--- license: other license_name: deepseek-license license_link: https://github.com/deepseek-ai/DeepSeek-Coder-V2/raw/main/LICENSE-MODEL tags: - code language: - code base_model: deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct model_creator: DeepSeek AI model_name: DeepSeek-Coder-V2-Lite-Instruct model_type: deepseek2 datasets: - m-a-p/CodeFeedback-Filtered-Instruction quantized_by: CISC --- # DeepSeek-Coder-V2-Lite-Instruct - SOTA GGUF - Model creator: [DeepSeek AI](https://huggingface.co/deepseek-ai) - Original model: [DeepSeek-Coder-V2-Lite-Instruct](https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct) ## Description This repo contains State Of The Art quantized GGUF format model files for [DeepSeek-Coder-V2-Lite-Instruct](https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct). Quantization was done with an importance matrix that was trained for ~250K tokens (64 batches of 4096 tokens) of answers from the [CodeFeedback-Filtered-Instruction](https://huggingface.co/datasets/m-a-p/CodeFeedback-Filtered-Instruction) dataset. Fill-in-Middle token metadata has been added, see [example](#simple-llama-cpp-python-example-fill-in-middle-code). NOTE: Due to some of the tensors in this model being oddly shaped a consequential portion of the quantization fell back to IQ4_NL instead of the specified method, causing somewhat larger (and "smarter"; even IQ1_M is quite usable) model files than usual! ## Prompt template: DeepSeek v2 ``` User: {prompt} Assistant: ``` ## Compatibility These quantised GGUFv3 files are compatible with llama.cpp from May 29th 2024 onwards, as of commit [fb76ec2](https://github.com/ggerganov/llama.cpp/commit/fb76ec31a9914b7761c1727303ab30380fd4f05c) They are also compatible with many third party UIs and libraries provided they are built using a recent llama.cpp. ## Explanation of quantisation methods
Click to see details The new methods available are: * GGML_TYPE_IQ1_S - 1-bit quantization in super-blocks with an importance matrix applied, effectively using 1.56 bits per weight (bpw) * GGML_TYPE_IQ1_M - 1-bit quantization in super-blocks with an importance matrix applied, effectively using 1.75 bpw * GGML_TYPE_IQ2_XXS - 2-bit quantization in super-blocks with an importance matrix applied, effectively using 2.06 bpw * GGML_TYPE_IQ2_XS - 2-bit quantization in super-blocks with an importance matrix applied, effectively using 2.31 bpw * GGML_TYPE_IQ2_S - 2-bit quantization in super-blocks with an importance matrix applied, effectively using 2.5 bpw * GGML_TYPE_IQ2_M - 2-bit quantization in super-blocks with an importance matrix applied, effectively using 2.7 bpw * GGML_TYPE_IQ3_XXS - 3-bit quantization in super-blocks with an importance matrix applied, effectively using 3.06 bpw * GGML_TYPE_IQ3_XS - 3-bit quantization in super-blocks with an importance matrix applied, effectively using 3.3 bpw * GGML_TYPE_IQ3_S - 3-bit quantization in super-blocks with an importance matrix applied, effectively using 3.44 bpw * GGML_TYPE_IQ3_M - 3-bit quantization in super-blocks with an importance matrix applied, effectively using 3.66 bpw * GGML_TYPE_IQ4_XS - 4-bit quantization in super-blocks with an importance matrix applied, effectively using 4.25 bpw * GGML_TYPE_IQ4_NL - 4-bit non-linearly mapped quantization with an importance matrix applied, effectively using 4.5 bpw Refer to the Provided Files table below to see what files use which methods, and how.
## Provided files | Name | Quant method | Bits | Size | Max RAM required | Use case | | ---- | ---- | ---- | ---- | ---- | ----- | | [DeepSeek-Coder-V2-Lite-Instruct.IQ1_S.gguf](https://huggingface.co/CISCai/DeepSeek-Coder-V2-Lite-Instruct-SOTA-GGUF/blob/main/DeepSeek-Coder-V2-Lite-Instruct.IQ1_S.gguf) | IQ1_S | 1 | 4.5 GB| 5.5 GB | smallest, significant quality loss | | [DeepSeek-Coder-V2-Lite-Instruct.IQ1_M.gguf](https://huggingface.co/CISCai/DeepSeek-Coder-V2-Lite-Instruct-SOTA-GGUF/blob/main/DeepSeek-Coder-V2-Lite-Instruct.IQ1_M.gguf) | IQ1_M | 1 | 4.7 GB| 5.7 GB | very small, significant quality loss | | [DeepSeek-Coder-V2-Lite-Instruct.IQ2_XXS.gguf](https://huggingface.co/CISCai/DeepSeek-Coder-V2-Lite-Instruct-SOTA-GGUF/blob/main/DeepSeek-Coder-V2-Lite-Instruct.IQ2_XXS.gguf) | IQ2_XXS | 2 | 5.1 GB| 6.1 GB | very small, high quality loss | | [DeepSeek-Coder-V2-Lite-Instruct.IQ2_XS.gguf](https://huggingface.co/CISCai/DeepSeek-Coder-V2-Lite-Instruct-SOTA-GGUF/blob/main/DeepSeek-Coder-V2-Lite-Instruct.IQ2_XS.gguf) | IQ2_XS | 2 | 5.4 GB| 6.4 GB | very small, high quality loss | | [DeepSeek-Coder-V2-Lite-Instruct.IQ2_S.gguf](https://huggingface.co/CISCai/DeepSeek-Coder-V2-Lite-Instruct-SOTA-GGUF/blob/main/DeepSeek-Coder-V2-Lite-Instruct.IQ2_S.gguf) | IQ2_S | 2 | 5.4 GB| 6.4 GB | small, substantial quality loss | | [DeepSeek-Coder-V2-Lite-Instruct.IQ2_M.gguf](https://huggingface.co/CISCai/DeepSeek-Coder-V2-Lite-Instruct-SOTA-GGUF/blob/main/DeepSeek-Coder-V2-Lite-Instruct.IQ2_M.gguf) | IQ2_M | 2 | 5.7 GB| 6.7 GB | small, greater quality loss | | [DeepSeek-Coder-V2-Lite-Instruct.IQ3_XXS.gguf](https://huggingface.co/CISCai/DeepSeek-Coder-V2-Lite-Instruct-SOTA-GGUF/blob/main/DeepSeek-Coder-V2-Lite-Instruct.IQ3_XXS.gguf) | IQ3_XXS | 3 | 6.3 GB| 7.3 GB | very small, high quality loss | | [DeepSeek-Coder-V2-Lite-Instruct.IQ3_XS.gguf](https://huggingface.co/CISCai/DeepSeek-Coder-V2-Lite-Instruct-SOTA-GGUF/blob/main/DeepSeek-Coder-V2-Lite-Instruct.IQ3_XS.gguf) | IQ3_XS | 3 | 6.5 GB| 7.5 GB | small, substantial quality loss | | [DeepSeek-Coder-V2-Lite-Instruct.IQ3_S.gguf](https://huggingface.co/CISCai/DeepSeek-Coder-V2-Lite-Instruct-SOTA-GGUF/blob/main/DeepSeek-Coder-V2-Lite-Instruct.IQ3_S.gguf) | IQ3_S | 3 | 6.8 GB| 7.8 GB | small, greater quality loss | | [DeepSeek-Coder-V2-Lite-Instruct.IQ3_M.gguf](https://huggingface.co/CISCai/DeepSeek-Coder-V2-Lite-Instruct-SOTA-GGUF/blob/main/DeepSeek-Coder-V2-Lite-Instruct.IQ3_M.gguf) | IQ3_M | 3 | 6.9 GB| 7.9 GB | medium, balanced quality - recommended | | [DeepSeek-Coder-V2-Lite-Instruct.IQ4_NL.gguf](https://huggingface.co/CISCai/DeepSeek-Coder-V2-Lite-Instruct-SOTA-GGUF/blob/main/DeepSeek-Coder-V2-Lite-Instruct.IQ4_NL.gguf) | IQ4_NL | 4 | 8.1 GB| 9.1 GB | small, substantial quality loss | Generated importance matrix file: [DeepSeek-Coder-V2-Lite-Instruct.imatrix.dat](https://huggingface.co/CISCai/DeepSeek-Coder-V2-Lite-Instruct-SOTA-GGUF/blob/main/DeepSeek-Coder-V2-Lite-Instruct.imatrix.dat) **Note**: the above RAM figures assume no GPU offloading with 4K context. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead. ## Example `llama.cpp` command Make sure you are using `llama.cpp` from commit [fb76ec3](https://github.com/ggerganov/llama.cpp/commit/fb76ec31a9914b7761c1727303ab30380fd4f05c) or later. ```shell ./llama-cli -ngl 28 -m DeepSeek-Coder-V2-Lite-Instruct.IQ4_NL.gguf --color -c 131072 --temp 0 --repeat-penalty 1.1 -p "User: {prompt}\n\nAssistant:" ``` Change `-ngl 28` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration. Change `-c 131072` to the desired sequence length. If you are low on V/RAM try quantizing the K-cache with `-ctk q8_0` or even `-ctk q4_0` for big memory savings (depending on context size). There is a similar option for V-cache (`-ctv`), however that requires Flash Attention [which is not working yet with this model](https://github.com/ggerganov/llama.cpp/issues/7343). 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) ## How to run from Python code You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) module. ### How to load this model in Python code, using llama-cpp-python For full documentation, please see: [llama-cpp-python docs](https://llama-cpp-python.readthedocs.io/en/latest/). #### First install the package Run one of the following commands, according to your system: ```shell # Prebuilt wheel with basic CPU support pip install llama-cpp-python --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cpu # Prebuilt wheel with NVidia CUDA acceleration pip install llama-cpp-python --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cu121 (or cu122 etc.) # Prebuilt wheel with Metal GPU acceleration pip install llama-cpp-python --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/metal # Build base version with no GPU acceleration pip install llama-cpp-python # With NVidia CUDA acceleration CMAKE_ARGS="-DLLAMA_CUDA=on" pip install llama-cpp-python # Or with OpenBLAS acceleration CMAKE_ARGS="-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS" pip install llama-cpp-python # Or with CLBLast acceleration CMAKE_ARGS="-DLLAMA_CLBLAST=on" pip install llama-cpp-python # Or with AMD ROCm GPU acceleration (Linux only) CMAKE_ARGS="-DLLAMA_HIPBLAS=on" pip install llama-cpp-python # Or with Metal GPU acceleration for macOS systems only CMAKE_ARGS="-DLLAMA_METAL=on" pip install llama-cpp-python # Or with Vulkan acceleration CMAKE_ARGS="-DLLAMA_VULKAN=on" pip install llama-cpp-python # Or with Kompute acceleration CMAKE_ARGS="-DLLAMA_KOMPUTE=on" pip install llama-cpp-python # Or with SYCL acceleration CMAKE_ARGS="-DLLAMA_SYCL=on -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx" pip install llama-cpp-python # In windows, to set the variables CMAKE_ARGS in PowerShell, follow this format; eg for NVidia CUDA: $env:CMAKE_ARGS = "-DLLAMA_CUDA=on" pip install llama-cpp-python ``` #### Simple llama-cpp-python example code ```python from llama_cpp import Llama # Chat Completion API llm = Llama(model_path="./DeepSeek-Coder-V2-Lite-Instruct.IQ4_NL.gguf", n_gpu_layers=28, n_ctx=131072) print(llm.create_chat_completion( repeat_penalty = 1.1, messages = [ { "role": "user", "content": "Pick a LeetCode challenge and solve it in Python." } ] )) ``` #### Simple llama-cpp-python example fill-in-middle code ```python from llama_cpp import Llama # Completion API prompt = "def add(" suffix = "\n return sum\n\n" llm = Llama(model_path="./DeepSeek-Coder-V2-Lite-Instruct.IQ4_NL.gguf", n_gpu_layers=28, n_ctx=131072) output = llm.create_completion( temperature = 0.0, repeat_penalty = 1.0, prompt = prompt, suffix = suffix ) # Models sometimes repeat suffix in response, attempt to filter that response = output["choices"][0]["text"] response_stripped = response.rstrip() unwanted_response_suffix = suffix.rstrip() unwanted_response_length = len(unwanted_response_suffix) filtered = False if unwanted_response_suffix and response_stripped[-unwanted_response_length:] == unwanted_response_suffix: response = response_stripped[:-unwanted_response_length] filtered = True print(f"Fill-in-Middle completion{' (filtered)' if filtered else ''}:\n\n{prompt}\033[32m{response}\033[{'33' if filtered else '0'}m{suffix}\033[0m") ```
DeepSeek-V2
Homepage Chat Hugging Face
Discord Wechat Twitter Follow
Code License Model License

API Platform | How to Use | License |

Paper LinkšŸ‘ļø

# DeepSeek-Coder-V2: Breaking the Barrier of Closed-Source Models in Code Intelligence ## 1. Introduction We present DeepSeek-Coder-V2, an open-source Mixture-of-Experts (MoE) code language model that achieves performance comparable to GPT4-Turbo in code-specific tasks. Specifically, DeepSeek-Coder-V2 is further pre-trained from DeepSeek-Coder-V2-Base with 6 trillion tokens sourced from a high-quality and multi-source corpus. Through this continued pre-training, DeepSeek-Coder-V2 substantially enhances the coding and mathematical reasoning capabilities of DeepSeek-Coder-V2-Base, while maintaining comparable performance in general language tasks. Compared to DeepSeek-Coder, DeepSeek-Coder-V2 demonstrates significant advancements in various aspects of code-related tasks, as well as reasoning and general capabilities. Additionally, DeepSeek-Coder-V2 expands its support for programming languages from 86 to 338, while extending the context length from 16K to 128K.

In standard benchmark evaluations, DeepSeek-Coder-V2 achieves superior performance compared to closed-source models such as GPT4-Turbo, Claude 3 Opus, and Gemini 1.5 Pro in coding and math benchmarks. The list of supported programming languages can be found in the paper. ## 2. Model Downloads We release the DeepSeek-Coder-V2 with 16B and 236B parameters based on the [DeepSeekMoE](https://arxiv.org/pdf/2401.06066) framework, which has actived parameters of only 2.4B and 21B , including base and instruct models, to the public.
| **Model** | **#Total Params** | **#Active Params** | **Context Length** | **Download** | | :-----------------------------: | :---------------: | :----------------: | :----------------: | :----------------------------------------------------------: | | DeepSeek-Coder-V2-Lite-Base | 16B | 2.4B | 128k | [šŸ¤— HuggingFace](https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Lite-Base) | | DeepSeek-Coder-V2-Lite-Instruct | 16B | 2.4B | 128k | [šŸ¤— HuggingFace](https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct) | | DeepSeek-Coder-V2-Base | 236B | 21B | 128k | [šŸ¤— HuggingFace](https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Base) | | DeepSeek-Coder-V2-Instruct | 236B | 21B | 128k | [šŸ¤— HuggingFace](https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Instruct) |
## 3. Chat Website You can chat with the DeepSeek-Coder-V2 on DeepSeek's official website: [coder.deepseek.com](https://coder.deepseek.com/sign_in) ## 4. API Platform We also provide OpenAI-Compatible API at DeepSeek Platform: [platform.deepseek.com](https://platform.deepseek.com/). Sign up for over millions of free tokens. And you can also pay-as-you-go at an unbeatable price.

## 5. How to run locally **Here, we provide some examples of how to use DeepSeek-Coder-V2-Lite model. If you want to utilize DeepSeek-Coder-V2 in BF16 format for inference, 80GB*8 GPUs are required.** ### Inference with Huggingface's Transformers You can directly employ [Huggingface's Transformers](https://github.com/huggingface/transformers) for model inference. #### Code Completion ```python from transformers import AutoTokenizer, AutoModelForCausalLM import torch tokenizer = AutoTokenizer.from_pretrained("deepseek-ai/DeepSeek-Coder-V2-Lite-Base", trust_remote_code=True) model = AutoModelForCausalLM.from_pretrained("deepseek-ai/DeepSeek-Coder-V2-Lite-Base", trust_remote_code=True, torch_dtype=torch.bfloat16).cuda() input_text = "#write a quick sort algorithm" inputs = tokenizer(input_text, return_tensors="pt").to(model.device) outputs = model.generate(**inputs, max_length=128) print(tokenizer.decode(outputs[0], skip_special_tokens=True)) ``` #### Code Insertion ```python from transformers import AutoTokenizer, AutoModelForCausalLM import torch tokenizer = AutoTokenizer.from_pretrained("deepseek-ai/DeepSeek-Coder-V2-Lite-Base", trust_remote_code=True) model = AutoModelForCausalLM.from_pretrained("deepseek-ai/DeepSeek-Coder-V2-Lite-Base", trust_remote_code=True, torch_dtype=torch.bfloat16).cuda() input_text = """<ļ½œfimā–beginļ½œ>def quick_sort(arr): if len(arr) <= 1: return arr pivot = arr[0] left = [] right = [] <ļ½œfimā–holeļ½œ> if arr[i] < pivot: left.append(arr[i]) else: right.append(arr[i]) return quick_sort(left) + [pivot] + quick_sort(right)<ļ½œfimā–endļ½œ>""" inputs = tokenizer(input_text, return_tensors="pt").to(model.device) outputs = model.generate(**inputs, max_length=128) print(tokenizer.decode(outputs[0], skip_special_tokens=True)[len(input_text):]) ``` #### Chat Completion ```python from transformers import AutoTokenizer, AutoModelForCausalLM import torch tokenizer = AutoTokenizer.from_pretrained("deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct", trust_remote_code=True) model = AutoModelForCausalLM.from_pretrained("deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct", trust_remote_code=True, torch_dtype=torch.bfloat16).cuda() messages=[ { 'role': 'user', 'content': "write a quick sort algorithm in python."} ] inputs = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt").to(model.device) # tokenizer.eos_token_id is the id of <|EOT|> token 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=tokenizer.eos_token_id) print(tokenizer.decode(outputs[0][len(inputs[0]):], skip_special_tokens=True)) ``` The complete chat template can be found within `tokenizer_config.json` located in the huggingface model repository. An example of chat template is as belows: ```bash <ļ½œbeginā–ofā–sentenceļ½œ>User: {user_message_1} Assistant: {assistant_message_1}<ļ½œendā–ofā–sentenceļ½œ>User: {user_message_2} Assistant: ``` You can also add an optional system message: ```bash <ļ½œbeginā–ofā–sentenceļ½œ>{system_message} User: {user_message_1} Assistant: {assistant_message_1}<ļ½œendā–ofā–sentenceļ½œ>User: {user_message_2} Assistant: ``` ### Inference with vLLM (recommended) To utilize [vLLM](https://github.com/vllm-project/vllm) for model inference, please merge this Pull Request into your vLLM codebase: https://github.com/vllm-project/vllm/pull/4650. ```python from transformers import AutoTokenizer from vllm import LLM, SamplingParams max_model_len, tp_size = 8192, 1 model_name = "deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct" tokenizer = AutoTokenizer.from_pretrained(model_name) llm = LLM(model=model_name, tensor_parallel_size=tp_size, max_model_len=max_model_len, trust_remote_code=True, enforce_eager=True) sampling_params = SamplingParams(temperature=0.3, max_tokens=256, stop_token_ids=[tokenizer.eos_token_id]) messages_list = [ [{"role": "user", "content": "Who are you?"}], [{"role": "user", "content": "write a quick sort algorithm in python."}], [{"role": "user", "content": "Write a piece of quicksort code in C++."}], ] prompt_token_ids = [tokenizer.apply_chat_template(messages, add_generation_prompt=True) for messages in messages_list] outputs = llm.generate(prompt_token_ids=prompt_token_ids, sampling_params=sampling_params) generated_text = [output.outputs[0].text for output in outputs] print(generated_text) ``` ## 6. License This code repository is licensed under [the MIT License](https://github.com/deepseek-ai/DeepSeek-Coder-V2/blob/main/LICENSE-CODE). The use of DeepSeek-Coder-V2 Base/Instruct models is subject to [the Model License](https://github.com/deepseek-ai/DeepSeek-Coder-V2/blob/main/LICENSE-MODEL). DeepSeek-Coder-V2 series (including Base and Instruct) supports commercial use. ## 7. Contact If you have any questions, please raise an issue or contact us at [service@deepseek.com](service@deepseek.com).