mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-06-26 06:10:19 +00:00
5ce013cd7e9cd3397150795c57ac477758691fcb
73 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 Galunid
|
5ce013cd7e | common : Disable backend sampling if reasoning budget is enabled (#21209) | ||
|
Aldehir Rojas
|
59d840209a |
common : inhibit lazy grammar sampler while reasoning is active (#20970)
* common : inhibit grammar while reasoning budget is active * cont : update force_pos in accept * cont : fix tests * cont : tweak should apply logic * cont : return early not using grammar sampler * Add tests * cont : prevent backend sampling when reasoning budget enabled * cont : fix typo --------- Co-authored-by: Piotr Wilkin <piotr.wilkin@syndatis.com> |
||
|
Piotr Wilkin (ilintar)
|
5e54d51b19 |
common/parser: add proper reasoning tag prefill reading (#20424)
* Implement proper prefill extraction * Refactor cli parameters, update docs, move reasoning budget sampler part to common/reasoning-budget.cpp * Update tools/server/server-task.cpp * refactor: move grammars to variant, remove grammar_external, handle exception internally * Make code less C++y Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> |
||
|
Piotr Wilkin (ilintar)
|
acb7c79069 |
common/parser: handle reasoning budget (#20297)
* v1 * Finished! * Handlie cli * Reasoning sampler * Apply suggestions from code review Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com> * Less explosive terminology :) * Add utf-8 case and tests * common : migrate reasoning budget sampler to common * cont : clean up * cont : expose state and allow passing as initial state * cont : remove unused imports * cont : update state machine doc string --------- Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com> Co-authored-by: Alde Rojas <hello@alde.dev> |
||
|
ddh0
|
13f1e4a9ca |
llama : add adaptive-p sampler (#17927)
* initial commit for branch * simplify constants * add params to `struct common_params_sampling`, add reference to PR * explicitly clamp `min_target` and `max_target` to `[0.0, 1.0]` * add args, rename `queue_size` -> `window_size` * improved comments * minor * remove old unused code from algorithm * minor * add power law case to `common_sampler_init`, add sampler name mappings * clarify behaviour when `window_size = 0` * add missing enums * remove `target_range` param, make `target == 1` no-op, cleanup code * oops, straggler * add missing parameters in `server-task.cpp` * copy from author ref: https://gist.github.com/MrJackSpade/9be99c7efbba7b95a41377e123b7b069 * remove old debug log, style nit * fix compiler warning, add commented-out logging per token * re-write + change parameters + simplify * oops forgot args.cpp * fix leftover `window_size` * add missing values to `common_params_sampling::print()` * with logging * does this fix it? * no, but does this? * update default decay * optimize * fix bad merge my git skills are lacking * silence `missing initializer for member` * update default decay to 0.9 * fix logging * format (double) * add power law to the new `samplers` vector * log sampler init values * improve logging messages in llama_sampler_power_law * remove extraneous logging * simplify target computation last commit with debug logging! * remove debug logging, explicitly clamp params at init * add `use_power_law` flag + logic, minor cleanup * update `power-law` -> `adaptive-p` * fix cold start EMA - `ctx->weighted_sum` is now initialized and reset to `target / (1.0f - clamped_decay)` - `ctx->total_weight` is now initialized and reset to `1.0f / (1.0f - clamped_decay)` this fixes a "cold start" problem with the moving average * update `SHARPNESS` constant to `10.0f` * minor style fixes no functional changes * minor style fixes cont. * update `llama_sampler_adaptive_p_i` for backend sampling (ref: #17004) * separate into `apply` + `accept` functions * `pending_token_idx`: switch from `llama_token` to `int32` functionally identical (`llama.h` has `typedef int32_t llama_token;`), but its more correct now * don't transform logits <= -1e9f * fix masking in backend top-p, min-p * address review comments * typo in comments `RND` -> `RNG` * add docs * add recommended values in completion docs * address PR feedback * remove trailing whitespace (for CI `editorconfig`) * add to adaptive-p to `common_sampler_types_from_chars` |
||
|
Georgi Gerganov
|
39173bcacb |
context : reserve new scheduler when graph topology changes (#18547)
* context : reserve new scheduler when graph topology changes * cont : fix * cont : fix reserve * cont : reserve only when changes occur + timing * context : add comments * llama : reserve on sampler changes * common : allow null common_sampler * server : task declares needs (embd, logits, sampling) * server : do not init sampler if not needed * llama : fix need_reserve when unsetting a sampler * server : consolidate slot reset/clear logic |
||
|
Daniel Bevenius
|
d3dce4e0a5 |
sampling : add support for backend sampling (#17004)
* sampling : add support for backend sampling This commit adds support for performing sampling operations on the backend (e.g. GPU) as part of the model computation graph. The motivation for this feature is to enable sampling to be performed directly on the backend as part of the computation graph being executed, allowing for some or all of the sampling to be done on the backend. For example, the backend sampler chain might select/sample a token directly in which case only the sampled token needs to be transferred from device memory to host memory. It is also possible for the backend samplers to perform filtering of the logits, or compute and filter the probability distribution, in which case only the filtered logits or probabilites need to be transferred back to system memory for further processing by CPU samplers. Currently the backend sampling works in a similar manner to how pooling works, it is a function that is called by build_graph and the sampler operations become part of the models computation graph. * llama-cli : add backend sampler configuration * server : add backend sampling options/configuration * webui : add backend sampling options * ggml : add initial cumsum implementation for CUDA * sampling : enable all backend sampler tests This commit enables all exisiting backend sampler tests in the test-backend-sampler. Previously, some tests were disabled because there were missing ggml operation implementations. * graph : do not include llama-model.h * sampling : always expose sampled_ids This commit precomputes and caches the full-vocab token id list in llama_context's constructor, so llama_get_backend_sampled_token_ids_ith always returns a valid pointer. The motivation for this is that this enables both common/sampling.cpp and src/llama-sampling.cpp can simplify their logic. Not all backends samplers that process logits need to set the sampled_tokens_id as they may not change the order of the logits, for example the temperature sampler only scales the logits but does not change their order. Simliar the logit bias sampler only adds bias to specific token ids but does not change the order of the logits. In these cases there will not be a device to host copy of the sampled token ids, and this is the use case where having this precomputed list is useful. * sampling : ensure at most one output token per seq This commit adds a check in the batch allocator to ensure that when backend sampling is enabled, at most one output token is specified per sequence. * CUDA: Optimize argsort for gpu-based token sampling Argsort is used for top-k currently. WE optimize argsort by 2 things: 1. Use `DeviceRadixSort` for single-row/sequence to parallelize it across our SMs 2. Use `DeviceSegmentedSort` for multi-row/sequence as this is the correct entrypoint (the function chooses different execution paths, it contains `DeviceSegmentedRadixSort` as one of the paths and will choose the best one according to heuristics. https://nvidia.github.io/cccl/cub/api/structcub_1_1DeviceSegmentedSort.html#overview Some perf numbers for a RTX PRO 6000: On the kernel level, tested with `GGML_CUDA_DISABLE_GRAPHS=1 ./test-backend-ops -o ARGSORT perf` Before: ``` ARGSORT(type=f32,ne=[65000,16,1,1],order=0): 4130 runs - 359.24 us/run ARGSORT(type=f32,ne=[200000,1,1,1],order=0): 8192 runs - 861.34 us/run ARGSORT(type=f32,ne=[200000,16,1,1],order=0): 1343 runs - 1020.01 us/run ``` After: ``` ARGSORT(type=f32,ne=[65000,16,1,1],order=0): 4130 runs - 312.41 us/run ARGSORT(type=f32,ne=[200000,1,1,1],order=0): 16384 runs - 63.48 us/run ARGSORT(type=f32,ne=[200000,16,1,1],order=0): 1343 runs - 874.36 us/run ``` --- On the model level, tested with `llama-cli -m gpt-oss-20b-mxfp4.gguf -n 200 -p "What is the Capital of Sweden?" -no-cnv -fa 1 --backend-sampling` Before: ``` llama_perf_sampler_print: sampling time = 0.25 ms / 207 runs ( 0.00 ms per token, 824701.20 tokens per second) llama_perf_context_print: load time = 18215.58 ms llama_perf_context_print: prompt eval time = 28.20 ms / 7 tokens ( 4.03 ms per token, 248.19 tokens per second) llama_perf_context_print: eval time = 714.79 ms / 199 runs ( 3.59 ms per token, 278.40 tokens per second) llama_perf_context_print: total time = 857.62 ms / 206 tokens ``` After ``` llama_perf_sampler_print: sampling time = 0.25 ms / 207 runs ( 0.00 ms per token, 828000.00 tokens per second) llama_perf_context_print: load time = 18366.92 ms llama_perf_context_print: prompt eval time = 35.92 ms / 7 tokens ( 5.13 ms per token, 194.87 tokens per second) llama_perf_context_print: eval time = 532.79 ms / 199 runs ( 2.68 ms per token, 373.50 tokens per second) llama_perf_context_print: total time = 683.65 ms / 206 tokens ``` * sampling : remove version from sampler chain This commit removes the version field from the sampler chain and instead used the sampler pointer itself for change detection. * sampling : always populate logits for sampled probs This commit updates common/sampler.cpp set_logits and src/llama-sampling.cpp llama_sampler_sample to always populate the logits field when backend sampled probabilities are available. The motivation for this is that this ensure that CPU sampler always have access to the logits values even when probabilites have been produced by backend samplers. * sampling : simplify backend sampling logic decode This commit tries to simplify the backend sampling logic in llama_context::decode. * squash! sampling : simplify backend sampling logic decode Fix condition to check if backend actually sampled tokens, not just that backend samplers are available. * common : fix regression caused by extra memory allocations during sampling * squash! sampling : simplify backend sampling logic decode The commit fixes a variable shadowing issue in the `llama_context::decode` function which was introduced in a previous refactoring. * squash! common : fix regression caused by extra memory allocations during sampling Apply the same changes to llama-sampling.cpp, llama_sampler_sample as were applied in commit |
||
|
Aldehir Rojas
|
cef1d23c5a |
common/grammar : replace problematic backtracking regex [\s\S]* (#18342)
* grammar : add support for std::regex_search() with trigger patterns * common : update hermes2 pro trigger to search instead of match * common : use regex_search with anchoring for partial matching * common : adjust regex partial tests to use new pattern * grammar : check pattern directly instead of adding a type * common : adjust existing patterns to match new semantics |
||
|
Georgi Gerganov
|
4301e27319 |
common : restore grammar-based rejection sampling (#18137)
* common : restart grammar-based rejection sampling * sampling : allow null samplers |
||
|
Georgi Gerganov
|
254098a279 |
common : refactor common_sampler + grammar logic changes (#17937)
* common : refactor common_sampler + grammar logic changes * tests : increase max_tokens to get needed response * batched : fix uninitialized samplers |
||
|
Georgi Gerganov
|
196f5083ef |
common : more accurate sampling timing (#17382)
* common : more accurate sampling timing * eval-callback : minor fixes * cont : add time_meas impl * cont : fix log msg [no ci] * cont : fix multiple definitions of time_meas * llama-cli : exclude chat template init from time measurement * cont : print percentage of unaccounted time * cont : do not reset timings |
||
|
Johannes Gäßler
|
e789095502 |
llama: print memory breakdown on exit (#15860)
* llama: print memory breakdown on exit |
||
|
Georgi Gerganov
|
e92d53b29e |
sampling : optimize samplers by reusing bucket sort (#15665)
* sampling : optimize sorting using bucket sort in more places ggml-ci * sampling : do not sort in dist sampler ggml-ci * sampling : avoid heap allocations for sort buffers ggml-ci * common : add option to sort sampling candidates by probability ggml-ci * sampling : revert the change for preserving sort buffers * sampling : use std::copy instead of memcpy * sampling : clarify purpose of partial sort helpers ggml-ci * cont : remove wrong comment [no ci] * common : update comment Co-authored-by: Johannes Gäßler <johannesg@5d6.de> --------- Co-authored-by: Johannes Gäßler <johannesg@5d6.de> |
||
|
Olivier Chafik
|
f5cd27b71d |
server: streaming of tool calls and thoughts when --jinja is on (#12379)
* add common_json w/ support for truncated json healing * add common_chat_msg_diff * partial common_chat_parse * refactor parser w/ optionals * server: wire chat diffs in stream mode * fix trigger of thinking models (must happen after thoughts are closed) * fix functionary v3.2 raw python! * rename: common_chat_syntax (now contains format) * rm common_regex.at_start * don't return empty <think></think> * accommodate yet another deepseek r1 distill fantasy syntax (`<｜tool▁calls｜>`) * fix QwQ 32B tool call parsing after thoughts (hermes2) * better logs for grammar triggers * consume spaces after parse_json_tool_calls * fix required tool calls w/ thinking models that have pre-opened thinking tags * fix thinking model's initial trigger + test qwq's template * run most test_tool_call tests in stream + non-stream modes * make functionary v3.2 parsing more strict (differentiate first match from others) * send final diff from server, to close off raw python arguments * support partial content streaming in Generic mode * tool-call: allow content prelude before hermes2 tool calls (for Qwen2.5) * Update function-calling.md * Update tool_bench.py * chat-parser: remove input from exception (llm output may contain PII) --------- Co-authored-by: ochafik <ochafik@google.com> Co-authored-by: Olivier Chafik <ochafik@users.noreply.github.com> |
||
|
Ycros
|
39e73ae0d6 | common : Add a warning when we can't match samplers from a string or char. (#13330) | ||
|
oobabooga
|
233461f812 |
sampling : Integrate Top-nσ into main sampling chain (and add it to the server) (#13264)
* sampling: add Top-nσ sampler to `llama-server` and sampler ordering * revert: sampler ordering * revert: VS' crappy auto-formatting * revert: VS' crappy auto-formatting pt.2 * revert: my crappy eye sight... * sampling: add XTC to Top-nσ sampler chain * sampling: add Dyna. Temp. to Top-nσ sampler chain * sampling: actually remove Top-nσ from sampler(oops) * Integrate top_n_sigma into main sampler chain * Define COMMON_SAMPLER_TYPE_TOP_N_SIGMA * Formatting * Lint * Exit early in the sampler if nsigma < 0 --------- Co-authored-by: CasualAutopsy <casual_autopsy@outlook.com> |
||
|
Johannes Gäßler
|
dd373dd3bf | llama: fix error on bad grammar (#12628) | ||
|
Olivier Chafik
|
669912d9a5 |
tool-call: fix Qwen 2.5 Coder support, add micro benchmarks, support trigger patterns for lazy grammars (#12034)
* sampler: turn lazy grammar trigger words to regexes * add scripts/tool_bench.sh & .py * constrain llama json output regardless of function name if matches at beginning * update relaxed newline space rule in grammar tests * support add_generation_prompt query parameter (useful for /apply_template) * Update src/llama-grammar.cpp Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> |
||
|
mgroeber9110
|
5bbe6a9fe9 |
ggml : portability fixes for VS 2017 (#12150)
* Add include files for std::min/max and std::toupper/tolower * win32: move _USE_MATH_DEFINES before includes to ensure M_PI is defined * Use GGML_RESTRICT instead of "restrict" keyword everywhere, and use "__restrict" in MSVC plain C mode * win32: only use __restrict in MSVC if C11/C17 support is not enabled --------- Co-authored-by: Marcus Groeber <Marcus.Groeber@cerence.com> |
||
|
Olivier Chafik
|
c7f460ab88 |
server: fix tool-call of DeepSeek R1 Qwen, return reasoning_content (Command 7RB & DeepSeek R1) unless --reasoning-format none (#11607)
* extract & return thoughts in reasoning_content field (unless --reasoning-format) for DeepSeek R1 & Command R7B * tool-calls: add deepseek r1 template (models/templates/llama-cpp-deepseek-r1.jinja) + hackommodate broken official template * tool-calls: accommodate variety of wrong tool call opening tags both R1 Qwen 32B and 7B distills like to spit out * server/oai: ensure content is null when there are tool calls, and reasoning_content appears before content for readability * tool-calls: add DeepSeek R1 Qwen distills to server/README.md & server tests Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> |
||
|
Vinesh Janarthanan
|
27e8a23300 |
sampling: add Top-nσ sampler (#11223)
* initial sampling changes: * completed top nsigma sampler implementation * apply parameter to only llama-cli * updated readme * added tests and fixed nsigma impl * cleaned up pr * format * format * format * removed commented tests * cleanup pr and remove explicit floats * added top-k sampler to improve performance * changed sigma to float * fixed string format to float * Update src/llama-sampling.cpp Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> * Update common/sampling.cpp Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> * Update src/llama-sampling.cpp Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> * Update src/llama-sampling.cpp Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> * Update src/llama-sampling.cpp Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> * Update src/llama-sampling.cpp Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> * added llama_sampler_init --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> |
||
|
Michał Moskal
|
ff227703d6 |
sampling : support for llguidance grammars (#10224)
* initial porting of previous LLG patch * update for new APIs * build: integrate llguidance as an external project * use '%llguidance' as marker to enable llg lark syntax * add some docs * clarify docs * code style fixes * remove llguidance.h from .gitignore * fix tests when llg is enabled * pass vocab not model to llama_sampler_init_llg() * copy test-grammar-integration.cpp to test-llguidance.cpp * clang fmt * fix ref-count bug * build and run test * gbnf -> lark syntax * conditionally include llguidance test based on LLAMA_LLGUIDANCE flag * rename llguidance test file to test-grammar-llguidance.cpp * add gh action for llg test * align tests with LLG grammar syntax and JSON Schema spec * llama_tokenizer() in fact requires valid utf8 * update llg * format file * add $LLGUIDANCE_LOG_LEVEL support * fix whitespace * fix warning * include <cmath> for INFINITY * add final newline * fail llama_sampler_init_llg() at runtime * Link gbnf_to_lark.py script; fix links; refer to llg docs for lexemes * simplify #includes * improve doc string for LLAMA_LLGUIDANCE * typo in merge * bump llguidance to 0.6.12 |
||
|
Olivier Chafik
|
8b576b6c55 |
Tool call support (generic + native for Llama, Functionary, Hermes, Mistral, Firefunction, DeepSeek) w/ lazy grammars (#9639)
--------- Co-authored-by: Xuan Son Nguyen <thichthat@gmail.com> Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> Co-authored-by: Xuan Son Nguyen <son@huggingface.co> |
||
|
Georgi Gerganov
|
afa8a9ec9b |
llama : add llama_vocab, functions -> methods, naming (#11110)
* llama : functions -> methods (#11110) * llama : add struct llama_vocab to the API (#11156) ggml-ci * hparams : move vocab params to llama_vocab (#11159) ggml-ci * vocab : more pimpl (#11165) ggml-ci * vocab : minor tokenization optimizations (#11160) ggml-ci Co-authored-by: Diego Devesa <slarengh@gmail.com> * lora : update API names (#11167) ggml-ci * llama : update API names to use correct prefix (#11174) * llama : update API names to use correct prefix ggml-ci * cont ggml-ci * cont ggml-ci * minor [no ci] * vocab : llama_vocab_add_[be]os -> llama_vocab_get_add_[be]os (#11174) ggml-ci * vocab : llama_vocab_n_vocab -> llama_vocab_n_tokens (#11174) ggml-ci --------- Co-authored-by: Diego Devesa <slarengh@gmail.com> |
||
|
Georgi Gerganov
|
644fd71b44 |
sampling : refactor + optimize penalties sampler (#10803)
* sampling : refactor + optimize penalties sampler ggml-ci * common : apply ignore_eos as logit bias ggml-ci * batched : remove penalties sampler * params : allow penalty_last_n == -1 to be equal to context size ggml-ci * common : by default, move the penalties at the end of the sampling chain ggml-ci * common : ignore all EOG tokens Co-authored-by: Diego Devesa <slarengh@gmail.com> * common : move back the penalties at the front of the sampling chain ggml-ci * readme : restore hint about --ignore-eos flag [no ci] * llama : minor ggml-ci * webui : update --------- Co-authored-by: Diego Devesa <slarengh@gmail.com> |
||
|
Georgi Gerganov
|
d9d54e498d |
speculative : refactor and add a simpler example (#10362)
* speculative : refactor and add a simpler example ggml-ci * speculative : clean-up and add comments and TODOs [no ci] * speculative : manage context in common_speculative ggml-ci * speculative : simplify ggml-ci * speculative : simplify (cont) ggml-ci * speculative : add --draft-min CLI arg * speculative : minor fixup * make : build fixes * speculative : do not redraft previous drafts ggml-ci * speculative : fix the draft sampling ggml-ci * speculative : fix compile warning * common : refactor args ggml-ci * common : change defaults [no ci] * common : final touches ggml-ci |
||
|
Georgi Gerganov
|
8d8ff71536 |
llama : remove Tail-Free sampling (#10071)
ggml-ci |
||
|
wwoodsTM
|
ff252ea48e |
llama : add DRY sampler (#9702)
* sampling : add DRY sampler (post-refactor) * DRY: Trying to fix coauthors, removed unneeded line * DRY: Fixed redundant code * DRY: Fixed crash issue due to DRY being in chain but uninitialized --------- Co-authored-by: l3utterfly <gc.pthzfoldr@gmail.com> Co-authored-by: pi6am <34464159+pi6am@users.noreply.github.com> |
||
|
Georgi Gerganov
|
55e47786e3 |
llama : default sampling changes + greedy update (#9897)
* llama : deprecate softmax sampler + fix dist sampler ggml-ci * tests : replace macros with functions ggml-ci * sampling : change temperature sampler logic For t <= 0.0f, keep the max logit intact and set the rest to -inf * cont : no need for special "greedy" logic top-k == 1 is the same * tests : init prob correctly * llama : handle temp <= 0.0 in the temp_ext sampler too ggml-ci * cont : avoid extra loop in temperature sampler for sub-zero temp ggml-ci |
||
|
Georgi Gerganov
|
755a9b2bf0 |
llama : add infill sampler (#9896)
ggml-ci |
||
|
MaggotHATE
|
fbc98b748e |
sampling : add XTC sampler (#9742)
* Initial XTC commit Adds XTC sampler, not activated by default, but recommended settings by default. * Cleanup * Simplified chances calculation To be more inline with the original implementation, chance is calculated once at the beginning. * First fixes by comments Still need to look into sorting * Fixed trailing backspaces * Fixed RNG to be reproduceable Thanks to @slaren for directions * Fixed forgotten header * Moved `min_keep` Moved from conditions to a simple check at the end. * Fixed broken randomization Thanks to @slaren for explanation * Swapped sorting for a custom algorithm Shifts tokens to remove the penalized ones, then puts the penalized at the back. Should make `min_keep` still viable. * Algorithm rework 1. Scan token from top till the first non-penalizable 2. Remove the last captured token (the least probable above threshold) 3. Shift all tokens to override the remaining penalizable 4. Penalize and put them at the the bottom. * Added XTC to `test-sampling` * Simplified algorithm and more tests * Updated info in common and args * Merged back lost commits in common and arg * Update dump info in common * Fixed incorrect min_keep check * Added XTC to README * Renamed parameters, fixed info and defaults * probability is at 0 by default, but XTC is included in sampling queue * threshold higher than 0.5 switches XTC off * Initial server support * Added XTC to server UIs * Fixed labels in old server UI * Made algorithm safer and more readable * Removed xtc_threshold_max * Fixed arg after update * Quick fixes by comments * Simplified algorithm since threshold_max is removed * Renamed random distribution * Fixed tests and outdated README * Small fixes |
||
|
Diego Devesa
|
7eee341bee |
common : use common_ prefix for common library functions (#9805)
* common : use common_ prefix for common library functions --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> |
||
|
Georgi Gerganov
|
b0f27361f3 |
sampling : avoid expensive softmax during greedy sampling (#9605)
* sampling : avoid expensive softmax during greedy sampling ggml-ci * speculative : fix default RNG seed + set sparams.n_probs * Update tests/test-sampling.cpp Co-authored-by: slaren <slarengh@gmail.com> * sampling : add clarifying comment [no ci] --------- Co-authored-by: slaren <slarengh@gmail.com> |
||
|
Georgi Gerganov
|
6262d13e0b |
common : reimplement logging (#9418)
https://github.com/ggerganov/llama.cpp/pull/9418 |
||
|
Georgi Gerganov
|
0abc6a2c25 |
llama : llama_perf + option to disable timings during decode (#9355)
* llama : llama_perf + option to disable timings during decode ggml-ci * common : add llama_arg * Update src/llama.cpp Co-authored-by: Xuan Son Nguyen <thichthat@gmail.com> * perf : separate functions in the API ggml-ci * perf : safer pointer handling + naming update ggml-ci * minor : better local var name * perf : abort on invalid sampler pointer ggml-ci --------- Co-authored-by: Xuan Son Nguyen <thichthat@gmail.com> |
||
|
slaren
|
49006c67b4 |
llama : move random seed generation to the samplers (#9398)
* llama_sampler_penalties : clamp penalty_last_n to zero |
||
|
Xuan Son Nguyen
|
bfe76d4a17 |
common : move arg parser code to arg.cpp (#9388)
* common : move arg parser to arg.cpp * better categorize args * add cmake * missing climits * missing cstdarg * common : more explicit includes * fix build * refactor gpt_params_parse * update server readme * fix test --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com> |
||
|
Georgi Gerganov
|
f12295b8a9 | llama : fix empty ring buffer push (#9358) | ||
|
Georgi Gerganov
|
df270ef745 |
llama : refactor sampling v2 (#9294)
- Add `struct llama_sampler` and `struct llama_sampler_i` - Add `llama_sampler_` API - Add `llama_sampler_chain_` API for chaining multiple samplers - Remove `LLAMA_API_INTERNAL` - Add `llama_perf_` API and remove old `llama_print_timings` and `llama_reset_timings` |
||
|
Georgi Gerganov
|
938943cdbf |
llama : move vocab, grammar and sampling into separate files (#8508)
* llama : move sampling code into llama-sampling ggml-ci * llama : move grammar code into llama-grammar ggml-ci * cont ggml-ci * cont : pre-fetch rules * cont ggml-ci * llama : deprecate llama_sample_grammar * llama : move tokenizers into llama-vocab ggml-ci * make : update llama.cpp deps [no ci] * llama : redirect external API to internal APIs ggml-ci * llama : suffix the internal APIs with "_impl" ggml-ci * llama : clean-up |
||
|
Kevin Wang
|
470939d483 |
common : preallocate sampling token data vector (#8363)
`emplace_back` repeatedly-called is slower than preallocating the vector to the vocab size and directly inserting the data. Some rudimentary profiling with `chrono` improves the performance of this block of code from ~500us/op to ~40us/op. Overall, this slightly improves the sampling performance which has a more substantial impact for the `examples/lookahead` implementation -- I am able to see a ~10% performance boost in lookahead inference. |
||
|
Kevin Wang
|
ffd00797d8 | common : avoid unnecessary logits fetch (#8358) | ||
|
Daniel Bevenius
|
e6bf007744 |
llama : return nullptr from llama_grammar_init (#8093)
* llama : return nullptr from llama_grammar_init This commit updates llama_grammar_init to return nullptr instead of throwing an exception. The motivation for this is that this function is declared inside an extern "C" block and is intended/may be used from C code which will not be able to handle exceptions thrown, and results in undefined behavior. On Windows and using MSVC the following warning is currently generated: ```console C:\llama.cpp\llama.cpp(13998,1): warning C4297: 'llama_grammar_init': function assumed not to throw an exception but does C:\llama.cpp\llama.cpp(13998,1): message : __declspec(nothrow), throw(), noexcept(true), or noexcept was specified on the function ``` Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com> * squash! llama : return nullptr from llama_grammar_init Add checks for nullptr when calling llama_grammar_init. Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com> --------- Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com> Co-authored-by: Clint Herron <hanclinto@gmail.com> |
||
|
Georgi Gerganov
|
6ff13987ad |
common : normalize naming style (#7462)
* common : normalize naming style ggml-ci * common : match declaration / definition order * zig : try to fix build |
||
|
Olivier Chafik
|
e402de364b |
grammars: fix resampling logic regression (#7424)
|
||
|
Johannes Gäßler
|
5ae3426b0b | server: fix reported top tokens for temperature 0 (#7203) | ||
|
Johannes Gäßler
|
af0a5b6163 |
server: fix incorrectly reported token probabilities (#7125)
* server: normalize token probabilities * fix temperature == 0.0f |
||
|
David Renshaw
|
3f167476b1 | sampling : use std::random_device{}() for default random seed (#6962) | ||
|
Johannes Gäßler
|
28103f4832 |
Server: fix seed for multiple slots (#6835)
* Server: add tests for consistent results * sampling: separate rng per sampling context |
||
|
Minsoo Cheong
|
586e7bc561 |
sampling : deduplicated code for probability distribution access (#6240)
* sampling: remove duplicated code for probability distribution access * free original_logits * fix original_logits allocation * fixes based on review @cebtenzzre * change function name to `llama_sampling_prepare` |