diff --git a/_static/images/llama_cpp.png b/_static/images/llama_cpp.png new file mode 100644 index 0000000..365c5b8 Binary files /dev/null and b/_static/images/llama_cpp.png differ diff --git a/_static/llamacpp_actions.js b/_static/llamacpp_actions.js new file mode 100644 index 0000000..98f3e61 --- /dev/null +++ b/_static/llamacpp_actions.js @@ -0,0 +1,49 @@ +$(document).ready(function () { + $.reset_selection = function (elem) { + elem.parent().children().each(function () { + $(this).removeClass("selected"); + }); + } + + $.get_options = function () { + var options = {}; + $('#col-values').children().each(function () { + var elem = $(this).find(".selected").each(function () { + var id = $(this).attr("id").split("-"); + var category = id[0]; + var value = id[1]; + options[category] = value; + }); + }); + return options; + } + + $.update_table = function () { + var options = $.get_options(); + } + + $("#col-values").on("click", ".values-element", function () { + id = $(this).attr("id"); + fields = id.split("-"); + + $.reset_selection($(this)); + $(this).addClass("selected"); + $.update_table(); + $.gen_content(); + }); + + + $.gen_content = function () { + var options = $.get_options(); + if (options['install_type'] == "docker") { + $('#install-llamacpp-pip-section').hide(); + $('#install-llamacpp-docker-section').show(); + } else if (options['install_type'] == "pip") { + $('#install-llamacpp-docker-section').hide(); + $('#install-llamacpp-pip-section').show(); + } + } + + $.update_table(); + $.gen_content(); +}); diff --git a/index.rst b/index.rst index e8b229b..d77d810 100644 --- a/index.rst +++ b/index.rst @@ -32,6 +32,7 @@ sources/lm_evaluation/index.rst sources/wenet/index.rst sources/whisper_cpp/index.rst + sources/llama_cpp/index.rst .. warning:: @@ -263,5 +264,23 @@ 快速上手 + +
+
+
+
+

llama.cpp

+

由C/C++实现的 Meta LLaMa 架构

+
+
+
+
+ 官方链接 + | + 安装指南 + | + 快速上手 +
+
diff --git a/sources/llama_cpp/index.rst b/sources/llama_cpp/index.rst new file mode 100644 index 0000000..3945d2e --- /dev/null +++ b/sources/llama_cpp/index.rst @@ -0,0 +1,10 @@ +llama.cpp +=========== + +.. toctree:: + :maxdepth: 2 + + install.rst + quick_start.rst + + diff --git a/sources/llama_cpp/install.rst b/sources/llama_cpp/install.rst new file mode 100644 index 0000000..2e34f54 --- /dev/null +++ b/sources/llama_cpp/install.rst @@ -0,0 +1,125 @@ +安装指南 +============== + +本教程面向使用 llama.cpp & 昇腾的开发者，帮助完成昇腾环境下 llama.cpp 的安装。 + +.. note:: + 目前 llama.cpp 仅支持 Atlas 300T A2 型号设备 + +llama.cpp 下载安装 +--------------------------- + +此处提供 docker 和 pip 两种安装方式，请按需选择: + +.. raw:: html + + +
+
+
+
+
安装方式
+
pip
+
Docker
+
+
+
+
+ +--------------- + +.. raw:: html + +
+

使用源代码安装

+
+

备注

+

请确保已经根据快速安装昇腾环境指引安装了对应的CANN-toolkit版本以及相应的固件和驱动，并应用了CANN-toolkit环境变量。

+
+
+

提示

+

LLAMA-Factory 支持的 CANN 最低版本为 8.0.rc1。安装 CANN 时，请同时安装 Kernel 算子包。

+
+

Python 环境创建

+
+

创建并激活 Python 环境：

+
+ 
conda create -y -n llamacpp python=3.10
+    conda activate llamacpp
+
+
+ +

获取源代码

+
+

使用以下 git 指令获取源码

+
+ 
git clone https://github.com/ggerganov/llama.cpp
+    cd llama.cpp
+
+
+ +

构建 llama.cpp

+
+
+ 
cmake -B build -DGGML_CANN=on -DCMAKE_BUILD_TYPE=release
+    cmake --build build --config release
+
+
+ +
+ +
+
+

使用 Docker

+
+

备注

+

请确保已经根据快速安装昇腾环境指引安装了对应的的固件和驱动。

+
+
+

提示

+

更多 CANN 的基础镜像选择见ascendai/cann。

+
+

构建 docker 镜像：

+
+ 
docker build -t llama-cpp-cann -f .devops/llama-cli-cann.Dockerfile .
+
+

找到所有卡的运行信息：

+
+ 
npu-smi info
+
+

启动 docker 容器：

+
+ 
docker run --name llamacpp \
+    --device /dev/davinci0  \
+    --device /dev/davinci_manager \
+    --device /dev/devmm_svm \
+    --device /dev/hisi_hdc \
+    -v /usr/local/dcmi:/usr/local/dcmi \
+    -v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
+    -v /usr/local/Ascend/driver/lib64/:/usr/local/Ascend/driver/lib64/ \
+    -v /usr/local/Ascend/driver/version.info:/usr/local/Ascend/driver/version.info \
+    -v /PATH_TO_YOUR_MODELS/:/app/models \
+    -it llama-cpp-cann -m /app/models/MODEL_PATH -ngl 32 \
+    -p "Building a website can be done in 10 simple steps:"
+
+
+
+ +安装校验 +----------------- + +安装完成后，无任何报错信息，即为安装成功，下面为部分回显信息: + +.. code-block:: shell + :linenos: + + [ 97%] Built target test-grammar-integration + [ 97%] Built target llama-speculative + [ 97%] Built target llama-perplexity + [ 98%] Linking CXX executable ../../bin/llama-bench + [ 98%] Linking CXX executable ../bin/test-json-schema-to-grammar + [ 98%] Built target llama-bench + [ 98%] Built target test-json-schema-to-grammar + [100%] Linking CXX executable ../../bin/llama-server + [100%] Built target llama-server + diff --git a/sources/llama_cpp/quick_start.rst b/sources/llama_cpp/quick_start.rst new file mode 100644 index 0000000..dd52aff --- /dev/null +++ b/sources/llama_cpp/quick_start.rst @@ -0,0 +1,179 @@ +快速开始 +============ + +.. note:: + 阅读本篇前，请确保已按照 :doc:`安装指南 <./install>` 准备好昇腾环境及 llama.cpp ！ + +本教程聚焦大语言模型（Large Language Model，LLM）的推理过程，以 Qwen2.5-7B 模型为例，讲述如何使用 llama.cpp 在昇腾 NPU 上进行推理。 + + +模型文件准备及量化 +--------------- + +llama.cpp 的推理需要使用 gguf 格式文件，llama.cpp 提供了两种方式转换 Hugging Face 模型文件： + +- 使用 `GGUF-my-repo `_ 将模型进行转换。 + +- 使用项目中的 `convert_hf_to_gguf.py` 文件将 Hugging Face 模型转换为 gguf 格式: + + .. code-block:: shell + :linenos: + + python convert_hf_to_gguf.py path/to/model + +详情请参考 `Prepare and Quantize `_ 。 + +注意：目前仅支持 FP16 精度及 Q4_0/Q8_0 量化模型。 + +推理 +------------ + +有两种设备选择模式: + +- 单设备：使用用户指定的一个设备目标。 +- 多设备：自动选择具有相同后端的设备。 + ++---------------+---------------------------------------------+ +| 设备选择 | 参数 | ++===============+=============================================+ +| 单设备 | --split-mode none --main-gpu DEVICE_ID | ++---------------+---------------------------------------------+ +| 多设备 | --split-mode layer (default) | ++---------------+---------------------------------------------+ + +使用单卡推理 +++++++++++++++++ + +.. code-block:: shell + :linenos: + + ./build/bin/llama-cli -m path_to_model -p "Building a website can be done in 10 simple steps:" -n 400 -e -ngl 33 -sm none -mg 0 + +使用多卡推理 +++++++++++++++++ + +.. code-block:: shell + :linenos: + + ./build/bin/llama-cli -m path_to_model -p "Building a website can be done in 10 simple steps:" -n 400 -e -ngl 33 -sm layer + +以下为正常推理结果： + +.. code-block:: shell + :linenos: + + Log start + main: build = 3520 (8e707118) + main: built with cc (Ubuntu 9.4.0-1ubuntu1~20.04.2) 9.4.0 for aarch64-linux-gnu + main: seed = 1728907816 + llama_model_loader: loaded meta data with 22 key-value pairs and 291 tensors from /home/jiahao/models/llama3-8b-instruct-fp16.gguf (version GGUF V3 (latest)) + llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. + llama_model_loader: - kv 0: general.architecture str = llama + llama_model_loader: - kv 1: general.name str = Meta-Llama-3-8B-Instruct + llama_model_loader: - kv 2: llama.block_count u32 = 32 + llama_model_loader: - kv 3: llama.context_length u32 = 8192 + llama_model_loader: - kv 4: llama.embedding_length u32 = 4096 + llama_model_loader: - kv 5: llama.feed_forward_length u32 = 14336 + llama_model_loader: - kv 6: llama.attention.head_count u32 = 32 + llama_model_loader: - kv 7: llama.attention.head_count_kv u32 = 8 + llama_model_loader: - kv 8: llama.rope.freq_base f32 = 500000.000000 + llama_model_loader: - kv 9: llama.attention.layer_norm_rms_epsilon f32 = 0.000010 + llama_model_loader: - kv 10: general.file_type u32 = 1 + llama_model_loader: - kv 11: llama.vocab_size u32 = 128256 + llama_model_loader: - kv 12: llama.rope.dimension_count u32 = 128 + llama_model_loader: - kv 13: tokenizer.ggml.model str = gpt2 + llama_model_loader: - kv 14: tokenizer.ggml.pre str = llama-bpe + llama_model_loader: - kv 15: tokenizer.ggml.tokens arr[str,128256] = ["!", "\"", "#", "$", "%", "&", "'", ... + llama_model_loader: - kv 16: tokenizer.ggml.token_type arr[i32,128256] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... + llama_model_loader: - kv 17: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "... + llama_model_loader: - kv 18: tokenizer.ggml.bos_token_id u32 = 128000 + llama_model_loader: - kv 19: tokenizer.ggml.eos_token_id u32 = 128009 + llama_model_loader: - kv 20: tokenizer.chat_template str = {% set loop_messages = messages %}{% ... + llama_model_loader: - kv 21: general.quantization_version u32 = 2 + llama_model_loader: - type f32: 65 tensors + llama_model_loader: - type f16: 226 tensors + llm_load_vocab: special tokens cache size = 256 + llm_load_vocab: token to piece cache size = 0.8000 MB + llm_load_print_meta: format = GGUF V3 (latest) + llm_load_print_meta: arch = llama + llm_load_print_meta: vocab type = BPE + llm_load_print_meta: n_vocab = 128256 + llm_load_print_meta: n_merges = 280147 + llm_load_print_meta: vocab_only = 0 + llm_load_print_meta: n_ctx_train = 8192 + llm_load_print_meta: n_embd = 4096 + llm_load_print_meta: n_layer = 32 + llm_load_print_meta: n_head = 32 + llm_load_print_meta: n_head_kv = 8 + llm_load_print_meta: n_rot = 128 + llm_load_print_meta: n_swa = 0 + llm_load_print_meta: n_embd_head_k = 128 + llm_load_print_meta: n_embd_head_v = 128 + llm_load_print_meta: n_gqa = 4 + llm_load_print_meta: n_embd_k_gqa = 1024 + llm_load_print_meta: n_embd_v_gqa = 1024 + llm_load_print_meta: f_norm_eps = 0.0e+00 + llm_load_print_meta: f_norm_rms_eps = 1.0e-05 + llm_load_print_meta: f_clamp_kqv = 0.0e+00 + llm_load_print_meta: f_max_alibi_bias = 0.0e+00 + llm_load_print_meta: f_logit_scale = 0.0e+00 + llm_load_print_meta: n_ff = 14336 + llm_load_print_meta: n_expert = 0 + llm_load_print_meta: n_expert_used = 0 + llm_load_print_meta: causal attn = 1 + llm_load_print_meta: pooling type = 0 + llm_load_print_meta: rope type = 0 + llm_load_print_meta: rope scaling = linear + llm_load_print_meta: freq_base_train = 500000.0 + llm_load_print_meta: freq_scale_train = 1 + llm_load_print_meta: n_ctx_orig_yarn = 8192 + llm_load_print_meta: rope_finetuned = unknown + llm_load_print_meta: ssm_d_conv = 0 + llm_load_print_meta: ssm_d_inner = 0 + llm_load_print_meta: ssm_d_state = 0 + llm_load_print_meta: ssm_dt_rank = 0 + llm_load_print_meta: model type = 8B + llm_load_print_meta: model ftype = F16 + llm_load_print_meta: model params = 8.03 B + llm_load_print_meta: model size = 14.96 GiB (16.00 BPW) + llm_load_print_meta: general.name = Meta-Llama-3-8B-Instruct + llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>' + llm_load_print_meta: EOS token = 128009 '<|eot_id|>' + llm_load_print_meta: LF token = 128 'Ä' + llm_load_print_meta: EOT token = 128009 '<|eot_id|>' + llm_load_print_meta: max token length = 256 + llm_load_tensors: ggml ctx size = 0.27 MiB + llm_load_tensors: CPU buffer size = 15317.02 MiB + llm_load_tensors: CANN buffer size = 13313.00 MiB + ......................................................................................... + llama_new_context_with_model: n_ctx = 8192 + llama_new_context_with_model: n_batch = 2048 + llama_new_context_with_model: n_ubatch = 512 + llama_new_context_with_model: flash_attn = 0 + llama_new_context_with_model: freq_base = 500000.0 + llama_new_context_with_model: freq_scale = 1 + llama_kv_cache_init: CANN KV buffer size = 1024.00 MiB + llama_new_context_with_model: KV self size = 1024.00 MiB, K (f16): 512.00 MiB, V (f16): 512.00 MiB + llama_new_context_with_model: CPU output buffer size = 0.49 MiB + llama_new_context_with_model: CANN compute buffer size = 1260.50 MiB + llama_new_context_with_model: CPU compute buffer size = 24.01 MiB + llama_new_context_with_model: graph nodes = 1030 + llama_new_context_with_model: graph splits = 4 + + system_info: n_threads = 192 / 192 | AVX = 0 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 1 | SVE = 0 | ARM_FMA = 1 | F16C = 0 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 0 | SSE3 = 0 | SSSE3 = 0 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | + sampling: + repeat_last_n = 64, repeat_penalty = 1.000, frequency_penalty = 0.000, presence_penalty = 0.000 + top_k = 40, tfs_z = 1.000, top_p = 0.950, min_p = 0.050, typical_p = 1.000, temp = 0.800 + mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000 + sampling order: + CFG -> Penalties -> top_k -> tfs_z -> typical_p -> top_p -> min_p -> temperature + generate: n_ctx = 8192, n_batch = 2048, n_predict = -1, n_keep = 1 + + + Building a website can be done in 10 simple steps: 1. Define your website's purpose and target audience 2. Choose a domain name and register it with a registrar 3. Select a web hosting service and set up your hosting account 4. Design your website's layout and structure 5. Create content for your website, including text, images, and other media 6. Build a responsive website design that adapts to different devices and screen sizes 7. Choose a Content Management System (CMS) and install it on your website 8. Customize your website's design and layout using a CMS + + llama_print_timings: load time = 9074.69 ms + llama_print_timings: sample time = 31.97 ms / 112 runs ( 0.29 ms per token, 3503.28 tokens per second) + llama_print_timings: prompt eval time = 238.53 ms / 13 tokens ( 18.35 ms per token, 54.50 tokens per second) + llama_print_timings: eval time = 13152.29 ms / 111 runs ( 118.49 ms per token, 8.44 tokens per second) + llama_print_timings: total time = 13623.53 ms / 124 tokens \ No newline at end of file