EngineX-Mthreads/enginex-mthreads-llama.cpp
4
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

Sync from upstream llama.cpp repository

Browse Source
This commit is contained in:
xiezhongtao
2026-01-16 10:43:34 +08:00
parent 3bc369a6f7
commit f4ae4cc7da
2053 changed files with 956010 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
tools/cli/CMakeLists.txt Normal file
View File

@@ -0,0 +1,10 @@
set(TARGET llama-cli)
add_executable(${TARGET} cli.cpp)
target_link_libraries(${TARGET} PRIVATE server-context PUBLIC common ${CMAKE_THREAD_LIBS_INIT})
target_compile_features(${TARGET} PRIVATE cxx_std_17)
include_directories(../server)
if(LLAMA_TOOLS_INSTALL)
install(TARGETS ${TARGET} RUNTIME)
endif()

192
tools/cli/README.md Normal file
View File

@@ -0,0 +1,192 @@
# llama.cpp/tools/cli
## Usage
<!-- HELP_START -->
<!-- IMPORTANT: The list below is auto-generated by llama-gen-docs; do NOT modify it manually -->
### Common params
| Argument | Explanation |
| -------- | ----------- |
| `-h, --help, --usage` | print usage and exit |
| `--version` | show version and build info |
| `--license` | show source code license and dependencies |
| `-cl, --cache-list` | show list of models in cache |
| `--completion-bash` | print source-able bash completion script for llama.cpp |
| `--verbose-prompt` | print a verbose prompt before generation (default: false) |
| `-t, --threads N` | number of CPU threads to use during generation (default: -1)<br/>(env: LLAMA_ARG_THREADS) |
| `-tb, --threads-batch N` | number of threads to use during batch and prompt processing (default: same as --threads) |
| `-C, --cpu-mask M` | CPU affinity mask: arbitrarily long hex. Complements cpu-range (default: "") |
| `-Cr, --cpu-range lo-hi` | range of CPUs for affinity. Complements --cpu-mask |
| `--cpu-strict <0\|1>` | use strict CPU placement (default: 0) |
| `--prio N` | set process/thread priority : low(-1), normal(0), medium(1), high(2), realtime(3) (default: 0) |
| `--poll <0...100>` | use polling level to wait for work (0 - no polling, default: 50) |
| `-Cb, --cpu-mask-batch M` | CPU affinity mask: arbitrarily long hex. Complements cpu-range-batch (default: same as --cpu-mask) |
| `-Crb, --cpu-range-batch lo-hi` | ranges of CPUs for affinity. Complements --cpu-mask-batch |
| `--cpu-strict-batch <0\|1>` | use strict CPU placement (default: same as --cpu-strict) |
| `--prio-batch N` | set process/thread priority : 0-normal, 1-medium, 2-high, 3-realtime (default: 0) |
| `--poll-batch <0\|1>` | use polling to wait for work (default: same as --poll) |
| `-c, --ctx-size N` | size of the prompt context (default: 0, 0 = loaded from model)<br/>(env: LLAMA_ARG_CTX_SIZE) |
| `-n, --predict, --n-predict N` | number of tokens to predict (default: -1, -1 = infinity)<br/>(env: LLAMA_ARG_N_PREDICT) |
| `-b, --batch-size N` | logical maximum batch size (default: 2048)<br/>(env: LLAMA_ARG_BATCH) |
| `-ub, --ubatch-size N` | physical maximum batch size (default: 512)<br/>(env: LLAMA_ARG_UBATCH) |
| `--keep N` | number of tokens to keep from the initial prompt (default: 0, -1 = all) |
| `--swa-full` | use full-size SWA cache (default: false)<br/>[(more info)](https://github.com/ggml-org/llama.cpp/pull/13194#issuecomment-2868343055)<br/>(env: LLAMA_ARG_SWA_FULL) |
| `-fa, --flash-attn [on\|off\|auto]` | set Flash Attention use ('on', 'off', or 'auto', default: 'auto')<br/>(env: LLAMA_ARG_FLASH_ATTN) |
| `-p, --prompt PROMPT` | prompt to start generation with; for system message, use -sys |
| `--perf, --no-perf` | whether to enable internal libllama performance timings (default: false)<br/>(env: LLAMA_ARG_PERF) |
| `-f, --file FNAME` | a file containing the prompt (default: none) |
| `-bf, --binary-file FNAME` | binary file containing the prompt (default: none) |
| `-e, --escape, --no-escape` | whether to process escapes sequences (\n, \r, \t, \', \", \\) (default: true) |
| `--rope-scaling {none,linear,yarn}` | RoPE frequency scaling method, defaults to linear unless specified by the model<br/>(env: LLAMA_ARG_ROPE_SCALING_TYPE) |
| `--rope-scale N` | RoPE context scaling factor, expands context by a factor of N<br/>(env: LLAMA_ARG_ROPE_SCALE) |
| `--rope-freq-base N` | RoPE base frequency, used by NTK-aware scaling (default: loaded from model)<br/>(env: LLAMA_ARG_ROPE_FREQ_BASE) |
| `--rope-freq-scale N` | RoPE frequency scaling factor, expands context by a factor of 1/N<br/>(env: LLAMA_ARG_ROPE_FREQ_SCALE) |
| `--yarn-orig-ctx N` | YaRN: original context size of model (default: 0 = model training context size)<br/>(env: LLAMA_ARG_YARN_ORIG_CTX) |
| `--yarn-ext-factor N` | YaRN: extrapolation mix factor (default: -1.0, 0.0 = full interpolation)<br/>(env: LLAMA_ARG_YARN_EXT_FACTOR) |
| `--yarn-attn-factor N` | YaRN: scale sqrt(t) or attention magnitude (default: -1.0)<br/>(env: LLAMA_ARG_YARN_ATTN_FACTOR) |
| `--yarn-beta-slow N` | YaRN: high correction dim or alpha (default: -1.0)<br/>(env: LLAMA_ARG_YARN_BETA_SLOW) |
| `--yarn-beta-fast N` | YaRN: low correction dim or beta (default: -1.0)<br/>(env: LLAMA_ARG_YARN_BETA_FAST) |
| `-kvo, --kv-offload, -nkvo, --no-kv-offload` | whether to enable KV cache offloading (default: enabled)<br/>(env: LLAMA_ARG_KV_OFFLOAD) |
| `--repack, -nr, --no-repack` | whether to enable weight repacking (default: enabled)<br/>(env: LLAMA_ARG_REPACK) |
| `--no-host` | bypass host buffer allowing extra buffers to be used<br/>(env: LLAMA_ARG_NO_HOST) |
| `-ctk, --cache-type-k TYPE` | KV cache data type for K<br/>allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1<br/>(default: f16)<br/>(env: LLAMA_ARG_CACHE_TYPE_K) |
| `-ctv, --cache-type-v TYPE` | KV cache data type for V<br/>allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1<br/>(default: f16)<br/>(env: LLAMA_ARG_CACHE_TYPE_V) |
| `-dt, --defrag-thold N` | KV cache defragmentation threshold (DEPRECATED)<br/>(env: LLAMA_ARG_DEFRAG_THOLD) |
| `-np, --parallel N` | number of parallel sequences to decode (default: 1)<br/>(env: LLAMA_ARG_N_PARALLEL) |
| `--mlock` | force system to keep model in RAM rather than swapping or compressing<br/>(env: LLAMA_ARG_MLOCK) |
| `--mmap, --no-mmap` | whether to memory-map model. Explicitly enabling mmap disables direct-io. (if mmap disabled, slower load but may reduce pageouts if not using mlock) (default: enabled)<br/>(env: LLAMA_ARG_MMAP) |
| `-dio, --direct-io, -ndio, --no-direct-io` | use DirectIO if available. Takes precedence over --mmap (default: enabled)<br/>(env: LLAMA_ARG_DIO) |
| `--numa TYPE` | attempt optimizations that help on some NUMA systems<br/>- distribute: spread execution evenly over all nodes<br/>- isolate: only spawn threads on CPUs on the node that execution started on<br/>- numactl: use the CPU map provided by numactl<br/>if run without this previously, it is recommended to drop the system page cache before using this<br/>see https://github.com/ggml-org/llama.cpp/issues/1437<br/>(env: LLAMA_ARG_NUMA) |
| `-dev, --device <dev1,dev2,..>` | comma-separated list of devices to use for offloading (none = don't offload)<br/>use --list-devices to see a list of available devices<br/>(env: LLAMA_ARG_DEVICE) |
| `--list-devices` | print list of available devices and exit |
| `-ot, --override-tensor <tensor name pattern>=<buffer type>,...` | override tensor buffer type<br/>(env: LLAMA_ARG_OVERRIDE_TENSOR) |
| `-cmoe, --cpu-moe` | keep all Mixture of Experts (MoE) weights in the CPU<br/>(env: LLAMA_ARG_CPU_MOE) |
| `-ncmoe, --n-cpu-moe N` | keep the Mixture of Experts (MoE) weights of the first N layers in the CPU<br/>(env: LLAMA_ARG_N_CPU_MOE) |
| `-ngl, --gpu-layers, --n-gpu-layers N` | max. number of layers to store in VRAM, either an exact number, 'auto', or 'all' (default: auto)<br/>(env: LLAMA_ARG_N_GPU_LAYERS) |
| `-sm, --split-mode {none,layer,row}` | how to split the model across multiple GPUs, one of:<br/>- none: use one GPU only<br/>- layer (default): split layers and KV across GPUs<br/>- row: split rows across GPUs<br/>(env: LLAMA_ARG_SPLIT_MODE) |
| `-ts, --tensor-split N0,N1,N2,...` | fraction of the model to offload to each GPU, comma-separated list of proportions, e.g. 3,1<br/>(env: LLAMA_ARG_TENSOR_SPLIT) |
| `-mg, --main-gpu INDEX` | the GPU to use for the model (with split-mode = none), or for intermediate results and KV (with split-mode = row) (default: 0)<br/>(env: LLAMA_ARG_MAIN_GPU) |
| `-fit, --fit [on\|off]` | whether to adjust unset arguments to fit in device memory ('on' or 'off', default: 'on')<br/>(env: LLAMA_ARG_FIT) |
| `-fitt, --fit-target MiB0,MiB1,MiB2,...` | target margin per device for --fit, comma-separated list of values, single value is broadcast across all devices, default: 1024<br/>(env: LLAMA_ARG_FIT_TARGET) |
| `-fitc, --fit-ctx N` | minimum ctx size that can be set by --fit option, default: 4096<br/>(env: LLAMA_ARG_FIT_CTX) |
| `--check-tensors` | check model tensor data for invalid values (default: false) |
| `--override-kv KEY=TYPE:VALUE,...` | advanced option to override model metadata by key. to specify multiple overrides, either use comma-separated values.<br/>types: int, float, bool, str. example: --override-kv tokenizer.ggml.add_bos_token=bool:false,tokenizer.ggml.add_eos_token=bool:false |
| `--op-offload, --no-op-offload` | whether to offload host tensor operations to device (default: true) |
| `--lora FNAME` | path to LoRA adapter (use comma-separated values to load multiple adapters) |
| `--lora-scaled FNAME:SCALE,...` | path to LoRA adapter with user defined scaling (format: FNAME:SCALE,...)<br/>note: use comma-separated values |
| `--control-vector FNAME` | add a control vector<br/>note: use comma-separated values to add multiple control vectors |
| `--control-vector-scaled FNAME:SCALE,...` | add a control vector with user defined scaling SCALE<br/>note: use comma-separated values (format: FNAME:SCALE,...) |
| `--control-vector-layer-range START END` | layer range to apply the control vector(s) to, start and end inclusive |
| `-m, --model FNAME` | model path to load<br/>(env: LLAMA_ARG_MODEL) |
| `-mu, --model-url MODEL_URL` | model download url (default: unused)<br/>(env: LLAMA_ARG_MODEL_URL) |
| `-dr, --docker-repo [<repo>/]<model>[:quant]` | Docker Hub model repository. repo is optional, default to ai/. quant is optional, default to :latest.<br/>example: gemma3<br/>(default: unused)<br/>(env: LLAMA_ARG_DOCKER_REPO) |
| `-hf, -hfr, --hf-repo <user>/<model>[:quant]` | Hugging Face model repository; quant is optional, case-insensitive, default to Q4_K_M, or falls back to the first file in the repo if Q4_K_M doesn't exist.<br/>mmproj is also downloaded automatically if available. to disable, add --no-mmproj<br/>example: unsloth/phi-4-GGUF:q4_k_m<br/>(default: unused)<br/>(env: LLAMA_ARG_HF_REPO) |
| `-hfd, -hfrd, --hf-repo-draft <user>/<model>[:quant]` | Same as --hf-repo, but for the draft model (default: unused)<br/>(env: LLAMA_ARG_HFD_REPO) |
| `-hff, --hf-file FILE` | Hugging Face model file. If specified, it will override the quant in --hf-repo (default: unused)<br/>(env: LLAMA_ARG_HF_FILE) |
| `-hfv, -hfrv, --hf-repo-v <user>/<model>[:quant]` | Hugging Face model repository for the vocoder model (default: unused)<br/>(env: LLAMA_ARG_HF_REPO_V) |
| `-hffv, --hf-file-v FILE` | Hugging Face model file for the vocoder model (default: unused)<br/>(env: LLAMA_ARG_HF_FILE_V) |
| `-hft, --hf-token TOKEN` | Hugging Face access token (default: value from HF_TOKEN environment variable)<br/>(env: HF_TOKEN) |
| `--log-disable` | Log disable |
| `--log-file FNAME` | Log to file<br/>(env: LLAMA_LOG_FILE) |
| `--log-colors [on\|off\|auto]` | Set colored logging ('on', 'off', or 'auto', default: 'auto')<br/>'auto' enables colors when output is to a terminal<br/>(env: LLAMA_LOG_COLORS) |
| `-v, --verbose, --log-verbose` | Set verbosity level to infinity (i.e. log all messages, useful for debugging) |
| `--offline` | Offline mode: forces use of cache, prevents network access<br/>(env: LLAMA_OFFLINE) |
| `-lv, --verbosity, --log-verbosity N` | Set the verbosity threshold. Messages with a higher verbosity will be ignored. Values:<br/> - 0: generic output<br/> - 1: error<br/> - 2: warning<br/> - 3: info<br/> - 4: debug<br/>(default: 3)<br/><br/>(env: LLAMA_LOG_VERBOSITY) |
| `--log-prefix` | Enable prefix in log messages<br/>(env: LLAMA_LOG_PREFIX) |
| `--log-timestamps` | Enable timestamps in log messages<br/>(env: LLAMA_LOG_TIMESTAMPS) |
| `-ctkd, --cache-type-k-draft TYPE` | KV cache data type for K for the draft model<br/>allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1<br/>(default: f16)<br/>(env: LLAMA_ARG_CACHE_TYPE_K_DRAFT) |
| `-ctvd, --cache-type-v-draft TYPE` | KV cache data type for V for the draft model<br/>allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1<br/>(default: f16)<br/>(env: LLAMA_ARG_CACHE_TYPE_V_DRAFT) |
### Sampling params
| Argument | Explanation |
| -------- | ----------- |
| `--samplers SAMPLERS` | samplers that will be used for generation in the order, separated by ';'<br/>(default: penalties;dry;top_n_sigma;top_k;typ_p;top_p;min_p;xtc;temperature) |
| `-s, --seed SEED` | RNG seed (default: -1, use random seed for -1) |
| `--sampler-seq, --sampling-seq SEQUENCE` | simplified sequence for samplers that will be used (default: edskypmxt) |
| `--ignore-eos` | ignore end of stream token and continue generating (implies --logit-bias EOS-inf) |
| `--temp N` | temperature (default: 0.8) |
| `--top-k N` | top-k sampling (default: 40, 0 = disabled)<br/>(env: LLAMA_ARG_TOP_K) |
| `--top-p N` | top-p sampling (default: 0.9, 1.0 = disabled) |
| `--min-p N` | min-p sampling (default: 0.1, 0.0 = disabled) |
| `--adaptive-target N` | adaptive-p: select tokens near this probability (valid range 0.0 to 1.0; negative = disabled) |
| `--adaptive-decay N` | adaptive-p: EMA decay for adaptation; effective history length ≈ 1/(1-decay) tokens (valid range 0.0 - 0.99) |
| `--top-nsigma N` | top-n-sigma sampling (default: -1.0, -1.0 = disabled) |
| `--xtc-probability N` | xtc probability (default: 0.0, 0.0 = disabled) |
| `--xtc-threshold N` | xtc threshold (default: 0.1, 1.0 = disabled) |
| `--typical N` | locally typical sampling, parameter p (default: 1.0, 1.0 = disabled) |
| `--repeat-last-n N` | last n tokens to consider for penalize (default: 64, 0 = disabled, -1 = ctx_size) |
| `--repeat-penalty N` | penalize repeat sequence of tokens (default: 1.0, 1.0 = disabled) |
| `--presence-penalty N` | repeat alpha presence penalty (default: 0.0, 0.0 = disabled) |
| `--frequency-penalty N` | repeat alpha frequency penalty (default: 0.0, 0.0 = disabled) |
| `--dry-multiplier N` | set DRY sampling multiplier (default: 0.0, 0.0 = disabled) |
| `--dry-base N` | set DRY sampling base value (default: 1.75) |
| `--dry-allowed-length N` | set allowed length for DRY sampling (default: 2) |
| `--dry-penalty-last-n N` | set DRY penalty for the last n tokens (default: -1, 0 = disable, -1 = context size) |
| `--dry-sequence-breaker STRING` | add sequence breaker for DRY sampling, clearing out default breakers ('\n', ':', '"', '*') in the process; use "none" to not use any sequence breakers |
| `--dynatemp-range N` | dynamic temperature range (default: 0.0, 0.0 = disabled) |
| `--dynatemp-exp N` | dynamic temperature exponent (default: 1.0) |
| `--mirostat N` | use Mirostat sampling.<br/>Top K, Nucleus and Locally Typical samplers are ignored if used.<br/>(default: 0, 0 = disabled, 1 = Mirostat, 2 = Mirostat 2.0) |
| `--mirostat-lr N` | Mirostat learning rate, parameter eta (default: 0.1) |
| `--mirostat-ent N` | Mirostat target entropy, parameter tau (default: 5.0) |
| `-l, --logit-bias TOKEN_ID(+/-)BIAS` | modifies the likelihood of token appearing in the completion,<br/>i.e. `--logit-bias 15043+1` to increase likelihood of token ' Hello',<br/>or `--logit-bias 15043-1` to decrease likelihood of token ' Hello' |
| `--grammar GRAMMAR` | BNF-like grammar to constrain generations (see samples in grammars/ dir) (default: '') |
| `--grammar-file FNAME` | file to read grammar from |
| `-j, --json-schema SCHEMA` | JSON schema to constrain generations (https://json-schema.org/), e.g. `{}` for any JSON object<br/>For schemas w/ external $refs, use --grammar + example/json_schema_to_grammar.py instead |
| `-jf, --json-schema-file FILE` | File containing a JSON schema to constrain generations (https://json-schema.org/), e.g. `{}` for any JSON object<br/>For schemas w/ external $refs, use --grammar + example/json_schema_to_grammar.py instead |
| `-bs, --backend-sampling` | enable backend sampling (experimental) (default: disabled)<br/>(env: LLAMA_ARG_BACKEND_SAMPLING) |
### CLI-specific params
| Argument | Explanation |
| -------- | ----------- |
| `--display-prompt, --no-display-prompt` | whether to print prompt at generation (default: true) |
| `-co, --color [on\|off\|auto]` | Colorize output to distinguish prompt and user input from generations ('on', 'off', or 'auto', default: 'auto')<br/>'auto' enables colors when output is to a terminal |
| `--ctx-checkpoints, --swa-checkpoints N` | max number of context checkpoints to create per slot (default: 8)[(more info)](https://github.com/ggml-org/llama.cpp/pull/15293)<br/>(env: LLAMA_ARG_CTX_CHECKPOINTS) |
| `-cram, --cache-ram N` | set the maximum cache size in MiB (default: 8192, -1 - no limit, 0 - disable)[(more info)](https://github.com/ggml-org/llama.cpp/pull/16391)<br/>(env: LLAMA_ARG_CACHE_RAM) |
| `--context-shift, --no-context-shift` | whether to use context shift on infinite text generation (default: disabled)<br/>(env: LLAMA_ARG_CONTEXT_SHIFT) |
| `-sys, --system-prompt PROMPT` | system prompt to use with model (if applicable, depending on chat template) |
| `--show-timings, --no-show-timings` | whether to show timing information after each response (default: true)<br/>(env: LLAMA_ARG_SHOW_TIMINGS) |
| `-sysf, --system-prompt-file FNAME` | a file containing the system prompt (default: none) |
| `-r, --reverse-prompt PROMPT` | halt generation at PROMPT, return control in interactive mode |
| `-sp, --special` | special tokens output enabled (default: false) |
| `-cnv, --conversation, -no-cnv, --no-conversation` | whether to run in conversation mode:<br/>- does not print special tokens and suffix/prefix<br/>- interactive mode is also enabled<br/>(default: auto enabled if chat template is available) |
| `-st, --single-turn` | run conversation for a single turn only, then exit when done<br/>will not be interactive if first turn is predefined with --prompt<br/>(default: false) |
| `-mli, --multiline-input` | allows you to write or paste multiple lines without ending each in '\' |
| `--warmup, --no-warmup` | whether to perform warmup with an empty run (default: enabled) |
| `-mm, --mmproj FILE` | path to a multimodal projector file. see tools/mtmd/README.md<br/>note: if -hf is used, this argument can be omitted<br/>(env: LLAMA_ARG_MMPROJ) |
| `-mmu, --mmproj-url URL` | URL to a multimodal projector file. see tools/mtmd/README.md<br/>(env: LLAMA_ARG_MMPROJ_URL) |
| `--mmproj-auto, --no-mmproj, --no-mmproj-auto` | whether to use multimodal projector file (if available), useful when using -hf (default: enabled)<br/>(env: LLAMA_ARG_MMPROJ_AUTO) |
| `--mmproj-offload, --no-mmproj-offload` | whether to enable GPU offloading for multimodal projector (default: enabled)<br/>(env: LLAMA_ARG_MMPROJ_OFFLOAD) |
| `--image, --audio FILE` | path to an image or audio file. use with multimodal models, use comma-separated values for multiple files |
| `--image-min-tokens N` | minimum number of tokens each image can take, only used by vision models with dynamic resolution (default: read from model)<br/>(env: LLAMA_ARG_IMAGE_MIN_TOKENS) |
| `--image-max-tokens N` | maximum number of tokens each image can take, only used by vision models with dynamic resolution (default: read from model)<br/>(env: LLAMA_ARG_IMAGE_MAX_TOKENS) |
| `-otd, --override-tensor-draft <tensor name pattern>=<buffer type>,...` | override tensor buffer type for draft model |
| `-cmoed, --cpu-moe-draft` | keep all Mixture of Experts (MoE) weights in the CPU for the draft model<br/>(env: LLAMA_ARG_CPU_MOE_DRAFT) |
| `-ncmoed, --n-cpu-moe-draft N` | keep the Mixture of Experts (MoE) weights of the first N layers in the CPU for the draft model<br/>(env: LLAMA_ARG_N_CPU_MOE_DRAFT) |
| `--chat-template-kwargs STRING` | sets additional params for the json template parser, must be a valid json object string, e.g. '{"key1":"value1","key2":"value2"}'<br/>(env: LLAMA_CHAT_TEMPLATE_KWARGS) |
| `--jinja, --no-jinja` | whether to use jinja template engine for chat (default: enabled)<br/>(env: LLAMA_ARG_JINJA) |
| `--reasoning-format FORMAT` | controls whether thought tags are allowed and/or extracted from the response, and in which format they're returned; one of:<br/>- none: leaves thoughts unparsed in `message.content`<br/>- deepseek: puts thoughts in `message.reasoning_content`<br/>- deepseek-legacy: keeps `<think>` tags in `message.content` while also populating `message.reasoning_content`<br/>(default: auto)<br/>(env: LLAMA_ARG_THINK) |
| `--reasoning-budget N` | controls the amount of thinking allowed; currently only one of: -1 for unrestricted thinking budget, or 0 to disable thinking (default: -1)<br/>(env: LLAMA_ARG_THINK_BUDGET) |
| `--chat-template JINJA_TEMPLATE` | set custom jinja chat template (default: template taken from model's metadata)<br/>if suffix/prefix are specified, template will be disabled<br/>only commonly used templates are accepted (unless --jinja is set before this flag):<br/>list of built-in templates:<br/>bailing, bailing-think, bailing2, chatglm3, chatglm4, chatml, command-r, deepseek, deepseek2, deepseek3, exaone3, exaone4, falcon3, gemma, gigachat, glmedge, gpt-oss, granite, grok-2, hunyuan-dense, hunyuan-moe, kimi-k2, llama2, llama2-sys, llama2-sys-bos, llama2-sys-strip, llama3, llama4, megrez, minicpm, mistral-v1, mistral-v3, mistral-v3-tekken, mistral-v7, mistral-v7-tekken, monarch, openchat, orion, pangu-embedded, phi3, phi4, rwkv-world, seed_oss, smolvlm, solar-open, vicuna, vicuna-orca, yandex, zephyr<br/>(env: LLAMA_ARG_CHAT_TEMPLATE) |
| `--chat-template-file JINJA_TEMPLATE_FILE` | set custom jinja chat template file (default: template taken from model's metadata)<br/>if suffix/prefix are specified, template will be disabled<br/>only commonly used templates are accepted (unless --jinja is set before this flag):<br/>list of built-in templates:<br/>bailing, bailing-think, bailing2, chatglm3, chatglm4, chatml, command-r, deepseek, deepseek2, deepseek3, exaone3, exaone4, falcon3, gemma, gigachat, glmedge, gpt-oss, granite, grok-2, hunyuan-dense, hunyuan-moe, kimi-k2, llama2, llama2-sys, llama2-sys-bos, llama2-sys-strip, llama3, llama4, megrez, minicpm, mistral-v1, mistral-v3, mistral-v3-tekken, mistral-v7, mistral-v7-tekken, monarch, openchat, orion, pangu-embedded, phi3, phi4, rwkv-world, seed_oss, smolvlm, solar-open, vicuna, vicuna-orca, yandex, zephyr<br/>(env: LLAMA_ARG_CHAT_TEMPLATE_FILE) |
| `--simple-io` | use basic IO for better compatibility in subprocesses and limited consoles |
| `--draft, --draft-n, --draft-max N` | number of tokens to draft for speculative decoding (default: 16)<br/>(env: LLAMA_ARG_DRAFT_MAX) |
| `--draft-min, --draft-n-min N` | minimum number of draft tokens to use for speculative decoding (default: 0)<br/>(env: LLAMA_ARG_DRAFT_MIN) |
| `--draft-p-min P` | minimum speculative decoding probability (greedy) (default: 0.8)<br/>(env: LLAMA_ARG_DRAFT_P_MIN) |
| `-cd, --ctx-size-draft N` | size of the prompt context for the draft model (default: 0, 0 = loaded from model)<br/>(env: LLAMA_ARG_CTX_SIZE_DRAFT) |
| `-devd, --device-draft <dev1,dev2,..>` | comma-separated list of devices to use for offloading the draft model (none = don't offload)<br/>use --list-devices to see a list of available devices |
| `-ngld, --gpu-layers-draft, --n-gpu-layers-draft N` | max. number of draft model layers to store in VRAM, either an exact number, 'auto', or 'all' (default: auto)<br/>(env: LLAMA_ARG_N_GPU_LAYERS_DRAFT) |
| `-md, --model-draft FNAME` | draft model for speculative decoding (default: unused)<br/>(env: LLAMA_ARG_MODEL_DRAFT) |
| `--spec-replace TARGET DRAFT` | translate the string in TARGET into DRAFT if the draft model and main model are not compatible |
| `--gpt-oss-20b-default` | use gpt-oss-20b (note: can download weights from the internet) |
| `--gpt-oss-120b-default` | use gpt-oss-120b (note: can download weights from the internet) |
| `--vision-gemma-4b-default` | use Gemma 3 4B QAT (note: can download weights from the internet) |
| `--vision-gemma-12b-default` | use Gemma 3 12B QAT (note: can download weights from the internet) |
<!-- HELP_END -->

393
tools/cli/cli.cpp Normal file
View File

@@ -0,0 +1,393 @@
#include "common.h"
#include "arg.h"
#include "console.h"
// #include "log.h"
#include "server-context.h"
#include "server-task.h"
#include <atomic>
#include <fstream>
#include <thread>
#include <signal.h>
#if defined(_WIN32)
#define WIN32_LEAN_AND_MEAN
#ifndef NOMINMAX
# define NOMINMAX
#endif
#include <windows.h>
#endif
const char * LLAMA_ASCII_LOGO = R"(
▄▄ ▄▄
██ ██
██ ██ ▀▀█▄ ███▄███▄ ▀▀█▄ ▄████ ████▄ ████▄
██ ██ ▄█▀██ ██ ██ ██ ▄█▀██ ██ ██ ██ ██ ██
██ ██ ▀█▄██ ██ ██ ██ ▀█▄██ ██ ▀████ ████▀ ████▀
██ ██
▀▀ ▀▀
)";
static std::atomic<bool> g_is_interrupted = false;
static bool should_stop() {
return g_is_interrupted.load();
}
#if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__)) || defined (_WIN32)
static void signal_handler(int) {
if (g_is_interrupted.load()) {
// second Ctrl+C - exit immediately
// make sure to clear colors before exiting (not using LOG or console.cpp here to avoid deadlock)
fprintf(stdout, "\033[0m\n");
fflush(stdout);
std::exit(130);
}
g_is_interrupted.store(true);
}
#endif
struct cli_context {
server_context ctx_server;
json messages = json::array();
std::vector<raw_buffer> input_files;
task_params defaults;
// thread for showing "loading" animation
std::atomic<bool> loading_show;
cli_context(const common_params & params) {
defaults.sampling = params.sampling;
defaults.speculative = params.speculative;
defaults.n_keep = params.n_keep;
defaults.n_predict = params.n_predict;
defaults.antiprompt = params.antiprompt;
defaults.stream = true; // make sure we always use streaming mode
defaults.timings_per_token = true; // in order to get timings even when we cancel mid-way
// defaults.return_progress = true; // TODO: show progress
defaults.oaicompat_chat_syntax.reasoning_format = COMMON_REASONING_FORMAT_DEEPSEEK;
}
std::string generate_completion(result_timings & out_timings) {
server_response_reader rd = ctx_server.get_response_reader();
{
// TODO: reduce some copies here in the future
server_task task = server_task(SERVER_TASK_TYPE_COMPLETION);
task.id = rd.get_new_id();
task.index = 0;
task.params = defaults; // copy
task.cli_input = messages; // copy
task.cli_files = input_files; // copy
rd.post_task({std::move(task)});
}
// wait for first result
console::spinner::start();
server_task_result_ptr result = rd.next(should_stop);
console::spinner::stop();
std::string curr_content;
bool is_thinking = false;
while (result) {
if (should_stop()) {
break;
}
if (result->is_error()) {
json err_data = result->to_json();
if (err_data.contains("message")) {
console::error("Error: %s\n", err_data["message"].get<std::string>().c_str());
} else {
console::error("Error: %s\n", err_data.dump().c_str());
}
return curr_content;
}
auto res_partial = dynamic_cast<server_task_result_cmpl_partial *>(result.get());
if (res_partial) {
out_timings = std::move(res_partial->timings);
for (const auto & diff : res_partial->oaicompat_msg_diffs) {
if (!diff.content_delta.empty()) {
if (is_thinking) {
console::log("\n[End thinking]\n\n");
console::set_display(DISPLAY_TYPE_RESET);
is_thinking = false;
}
curr_content += diff.content_delta;
console::log("%s", diff.content_delta.c_str());
console::flush();
}
if (!diff.reasoning_content_delta.empty()) {
console::set_display(DISPLAY_TYPE_REASONING);
if (!is_thinking) {
console::log("[Start thinking]\n");
}
is_thinking = true;
console::log("%s", diff.reasoning_content_delta.c_str());
console::flush();
}
}
}
auto res_final = dynamic_cast<server_task_result_cmpl_final *>(result.get());
if (res_final) {
out_timings = std::move(res_final->timings);
break;
}
result = rd.next(should_stop);
}
g_is_interrupted.store(false);
// server_response_reader automatically cancels pending tasks upon destruction
return curr_content;
}
// TODO: support remote files in the future (http, https, etc)
std::string load_input_file(const std::string & fname, bool is_media) {
std::ifstream file(fname, std::ios::binary);
if (!file) {
return "";
}
if (is_media) {
raw_buffer buf;
buf.assign((std::istreambuf_iterator<char>(file)), std::istreambuf_iterator<char>());
input_files.push_back(std::move(buf));
return mtmd_default_marker();
} else {
std::string content((std::istreambuf_iterator<char>(file)), std::istreambuf_iterator<char>());
return content;
}
}
};
int main(int argc, char ** argv) {
common_params params;
params.verbosity = LOG_LEVEL_ERROR; // by default, less verbose logs
if (!common_params_parse(argc, argv, params, LLAMA_EXAMPLE_CLI)) {
return 1;
}
// TODO: maybe support it later?
if (params.conversation_mode == COMMON_CONVERSATION_MODE_DISABLED) {
console::error("--no-conversation is not supported by llama-cli\n");
console::error("please use llama-completion instead\n");
}
common_init();
// struct that contains llama context and inference
cli_context ctx_cli(params);
llama_backend_init();
llama_numa_init(params.numa);
// TODO: avoid using atexit() here by making `console` a singleton
console::init(params.simple_io, params.use_color);
atexit([]() { console::cleanup(); });
console::set_display(DISPLAY_TYPE_RESET);
#if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__))
struct sigaction sigint_action;
sigint_action.sa_handler = signal_handler;
sigemptyset (&sigint_action.sa_mask);
sigint_action.sa_flags = 0;
sigaction(SIGINT, &sigint_action, NULL);
sigaction(SIGTERM, &sigint_action, NULL);
#elif defined (_WIN32)
auto console_ctrl_handler = +[](DWORD ctrl_type) -> BOOL {
return (ctrl_type == CTRL_C_EVENT) ? (signal_handler(SIGINT), true) : false;
};
SetConsoleCtrlHandler(reinterpret_cast<PHANDLER_ROUTINE>(console_ctrl_handler), true);
#endif
console::log("\nLoading model... "); // followed by loading animation
console::spinner::start();
if (!ctx_cli.ctx_server.load_model(params)) {
console::spinner::stop();
console::error("\nFailed to load the model\n");
return 1;
}
console::spinner::stop();
console::log("\n");
std::thread inference_thread([&ctx_cli]() {
ctx_cli.ctx_server.start_loop();
});
auto inf = ctx_cli.ctx_server.get_meta();
std::string modalities = "text";
if (inf.has_inp_image) {
modalities += ", vision";
}
if (inf.has_inp_audio) {
modalities += ", audio";
}
if (!params.system_prompt.empty()) {
ctx_cli.messages.push_back({
{"role", "system"},
{"content", params.system_prompt}
});
}
console::log("\n");
console::log("%s\n", LLAMA_ASCII_LOGO);
console::log("build : %s\n", inf.build_info.c_str());
console::log("model : %s\n", inf.model_name.c_str());
console::log("modalities : %s\n", modalities.c_str());
if (!params.system_prompt.empty()) {
console::log("using custom system prompt\n");
}
console::log("\n");
console::log("available commands:\n");
console::log(" /exit or Ctrl+C stop or exit\n");
console::log(" /regen regenerate the last response\n");
console::log(" /clear clear the chat history\n");
console::log(" /read add a text file\n");
if (inf.has_inp_image) {
console::log(" /image <file> add an image file\n");
}
if (inf.has_inp_audio) {
console::log(" /audio <file> add an audio file\n");
}
console::log("\n");
// interactive loop
std::string cur_msg;
while (true) {
std::string buffer;
console::set_display(DISPLAY_TYPE_USER_INPUT);
if (params.prompt.empty()) {
console::log("\n> ");
std::string line;
bool another_line = true;
do {
another_line = console::readline(line, params.multiline_input);
buffer += line;
} while (another_line);
} else {
// process input prompt from args
for (auto & fname : params.image) {
std::string marker = ctx_cli.load_input_file(fname, true);
if (marker.empty()) {
console::error("file does not exist or cannot be opened: '%s'\n", fname.c_str());
break;
}
console::log("Loaded media from '%s'\n", fname.c_str());
cur_msg += marker;
}
buffer = params.prompt;
if (buffer.size() > 500) {
console::log("\n> %s ... (truncated)\n", buffer.substr(0, 500).c_str());
} else {
console::log("\n> %s\n", buffer.c_str());
}
params.prompt.clear(); // only use it once
}
console::set_display(DISPLAY_TYPE_RESET);
console::log("\n");
if (should_stop()) {
g_is_interrupted.store(false);
break;
}
// remove trailing newline
if (!buffer.empty() &&buffer.back() == '\n') {
buffer.pop_back();
}
// skip empty messages
if (buffer.empty()) {
continue;
}
bool add_user_msg = true;
// process commands
if (string_starts_with(buffer, "/exit")) {
break;
} else if (string_starts_with(buffer, "/regen")) {
if (ctx_cli.messages.size() >= 2) {
size_t last_idx = ctx_cli.messages.size() - 1;
ctx_cli.messages.erase(last_idx);
add_user_msg = false;
} else {
console::error("No message to regenerate.\n");
continue;
}
} else if (string_starts_with(buffer, "/clear")) {
ctx_cli.messages.clear();
ctx_cli.input_files.clear();
console::log("Chat history cleared.\n");
continue;
} else if (
(string_starts_with(buffer, "/image ") && inf.has_inp_image) ||
(string_starts_with(buffer, "/audio ") && inf.has_inp_audio)) {
// just in case (bad copy-paste for example), we strip all trailing/leading spaces
std::string fname = string_strip(buffer.substr(7));
std::string marker = ctx_cli.load_input_file(fname, true);
if (marker.empty()) {
console::error("file does not exist or cannot be opened: '%s'\n", fname.c_str());
continue;
}
cur_msg += marker;
console::log("Loaded media from '%s'\n", fname.c_str());
continue;
} else if (string_starts_with(buffer, "/read ")) {
std::string fname = string_strip(buffer.substr(6));
std::string marker = ctx_cli.load_input_file(fname, false);
if (marker.empty()) {
console::error("file does not exist or cannot be opened: '%s'\n", fname.c_str());
continue;
}
cur_msg += marker;
console::log("Loaded text from '%s'\n", fname.c_str());
continue;
} else {
// not a command
cur_msg += buffer;
}
// generate response
if (add_user_msg) {
ctx_cli.messages.push_back({
{"role", "user"},
{"content", cur_msg}
});
cur_msg.clear();
}
result_timings timings;
std::string assistant_content = ctx_cli.generate_completion(timings);
ctx_cli.messages.push_back({
{"role", "assistant"},
{"content", assistant_content}
});
console::log("\n");
if (params.show_timings) {
console::set_display(DISPLAY_TYPE_INFO);
console::log("\n");
console::log("[ Prompt: %.1f t/s | Generation: %.1f t/s ]\n", timings.prompt_per_second, timings.predicted_per_second);
console::set_display(DISPLAY_TYPE_RESET);
}
if (params.single_turn) {
break;
}
}
console::set_display(DISPLAY_TYPE_RESET);
console::log("\nExiting...\n");
ctx_cli.ctx_server.terminate();
inference_thread.join();
// bump the log level to display timings
common_log_set_verbosity_thold(LOG_LEVEL_INFO);
llama_memory_breakdown_print(ctx_cli.ctx_server.get_llama_context());
return 0;
}