+
215
+ If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
216
+
217
+ For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)
218
+
219
+ ## How to run in `text-generation-webui`
220
+
221
+ Further instructions here: [text-generation-webui/docs/llama.cpp.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp.md).
222
+
223
+ ## How to run from Python code
224
+
225
+ You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries.
226
+
227
+ ### How to load this model in Python code, using ctransformers
228
+
229
+ #### First install the package
230
+
231
+ Run one of the following commands, according to your system:
232
+
233
+ ```shell
234
+ # Base ctransformers with no GPU acceleration
235
+ pip install ctransformers
236
+ # Or with CUDA GPU acceleration
237
+ pip install ctransformers[cuda]
238
+ # Or with AMD ROCm GPU acceleration (Linux only)
239
+ CT_HIPBLAS=1 pip install ctransformers --no-binary ctransformers
240
+ # Or with Metal GPU acceleration for macOS systems only
241
+ CT_METAL=1 pip install ctransformers --no-binary ctransformers
242
+ ```
243
+
244
+ #### Simple ctransformers example code
245
+
246
+ ```python
247
+ from ctransformers import AutoModelForCausalLM
248
+
249
+ # Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
250
+ llm = AutoModelForCausalLM.from_pretrained("TheBloke/deepseek-coder-33B-instruct-GGUF", model_file="deepseek-coder-33b-instruct.Q4_K_M.gguf", model_type="deepseek", gpu_layers=50)
251
+
252
+ print(llm("AI is going to"))
253
+ ```
254
+
255
+ ## How to use with LangChain
256
+
257
+ Here are guides on using llama-cpp-python and ctransformers with LangChain:
258
+
259
+ * [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
260
+ * [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
261
+
262
+ <!-- README_GGUF.md-how-to-run end -->
263
+
264
+ <!-- footer start -->
265
+ <!-- 200823 -->
266
+ ## Discord
267
+
268
+ For further support, and discussions on these models and AI in general, join us at:
269
+
270
+ [TheBloke AI's Discord server](https://discord.gg/theblokeai)
271
+
272
+ ## Thanks, and how to contribute
273
+
274
+ Thanks to the [chirper.ai](https://chirper.ai) team!
275
+
276
+ Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
277
+
278
+ 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.
279
+
280
+ 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.
281
+
282
+ Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
283
+
284
+ * Patreon: https://patreon.com/TheBlokeAI
285
+ * Ko-Fi: https://ko-fi.com/TheBlokeAI
286
+
287
+ **Special thanks to**: Aemon Algiz.
288
+
289
+ **Patreon special mentions**: Brandon Frisco, LangChain4j, Spiking Neurons AB, transmissions 11, Joseph William Delisle, Nitin Borwankar, Willem Michiel, Michael Dempsey, vamX, Jeffrey Morgan, zynix, jjj, Omer Bin Jawed, Sean Connelly, jinyuan sun, Jeromy Smith, Shadi, Pawan Osman, Chadd, Elijah Stavena, Illia Dulskyi, Sebastain Graf, Stephen Murray, terasurfer, Edmond Seymore, Celu Ramasamy, Mandus, Alex, biorpg, Ajan Kanaga, Clay Pascal, Raven Klaugh, 阿明, K, ya boyyy, usrbinkat, Alicia Loh, John Villwock, ReadyPlayerEmma, Chris Smitley, Cap'n Zoog, fincy, GodLy, S_X, sidney chen, Cory Kujawski, OG, Mano Prime, AzureBlack, Pieter, Kalila, Spencer Kim, Tom X Nguyen, Stanislav Ovsiannikov, Michael Levine, Andrey, Trailburnt, Vadim, Enrico Ros, Talal Aujan, Brandon Phillips, Jack West, Eugene Pentland, Michael Davis, Will Dee, webtim, Jonathan Leane, Alps Aficionado, Rooh Singh, Tiffany J. Kim, theTransient, Luke @flexchar, Elle, Caitlyn Gatomon, Ari Malik, subjectnull, Johann-Peter Hartmann, Trenton Dambrowitz, Imad Khwaja, Asp the Wyvern, Emad Mostaque, Rainer Wilmers, Alexandros Triantafyllidis, Nicholas, Pedro Madruga, SuperWojo, Harry Royden McLaughlin, James Bentley, Olakabola, David Ziegler, Ai Maven, Jeff Scroggin, Nikolai Manek, Deo Leter, Matthew Berman, Fen Risland, Ken Nordquist, Manuel Alberto Morcote, Luke Pendergrass, TL, Fred von Graf, Randy H, Dan Guido, NimbleBox.ai, Vitor Caleffi, Gabriel Tamborski, knownsqashed, Lone Striker, Erik Bjäreholt, John Detwiler, Leonard Tan, Iucharbius
290
+
291
+
292
+ Thank you to all my generous patrons and donaters!
293
+
294
+ And thank you again to a16z for their generous grant.
295
+
296
+ <!-- footer end -->
297
+
298
+ <!-- original-model-card start -->
299
+ # Original model card: DeepSeek's Deepseek Coder 33B Instruct
300
+
301
+
302
+
303
+ <p align="center">
304
+ <img width="1000px" alt="DeepSeek Coder" src="https://github.com/deepseek-ai/DeepSeek-Coder/blob/main/pictures/logo.png?raw=true">
305
+ </p>
306
+ <p align="center"><a href="https://www.deepseek.com/">[🏠Homepage]</a> | <a href="https://coder.deepseek.com/">[🤖 Chat with DeepSeek Coder]</a> | <a href="https://discord.gg/Tc7c45Zzu5">[Discord]</a> | <a href="https://github.com/guoday/assert/blob/main/QR.png?raw=true">[Wechat(微信)]</a> </p>
307
+ <hr>
308
+
309
+
310
+
311
+ ### 1. Introduction of Deepseek Coder
312
+
313
+ Deepseek Coder is composed of a series of code language models, each trained from scratch on 2T tokens, with a composition of 87% code and 13% natural language in both English and Chinese. We provide various sizes of the code model, ranging from 1B to 33B versions. Each model is pre-trained on project-level code corpus by employing a window size of 16K and a extra fill-in-the-blank task, to support project-level code completion and infilling. For coding capabilities, Deepseek Coder achieves state-of-the-art performance among open-source code models on multiple programming languages and various benchmarks.
314
+
315
+ - **Massive Training Data**: Trained from scratch on 2T tokens, including 87% code and 13% linguistic data in both English and Chinese languages.
316
+
317
+ - **Highly Flexible & Scalable**: Offered in model sizes of 1.3B, 5.7B, 6.7B, and 33B, enabling users to choose the setup most suitable for their requirements.
318
+
319
+ - **Superior Model Performance**: State-of-the-art performance among publicly available code models on HumanEval, MultiPL-E, MBPP, DS-1000, and APPS benchmarks.
320
+
321
+ - **Advanced Code Completion Capabilities**: A window size of 16K and a fill-in-the-blank task, supporting project-level code completion and infilling tasks.
322
+
323
+
324
+
325
+ ### 2. Model Summary
326
+ deepseek-coder-33b-instruct is a 33B parameter model initialized from deepseek-coder-33b-base and fine-tuned on 2B tokens of instruction data.
327
+ - **Home Page:** [DeepSeek](https://deepseek.com/)
328
+ - **Repository:** [deepseek-ai/deepseek-coder](https://github.com/deepseek-ai/deepseek-coder)
329
+ - **Chat With DeepSeek Coder:** [DeepSeek-Coder](https://coder.deepseek.com/)
330
+
331
+
332
+ ### 3. How to Use
333
+ Here give some examples of how to use our model.
334
+ #### Chat Model Inference
335
+ ```python
336
+ from transformers import AutoTokenizer, AutoModelForCausalLM
337
+ tokenizer = AutoTokenizer.from_pretrained("deepseek-ai/deepseek-coder-33b-instruct", trust_remote_code=True)
338
+ model = AutoModelForCausalLM.from_pretrained("deepseek-ai/deepseek-coder-33b-instruct", trust_remote_code=True).cuda()
339
+ messages=[
340
+ { 'role': 'user', 'content': "write a quick sort algorithm in python."}
341
+ ]
342
+ inputs = tokenizer.apply_chat_template(messages, return_tensors="pt").to(model.device)
343
+ # 32021 is the id of <|EOT|> token
344
+ outputs = model.generate(inputs, max_new_tokens=512, do_sample=False, top_k=50, top_p=0.95, num_return_sequences=1, eos_token_id=32021)
345
+ print(tokenizer.decode(outputs[0][len(inputs[0]):], skip_special_tokens=True))
346
+ ```
347
+
348
+ ### 4. License
349
+ This code repository is licensed under the MIT License. The use of DeepSeek Coder models is subject to the Model License. DeepSeek Coder supports commercial use.
350
+
351
+ See the [LICENSE-MODEL](https://github.com/deepseek-ai/deepseek-coder/blob/main/LICENSE-MODEL) for more details.
352
+
353
+ ### 5. Contact
354
+
355
+ If you have any questions, please raise an issue or contact us at [[email protected]](mailto:[email protected]).
356
+
357
+
358
+ <!-- original-model-card end -->