Skip to content

common : improve -ctv -ctk CLI arguments #10806

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Dec 12, 2024
Merged

common : improve -ctv -ctk CLI arguments #10806

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
ba35f29
common : improve ctv ctk cli argument
ngxson Dec 12, 2024
6b36960
regenerate docs
ngxson Dec 12, 2024
42a9c5a
even better approach
ngxson Dec 12, 2024
3fec232
use std::vector
ngxson Dec 12, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
51 changes: 45 additions & 6 deletions common/arg.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -145,6 +145,35 @@ static void common_params_handle_model_default(common_params & params) {
}
}

const std::vector<ggml_type> kv_cache_types = {
GGML_TYPE_F32,
GGML_TYPE_F16,
GGML_TYPE_BF16,
GGML_TYPE_Q8_0,
GGML_TYPE_Q4_0,
GGML_TYPE_Q4_1,
GGML_TYPE_IQ4_NL,
GGML_TYPE_Q5_0,
GGML_TYPE_Q5_1,
};

static ggml_type kv_cache_type_from_str(const std::string & s) {
for (const auto & type : kv_cache_types) {
if (ggml_type_name(type) == s) {
return type;
}
}
throw std::runtime_error("Unsupported cache type: " + s);
}

static std::string get_all_kv_cache_types() {
std::ostringstream msg;
for (const auto & type : kv_cache_types) {
msg << ggml_type_name(type) << (&type == &kv_cache_types.back() ? "" : ", ");
}
return msg.str();
}

//
// CLI argument parsing functions
//
Expand Down Expand Up @@ -1174,18 +1203,28 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
).set_env("LLAMA_ARG_NO_KV_OFFLOAD"));
add_opt(common_arg(
{"-ctk", "--cache-type-k"}, "TYPE",
string_format("KV cache data type for K (default: %s)", params.cache_type_k.c_str()),
string_format(
"KV cache data type for K\n"
"allowed values: %s\n"
"(default: %s)",
get_all_kv_cache_types().c_str(),
ggml_type_name(params.cache_type_k)
),
[](common_params & params, const std::string & value) {
// TODO: get the type right here
params.cache_type_k = value;
params.cache_type_k = kv_cache_type_from_str(value);
}
).set_env("LLAMA_ARG_CACHE_TYPE_K"));
add_opt(common_arg(
{"-ctv", "--cache-type-v"}, "TYPE",
string_format("KV cache data type for V (default: %s)", params.cache_type_v.c_str()),
string_format(
"KV cache data type for V\n"
"allowed values: %s\n"
"(default: %s)",
get_all_kv_cache_types().c_str(),
ggml_type_name(params.cache_type_v)
),
[](common_params & params, const std::string & value) {
// TODO: get the type right here
params.cache_type_v = value;
params.cache_type_v = kv_cache_type_from_str(value);
}
).set_env("LLAMA_ARG_CACHE_TYPE_V"));
add_opt(common_arg(
Expand Down
36 changes: 2 additions & 34 deletions common/common.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1015,38 +1015,6 @@ struct llama_model_params common_model_params_to_llama(common_params & params) {
return mparams;
}

static ggml_type kv_cache_type_from_str(const std::string & s) {
if (s == "f32") {
return GGML_TYPE_F32;
}
if (s == "f16") {
return GGML_TYPE_F16;
}
if (s == "bf16") {
return GGML_TYPE_BF16;
}
if (s == "q8_0") {
return GGML_TYPE_Q8_0;
}
if (s == "q4_0") {
return GGML_TYPE_Q4_0;
}
if (s == "q4_1") {
return GGML_TYPE_Q4_1;
}
if (s == "iq4_nl") {
return GGML_TYPE_IQ4_NL;
}
if (s == "q5_0") {
return GGML_TYPE_Q5_0;
}
if (s == "q5_1") {
return GGML_TYPE_Q5_1;
}

throw std::runtime_error("Unsupported cache type: " + s);
}

struct llama_context_params common_context_params_to_llama(const common_params & params) {
auto cparams = llama_context_default_params();

Expand Down Expand Up @@ -1081,8 +1049,8 @@ struct llama_context_params common_context_params_to_llama(const common_params &
cparams.pooling_type = LLAMA_POOLING_TYPE_RANK;
}

cparams.type_k = kv_cache_type_from_str(params.cache_type_k);
cparams.type_v = kv_cache_type_from_str(params.cache_type_v);
cparams.type_k = params.cache_type_k;
cparams.type_v = params.cache_type_v;

return cparams;
}
Expand Down
4 changes: 2 additions & 2 deletions common/common.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -286,8 +286,8 @@ struct common_params {
bool warmup = true; // warmup run
bool check_tensors = false; // validate tensor data

std::string cache_type_k = "f16"; // KV cache data type for the K
std::string cache_type_v = "f16"; // KV cache data type for the V
ggml_type cache_type_k = GGML_TYPE_F16; // KV cache data type for the K
ggml_type cache_type_v = GGML_TYPE_F16; // KV cache data type for the V

// multimodal models (see examples/llava)
std::string mmproj = ""; // path to multimodal projector // NOLINT
Expand Down
1 change: 1 addition & 0 deletions examples/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,6 +51,7 @@ else()
add_subdirectory(speculative)
add_subdirectory(speculative-simple)
add_subdirectory(tokenize)
add_subdirectory(gen-docs)
if (NOT GGML_BACKEND_DL)
# these examples use the backends directly and cannot be built with dynamic loading
add_subdirectory(convert-llama2c-to-ggml)
Expand Down
19 changes: 10 additions & 9 deletions examples/server/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -62,8 +62,8 @@ The project is under active development, and we are [looking for feedback and co
| `--yarn-beta-fast N` | YaRN: low correction dim or beta (default: 32.0)<br/>(env: LLAMA_ARG_YARN_BETA_FAST) |
| `-dkvc, --dump-kv-cache` | verbose print of the KV cache |
| `-nkvo, --no-kv-offload` | disable KV offload<br/>(env: LLAMA_ARG_NO_KV_OFFLOAD) |
| `-ctk, --cache-type-k TYPE` | KV cache data type for K (default: f16)<br/>(env: LLAMA_ARG_CACHE_TYPE_K) |
| `-ctv, --cache-type-v TYPE` | KV cache data type for V (default: f16)<br/>(env: LLAMA_ARG_CACHE_TYPE_V) |
| `-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 (default: 0.1, < 0 - disabled)<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) |
Expand Down Expand Up @@ -138,6 +138,7 @@ The project is under active development, and we are [looking for feedback and co
| -------- | ----------- |
| `--no-context-shift` | disables context shift on inifinite text generation (default: disabled)<br/>(env: LLAMA_ARG_NO_CONTEXT_SHIFT) |
| `-sp, --special` | special tokens output enabled (default: false) |
| `--no-warmup` | skip warming up the model with an empty run |
| `--spm-infill` | use Suffix/Prefix/Middle pattern for infill (instead of Prefix/Suffix/Middle) as some models prefer this. (default: disabled) |
| `--pooling {none,mean,cls,last,rank}` | pooling type for embeddings, use model default if unspecified<br/>(env: LLAMA_ARG_POOLING) |
| `-cb, --cont-batching` | enable continuous batching (a.k.a dynamic batching) (default: enabled)<br/>(env: LLAMA_ARG_CONT_BATCHING) |
Expand All @@ -146,7 +147,7 @@ The project is under active development, and we are [looking for feedback and co
| `--host HOST` | ip address to listen (default: 127.0.0.1)<br/>(env: LLAMA_ARG_HOST) |
| `--port PORT` | port to listen (default: 8080)<br/>(env: LLAMA_ARG_PORT) |
| `--path PATH` | path to serve static files from (default: )<br/>(env: LLAMA_ARG_STATIC_PATH) |
| `--no-webui` | disable the Web UI<br/>(env: LLAMA_ARG_NO_WEBUI) |
| `--no-webui` | Disable the Web UI (default: enabled)<br/>(env: LLAMA_ARG_NO_WEBUI) |
| `--embedding, --embeddings` | restrict to only support embedding use case; use only with dedicated embedding models (default: disabled)<br/>(env: LLAMA_ARG_EMBEDDINGS) |
| `--reranking, --rerank` | enable reranking endpoint on server (default: disabled)<br/>(env: LLAMA_ARG_RERANKING) |
| `--api-key KEY` | API key to use for authentication (default: none)<br/>(env: LLAMA_API_KEY) |
Expand All @@ -164,13 +165,13 @@ The project is under active development, and we are [looking for feedback and co
| `--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/>list of built-in templates:<br/>chatglm3, chatglm4, chatml, command-r, deepseek, deepseek2, exaone3, gemma, granite, llama2, llama2-sys, llama2-sys-bos, llama2-sys-strip, llama3, minicpm, mistral-v1, mistral-v3, mistral-v3-tekken, mistral-v7, monarch, openchat, orion, phi3, rwkv-world, vicuna, vicuna-orca, zephyr<br/>(env: LLAMA_ARG_CHAT_TEMPLATE) |
| `-sps, --slot-prompt-similarity SIMILARITY` | how much the prompt of a request must match the prompt of a slot in order to use that slot (default: 0.50, 0.0 = disabled)<br/> |
| `--lora-init-without-apply` | load LoRA adapters without applying them (apply later via POST /lora-adapters) (default: disabled) |
| `--draft-max, --draft, --draft-n N` | number of tokens to draft for speculative decoding (default: 16) |
| `--draft-min, --draft-n-min N` | minimum number of draft tokens to use for speculative decoding (default: 5) |
| `--draft-p-min P` | minimum speculative decoding probability (greedy) (default: 0.9) |
| `-cd, --ctx-size-draft N` | size of the prompt context for the draft model (default: 0, 0 = loaded from model) |
| `--draft-max, --draft, --draft-n 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: 5)<br/>(env: LLAMA_ARG_DRAFT_MIN) |
| `--draft-p-min P` | minimum speculative decoding probability (greedy) (default: 0.9)<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` | number of layers to store in VRAM for the draft model |
| `-md, --model-draft FNAME` | draft model for speculative decoding (default: unused) |
| `-ngld, --gpu-layers-draft, --n-gpu-layers-draft N` | number of layers to store in VRAM for the draft model<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) |


Note: If both command line argument and environment variable are both set for the same param, the argument will take precedence over env var.
Expand Down
Loading